OAQPS CONTROL COST MANUAL
Fifth Edition
EPA 453/B-96-001
February 1996
U.S. Er,v;ron'TV2nt3! Fraction Agency
Region C, ,!j> ;r> vF'L-I?.n
77 \Vcii j«r.c\5in Liuuievcird, 12th Floor
Chicago, IL 60604-3590
United States Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
-------
This fifth edition of the OAQPS Control Cost Manual was prepared by the Emissions
Standards Division of the Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711. Mention of trade names or commercial
products is not intended to constitute endorsement or recommendation for use. Copies of this
report are available through the OAQPS Control Technology Center (MD-13), U.S.
Environmental Protection Agency, Research Triangle Park NC 27711, or from the National
Technical Information Service, 5285 Port Royal Road, Springfield VA 22161, (phone: 1-800-
553-6847.)
Questions and comments should be addressed to the principal author, William M. Vatavuk,
OAQPS, phone 919-541-5309, fax 919-541-0839.
-------
Chapter 1
INTRODUCTION
William M. Vatavuk
Innovative Strategies and Economics Group, OAQPS
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
December 1995
Contents
1.1 Role of Cost in Setting of Regulations 1-2
1.2 Purpose of Manual 1-2
1.3 Organization of the Manual 1-3
1.4 Intended Users of the Manual 1-4
1.5 "Uniqueness" of the Manual 1-4
References 1-6
1-1
-------
1.1 Role of Cost in Setting of Regulations
Cost has an important role in setting many state and federal air pollution control regulations.
The extent of this role varies with the type of regulation. For some types of regulations, such
as Maximum Achievable Control Technology (MACT) standards, cost is explicitly used in
determining their stringency. This use may involve a balancing of costs and environmental
impacts, costs and dollar valuation of benefits, or environmental impacts and economic
consequences of control costs.
For other types of regulations (e.g., National Ambient Air Quality Standards), cost analysis
is used to choose among alternative regulations with the same level of stringency. For these
regulations, the environmental goal is determined by some set of criteria which do not include
costs. However, cost-effectiveness analysis is employed to determine the minimum cost way
of achieving the goal.
For some regulations, cost influences enforcement procedures or requirements for
demonstration of progress towards compliance with an air quality standard. For example, the
size of any monetary penalty assessed for noncompliance as part of an enforcement action needs
to be set with awareness of the magnitude of the control costs being postponed by the
noncomplying facility. For regulations without a fixed compliance schedule, demonstration of
reasonable progress towards the goal is sometimes tied to the cost of attaining the goal on
different schedules.
Cost is a vital input to two other types of analyses that also sometimes have a role in standard
setting. Cost is needed for a benefit-cost analysis that addresses the economic efficiency of
alternative regulations. Cost is also an input into any analysis of the economic impact of each
regulatory alternative. An economic impact analysis deals with the consequences of the
regulation for small businesses, employment, prices, and market and industry structure.
1.2 Purpose of Manual
The purpose of this Manual is two-fold: (1) to compile up-to-date capital costs, operating and
maintenance expenses, and other costs for "add-on" air pollution control systems and (2) to
provide a comprehensive, concise, consistent, and easy-to-use procedure for estimating and
(where appropriate) escalating these costs. The EPA report, Escalation Indexes for Air Pollution
Control Costs (EPA-452/R-95-006, October 1995) provides data and a procedure for escalating
control device equipment costs. The data are in the form of indexes ("Vatavuk Air Pollution
Control Cost Indexes" or "VAPCCI") that are updated quarterly. This report and the updated
indexes are installed on the OAQPS Technology Transfer Network (TTN), Control Technology
Center (CTC) bulletin board. ("Add-on" systems are those installed downstream of an air
pollution source to control its emissions.)
1-2
-------
The Manual estimating procedure rests on the notion of the "factored" or "study" estimate,
nominally accurate to within ± 30%. This type of estimate is well suited to estimating control
system costs intended for use in regulatory development. Study estimates are sufficiently
accurate, yet do not require the detailed, site-specific data inputs needed to make definitive or
other more accurate types of estimates.
1.3 Organization of the Manual
This Manual is a revision of the 1990 edition of the OAQPS Control Cost Manual, [1] which,
in turn, was a revision of the edition completed in 1987. This fifth edition of the Manual
includes sizing and costing procedures and data for eight types of add-on control devices and
three kinds of auxiliary equipment (See Table 1.1). This edition includes revisions to all ten
Chapters in the fourth edition (i.e., the base Manual plus Supplements 1-3). The revisions
mainly consist of corrections and updates, most of which are minor.
As with the third edition, this edition has been issued in self-contained chapters. Each
chapter addresses a logically separate topic, which can be either of a general nature (e.g., this
introduction) or of a more specific, equipment-oriented nature (e.g., fabric filters). The chapters
which comprise this portion of the Manual are listed in Table 1.1, alongside the portions of the
1990 Manual they replace.
As in the fourth edition, each type of equipment (control device or auxiliary), background
topic, etc., is given its own chapter number for ease of identification and to reinforce the intent
that each chapter should stand alone.
Table 1.1: Contents of the OAQPS Control Cost Manual (Fifth Edition)
Chapter
No. Title
4th Edition Manual Portion
Replaced (Date)
1 "Introduction"
2 "Cost Estimating Methodology"
3 "Thermal and Catalytic Incinerators"
4 "Carbon Adsorbers"
5 "Fabric Filters"
6 "Electrostatic Precipitators"
7 "Flares"
8 "Refrigerated Condensers"
9 "Gas Absorbers"
10 "Hoods, Ductwork, and Stacks"
Base Manual (1990)
Supplement 1 (Apr. 1991)
Supplement 1 (Nov. 1991)
Supplement 2 (Oct. 1992)
Supplement 3 (Mar. 1994)
1-3
-------
Each of these stand-alone chapters contains a:
• Process description, where the types, uses, and operating modes of the equipment item
and (if applicable) its auxiliaries are discussed;
• Sizing (design) procedure, which enables one to use the parameters of the pollution
source (e.g., gas volumetric flow rate) to size the equipment item(s) in question;
• Capital and annual costing procedure and data for the equipment and suggested factors
to use in estimating these costs from equipment design and operational (e.g., operating
hours) parameters. These costs are presented in both graphical and equational forms
wherever possible. To complement the Manual costs, Lotus spreadsheets have been
developed. These spreadsheets (filename: COST.AIR.ZIP), which have been installed
on the CTC bulletin board of the TTN, allow the user to size and cost any of the control
devices covered in the Manual, plus several others (e.g., venturi scrubbers).
• Example problem to illustrate the sizing and costing procedures presented in the chapter.
1.4 Intended Users of the Manual
As explained in Section 1.2, the Manual provides comprehensive procedures and data for sizing
and costing control equipment. Some of these procedures are based on rigorous engineering
principles, such as the material and engineering balances in Chapter 3. To fully appreciate, and
correctly apply, these procedures the user should be able to understand them. Moreover, the user
should be able to exercise "engineering judgement" on those occasions when the procedures may
need to be modified or disregarded. Typically, only engineers and others with strong technical
backgrounds possess this kind of knowledge. Hence, the Manual is oriented toward the
technical not the non-technical user.
1.5 "Uniqueness'' of the Manual
The Manual presents a different perspective on estimating air pollution control system costs than
other cost-oriented reports, such as:
• The Cost Digest: Cost Summaries of Selected Environmental Control Technologies [2]
• A Standard Procedure for Cost Analysis of Pollution Control Operations [3]
• Handbook: Control Technologies for Hazardous Air Pollutants [4]
Although these reports (as well as many of the MACT Regulatory Impact Analyses and other
standards-supporting documents) contain costs for add-on control systems, they do not duplicate
1-4
-------
the Manual for one or more of the following reasons: (1) their costs have been based either
wholly or partly on data in the previous Manuals; (2) they apply to specific source categories
only, whereas the Manual data may be applied generally; (3) their estimating procedures and
costs are of less than study estimate quality; or (4) they are not intended for estimating costs used
in regulatory development.
Reason (3) applies to the Cost Digest, for example, as this report, designed for use by non-
technical personnel, contains procedures for making order-of-magnitude estimates (± 30%
accuracy or worse). A Standard Procedure, conversely, was primarily intended for estimating
costs for R&D cases (e.g., demonstration projects), where some site-specific data are available.
Further, although the latter report contains a thorough list of equipment installation factors, it
contains few equipment costs. The report, Handbook: Control Technologies, used data and
estimating procedures from the 1990 Manual to provide sound generalized procedures for
estimating costs for various types of control equipment. The fifth edition of the Manual
supplements this information.
Finally, since its inception, the Manual has been extensively used to support Agency
regulatory development, state permitting programs, and other activities where current, consistent,
and comprehensive control cost data are required. Accordingly, the Manual's role in the
specialty of air pollution control system cost estimating is both unique and secure.
1-5
-------
References
[1] OAQPS Control Cost Manual (Fourth Edition), EPA, Office of Air Quality Planning and
Standards, Emissions Standards Division, January 1990 (EPA 450/3-90-006 [NTIS/PB90-
169954]).
[2] DeWolf, Glenn, et al. (Radian, Inc.), The Cost Digest: Cost Summaries of Selected
Environmental Control Technologies, EPA, ORD, Office of Environmental Engineering and
Technology, October 1984 (EPA-600/884-010).
[3] Uhl, Vincent W., A Standard Procedure for Cost Analysis of Pollution Control Operations,
Volumes I and II, EPA, ORD, Industrial Environmental Research Laboratory, June 1979
(EPA-600/8-79-018a).
[4] Handbook: Control Technologies for Hazardous Air Pollutants, EPA, Office of Research
and Development, Air and Energy Engineering Research Laboratory, June 1991 (EPA-
625/6-91-014).
1-6
-------
Chapter 2
COST ESTIMATING
METHODOLOGY
William M. Vatavuk
Innovative Strategies and Economics Group. OAQPS
U.S. Environmental Protection Agency
Research Triangle Park. NC 27711
December 1995
-------
Contents
2.1 Types of Cost Estimates 2-4
2.2 Cost Categories Defined 2-5
2.2.1 Elements of Total Capital Investment 2-6
2.2.2 Elements of Total Annual Cost 2-7
2.3 Engineering Economy Concepts 2-10
2.3.1 Time Value of Money 2-10
2.3.2 Cash Flow 2-10
2.3.3 Annualization and Discounting Methods 2-13
2.4 Estimating Procedure 2-14
2.4.1 Facility Parameters and Regulatory Options 2-14
2.4.2 Control System Design 2-15
2.4.3 Sizing the Control System 2-16
2.4.4 Estimating Total Capital Investment 2-20
2.4.4.1 General Considerations 2-20
2.4.4.2 Retrofit Cost Considerations 2-21
2.4.5 Estimating Annual Costs 2-23
2.4.5.1 Raw Materials 2-23
2.4.5.2 Operating Labor 2-23
2.4.5.3 Maintenance 2-24
2.4.5.4 Utilities 2-24
2-2
-------
2.4.5.5 Waste Treatment and Disposal 2-25
2.4.5.6 Replacement Parts 2-25
2.4.5.7 Overhead 2-26
2.4.5.8 Property Taxes, Insurance, and Administrative Charges 2-27
2.4.5.9 Capital Recovery 2-27
References 2-28
2-3
-------
This chapter presents a methodology that will enable the user, having knowledge of the source being
controlled, to produce study-level cost estimates for a control system to control that source. The
methodology, which applies to each of the control systems included in this Manual, is general enough
to be used with other "add-on" systems as well. Further, the methodology may also be applicable to
estimating costs of fugitive emission controls and of other non-stack abatement methods.
Before presenting this methodology in detail, we should first discuss the various kinds of cost
estimates and then define the cost categories and engineering economy concepts employed in making
the estimates.
2.1 Types of Cost Estimates
As noted above, the costs and estimating methodology in this Manual are directed toward the "study"
estimate, of ± 30% accuracy. According to Perry's Chemical Engineer's Handbook, a study estimate
is "... used to estimate the economic feasibility of a project before expending significant funds for
piloting, marketing, land surveys, and acquisition ... [However] it can be prepared at relatively low
cost with minimum data."[l] Specifically, to make a study estimate, the following must be known:
• Location of the source within the plant;
• Rough sketch of the process flow sheet (i.e., the relative locations of the equipment in the
system);
• Preliminary sizes of, and material specifications for, the system equipment items;
• Approximate sizes and types of construction of any buildings required to house the
control system;
• Rough estimates of utility requirements (e.g., electricity);
• Preliminary flow sheet and specifications for ducting and piping;
• Approximate sizes of motors required.fl]
In addition, an estimate of the labor hours required for engineering and drafting is needed, as the
accuracy of an estimate (study or otherwise) is highly dependent on the amount of engineering work
expended on the project.
There are, however, four other types of estimates, three of which are more accurate than the
study estimate. These are:[l]
2-4
-------
Order-of-magnitude. This estimate provides "a rule-of-thumb procedure applied only to
repetitive types of plant installations for which there exists good cost history". Its error
bounds are greater than ± 30%. (However, according to Perry's, "... no limits of accuracy
can safely be applied to it.") The sole input required for making this level of estimate is the
control system's capacity (often measured by the maximum volumetric flow rate of the gas
passing through the system). So-called "six-tenths factor" estimates (not to be confused with
factored estimates) are examples of this type.
Scopeor Budget authorization or Preliminary. This estimate, nominally of ± 20% accuracy,
requires more detailed knowledge than the study estimate regarding the site, flow sheet,
equipment, buildings, etc. In addition, rough specifications for the insulation and
instrumentation are also needed.
Project control or Definitive. These estimates, accurate to within ± 10%, require yet more
information than the scope estimates, especially concerning the site, equipment, and electrical
requirements.
Firm or Contractor's or Detailed. This is the most accurate (± 5%) of the estimate types,
requiring complete drawings, specifications, and site surveys. Further, "[t]ime seldom permits
the preparation of such estimates prior to an approval to proceed with the project."[1]
For the purposes of regulatory development, study estimates have been found to be acceptable,
as they represent a compromise between the less accurate order-of-magnitude and the more accurate
estimate types. The former are too imprecise to be of much value, while the latter are not only very
expensive to make, but require detailed site and process-specific knowledge that most Manual users
will not have available to them.
2.2 Cost Categories Defined
The names given certain categories of costs and what they contain vary considerably throughout the
literature. Certain words like "capital cost" can have vastly different meanings, which can often lead
to confusion, even among cost estimators. To avoid this confusion and, at the same time, provide
uniformity in the Manual basic terms are defined in this chapter and will be used throughout. The
terminology used is adapted from that of the American Association of Cost Engineers. [2] Although
it has been developed for general use, it is readily adaptable to air pollution control system costing.
First, two general kinds of costs are estimated, total capital investment (TCI) and total annual
cost (TAC). These are discussed below.
2-5
-------
2.2.1 Elements of Total Capital Investment
The total capital investment includes all costs required to purchase equipment needed for the control
system (termed purchased equipment costs), the costs of labor and materials for installing that
equipment (termed direct installation costs), costs for site preparation and buildings, and certain
other costs which are termed indirect installation costs. The TCI also includes costs for land,
working capital, and off-site facilities.
Direct installation costs include costs for foundations and supports, erecting and handling the
equipment, electrical work, piping, insulation, and painting. Indirect installation costs include such
costs as engineering costs; construction and field expenses (i.e., costs for construction supervisory
personnel, office personnel, rental of temporary offices, etc.); contractor fees (for construction and
engineering firms involved in the project); start-up and performance test costs (to get the control
system running and to verify that it meets performance guarantees); and contingencies. Contingencies
is a catch-all category that covers unforeseen costs that may arise, including (but certainly not limited
to)"... possible redesign and modification of equipment, escalation increases in cost of equipment,
increases in field labor costs, and delays encountered in start-up. "[2]
These elements of total capital investment are displayed in Figure 2.1. Note that the sum of the
purchased equipment cost, direct and indirect installation costs, site preparation, and buildings costs
comprises the battery limits estimate. By definition, this is the total estimate "... for a specific job
without regard to required supporting facilities which are assumed to already exist... "[2] at the plant.
This would mainly apply to control systems installed in existing plants, though it could also apply to
those systems installed in new plants when no special facilities for supporting the control system (/'. e.,
off-site facilities) would be required.
Where required, these off-site facilities would encompass units to produce steam, electricity, and
treated water; laboratory buildings, railroad spurs, roads, and the like. It is unusual, however, for a
pollution control system to have one of these units (e.g., a power plant) dedicated to it. The system
needs are rarely that great. However, it may be necessary—especially in the case of control systems
installed in new or "grass roots" plants—for extra capacity to be built into the site generating plant
to service the system. (A venturi scrubber, which often requires large amounts of electricity, is a
good example of this.) It is customary for the utility costs to be charged to the project as operating
costs at a rate which covers both the investment and operating and maintenance costs for the utility.
As Figure 2.1 shows, there are two other costs which may be included in the total capital
investment for a control system. These are working capital and land. The first of these, working
capital, is a fund set aside to cover the initial costs of fuel, chemicals, and other materials, as well as
labor and maintenance. It usually does not apply to control systems, for the quantities of utilities,
materials, labor, etc., they require are usually small. (An exception might be an oil-fired thermal
incinerator, where a small supply (e.g., 30-day) of distillate fuel would have to be available during its
initial period of operation.)
Land may also be required. But, since most add-on control systems take up very little space (a
quarter-acre or less) this cost would be relatively small. (Certain control systems, such as those used
2-6
-------
for flue gas desulfurization, require larger quantities of land for the process equipment, chemicals
storage, and waste disposal.)
Note also in Figure 2.1 that the working capital and land are nondepreciable expenses. In other
words, these costs are "recovered" when the control system reaches the end of its useful life
(generally in 10 to 20 years). Conversely, the other capital costs are depreciable, in that they cannot
be recovered and are included in the calculation of income tax credits (if any) and depreciation
allowances, whenever income taxes are considered in a cost analysis. (In the Manual methodology,
however, income taxes are not considered. See Section 2.3.)
Notice that when 100% of the system costs are depreciated, no salvage value is taken for the
system equipment at the conclusion of its useful life. This is a reasonable assumption for add-on
control systems, as most of the equipment, which is designed for a specific source, cannot be used
elsewhere without modifications. Even if it were reusable, the cost of disassembling the system into
its components (i.e., "decommissioning cost") could be as high (or higher) than the salvage value.
2.2.2 Elements of Total Annual Cost
The Total Annual Cost (TAG) for control systems is comprised of three elements: direct costs (DC),
indirect costs (1C), and recovery credits (RC),
which are related by the following equation:
TAC = DC + IC-RC (2.1)
Clearly, the basis of these costs is one year, as this period allows for seasonal variations in production
(and emissions generation) and is directly usable in profitability analyses. (See Section 2.3.)
Direct costs are those which tend to be proportional or partially proportional to the quantity of
exhaust gas processed by the control system per unit time. These include costs for raw materials,
utilities (steam, electricity, process and cooling water, etc.), waste treatment and disposal.
maintenance materials, replacement parts, and operating, supervisory, and maintenance labor. Of
these direct costs, costs for raw materials, utilities, and waste treatment and disposal are variable,
in that they tend to be a direct function of the exhaust flow rate. That is, when the flow rate is at its
maximum rate, these costs are highest. Conversely, when the flow rate is zero, so are the costs.
Semivariable direct costs are only partly dependent upon the exhaust flow rate. These include
all kinds of labor, overhead, maintenance materials, and replacement parts. Although these costs are
a function of the gas flow rate, they are not linear functions. Even while the control system is not
operating, some of the semivariable costs continue to be incurred.
Indirect, or "fixed", annual costs are those whose values are totally independent of the exhaust
flow rate and, in fact, would be incurred even if the control system were shut down. They include
such categories as administrative charges, property taxes, insurance, and capital recovery.
2-7
-------
Finally, the direct and indirect annual costs are offset by recovery credits, taken for materials or
energy recovered by the control system, which may be sold, recycled to the process, or reused
• Primary Control Device
• Auxiliary Equipment
(including ductwork)
• Instrumentation
* S ales Taxes
• Fieignt
Foundations
and Supports
Handling and Erection
Electrical
Piping
Insulation
Painting
Site
Purchased Direct Preparation fi°
Equipment Installation
Cost Cost* Buildings
Total Direct
Cost
Land'
Wbrking
Capital^
"Battery Limits"
Cost
Engineering
C onstroction and
Field Expenses
Contractor Fees
S tart-up
Performance Test
Contingencies
Indirect
Installation
Cost*
Total Indirect
Cost
Ofif-Site
Facilities '
Total Non-depreciab k
Investment
Total Depreciable
Investment
Total Capital
Investment
"Typically factored from (hesumct heprimar^ control deuce aid aiKiliary eqifmait costs.
Typically factored from (repurchased eqtipmait cost.
^Usually required only at "grass roots" installations.
Unl ike the o(her d rect aid rnd rect costs; costs Eor Aiese item usual! y * e not factored from (he
^purchased ecjuipmat cost. Bather, they are si zed did costed sqiaratelu
'Normally not required wi8i add-on control systems.
Figure 2.1: Elements of Total Capital Investment
2-8
-------
Raw material
Utilities
Electricity
Fuel
Steam
Water
Compressed air
Water treatment/
disposal
Variable
Labor
Operating
Supervisory-
Maintenance
Maintenance
Materials
Replacement parts
Semivariable
Direct Costs
Overhead
Property taxes
Insurance
Administrative
Charges
Capital recovery
.Indirect.
Costs
Materials
Energy
.Recovery.
Credits
Total
.Annual
Cost
Figure 2.2: Elements of Total Annual Cost
2-9
-------
elsewhere at the site. These credits, in turn, must be offset by the costs necessary for their
processing, storage, transportation, and any other steps required to make the recovered materials
or energy reusable or resalable. Great care and judgement must be exercised in assigning values
to recovery credits, since materials recovered may be of small quantity or of doubtful purity,
resulting in their having less value than virgin material. Like direct annual costs, recovery credits
are variable, in that their magnitude is directly proportional to the exhaust flow rate. The various
annual costs and their interrelationships are displayed in Figure 2.2. A more thorough description
of these costs and how they may be estimated is given in Section 2.4.
2.3 Engineering Economy Concepts
As mentioned previously, the estimating methodology presented in Section 2.4 rests upon the
notion of the "factored" or "study" estimate. However, there are other concepts central to the
cost analyses which must be understood. These are (1) the time value of money, (2) cash flow,
and (3) annualization.
2.3.1 Time Value of Money
The time value of money is based on the truism that".. .a dollar now is worth more than the
prospect of a dollar... at some later date."[3] A measure of this value is the interest rate which
"... may be thought of as the return obtainable by the productive investment of capital. "[3]
2.3.2 Cash Flow
During the lifetime of a project, various kinds of cash expenditures are made and various incomes
are received. The amounts and timing of these expenditures and incomes constitute the cash
flows for the project. In control system costing it is normal to consider expenditures (negative
cash flows) and unusual to consider income (positive cash flows), except for product or energy
recovery income. By the simplifying convention recommended by Grant, Ireson, and
Leavenworth[3], each annual expenditure (or payment) is considered to be incurred at the end of
the year, even though the payment will probably be made sometime during the year in question.
(The error introduced by this assumption is minimal, however.) Figure 2.3, which shows three
hypothetical cash flow diagrams, illustrates these end-of-year payments. In these diagrams, P
represents the capital investment, while the A's denote the end-of-year annual payments. Note
that in all diagrams, the cash flows are in constant (real) dollars, meaning that they do not reflect
the effects of inflation. Also note that in the top diagram (I), the annual payments are different for
each year. (These represent the control system annual costs (exclusive
2-10
-------
I.
Ye:
G
P
IT:
1234567891
A,
i
A,
i
A3
A4
1
A,
i
A«
Ar
A.
A.
' t
AlO
n.
Year:
0123456789 10
A
A
A
t i
A
A
A
A
A
> i
A
A
i <
Year:
01
in.
23456789 10
A1
A1
A1
A1
A1
A1
i t
A1
r
Av
A1
i
A*
• '
•All Values Are Constant Year (Real) Dollars
Figure 2.3: Hypothetical Cash Flow Diagrams
2-11
-------
of capital recovery) described in Section 2.2.) In reality, these payments would be different, as
labor and maintenance requirements, labor and utility costs, etc., would vary from year to year. A
generally upward trend in annual costs would be seen, however.
In diagram II, these fluctuating annual payments have been converted to equal payments. This
can be done by calculating the sum of the present values of each of the annual payments shown in
diagram I and annualizing the total net present value to equivalent equal annual payments via a
capital recovery factor. (See discussion in the following paragraphs and in Section 2.3.3.)
Alternatively, it is adequate to choose a value of A equal to the sum of the direct and indirect
annual costs estimated for the first year of the project. This assumption is in keeping with the
overall accuracy of study estimates and allows for easier calculations.
Finally, notice diagram III. Here, the annual costs (A1) are again equal, while the capital
investment (P) is missing. Put simply, P has been incorporated into A1, so that A1 reflects not
only the various annual costs but the investment as well. This was done by introducing another
term, the capital recovery factor (CRF), defined as follows: "when multiplied by a present debt or
investment, [the CRF] gives the uniform end-of-year payment necessary to repay the debt or
investment in n years with interest rate i."[3] The product of the CRF and the investment (P) is
the capital recovery cost (CRC):
CRC = CRFxP (2.2)
where
(2.3)
v '
Therefore, A1 is the sum of A and the CRC, or:
A]=A + CRFxP (2.4)
In this context, n is the control system economic life, which, as stated above, typically varies
from 10 to 20 years. The interest rate (/) used in this Manual is a pretax marginal rate of return
on private investment of 7% (annual). This value, which could also be thought of as a "real
private rate of return", is used in most of the OAQPS cost analyses and is in keeping with current
OAQPS guidelines and the Office of Management and Budget recommendation for use in
regulatory analyses. [4]
It may be helpful to illustrate the difference between real and nominal interest rates. The
mathematical relationship between them is straight forward:[4]
2-12
-------
(!+/„) = (l+i)(l+r) (2.5)
where
/„, / = the annual nominal and real interest rates, respectively
r = the annual inflation rate
Clearly, the real rate does not consider inflation and is in keeping with the expression of
annual costs in constant (i.e., real) dollars.
The above procedure using the pre-tax marginal (or real) rate of return on private investment
is the appropriate method for assessing the costs from the perspective of the entity having to
install the pollution control equipment. For example, costs developed with the above procedure
can appropriately be used for answering questions concerning the market response to regulation
like price increases, quantity adjustments, and reduced profitability.
In an idealized economy with perfectly competitive and complete markets, this private cost
and the social cost would be equal. However, in a more realistic economy in which allocation of
resources is distributed by taxes, credit restrictions, and other market imperfections, the cost to
society is different than the private costs for capital expenditures. The costs to society are the
relevant costs for use in answering questions about economic efficiency. For example, benefit
cost analysis and cost-effectiveness analyses should focus on cost to society, not just the cost to
the entity facing additional pollution control costs.
EPA has adopted a new approach, a two-stage approach, to discounting for social costs. This
new approach begins with the same capital recovery costs (CRC) described above using the same
7% pre-tax marginal rate of return on private investment. The second step of the two-stage
approach involves "discounting" both direct and indirect annual costs and CRC back to an initial
date ("year 0") using a consumption rate of interest of 3%. (See Section 2.3.3 for an explanation
of the discounting concept.) This results in a relatively higher cost of capital from society's
perspective than from the perspective of the entity facing additional control cost. A detailed
explanation of this procedure and when it should be employed is beyond the scope of this
document. A fuller explanation is given in draft EPA guidelines [5]. However, it is mentioned
here because the CRC and direct and indirect annual costs are inputs to the two-stage procedure
and must be sufficiently itemized to allow use in the two-stage procedure.
2.3.3 Annualization and Discounting Methods
The above method of smoothing out the investment into equal end-of-year payments, is termed
the equivalent uniform annual cashflow (EUAC) method.[3] In addition to its inherent
simplicity, this method is very useful when comparing the costs of two or more alternative control
systems (i.e., those which are designed to control the same source to an equivalent degree). In
fact, the EUAC's—or simply the total annual costs—of two competing systems may be compared
even if both the systems have different economic lives, say 10 and 20 years. We recommend that
2-13
-------
the EUAC method be used for estimating control costs unless particular circumstances preclude
its use.
Comparisons of systems with different economic lives cannot be made, however, using the
other two annualization (i.e., profitability analysis) methods—present worth (PW) and internal
rate of return (IRR). The present worth (or discounted cash flow) method involves the
discounting of all cash flows occurring after year 0 (i.e., the system startup date) back to year 0.
These cash flows are discounted by multiplying each by a discount factor, —L_, where m is the
number of years from year 0 to the year in which the cash flow is incurred. The sum of these
discounted cash flows is then added to the capital investment to yield the present worth of the
project. The alternative having the highest present worth would be selected (in control system
costing this is usually a negative number). But when comparing the present worths of alternative
systems, the system lifetimes must be equal for the comparison to be valid. [3]
The third annualization method, internal rate of return (IRR), is similar to the present worth
method, in that it involves the discounting of a series of unequal cash flows. However, where
with the PW method the interest rate, /, is set beforehand, in the IRR method the interest rate is
solved for (usually via trial-and-error) after arbitrarily setting the PW to zero. When comparing
alternative systems, the one with the highest "IRR" (interest rate) is selected. [3] But here again,
the alternative systems compared must have equal economic lives.
2.4 Estimating Procedure
The estimating procedure used in the Manual consists of five steps: (1) obtaining the facility
parameters and regulatory options for a given facility; (2) roughing out the control system design;
(3) sizing the control system components; (4) estimating the costs of these individual components;
and (5) estimating the costs (capital and annual) of the entire system.
2.4.1 Facility Parameters and Regulatory Options
Obtaining the facility parameters and regulatory options involves not only assembling the
parameters of the air pollution source (i.e., the quantity, temperature, and composition of the
emission stream(s)), but also compiling data for the facility's operation. (Table 2.1 lists examples
of these.) Note that two kinds of facility parameters are identified—intensive and extensive. The
former are simply those variables whose values are independent of quantity or dimensions—i.e.,
the extent of the system. Conversely, extensive parameters encompass all size-dependent
variables, such as the gas volumetric flow rate.
Like the facility parameters, the regulatory options are usually specified by others. These
options are ways to achieve a predetermined emission limit. They range from no control to
maximum control technically achievable. The option provided will depend, firstly, on whether the
2-14
-------
emission source is a stack (point source), a process leak (process fugitives source) or an
unenclosed or partly enclosed area, such as a storage pile (area fugitives source). Stacks are
normally controlled by "add-on" devices. As discussed above, this Manual will deal primarily
with these add-on devices. (However, some of these devices can be used to control process
fugitives in certain cases, such as a fabric filter used in conjunction with a building evacuation
system.) Add-ons are normally used to meet a specified emission level, although in the case of
particulate emissions, they may also be required to meet an opacity level.
2.4.2 Control System Design
Step 2—roughing out the control system design—first involves deciding what kinds of systems
will be priced (a decision that will depend on the pollutants to be controlled, exhaust gas stream
conditions, and other factors), and what auxiliary equipment will be needed. When specifying the
auxiliary equipment, several questions need to be answered:
• What type of hood (if any) will be needed to capture the emissions at the source?
• Will a fan be needed to convey the exhaust through the system?
• Is a cyclone or another pre-cleaner needed to condition the exhaust before it enters the
control device?
• Will the captured pollutants be disposed of or recycled? How will this be done?
• Can the on-site utility capacity (e.g., electricity) accommodate the added requirements of
the control system?
The kinds of auxiliary equipment selected will depend on the answers to these and other site-
specific questions. However, regardless of the source being controlled, each system will likely
contain, along with the control device itself, the following auxiliaries:
• Hood, or other means for capturing the exhaust;
• Ductwork, to convey the exhaust from the source to, through, and from the control
system;
• Fan system (fan, motor, starter, inlet/outlet dampers, etc.), to move the exhaust through
the system;
2-15
-------
Stack, for dispersing the cleaned gas into the atmosphere.
2.4.3 Sizing the Control System
Once the system components have been selected, they must be sized. Sizing is probably the most
critical step, because the assumptions made in this step will more heavily influence the capital
investment than any other. Before discussing how to size equipment, we need to define the term.
parameter. (Table 2.2 lists examples of these parameters. For a full description of the ESP sizing
procedure, see Chapter 6.)
Also listed in Table 2.2 are general parameters which must also be specified before the
purchased cost of the system equipment can be estimated. Note that, unlike the control device
parameters,
these may apply to any kind of control system. These parameters include materials of
construction (which may range from carbon steel to various stainless steels to fiberglass-
reinforced plastic), presence or absence of insulation, and the economic or useful life of the
system. As indicated in Section 2.3.2, this last parameter is required for estimating the annual
capital recovery costs. The lifetime not only varies according to the type of the control system,
but with the severity of the environment in which it is installed. (Representative values for the
system life and the other control device parameters will be presented in those chapters of the
Manual covering them.)
2-16
-------
Table 2.1: Facility Parameters and Regulatory Options
Facility Parameters
• Intensive
- Facility status (new or existing, location)
- Gas characteristics (temperature, pressure, moisture content)
- Pollutant concentrations) and/or particle size distribution
• Extensive
- Facility capacity
- Facility life
- Exhaust gas flow rate
- Pollutant emission rate(s)
Regulatory Options
No control
"Add-on" devices
- Emission limits
- Opacity limits
2-17
-------
Table 2.2: Examples of Typical Control Device Parameters [6]
General
Material of construction: carbon steel
Insulated? Yes
Economic life: 20 yr
Redundancy3: none
Device-Specific
Gas-to-cloth ratio ("critical parameter"): 3.0 to 1
Pressure drop: 6.0 in w.c. (inches water column)
Construction: standard (vs. custom)
Duty: continuous (vs. intermittent)
Filter type: shaker
Bag material: polyester, 16-oz.
aRefers to whether there are any extra equipment items installed (e.g., fans) to function in case the basic item
becomes inoperative, so as to avoid shutting down the entire system.
2-18
-------
QUOTATION
(Note: Company name and address have been deleted.)
Mail drop #12
U. s. EPA
RESEARCH TRIANGLE PARK
DURHAM, NC 27711
ATTN: MR. BILL VATAVUK
QUOTATION NO 85523382
DATE 9-23-85
REFERENCE
VERBAL - BUDGET
Thank you for your inquiry. We are pleased to submit our quotation as follows:
QUANTITY
DESCRIPTION
ITEM PI PREHEATER
MODEL 191-19 SIZE £9 IMPERVITE SHELL & TUBE HEAT EXCHANGER WITH
55 8 SQ FT OF HEAT TRANSFER AREA AND CODE STAMPED
ITEM #2 CONDENSER
MODEL 191-19 SIZE *12 IMPERVITE SHELL & TUBE HEAT EXCHANGER WITH
74 5 SQ FT OF HEAT TRANSFER AREA AND CODE STAMPED
APPROVAL DWG'S 2 -3 WEEKS AFTER RECEIPT OF ORDER
THIS QUOTATION IS IN CONFIRMATION OF OUR PHONE CONVERSATION OF
9/18/85
PRICE
$7,147 00 EA
7,430 00 EA
ESTIMATED SHIPMENT 6 to 8 weeds AFTER RECEIPT OF ORDER RECEIPT OF DRAWING APPROVAL
PRICES are F O B Net 30 Days
Unless otherwise stated these pnces are subject to acceptance within 30 days from date
By-
ANY PlIRCHASF ORDER RhstJlTD*, FROM THIS QUOTATION WH1 BE SUBJECT TO TO CONTRAC1 TERMS AND CONDI! IONS PRINTED ON THF RFVtRSh SIDF OF THIS PAGI-
Figure 2.4: Typical Vendor quotation
2-19
-------
2.4.4 Estimating Total Capital Investment
2.4.4.1 General Considerations
The fourth step is estimating the purchased equipment cost of the control system equipment.
These costs are available from this Manual for the most commonly used add-on control devices
and auxiliary equipment. Each type of equipment is covered in a separate chapter. (See Table of
Contents.)
Most of these costs, in turn, have been based on data obtained from control equipment
vendors. There are scores of these firms, many of whom fabricate and erect a variety of control
systems. [7] They have current price lists of their equipment, usually indexed by model
designation. If the items for which costs are requested are fabricated, "off-the-shelf1 equipment,
then the vendor can provide a written quotation listing their costs, model designations, date of
quotation, estimated shipment date, and other information. (See Figure 2.4 for a sample
quotation.) Moreover, the quote is usually "F.O.B." (free-on-board) the vendor, meaning that no
taxes, freight, or other charges are included. However, if the items are not off-the-shelf, they
must be custom fabricated or, in the case of very large systems, constructed on-site. In such
cases, the vendor can still give quotations—but will likely take much longer to do so and may
even charge for this service, to recoup the labor and overhead expenses of his estimating
department.
As discussed in Section 2.2 in this Manual, the total capital investment is factored from the
purchased equipment cost, which in turn, is the sum of the base equipment cost (control device
plus auxiliaries), freight, instrumentation, and sales tax. The values of these installation factors
depend on the type of the control system installed and are, therefore, listed in the individual
Manual chapters dedicated to them.
The costs of freight, instrumentation, and sales tax are calculated differently from the direct
and indirect installation costs. These items are factored also, but from the base equipment cost
(F.O.B. the vendor(s)). But unlike the installation factors, these factors are essentially equal for
all control systems. Values for these are as follows:
2-20
-------
Cost
Freight
Sales Tax
Instrumentation
Range
0.01 -0.10
0 - 0.08
0.05 - 0.30
Typical
0.05
0.03
0.10
The range in freight costs reflects the distance between the vendor and the site. The lower end is
typical of major U.S. metropolitan areas, while the latter would reflect freight charges to remote
locations such as Alaska and Hawaii. [6] The sales tax factors simply reflect the range of local and
state tax rates currently in effect in the United States. [8]
The range of instrumentation factors is also quite large. For systems requiring only simple
continuous or manual control, the lower factor would apply. However, if the control is
intermittent and/or requires safety backup instrumentation, the higher end of the range would be
applicable.[6] Finally, some "package" control systems (e.g., incinerators covered in Chapter 3)
have built-in controls, whose cost is included in the base equipment cost. In those cases, the
instrumentation factor to use would, of course, be zero.
2.4.4.2 Retrofit Cost Considerations
The installation factors listed elsewhere in the Manual apply primarily to systems installed in new
facilities. These factors must be adjusted whenever a control system is sized for, and installed in
(i.e., "retrofitted") an existing facility. However, because the size and number of auxiliaries are
usually the same in a retrofit situation, the purchased equipment cost of the control system would
probably not be different from the new plant purchased cost. An exception is the ductwork cost,
for in many retrofit situations exceptionally long duct runs are required to tie the control system
into the existing process.
Each retrofit installation is unique; therefore, no general factors can be developed.*
Nonetheless, some general information can be given concerning the kinds of system modifications
one might expect in a retrofit:
1. Auxiliaries. Again, the most important component to consider is the ductwork cost. In
addition, to requiring very long duct runs, some retrofits require extra tees, elbows,
dampers, and other fittings.
Retrofit factors for specific applications (coal-fired boiler controls) have been developed.
See references [9] and [10].
2-21
-------
2. Handling and Erection. Because of a "tight fit," special care may need to be taken when
unloading, transporting, and placing the equipment. This cost could increase significantly
if special means (e.g., helicopters) are needed to get the equipment on roofs or to other
inaccessible places.
3. Piping, Insulation, and Painting. Like ductwork, large amounts of piping may be needed
to tie in the control device to sources of process and cooling water, steam, etc. Of course,
the more piping and ductwork required, the more insulation and painting will be needed.
4. Site Preparation. Unlike the other categories, this cost may actually decrease, for most of
this work would have been done when the original facility was built.
5. Off-Site Facilities. Conceivably, retrofit costs for this category could be the largest. For
example, if the control system requires large amounts of electricity (e.g., a venturi
scrubber), the source's power plant may not be able to service it. In such cases, the source
would have to purchase the additional power from a public utility, expand its power plant,
or build another one. In any case, the cost of electricity supplied to that control system
would likely be higher than if the system were installed in a new source where adequate
provision for its electrical needs would have been made.
6. Engineering. Designing a control system to fit into an existing plant normally requires
extra engineering, especially when the system is exceptionally large, heavy, or utility-
consumptive. For the same reasons, extra supervision may be needed when the
installation work is being done.
7. Lost Production. This cost is incurred whenever a retrofit control system cannot be tied
into the process during normally scheduled maintenance periods. Then, part or all of the
process may have to be temporarily shut down. The net revenue (i.e., gross revenue
minus the direct costs of generating it) lost during this shutdown period is a bonafide
retrofit expense.
8. Contingency. Due to the uncertain nature of many retrofit estimates, the contingency
(i.e., uncertainty) factor in the estimate should be increased. From the above points, it is
apparent that some or most of these installation costs would increase in a retrofit
situation.However,there may be other cases where the retrofitted installation cost would
be less than the cost of installing the system in a new plant. This could occur when one
control device, say an ESP, is being replaced by a more efficient unit—a baghouse, for
example. The ductwork, stack, and other auxiliaries for the ESP might be adequate for
the new system, as perhaps would be the support facilities (power plant, etc.).
2-22
-------
2.4.5 Estimating Annual Costs
Determining the total annual cost is the last step in the estimating procedure. As mentioned in
Section 2.2 the TAG is comprised of three components—direct and indirect annual costs and
recovery credits. Unlike the installation costs, which are factored from the purchased equipment
cost, annual cost items are usually computed from known data on the system size and operating
mode, as well as from the facility and control device parameters.
Following is a more detailed discussion of the items comprising the total annual cost.
(Values/factors for these costs are also given in the chapters for the individual devices.)
2.4.5.1 Raw Materials
Raw materials are generally not required with control systems. Exceptions would be chemicals
used in gas absorbers or venturi scrubbers as absorbents or to neutralize acidic exhaust gases
(e.g., hydrochloric acid). Chemicals may also be required to treat wastewater discharged by
scrubbers or absorbers before releasing it to surface waters. But, these costs are only considered
when a wastewater treatment system is exclusively dedicated to the control system. In most
cases, a pro-rata waste treatment charge is applied. (See also discussion below on Waste
Treatment and Disposal.)
Quantities of chemicals required are calculated via material balances, with an extra 10 to 20%
added for miscellaneous losses. Costs for chemicals are available from the Chemical Marketing
Reporter and similar publications.
2.4.5.2 Operating Labor
The amount of labor required for a system depends on its size, complexity, level of automation,
and operating mode (i.e., batch or continuous). The labor is usually figured on an hours-per-shift
basis. As a rule, though, data showing explicit correlations between the labor requirement and
capacity are hard to obtain. One correlation found in the literature is logarithmic:[11]
2
1
V.
V,
(2.6)
where
L2
= labor requirements for systems 1 and 2
V,, V2 = capacities of systems 1 and 2 (as measured by the gas flow rate, for instance)
y = 0.2 to 0.25 (typically)
2-23
-------
The exponent in Equation 2.6 can vary considerably, however. Conversely, in many cases, the
amount of operator labor required for a system will be approximately the same regardless of its
size.
A certain amount must be added to operating labor to cover supervisory requirements.
Fifteen per cent of the operating labor requirement is representative. [12]
To obtain the annual labor cost, multiply the operating and supervisory labor
requirements(labor-hr/operating-hr) by the respective wage rates (in $/labor-hr) and the system
operating factor (number of hours per year the system is in operation). The wage rates also vary
widely, depending upon the source category, geographical location, etc. These data are tabulated
and periodically updated by the U.S. Department of Labor, Bureau of Labor Statistics, in its
Monthly Labor Review and in other publications. Finally, note that these are base labor rates,
which do not include payroll and plant overhead. (See Overhead discussion below.)
2.4.5.3 Maintenance
Maintenance labor is calculated in the same way as operating labor and is influenced by the same
variables. The maintenance labor rate, however, is normally higher than the operating labor rate,
mainly because more skilled personnel are required. A 10% wage rate premium is typical. [12]
Further, there are expenses for maintenance materials—oil, other lubricants, duct tape, etc.,
and a host of small tools. Costs for these items can be figured individually, but since they are
normally so small, they are usually factored from the maintenance labor. Reference [11] suggests
a factor of 100% of the maintenance labor to cover the maintenance materials cost.
2.4.5.4 Utilities
This cost category covers many different items, ranging from electricity to compressed air. Of
these, only electricity is common to all control devices, where fuel oil and natural gas are
generally used only by incinerators; water and water treatment, by venturi scrubbers, quenchers,
and spray chambers; steam, by carbon adsorbers; and compressed air, by pulse-jet fabric filters.
Techniques and factors for estimating utility costs for specific devices are presented in their
respective sections. However, because nearly every system requires a fan to convey the exhaust
gases to and through it, a general expression for computing the fan electricity cost (Ce) is given
here: [6]
0.746 Q AP s0 p
C = (1 7)
6356H *• }
2-24
-------
where
Q = gas flow rate (actual ft3 /min)
AP = pressure drop through system (inches of water, column) (Values for AP are
given in the chapters covering the equipment items.)
s = specific gravity of gas relative to air (1.000, for all practical purposes)
6 = operating factor (hr/yr)
77 = combined fan and motor efficiency (usually 0.60 to 0.70)
pe = electricity cost ($/kwhr)
A similar expression can be developed for calculating pump motor electricity requirements.
2.4.5.5 Waste Treatment and Disposal
Though often overlooked, there can be a significant cost associated with treating and/or disposing
of waste material captured by a control system that neither can be sold nor recycled to the
process.
Liquid waste streams, such as the effluent from a gas absorber, are usually processed before
being released to surface waters. The type and extent of this processing will, of course, depend
on the characteristics of the effluent. For example, the waste can first be sent to one (or more)
clarifiers, for coagulation and removal of suspended solids. The precipitate from the clarifier is
then conveyed to a rotary filter, where most of the liquid is removed. The resulting filter cake is
then disposed of, via landfilling, for example.
The annual cost of this treatment can be relatively high—$1.00 to $2.00/thousand gallons
treated or more.[13] The (non-hazardous) solid waste disposal costs (via landfilling, for example)
typically would add another $20 to $30/ton disposed of.(14] This, however, would not include
transportation to the disposal site. Disposal of hazardous waste (which may not be landfilled) can
be much more costly—$200 to $300/ton or more. More information on these technologies and
their costs is found in References [13] and [14].
2.4.5.6 Replacement Parts
This cost is computed separately from maintenance, because it is a large expenditure, incurred one
or more times during the useful life of a control system. This category includes such items as
carbon (for carbon absorbers), bags (for fabric filters) and catalyst (for catalytic incinerators),
along with the labor for their installation.
The annual cost of the replacement materials is a function of the initial parts cost, the parts
replacement labor cost, the life of the parts, and the interest rate, as follows:
(2.8)
2-25
-------
where
CRCp = capital recovery cost of replacement parts ($/yr)
Cp = initial cost of replacement parts, including sales taxes and freight ($)
Cp/ = cost of parts-replacement labor ($)
CRFp = capital recovery factor for replacement parts (defined in Section 2.3).
In the Manual methodology, replacement parts are treated the same as any other investment,
in that they are also considered an expenditure that must be amortized over a certain period.
Also, the useful life of the parts (typically 2 to 5 years) is generally less than the useful life of the
rest of the control system.
Replacement-part labor will vary, depending upon the amount of the material, its workability,
accessibility of the control device, and other factors.
2.4.5.7 Overhead
This cost is easy to calculate, but often difficult to comprehend. Much of the confusion
surrounding overhead is due to the many different ways it is computed and to the several costs it
includes, some of which may appear to be duplicative.
There are, generally, two categories of overhead, payroll and plant. Payroll overhead
includes expenses directly associated with operating, supervisory, and maintenance labor, such as:
workmen's compensation, Social Security and pension fund contributions, vacations, group
insurance, and other fringe benefits. Some of these are fixed costs (i.e., they must be paid
regardless of how many hours per year an employee works). Payroll overhead is traditionally
computed as a percentage of the total annual labor cost (operating, supervisory, and
maintenance).
Conversely, plant (or "factory") overhead accounts for expenses not necessarily tied to the
operation and maintenance of the control system, including: plant protection, control
laboratories, employee amenities, plant lighting, parking areas, and landscaping. Some estimators
compute plant overhead by taking a percentage of all labor plus maintenance materials [11], while
others factor it from the total labor costs alone.[2]
For study estimates, it is sufficiently accurate to combine payroll and plant overhead into a
single indirect cost. This is done in this Manual. Also, overhead is factored from the sum of all
labor (operating, supervisory, and maintenance) plus maintenance materials, the approach
recommended in reference [11]. The factors recommended therein range from 50 to 70% [11] An
average value of 60% is used in this Manual.
2-26
-------
2.4.5.8 Property Taxes, Insurance, and Administrative Charges
These three indirect operating costs are factored from the system total capital investment, and
typically comprise 1,1, and 2% of it, respectively. Property taxes and insurance are self-
explanatory. Administrative charges covers sales, research and development, accounting, and
other home office expenses. (It should not be confused with plant overhead, however.) For
simplicity, the three items are usually combined into a single, 4% factor. This value, incidentally,
is standard in all OAQPS cost analyses.
2.4.5.9 Capital Recovery
As discussed in Section 2.3, the annualization method used in the Manual is the equivalent
uniform annualized cost method. Recall that the cornerstone of this method is the capital
recovery factor which, when multiplied by the total capital investment, yields the capital recovery
cost. (See Equation 2.2.)
However, whenever there are parts in the control system that must be replaced before the end
of its useful life, Equation 2.2 must be adjusted, to avoid double-counting.
That is:
= CRFg [TCI- (Cp + Cpl)} (2.9)
where
CRCS = capital recovery cost for control system ($/yr)
TCI = total capital investment for entire system ($)
CRF, = capital recovery factor for control system.
The term (Cp + Cp/) accounts for the cost of those parts (including sales taxes and freight) that
would be replaced during the useful life of the control system and the labor for replacing them.
Clearly, CRFS and CRFp will not be equal unless the control system and replacement part lives are
equal.
2-27
-------
References
[1] Perry, Robert H., and Chilton, Cecil H., Perry's Chemical Engineers' Handbook (Fifth
Edition), McGraw-Hill, New York, 1973, pp. 25-12 to 25-16.
[2] Humphries, K. K. and Katell, S., Basic Cost Engineering, Marcel Dekker, New York,
1981, pp. 17-33.
[3] Grant, E.L., Ireson, W.G., and Leavenworth, R.S., Principles of Engineering Economy,
Sixth Edition, John Wiley & Sons, New York, 1976.
[4] CEIS (OAQPS) Cost Guidance Memo #2: "Implementing OMB Circular A-94: Guidelines
and Discount Rates for Benefit Cost Analysis of Federal Programs" March 18,1993.
[5] Scheraga, Joel D., Draft of "Supplemental Guidelines on Discounting in the Preparation of
Regulatory Impact Analyses", Office of Policy, Planning and Evaluation, U.S. EPA, March
31, 1989.
[6] Vatavuk, W.M. and Neveril, R.B., "Estimating Costs of Air-Pollution Control
Systems—Part I: Parameters for Sizing Systems," Chemical Engineering, October 6, 1980,
pp.165-168.
[7] Pollution Equipment News 1996 Buyer's Guide, Rimbach Publishing, Pittsburgh, 1996.
[8] Internal Revenue Service, Form 1040, 1985.
[9] Shattuck, D.M., et al., Retrofit FGD Cost-Estimating Guidelines. Electric Power Research
Institute, Palo Alto, CA (CS-3696, Research Project 1610-1), October 1984.
[10] Kaplan, N., et al., "Retrofit Costs of SO2 and NOX Control at 200 U.S. Coal-Fired Power
Plants," Pittsburgh Coal Conference, 1990.
[11] Peters, M.S. and Timmerhaus, K.D., Plant Design and Economics for Chemical Engineers
(Third Edition), McGraw-Hill, New York, 1980.
[12] Vatavuk, W.M. and Neveril, R.B., "Estimating Costs of Air Pollution Control
Systems—Part II: Factors for Estimating Capital and Operating Costs," Chemical
Engineering, November 3, 1980, pp. 157-162.
[13] Vatavuk, W.M. and Neveril, R.B., "Estimating Costs of Air-Pollution Control
Systems—Part XVII: Particle Emissions Control," Chemical Engineering, April 2, 1984,
pp. 97-99.
2-28
-------
[14] The RCRA Risk-Cost Analysis Model, U.S. Environmental Protection Agency, Office of
Solid Waste, January 13, 1984.
2-29
-------
Chapter 3
THERMAL and CATALYTIC
INCINERATORS
Donald R. van der Vaart
James J. Spivey
Research Triangle Institute
Research Triangle Park, NC 27709
William M. Vatavuk
Innovative Strategies and Economics Group, OAQPS
Albert H. Wehe
Policy Planning and Standards Group, OAQPS
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
December 1995
Contents
3.1 Introduction 3-4
3.2 Process Description 3-4
3.2.1 Thermal Incinerators 3-7
3.2.1.1 Direct Flame Incinerators 3-9
3-1
-------
3.2.1.2 Recuperative Incinerators 3-9
3.2.1.3 Regenerative Incinerators 3-10
3.2.2 Catalytic Incinerators 3-10
3.2.2.1 Fixed-Bed Catalytic Incinerators 3-13
3.2.2.2 Fluid-Bed Catalytic Incinerators 3-14
3.2.3 Other Considerations: Packaged versus Field-Erected Units, Auxiliary Equipment
3-14
3.2.3.1 Packaged vs. Field-Erected Units 3-14
3.2.3.2 Acid Gas Scrubbers 3-15
3.2.3.3 Heat Exchangers (Preheaters and Other Waste Energy Recovery Units) . 3-15
3.2.3.4 Other Auxiliary Equipment 3-15
3.2.4 Technology Comparison 3-16
3.3 General Treatment of Material and Energy Balances 3-17
3.4 Design Procedures 3-19
3.4.1 Steps Common to Thermal and Catalytic Units 3-19
3.4.2 Steps Specific to Thermal Units 3-24
3.4.3 Steps Specific to Catalytic Units 3-30
3.5 Cost Analysis 3-35
3.5.1 Estimating Total Capital Investment 3-35
3.5.1.1 Equipment Costs, EC 3-35
3.5.1.2 Installation Costs 3-42
3.5.2 Estimating Total Annual Cost 3-42
3.5.2.1 Direct Annual Costs 3-42
3-2
-------
Appendix 3A 3-52
Appendix 3B 3-56
References 3-61
3-3
-------
3.1 Introduction
Incineration, like carbon adsorption, is one of the best known methods of industrial gas waste
disposal. Unlike carbon adsorption, however, incineration is an ultimate disposal method in that
the objectionable combustible compounds in the waste gas are converted rather than collected.
On the other hand, carbon adsorption allows recovery, of organic compounds which may have
more value as chemicals than just their heating value. A major advantage of incineration is that
virtually any gaseous organic stream can be incinerated safely and cleanly, provided proper
engineering design is used.
The particular application of both thermal and catalytic incineration to gaseous waste streams
containing volatile organic compounds (VOCs) is discussed here. The U.S. Environmental
Protection Agency defines any organic compound to be a VOC unless it is specifically determined
not to be a VOC. Indeed, a number of organics (e.g., methane) are specified as not being VOCs.
Although both VOC and non-VOC organic compounds are combustible and are therefore
important in the design of the incinerator, this distinction is important since it is the control of
VOCs that is regulated.
3.2 Process Description
Seldom is the waste stream to be combusted a single organic compound. Rather, it is common
to have a complex mixture of organic compounds. This mixture is typically analyzed for carbon,
hydrogen, oxygen, and other elements; and an empirical formula is developed which represents
the mixture. Combustion of such a mixture of organic compounds containing carbon, hydrogen,
and oxygen is described by the overall exothermic reaction:
CHO.
y
°2 * XC°2 + ~H2° (3.1)
The complete combustion products CO2 and H2O are relatively innocuous, making incineration
an attractive waste disposal method. When chlorinated sulfur-containing compounds are present
in the mixture, the products of complete combustion include the acid components HC1 or SO2,
respectively, in addition to H2O and CO2. In general, these streams would require removal of the
acid components by a scrubber unit, which could greatly affect the cost of the incineration system.
(The sizing and costing of these scrubbers is covered in the "Wet Scrubbers" chapter of this
Manual.)
The heart of an incinerator system is a combustion chamber in which the VOC-containing
waste stream is burned. Since the inlet waste gas stream temperature is generally much lower
than that required for combustion, energy must be supplied to the incinerator to raise the waste
gas temperature. Seldom, however, is the energy released by the combustion of the total organics
3-4
-------
(VOCs and others) in the waste gas stream sufficient to raise its own temperature to the desired
levels, so that auxiliary fuel (e.g., natural gas) must be added.
The combustion of the waste gases may be accomplished in a thermal incinerator or in a
catalytic incinerator. In the catalytic incinerator a catalyst is used to increase the rate of the
combustion reaction, allowing the combustion to occur at lower temperatures. Because the
catalytic process operates at a lower temperature than the thermal process, less auxiliary fuel may
be required in the catalytic process to preheat the waste gas.
Auxiliary fuel requirements may also be decreased, and energy efficiency improved, by
providing heat exchange between selected inlet streams and the effluent stream. The effluent
stream containing the products of combustion, along with any inerts that may have been present
in or added to the inlet streams, can be used to preheat the incoming waste stream, auxiliary air,
or both via a "primary", or recuperative, heat exchanger. It is useful to define the fractional
energy recovery by the preheater, or primary, heat exchanger as follows:
Fractional _ Energy actually recovered flue gas
Energy Maximum energy recoverable if flue gas approaches 3.2
Recovery lowest temperature available to heat exchanger
The energy actually recovered, the numerator of Equation 3.2, is the increase in sensible heat of
the gas, i.e., waste gas or waste gas plus dilution air, being heated. The maximum energy
recoverable would be the decrease in sensible heat of the flue gas, if it were cooled to the
temperature of the incoming waste gas. While this maximum energy recovery would be attained
only with a very large heat exchanger, the concept of fractional energy recovery is useful in
expressing the extent of the improvement in energy efficiency using a "primary" heat exchanger.
Energy efficiency can be further improved by placing another ("secondary") exchanger
downstream of the primary exchanger to recover additional energy from the effluent stream (e.g.,
to generate low pressure process steam or hot water). However, secondary energy recovery is
generally not used, unless there is a specific on site use for the steam or hot water.
The majority of industrial gases that contain VOCs are dilute mixtures of combustible gases
in air. In some applications, such as air oxidation processes, the waste gas stream is very deficient
in oxygen. Depending on the oxygen content of the waste stream, auxiliary air may be required
to combust the total organic content of the waste gas as well as any auxiliary fuel that has been
used.
The concentration of combustible gas in the waste gas stream plays an integral role in the
design and operation of an incinerator. From a cost standpoint, the amount of air in excess of the
stoichiometric amounts should be minimized. For safety reasons, however, any mixture within
the flammability limits, on either the fuel-rich or fuel-lean side of the stoichiometric mixture,
presents an unacceptable fire hazard as a feed stream to the incinerator. The lower, or fuel-lean,
explosive limit (LEL) of a given organic compound defines the minimum concentration of that
3-5
-------
compound in air that can produce more energy than is needed to raise its own temperature to the
ignition point (i.e., ignite). Similarly, the upper, or fuel-rich, explosive limit (UEL) represents the
highest concentration of the organic in air that is ignitable. In the latter case, air is limiting the
reaction. Both the LEL and the UEL are measured at ambient conditions. Empirically, it has
been found that mixtures of hydrocarbons in air at their LEL have a heating value of
approximately 50 Btu/scf.
Since the majority of industrial waste gases that contain VOCs are dilute mixtures of
combustible gases in air, their heating value is low and their oxygen content exceeds that required
to combust both the waste organics (VOCs and others) and the auxiliary fuel. If a waste gas
above 50 percent LEL (about 25 Btu/scf) is encountered, it must be diluted to satisfy fire
insurance regulations. Generally, the streams are brought to below 25 percent LEL, although
concentrations from 25 percent to 50 percent are permitted provided the waste stream is
continuously monitored by LEL monitors. Because air is the usual diluent gas, care must be
taken with preheating the diluted stream so that it remains below about 1200°F. (See discussion
below on preheating.) A table showing LEL, UEL, and heats of combustion for selected organic
compounds is given in Appendix 3A.
The goal of any incineration system is to control the amount of VOCs released to the
environment. Performance of a control device such as an incinerator can be described by a
control efficiency, defined according to the following equation:
Control Eff., % =
Inlet mass rate VOC - Outlet mass rate VOC
Inlet mass rate VOC
100
3.3
It is important to note, however, that incomplete combustion of the inlet VOCs could result in
the formation of other VOCs not originally present. For example, the incomplete oxidation of
dichloroethane can yield vinyl chloride. Both of these compounds are VOCs. The definition
given in Equation 3.3 would still be meaningful, however, as long as the newly formed VOC (e.g.,
vinyl chloride) is detected. This situation necessitates the complete chemical analysis of the inlet
and outlet gas streams to confirm compliance with State and Federal regulations.
Performance of an incinerator can also be measured solely by the outlet VOC concentration,
usually in ppmv.
There are a number of different incinerator designs. These designs can be broadly classified
as thermal systems and catalytic systems. Thermal systems may be direct flame incinerators with
no energy recovery, flame incinerators with a recuperative heat exchanger, or regenerative
systems which operate in a cyclic mode to achieve high energy recovery. Catalytic systems
include fixed-bed (packed-bed or monolith) systems and fluid-bed systems,
3-6
-------
Source
Dilution Air
Stack
Figure 3.1: Thermal Incinerator - General Case
both of which provide for energy recovery. The following sections discuss design aspects of
these systems.
3.2.1 Thermal Incinerators
The heart of the thermal incinerator is a nozzle-stabilized flame maintained by a combination of
auxiliary fuel, waste gas compounds and supplemental air added when necessary (see Figure 3.1).
Upon passing through the flame, the waste gas is heated from its inlet temperature (e.g., 100°F)
to its ignition temperature. The ignition temperature varies for different compounds and is usually
determined empirically. It is the temperature at which the combustion reaction rate (and
consequently the energy production rate) exceeds the rate of heat losses, thereby raising the
temperature of the gases to some higher value. Thus, any organic/air mixture will ignite if its
temperature is raised to a sufficiently high level.
The organic-containing mixture ignites at some temperature between the preheat temperature
and the reaction temperature. That is, ignition, as defined in this section, occurs at some point
during the heating of a waste stream as it passes through the nozzle-stabilized flame regardless
of its concentration. The mixture continues to react as it flows through the combustion chamber.
The required level of VOC control of the waste gas that must be achieved within the time that
it spends in the thermal combustion chamber dictates the reactor temperature. The shorter the
residence time, the higher the reactor temperature must be. The nominal residence time of the
reacting waste gas in the combustion chamber is defined as the combustion chamber volume
divided by the volumetric flow rate of the gas. Most thermal units are designed to provide no
more than 1 second of residence time to the waste gas with typical temperatures of 1,200 to
2,000 °F. Once the unit is designed and built, the residence time is not easily changed, so that the
required reaction temperature becomes a function of the particular gaseous species and the
desired level of control. Table 3.1 illustrates the variability in (theoretical) reactor temperatures
3-7
-------
that is required to destroy 99.99 percent of the inlet mass of various noxious compounds with
excess air for a 1-second reactor residence time [1],
Table 1.1: Theoretical Reactor Temperatures Required for 99.99 Percent Destruction by
Thermal Incineration for a 1-Second Residence Time*
Compound
acrylonitrile
allyl chloride
benzene
chlorobenzene
1 ,2-dichloroethane
methyl chloride
toluene
vinyl chloride
Temperature, °F
1,344
1,276
1,350
1,407
1,368
1,596
1,341
1,369
*Reference [1]
These temperatures cannot be calculated a priori, although incinerator vendors can provide
guidelines based on their extensive experience. In practice, most streams are mixtures of
compounds, thereby further complicating the prediction of this temperature. Other studies
[2,3,4], which are based on actual field test data, show that commercial incinerators should
generally be run at 1600°F with a nominal residence time of 0.75 seconds to ensure 98%
destruction of non-halo genated organics. In some States the reactor temperature and residence
time of the unit are specified rather than attempting to measure actual levels of VOC control. The
selected temperature must be maintained for the full, selected residence time for combustion to
be complete.
These three studies also conclude that mixing is a critical factor in determining the destruction
efficiency. Even though it cannot be measured, mixing is a factor of equal or even greater
importance than other parameters, such as temperature. The most feasible and efficient way to
improve the mixing in an incinerator is to adjust it after start-up. The 98% control level discussed
in the previous paragraph presumes such an adjustment.
Ultimately, once the unit is built, it is the responsibility of the user to operate and maintain the
incinerator to insure compliance with applicable regulations.
3-8
-------
3.2.1.1 Direct Flame Incinerators
Many configurations of thermal incinerators exist with the same goal—to raise the VOC-
containing stream to the desired reaction temperature and hold it there for the given reaction time
to achieve the required destruction efficiency. The simplest example of such a system is the direct
flame incinerator. With reference to Figure 3.1, the direct flame incinerator is comprised only of
the combustion chamber. The waste gas preheater and the secondary energy recovery heat
exchanger are energy recovery devices and are not included as part of the direct flame incinerator.
3.2.1.2 Recuperative Incinerators
Recuperative incinerators have improved energy efficiency as a result of placing heat exchangers
in the hot outlet gas streams. With reference to Figure 3.1, the recuperative incinerator is
comprised of the combustion chamber, the waste gas preheater, and, if appropriate, the
secondary, energy recovery heat exchanger.
Primary Energy Recovery (Preheating Inlet Streams) Considerable fuel savings can be
realized by using the exit (product) gas to preheat the incoming feed stream, combustion air, or
both via a heat exchanger, as shown in Figure 3.1 in the so-called "recuperative" incinerator.
These heat exchangers can recover up to 70% of the energy (enthalpy) in the product gas.
The two types of heat exchangers most commonly used sue plate-to-plate and shell-and-tube.
Plate-to-plate exchangers offer high efficiency energy recovery at lower cost than shell-and-tube
designs. Also, because of their modular configuration, plate-to-plate units can be built to achieve
a variety of efficiencies. But when gas temperatures exceed 1000°F, shell-and-tube exchangers
usually have lower purchase costs than plate-to-plate designs. Moreover, shell-and-tube
exchangers offer better long-term structural reliability than plate-to-plate units.[5] In any case,
because most incinerators installed are packaged units, the design (and cost) of the recuperative
heat exchangers have already been incorporated.
Most heat exchangers are not designed to withstand high temperatures, so that most of the
energy needed to reach ignition is supplied by the combustion of fuel in the combustion chamber
and only moderate preheat temperatures are sought in practice (<1200°F).
Secondary Energy Recovery (Additional Waste Energy Recovery) It should be noted,
however, that at least some of the energy added by auxiliary fuel in the traditional thermal units
(but not recovered in preheating the feed stream) can still be recovered. Additional heat
exchangers can be added to provide process heat in the form of low pressure steam or hot water
for on-site application. Obviously, an in-plant use for such low level energy is needed to realize
these savings.
The need for this higher level of energy recovery will be dependent upon the plant site. The
additional heat exchanger is often provided by the incineration unit vendor. The cost of this
additional heat exchanger may be estimated via standard heat exchanger correlations and should
be added to the costs estimated using the cost correlations in this chapter.
3-9
-------
Zmlmlon Source
Dilution Air
_^
H^-
|
Figure 3.3: Catalytic Incinerator
Until recently, the use of catalytic oxidation for control of gaseous pollutants has really been
restricted to organic compounds containing only carbon, hydrogen and oxygen. Gases containing
compounds with chlorine, sulfur, and other atoms that may deactivate the supported noble metal
catalysts often used for VOC control were not suitably controlled by catalytic oxidation systems.
Catalysts now exist, however, that are tolerant of such compounds. Most of these catalysts are
single or mixed metal oxides, often supported by a mechanically strong carrier such as T-alumina.
Perhaps most of the development of poison-tolerant catalysts has focused on the oxidation of
chlorine-containing VOCs. These compounds are widely used as solvents and degreasers and are
often the subject of concern in VOC control. Catalysts such as chromia/alumina [6,7], cobalt
oxide [8], and copper oxide/manganese oxide [9] have been used for oxidation of gases
containing chlorinated compounds. Platinum-based catalysts are active for oxidation of sulfur
containing VOCs, although they are rapidly deactivated by the presence of chlorine. Compounds
containing atoms such as lead, arsenic, and phosphorous should, in general, be considered poisons
for most oxidation catalysts. Nevertheless, their concentration may be sufficiently low so that the
rate of deactivation and therefore, the catalyst replacement costs, could be low enough to
consider catalytic oxidation.
As was the case for thermal units, it is impossible to predict a priori the temperature and
residence time (i.e., inverse space velocity) needed to obtain a given level of conversion of a VOC
mixture in a catalytic oxidation system. For example, Table 3.2 from Pope et al. [8] shows the
temperature needed for 80% conversion of a number of VOCs over two oxidation catalysts in a
specific reactor design. This table shows that the temperature required for this level of conversion
of different VOCs on a given catalyst and of the same VOC on different catalysts can vary
significantly.
Paniculate matter, including dissolved minerals in aerosols, can rapidly blind the pores of
catalysts and deactivate them over time. Because essentially all the active surface of the catalyst
is contained in relatively small pores, the participate matter need not be large to blind the catalyst.
No general guidelines exist as to particulate concentration and paniculate size that can be
tolerated by catalysts because the pore size and volume of catalysts vary greatly.
3-12
-------
Table 3.2: Catalyst Temperatures Required for Oxidizing 80% of Inlet VOC to CO2, °F for
Two Catalysts
Compound Temperature, °F
CO,O4 Pt - Honeycomb
acrolein
n-butanol
n-propylamine
toluene
n-butyric acid
1,1,1 -trichloroethane
dimethyl sulfide
382
413
460
476
517
661
—
294
440
489
373
451
»661
512
The volumetric gas flow rate and the concentration of combustibles in the gas flowing to the
catalytic incinerator should be constant for optimal operation. Large fluctuations in the flow rate
will cause the conversion of the VOCs to fluctuate also. Changes in the concentration or type
of organics in the gas stream can also affect the overall conversion of the VOC contaminants.
These changes in flow rate, organics concentration, and chemical composition are generally the
result of upsets in the manufacturing process generating the waste stream. It may be
uneconomical to change the process for the sake of making the operation of the catalytic
incinerator feasible. In such cases, thermal incinerators (discussed earlier in this chapter) or
carbon adsorption (discussed in Chapter 4 of this Manual) should be evaluated as alternative
control technology.
The method of contacting the VOC-containing stream with the catalyst serves to distinguish
catalytic incineration systems. Both fixed-bed and fluid-bed systems are used.
3.2.2.1 Fixed-Bed Catalytic Incinerators
Fixed-bed catalytic incinerators may use a monolith catalyst or a packed-bed catalyst. Each of
these is discussed below.
Monolith Catalyst Incinerators The most widespread method of contacting the VOC-
containing stream with the catalyst is the catalyst monolith. In this scheme the catalyst is a porous
solid block containing parallel, non-intersecting channels aligned in the direction of the gas flow.
Monoliths offer the advantages of minimal attrition due to thermal expansion/ contraction during
startup/shutdown and low overall pressure drop.
3-13
-------
3.2.1.3 Regenerative Incinerators
A distinction in thermal incinerators can now be made on the basis of this limitation in the preheat
temperature. The traditional approach to energy recovery in the units (shown schematically in
Figure 3.1) still requires a significant amount of auxiliary fuel to be burned in the combustion
chamber when the waste gas heating values are too low to sustain the desired reaction
temperature at the moderate preheat temperature employed. Additional savings can, under these
conditions, be realized in units with more complete transfer of exit-stream energy. This is the
concept behind the so-called excess-enthalpy or regenerable burner systems. These systems use
direct contact heat exchangers constructed of a ceramic material that can tolerate the high
temperatures needed to achieve ignition of the waste stream.
The operation of the regenerative system is illustrated in Figure 3.2. The inlet gas first passes
through a hot ceramic bed thereby heating the stream (and cooling the bed) to its ignition
temperature. If the desired temperature is not attainable, a small amount of auxiliary fuel is added
in the combustion chamber. The hot gases then react (releasing energy) in the combustion
chamber and while passing through another ceramic bed. thereby heating it to the combustion
chamber outlet temperature. The process flows are then switched, now feeding the inlet stream
to the hot bed. This cyclic process affords very high energy recovery (up to 95%).
The higher capital costs associated with these high-performance heat exchangers and
combustion chambers may be offset by the increased auxiliary fuel savings to make such a system
economical. The costs of these regenerative units will be given separately in the cost correlations
presented in Section 3.5. Regenerative incinerators are not packaged units but are field-erected
only. Accordingly, the costs given in Section 3.5 for regenerative units are for field-erected units.
3.2.2 Catalytic Incinerators
Catalytic incinerators employ a bed of active material (catalyst) that facilitates the overall
combustion reaction given in Equation 3.1. The catalyst has the effect of increasing the
reaction rate, enabling conversion at lower reaction temperatures than in thermal incinerator
units. Nevertheless, the waste stream must be preheated to a temperature sufficiently high
(usually from 300 to 900 °F) to initiate the oxidation reactions. The waste stream is preheated
either directly in a preheater combustion chamber or indirectly by heat exchange with the
incinerator's effluent or other process heat or both (Figure 3.3). The preheated gas stream is
then passed over the catalyst bed. The chemical reaction (combustion) between the oxygen in
the gas stream and the gaseous pollutants takes place at the catalyst surface. Catalytic
incineration can, in principle, be used to destroy essentially any oxidizable compound in an air
stream. However, there are practical limits to the types of compounds that can be oxidized
due to the poisoning effect some species have on the catalyst. These limits are described
below. In addition, most configurations require a low heating value of the inlet gas and a
paniculate content which is less than some small value.
3-10
-------
Atr
Mod. A
Mod* B
Ant
Figure 3.2: Regenerable-type Thermal Incinerator
3-11
-------
Packed-Bed Catalytic Incinerators A second contacting scheme is a simple packed-bed in
which catalyst particles are supported either in a tube or in shallow trays through which the gases
pass. The first scheme is not in widespread use due to its inherently high pressure drop,
compared to a monolith, and the breaking of catalyst particles due to thermal expansion when the
confined catalyst bed is heated/cooled during startup/shutdown. However, the tray type
arrangement, where the catalyst is pelletized is used by several industries (e.g., heat-set web-offset
printing). Pelletized catalyst is advantageous where large amounts of such contaminants as
phosphorous or silicon compounds are present.[10]
3.2.2.2 Fluid-Bed Catalytic Incinerators
A third contacting pattern between the gas and catalyst is a fluid-bed. Fluid-beds have the
advantage of very high mass transfer rates, although the overall pressure drop is somewhat higher
than for a monolith. An additional advantage of fluid-beds is a high bed-side heat transfer as
compared to a normal gas heat transfer coefficient. This higher heat transfer rate to heat transfer
tubes immersed in the bed allows higher heat release rates per unit volume of gas processed and
therefore may allow waste gas with higher heating values to be processed without exceeding
maximum permissible temperatures in the catalyst bed. In these reactors the gas phase
temperature rise from gas inlet to gas outlet is low, depending on the extent of heat transfer
through imbedded heat transfer surfaces. The catalyst temperatures depend on the rate of
reaction occurring at the catalyst surface and the rate of heat exchange between the catalyst and
imbedded heat transfer surfaces.
As a general rule, fluid-bed systems are more tolerant of particulates in the gas stream than
either fixed-bed or monolithic catalysts. This is due to the constant abrasion of the fluidized
catalyst pellets, which helps remove these particulates from the exterior of the catalysts in a
continuous manner.
A disadvantage of a fluid-bed is the gradual loss of catalyst by attrition. Attrition-resistant
catalysts have been developed to overcome this disadvantage. [11]
3.2.3 Other Considerations: Packaged versus Field-Erected Units,
Auxiliary Equipment
3.2.3.1 Packaged vs. Field-Erected Units
With the exception of regenerative incinerators, the equipment cost correlations included in this
chapter are for packaged units only. They are not valid for field-erected units. For regenerative
incinerators, the correlations are valid for field-erected units only. Packaged units are units that
have been shop fabricated and contain all elements necessary for operation, except for connection
to facilities at the site, e.g., utilities. The elements include the combustion chamber, preheater,
instrumentation, fan, and the necessary structural steel, piping, and electrical equipment. This
equipment is assembled and mounted on a "skid" to facilitate installation on a foundation at the
plant site. Tie-in to the local emission source is not part of the packaged unit. Units are usually
3-14
-------
sized to handle flow rates of <20,000 scfin, but can be built to accommodate flow rates up to
50,000 scfin. The cost correlations in this chapter are valid to 50,000 scfin for packaged units,
except for fluid-bed units which are valid to 25,000 scfin.
Conversely, field-erected units may be built to any desired size. The combustion chamber,
preheater, and other equipment items are designed and fabricated individually, and assembled at
the site. However, both the equipment and installation costs of field-erected units are typically
higher than those for equivalent-sized packaged units because the factors that improve efficiency
of shop-fabrication, such as uniform working environment, availability of tools and equipment,
and more efficient work scheduling, are generally not available in the field.
3.2.3.2 Acid Gas Scrubbers
The final outlet stream of any incineration system may contain certain pollutants that must be
removed. The combustion of sulfur-containing compounds results in SO2, while chlorinated
compounds yield C12 and HC1 in the product stream. These acid gases must be removed from the
gas stream if they are present at significant concentrations (regulations for limits on these gases
vary from state to state). This removal can be effected in, for instance, a packed-bed gas absorber
(vertical scrubber) in which the flue gas is contacted with a caustic scrubbing liquid. For fluid-bed
catalytic reactors, venturi scrubbers are often used because they provide for particulate removal
as well as acid gas scrubbing. In most cases adding a scrubber or absorber significantly increases
the cost of the incineration unit, sometimes by a factor of two. Costing of absorbers is discussed
in the "Gas Absorbers" chapter (Chapter 9) of this Manual.
If chlorinated VOCs are present in the waste gas, heat exchangers may require special
materials of construction. This added expense is not included in the costing procedures outlined
in this chapter.
3.2.3.3 Heat Exchangers (Preheaters and Other Waste Energy Recovery Units)
For the thermal and catalytic units having some degree of energy recovery, the cost of the primary
heat exchanger is included in the cost, and its design is usually done by the incineration unit
vendor. The cost correlations presented in this chapter include units both with and without
energy recovery. Secondary energy recovery, if desired, requires an additional heat exchanger,
which is also often provided by the incineration unit vendor. Costing procedures for secondary
energy recovery are not included in this chapter.
3.2.3.4 Other Auxiliary Equipment
Additional auxiliary equipment such as hoods, ductwork, precoolers, cyclones, fans, motors, and
stacks are addressed separately in other chapters of this Manual.
3-15
-------
3.2.4 Technology Comparison
Both the thermal and catalytic incineration systems are designed to provide VOC control through
combustion at a level in compliance with applicable state and federal requirements. Given the
wide range of options available, however, it is obvious that not every incinerator will fulfill these
requirements at the same cost. This section presents a first step toward deciding how best to deal
with VOC emission abatement using incinerators considering some qualitative factors pertinent
to the types of incinerators described in this chapter. It is the intent of the remainder of Chapter
3 to provide a method by which the cost of VOC control for a particular application can be
calculated.
Table 1.3: Principal VOC Incinerator Technologies
Thermal Systems
Direct Flame Incinerator
Recuperative Incinerator (Direct Flame with Recuperative Heat
Exchanger)
Regenerative Incinerator Operating in a Cyclic Mode
Catalytic Systems
Fixed-Bed
- Monolith
Packed-Bed
Fluid-Bed
A summary of the principal types of incinerators is presented in Table 3.3. From the earlier
discussions, the following factors relating to the presence of contaminants should be considered
by potential users [12]:
3-16
-------
The fouling of the catalyst in a catalytic system is a possibility. Poisons to the system
include heavy metals, phosphorous, sulfur and most halogens, although catalysts have
been developed that are chlorine resistant.
The possibility of process upsets that could release any of the above poisons or cause
fluctuations in the heating value to the incinerator would favor a thermal system.
Except for the No.2 grade, fuel oil should not be considered as auxiliary fuel to a catalytic
system due to the sulfur and vanadium it may contain. [10]
All of the above factors would serve to increase the operating expense of a catalytic unit through
replacement costs of the catalyst. An additional factor relates to relative energy efficiency of the
various types of incinerators:
Thermal units generally require more auxiliary fuel than catalytic units and operate at
temperatures that are roughly 1000°F higher. This difference in fuel requirement
increases as the heating value of the waste stream decreases.
In general, a trade-off exists between the higher capital costs of catalytic incinerators and the
higher operating costs of thermal incinerators. This difference will be illustrated by a design
example presented in Section 3.4 which treats both technologies.
3.3 General Treatment of Material and Energy Balances
In the sizing and costing of the incinerator and the calculation of the auxiliary fuel requirements,
it is necessary to make material and energy balances around the entire incinerator unit and around
certain parts of the unit, such as the combustion chamber or the preheater. This section presents
a general approach to making these balances.
These balances are based on the law of conservation of mass and energy. They can be stated
in general equation form as
In - Out + Generation = Accumulation (3.4)
Because the incineration process is a steady-state process, the accumulation term is zero and the
equation becomes
In - Out + Generation = 0
3-17
-------
For mass balances it is useful to restrict the balances to be made on the mass of each atomic
species so that for mass balances the generation term becomes zero. However, because the
combustion reaction liberates energy, the energy balances around equipment where combustion
takes place would include a generation term. Thus, the simplified equations are
In - Out = 0 , for steady-state mass balances (3.5)
In - Out + Generation = 0 , for steady-state energy balances (3.6)
For the incineration process the two terms In and Out are generally mass terms (for a mass
balance) of the form,
pQ
where
p = density (mass per unit volume)
Q = volumetric flow rate (volume per unit time)
or sensible heat terms (for an energy balance) of the form,
pQCp(T-Tref)
where
Cp - heat capacity
T = temperature
The reference temperature, Tref, is often taken to be zero or the temperature of a
convenient stream, e.g., the inlet gas stream, in whatever units T is in, so the T^term may not
appear in the equations. When the reference temperature is taken as zero, the sensible heat terms
become
PQCPT.
Energy losses, Hi5 are also part of the Out term and, for the incinerator process, are taken here
to be 10% of the total energy input to the incinerator.
For the incineration process, the generation term for energy balances accounts for the
energy released through the combustion reactions. This term is generally of the form
pQ(-Ahc)
where
(-A/zc) = heat of combustion.
3-18
-------
3.4 Design Procedures
The following procedure is designed to provide parameters for use in developing a study cost
estimate (accuracy ± 30%). The principal parameters of interest are
flue gas flow rate, upon which all the equipment cost correlations are based.
auxiliary fuel requirement, which is important in estimating annual operating costs.
For applications which involve control of waste gas streams that are dilute mixtures of VOCs
in air (>20% oxygen in the waste gas stream), the flue gas flow rate is greater than the inlet waste
gas flow rate by the amount of auxiliary iuel and the increase in the moles of gas as a result of the
combustion reaction. Because these two factors usually cause only small increases in flow rate,
a number of simplifying assumptions can be made in the design calculations. For applications
where diluent air must be used to adjust the combustible concentration in the waste gas to 25%
LEL and where auxiliary fuel and auxiliary combustion air are needed, more complete mass and
energy balances must be made.
The design procedure illustrated below is for waste gas streams that are dilute mixtures of
VOCs in air (>20% oxygen in the waste gas stream). In this discussion the design procedure will
be illustrated by a sample problem that will be solved step-by-step.
3.4.1 Steps Common to Thermal and Catalytic Units
Step 1 - Establish design specifications The first step in the design procedure is to determine
the specifications of the incinerator and the waste gas to be processed. The following parameters
of the waste gas stream at the emission source must be available:
• Volumetric flow rate, scfin—Standard conditions are normally 77 °F and 1 atm. pressure
• Temperature
• Oxygen content
• Chemical composition of the combustibles
• Inerts content
3-19
-------
Table 3.4: Specification of Sample Problem
Variable Value
•
Preheater Inlet Waste Gas Vol Flow Rate, Q ^fa 20,000
Preheater Inlet Waste Gas Temp., T °F 100
Composition
Benzene Content, ppmv 1000
Methyl Chloride Content, ppmv 1000
Air Content Balance
Paniculate Content Negligible
Moisture Content Negligible
Desired Control Efficiency, % 98
Desired Percent Energy Recovery, HR% 70
Heating value — In cases the heating value may act as a surrogate for the chemical
composition of the combustibles. This is particularly true for dilute mixtures of
combustibles in air.
Particulate content — The paniculate content is important if catalytic incinerators are to
be coated. An upstream filter may suffice if particulate content is too high. Fluid-bed
catalytic incinerators can tolerate higher particulate contents than fixed-bed catalytic
incinerators.
The following parameters must be specified for the incinerator:
• Desired control efficiency — This efficiency should be based on requirements dictated by
relevant state and federal regulations.
• Combustion chamber outlet temperature— This temperature may also be based on
requirements of a regulation or on recommendations developed during regulatory
development.
• Desired percent energy recovery — The desired percent energy recovery should be the
result of a process optimization in which costs of incinerators with several different levels
of energy recovery are estimated and the minimum cost design selected. The tradeoff is
between the capital cost of the energy recovery equipment and the operating (fuel) cost.
Specifications for the sample problem are given in Table 3.4.
Step 2 - Verify that the oxygen content of the waste gas exceeds 20% There must be
sufficient oxygen in the waste gas to support the combustion of the waste organics (including
3-20
-------
VOCs) and the auxiliary iuel, if auxiliary fuel is needed. It may be necessary to add auxiliary air
if the oxygen content is less than about 20%. This example is based on streams that contain
>20% oxygen, as shown below:
Air Content, Vol. % = 100.0 - x 100 - x 100
106 106 3. .7
= 99.8%
Oxygen Content, % = Air Content x 0.209
- 20.86%
Step 3 - Calculate the LEL and the Percent of the LEL of the gas mixture Note: If the waste
stream contains a significant amount of inerts in addition to the nitrogen associated with the
oxygen in air, the calculation of LEL (and UEL) loses meaning since the LEL (and UEL) is
measured in mixtures of organic with air only. A complete chemical analysis is necessary to
complete the design procedure in such a case.
The example chosen here is typical, in that there is more than one VOC component in the gas
stream. An approximate method to calculate the LEL of a mixture of compounds, LELmtr, is
given by Grelecki [13] as
LEL_.,. - / , ., n
LELj
where
x, — volume fraction of combustible component /
LEL, = lower explosive limits of combustible component y (ppmv)
n = number of combustible components in mixture
For the example case,
n
-6
, = (1,000 + 1,000) x 10-
1=1
= 2,000 x 1Q-6
From standard references [14] or from Appendix 3 A,
LELBz = 14,000ppnv fir benzene
3-21
-------
LELMC = 82,5000 ppmv for methyl chloride
LEL
1000 1000
2,000 x 14,000 2,000 x 82,500
= 23,938 ppmv
-i
3.11
- total combustible cone, in mixture
mix
o/ TFT - , x
3.12
2'000 x 100 - 8.4% 3 13
23,938
The percent LEL of the mixture is therefore 8.4%. Because this is well below 25%, no
dilution air is needed in this example. If the mixture had been above 25% LEL, sufficient dilution
air would have been needed to bring the concentration of the mixture to less than 25% to satisfy
fire insurance regulations.
Step 4 - Calculate the volumetric heat of combustion of the waste gas streams, (-Ahcw),
Btu/scf The energy content of the gas stream, expressed in terms of the heat of combustion, is
calculated as follows:
3.14
where
(-A/i ) = heat of combustion of the waste stream (Btu/scf)
cw
(-A/7 ) = volumetric heat of combustion of component i at 25 °C (Btu/scf)
ci
xt = volume fraction of component i in the waste gas
n = number of combustible components in the waste gas
The heat of combustion that should be used in these calculations is the "lower" heat of
combustion, i.e., with gaseous water, rather than liquid water, as a reaction product since water
leaves the incinerator in the vapor state. From Appendix 3 A or standard references [14,15] with
3-22
-------
appropriate conversion of units, the volumetric heat of combustion at 25 °C for the two
components is calculated to be as follows:
(-A/I, ) = 3,475 Btu/scf for benzene
B:
(-A/i ) = 705 Btu/scf for methyl chloride
cuc J
The compositions specified earlier as ppmv are converted to volume fractions as follows:
Xg. = 1,000 ppmv x 10 6 = 10'3 for benzene
XMC = 1,000 ppmv x 1Q-6 = 10"3 for methyl chloride
Using these values of heat of combustion and composition, the heat of combustion of the
waste gas stream per standard cubic foot of incoming gas is
(-A/I, ) = (3,475) (10-3) + (705) (lO'3)
= 4.18 Btu/scf
Assuming the waste gas is principally air, with a molecular weight of 28.97 and a
corresponding density of 0.0739 Ib/scf, the heat of combustion per pound of incoming waste gas
is
(-A/7c ) = 56.6 Btu/lb
The negative heat of combustion, by convention, denotes an exothermic reaction. Also by
convention, if one refers to heat of reaction rather than heat of combustion, then a positive value
denotes an exothermic reaction.
Empirically, it has been found that 50 Btu/scf roughly corresponds to the LEL of organic/air
mixtures. Insurance codes require a value below 25% LEL, which corresponds to about 13
Btu/scf. However, if LEL sensors and monitors are installed, one can incinerate a waste gas with
a combustible organic content between 25 and 50% LEL, which corresponds to 13 to 25 Btu/scf.
For catalytic applications the heat of combustion must normally be less than 10 Btu/scf (for
VOCs in air) to avoid excessively high temperatures in the catalyst bed. This is, of course, only
an approximate guideline and may vary from system to system.
After Step 4, determination of the (-A^j design procedure for thermal and catalytic
incinerators is discussed separately, beginning with Step 5 for each type of incinerator.
3-23
-------
3.4.2 Steps Specific to Thermal Units
Figure 3.1 shows a generic thermal incinerator with the appropriate streams labeled.
Step 5t - Establish the temperature at which the incinerator will operate As mentioned in
Section 3.2.1, both the reactor temperature and residence time of the waste gas in the reactor
determine the level of VOC destruction. In general, state and local regulations specify the
required level of destruction that the customer must meet. In this example a destruction efficiency
of 98 percent is specified. Studies by Mascone [2,3,4] show that this destruction efficiency can
be met in a thermal incinerator operated at a temperature, Tf,, of 1,600 °F and a residence time
of 0.75 second. (Note: This higher efficiency level is the minimum achievable by any new properly
designed and operated incinerator. Many incinerators can achieve destruction efficiencies of 99%
or higher.)
Step 6t - Calculate the waste gas temperature at the exit of the preheater The extent of
the heat exchange to be carried out in the preheater is the result of a technical and economic
optimization procedure that is not illustrated in this example. As the VOC stream temperature
leaving the heat exchanger, TWQJ increases, the auxiliary fuel requirement decreases, but at the
expense of a larger heat exchanger. However, there are several important limits on Tw _ First,
TWO must not be close to the ignition temperature of the organic-containing gas to prevent
damaging temperature excursions inside the heat exchanger should the gas ignite. Second, for
gases containing halogens, sulfur, and phosphorous (or other acid-forming atoms), the flue
gas temperature after the heat exchanger, Tfo> must not drop below the acid dew point. Both
limitations limit the amount of heat exchange and thus the maximum value of Tw jne
calculation of the acid dew point is not simple. It is recommended that vendor guidance be
sought to ensure that the dew point is not reached. Condensation of acid gases will result in
corrosion of many of the metals used in heat exchangers. As an example, fuel sulfur contents
of 1 to 2 percent can give acid dew points of about 200 to 270°F. Increasing the sulfur
content to 4 percent can raise the dew to about 290°F. Chlorine and phosphorous have a
much smaller effect on acid dew elevation.
With the following assumptions, one can estimate TWQ using equation 3.2, the definition of
fractional energy, recovery for a heat exchanger.
• The fractional energy recovery is specified.
• The amount of auxiliary fuel, Qaf, and auxiliary combustion air, Qa, are small relative to
the waste gas, Qw, so that the mass flow rates of gases, pwQw and p,Qf, on both sides of
the preheater are approximately the same, or
3-24
-------
The heat capacities of the gases on both sides of the preheater are approximately the
same, regardless of composition. This is true for waste streams which are dilute mixtures
of organics in air, the properties of the streams changing only slightly on combustion.
The mean heat capacities above the reference temperature of the gases on both sides of
the preheater are approximately the same regardless of temperature.
With these assumptions, the equation for fractional energy recovery for a heat exchanger becomes
T - T
W W
Fractional Energy Recovery = — 3 16
Tfi ~ T",
For this example with a fractional energy recovery of 0.70, an incinerator operating temperature,
Tfp of 1600°F, and a waste gas inlet temperature, Tw, of 100°F, the waste gas temperature at the
end of the preheater becomes
7; - 1,150°F
The temperature of the exhaust gas, Tfoj can be determined by an energy balance on the
preheater, which, with the same assumptions as used in deriving Equation 3.16 regarding the mass
flow rates and average heat capacities of the gases involved, results in the following equation:
Tf, ~
i.e., the temperature rise in the waste gas is approximately equal to the temperature decrease in
the flue gas with which it is exchanged. For this example, this results in
Tf = 550°F
Jo
This value of Tfo should be well above the acid dew point of the flue gas stream.
It should be remembered that TWo should be well below the ignition temperature of the
VOC stream to prevent unwanted temperature excursions in the preheater. This must be
verified even if the stream is well below the LEL because flammability limits can be expanded
3-25
-------
by raising the reactant stream temperature. A sufficiently high preheat temperature, Tw
could initiate reaction (with heat release) in the preheater. This would ordinarily be
detrimental to the materials of construction in the heat exchanger. The one exception is the
thermal incinerator of the regenerable type described in Section 3.2. The 95-percent energy
recovery, obtainable in regenerable systems would result in this example in a Tw of \ ^525 °F.
The significant reaction rate that would occur at this temperature in the ceramic packing of
the heat exchanger/reactor is by design.
Step 7t - Calculate the auxiliary fuel requirement, Qa/ Auxiliary fuel will almost invariably
be needed for startup of the unit. However, at steady state, if the energy released by combustion
of the organics present in the waste stream is sufficient to maintain the reactor temperature
(1,600°F in the example), only a small amount of auxiliary fuel (< 5% of the total energy input)
is needed to stabilize the flame. In most cases, however, more fuel than just this stabilizing fuel
will be required to maintain the reactor temperature.
With the following assumptions, one can estimate Q^-using a mass and energy balance around
the combustion chamber and following the principles discussed in Section 3.3, with reference to
Figure 3.1.
• The reference temperature, Tr^ is taken as the inlet temperature of the auxiliary fuel, 1af.
• No auxiliary air, Qa, is required.
• Energy losses, HL, are assumed to be 10% of the total energy input to the incinerator
above ambient conditions. [16,17] Thus, if the reference temperature is near ambient
conditions,
HL = O.lpjOjC,. (Tft - Tre) 3j7
The heat capacities of the waste gases entering and leaving the combustion chamber are
approximately the same, regardless of composition. This is true for waste streams which
are dilute mixtures of organics in air, the properties of the streams changing only slightly
on combustion.
3-26
-------
• The mean heat capacities above the reference temperature of the waste gases entering and
leaving the combustion chamber are approximately the same regardless of temperature.
Thus the mean heat capacity for the waste gas stream entering or leaving the combustion
chamber should be evaluated at the average of TW{> ^4 TJ, For & this assumption
introduces an error of, at most, 5% over the temperatures of interest.
With these assumptions, the mass and energy balance around the combustion chamber reduces
to the following equation:
Input data for this equation are summarized below:
• The waste stream is essentially air so that
w = 0.0739 Ib/scf, air at 77°F, 1 atm.
Table 3.5: Summary of Example Problem Variable Valuation
(-Ahc), waste gas = 56.6 Btu/lb
(-Ahc), auxiliary fuel = 21,502 Btu/lb
*Not applicable.
**Not used because reference temperature is taken equal to auxiliary fuel temperature.
3-18
Stream
IN - Sensible heat
Auxiliary Air
Auxiliary Fuel
Waste Gas
OUT - Sensible Heat
Waste Stream
Pfi
Subscript, j Ib/scf
a na*
af 0.0408
w0 0.0739
/ 0.0739
scfm
na*
167
20,000
20,167
Btu/lb °F
na*
**
0.255
0.255
TP
°F
na*
77
1,150
1,600
3-27
-------
cPmair = 0.255Btu/lb °F, the mean heat capacity of air between 77°F and 1,375°F (the
average temperature of the waste gas entering and leaving the
combustion chamber)
• Other input data to Equation 3.18 include:
Qy>o = QV: = 20,000 scfm
(-A/zc ) = 21,502 Btu/lb, for methane
Taf = Tref = 77°F, assume ambient conditions
pa/ = 0.0408 lb/ft3, methane at 77 °F ,1 atm.
Tfl = 1,600°F, Step 5t
Tw = 1,150°F, Step 6t
(-AAC ) = 56.6 Btu/lb. Step4
substituting the above values into Equation 3.18 results in:
Qaf- 167 scfhi
The values of the parameters in the energy balance are summarized in Table 3.5.
It is instructive to examine the magnitude of the various terms in the energy balance around
the combustor for the sample problem. This is done in Table 3.6. The energy balance shown
does not quite add to zero due to round-off-error and simplifying assumptions. Table 3.6 shows
that the largest inlet term is the sensible heat of the incoming waste gas. The heat of combustion
of the organics contained in the waste gas stream is somewhat smaller than that of the auxiliary
methane because of the relatively small amount of organics in the waste gas stream. The largest
term in the outlet stream is the sensible heat of the outgoing waste stream. The overall energy
losses are based on an assumption, but are relatively small. Because the sensible heat contents
of the entering and leaving waste stream are so large, it is apparent that energy recovery is an
important factor in achieving energy efficiency. In fact, with zero energy recovery in the sample
problem, the auxiliary fuel requirements would be 605 scfrn, about four times the energy
requirements based on 70% energy recovery.
Step 8t - Verify that the auxiliary fuel requirement is sufficient to stabilize the burner flame
Only a small amount of auxiliary fuel (< 5% of the total energy input) is needed to stabilize the
burner flame. In general, more fuel than just this stabilizing fuel will be required to maintain the
reactor temperature. It is wise to verify that the auxiliary fuel requirement calculated in Step 7t
is sufficient for stabilization. If it is insufficient, then a minimum amount of auxiliary fuel must
be used, and the amount of energy recovery, specified earlier must be reduced to avoid exceeding
the specified temperature at which the incinerator will operate (Step 5t).
3-28
-------
Table 3.6: Terms in Energy Balance Around Combustor—Example Problem
Value,
Stream Subscript, j Btu/min
IN - Sensible Heat, p,Q,Cw (T, - 7^)
Auxiliary Air a 0
Waste Gas w0 404,403
OUT - Sensible Heat, p,Q,Cw (T, - Tref)
Waste Stream / 578,796
OUT - Losses
10% of total energy input 57,880
GENERATION -
Heat of Combustion, p^Q, (-A/i9)
Waste Gas w0 83,655
Auxiliary Fuel af 146,506
This check is made by calculating 5% of the total energy input to the incinerator and
comparing it with the auxiliary fuel energy input. The total energy input is given as follows:
Total Energy Input - VfQfpmfi (Tfl - Tre) 3.19
Auxiliary Fuel Energy Input = pa/ Qaf (~A/»CJ 3.20
The auxiliary fuel used in the design, Q^ should be the larger of 5% of the total energy input
(28,900 Btu/min.) and the auxiliary fuel energy input (146,500 Btu/min.). The auxiliary fuel used
easily meets this criterion.
Step 9t - Calculate the total volumetric flow rate of gas through the incinerator, Qfl The
total volumetric flow rate of gas leaving the incinerator is referred to as the flue gas flow rate, Qfr
and is the gas rate on which the incinerator sizing and cost correlations are based. The flue gas
flow rate measured at the standard conditions of 77 °F and 1 atmosphere, where the increase in
volumetric throughput due to an increase in the number of moles of gas as a result of combustion
is neglected, is the sum of the inlet streams to the incinerator.
Qfi - QWo - Qa - Qaf
= 20,000 +0 +167
= 20,167 scfin
This result conforms with the assumptions stated in Step 6t, i.e., the mass (and volume) flow rates
on both sides of the preheater are approximately equal. Finally, it must be emphasized that steps
5t to 9t apply to thermal recuperative incinerators, only. To calculate the auxiliary- fael
3-29
-------
requirements for other types of thermal incinerators (e.g., regenerative), a different procedure
must be used. (See Appendix 3B.)
3.4.3 Steps Specific to Catalytic Units
Figure 3.3 shows a generic catalytic incinerator with the appropriate streams labeled. The
approach used in the calculations on the catalytic incinerator is somewhat different than that used
in the thermal incinerator. This difference arises because of additional constraints which are
placed on the catalytic incinerator. These constraints are as follows:
• The desired catalyst bed outlet temperature is typically 700 to 900°F. The maximum
temperature to which the catalyst bed can be exposed continuously is limited to about
1,200°F. Therefore, the combustible content of the waste gas is limited, and the amount
of heat exchange that occurs in the primary heat exchanger may be limited.
• The inlet temperature to the catalyst bed itself must be above the catalytic ignition
temperature required to give the desired destruction efficiency in the incinerator.
Therefore, the combustible content of the waste gas is further limited to that which, when
combusted, will raise the temperature in the catalyst bed no more than the AT between
the required reactor bed inlet temperature, and the desired reactor bed outlet temperature.
• Auxiliary fuel, in combination with the preheat from the primary heat exchanger, is used
to preheat the waste gas to the reactor inlet temperature. A minimum amount of auxiliary
fuel (< 5% of the total energy input) must be used to stabilize the flame in the preheat
combustion chamber. This has the effect of further limiting the combustible content of
the waste gas stream and the amount of heat exchange permissible in the primary heat
exchanger.
The steps outlined below represent one approach to recognizing these constraints and
incorporating them into the calculation procedures.
Step 5c - Establish the desired outlet temperature of the catalyst bed, Tfl The energy
released by the oxidation of the VOCs in the catalyst bed will raise the temperature of the gases
by an amount, AT, as the gases pass through the catalyst bed. An outlet temperature from the
catalyst, and thus from the reactor, must be specified that will ensure the desired level of
destruction of the VOC stream. As in thermal incinerators, this value varies from compound to
compound and also varies from catalyst to catalyst. Final design of the incinerator should be done
by firms with experience in incinerator design. Guidelines given by Combustion Engineering [12]
indicate that values from 300 to 900°F result in destruction efficiencies between 90 and 95
percent. To prevent deactivation of the catalyst a maximum bed temperature of 1,200°F should
not be exceeded. In the example problem the catalyst outlet temperature, Tfi, is selected to be
900°F.
Step 6c - Calculate the waste gas temperature at the exit of the preheater (primary) heat
exchanger The waste gas temperature at the exit of the primary heat exchanger is estimated in
3-30
-------
the same manner as for the thermal incinerator. The equation for fractional energy recovery
Equation 3.16, is used, with the same assumptions as used for the thermal incinerator. For the
example problem with a fractional energy recovery of 0.70, a catalyst bed outlet temperature, Tfl,
of 900°F, and a waste gas inlet temperature, TWjs of 100°F, the gas temperature at the exit of the
preheater becomes
T = 660°F
The same considerations regarding the closeness of the temperature of the exhaust gas, T/a,
to its dew point apply to the catalytic incinerator as they did to the thermal incinerator.
Step 7c - Calculate the auxiliary fuel requirement, Qa/ The auxiliary fuel requirement, Q^ is
calculated by making mass and energy balances around the preheater combustion chamber and
the catalyst chamber. The auxiliary fuel requirement calculated in this manner must be checked
to insure that it falls within the constraints imposed by design considerations of the catalytic
incinerator. These constraints are as follows:
• The auxiliary fuel requirement must be positive. A negative fuel requirement indicates
that the heat of combustion of the waste gas, (-Ahc), is too high for the fractional energy
recovery in the primary heat exchanger that was selected.
• The auxiliary fuel amount must be high enough to provide a stable flame in the preheater
combustion chamber (See Step 8c below).
An energy balance around the preheater combustion chamber and the catalyst chamber, taken
together, results in Equation 3.18, the same equation used in the thermal incinerator calculations.
The input data for Equation 3.18 for the catalytic incinerator example problem are summarized
below:
• The waste stream is essentially air so that
Pw0 = pw, = 0.0739 Ib/scf, air at 77°F, 1 atm
cP»,a,r = 0.248 Btu/lb °F, the mean heat capacity of air between 77 °F and 780 °F (the
average of the preheater exit and catalyst bed outlet
temperatures)
• Other input data to Equation 3.18 include
3-31
-------
2W = Qv = 20,000 scfm
(-AA ) = 21,502 Btu/lb, for methane
a/
J , = 77 °F, assumeambientconditions
p0/ = 0.0408 lb/ft3, methane at 77 °F, latm.
Tfl = 900°F, from Step 5c
Tw = 660°F, from Step 6c
(-A/2f ) = 56.6 Btu/lb, from Step 4
Substituting the above values into Equation 3.18 results in
£>a/= 40 scfin
If the outlet temperature of the catalyst bed, T^, is 800°F, then Q^ decreases to -6.7 scfin.
In other words, no auxiliary fuel would, theoretically, be required at this bed temperature.
However, as discussed above in Step 8t, a certain quantity of auxiliary fuel would be required to
maintain burner stability.
At 70% energy recovery and 900 °F outlet catalyst bed temperature, a waste gas with a heat
of combustion, (-A/i.^ pf about 79 9 Rtu/lb would cause the auxiliary, fuel requirement, Q"/ to
become negative, indicating the catalyst bed would exceed 900 F. At 70% energy recovery and
800 °F outlet catalyst bed temperature, this same result occurs with a (~^cw \ Of 57 7 Btu/lb
Both of these heats of combustion are relatively low for typical waste gases. These results are^
of course, dependent on the assumption of energy losses from the combustion chamber. The
lower the energy losses, the lower the allowable waste gas heat of combustion before overheating
occurs in the catalyst bed.
Step 8c - Verify that the auxiliary fuel requirement is sufficient to stabilize the burner
flame Only a small amount of auxiliary fuel (< 5% of the total energy input) is needed to stabilize
the burner flame. In general, more fuel than just this stabilizing fuel will be required to maintain
the reactor temperature. It is wise to verify that the auxiliary fuel requirement calculated in Step
7c is sufficient for stabilization. If it is insufficient, then a minimum amount of auxiliary fuel must
be used and the amount of energy recovery specified earlier must be reduced to avoid exceeding
the specified temperature at which the incinerator will operate (Step 5c).
This check is made in the same manner as that in Step 8t of the thermal incinerator
calculation. The results of this check indicate that the auxiliary fuel requirement is more than
sufficient to stabilize the burner flame.
3-32
-------
Step 9c - Estimate the inlet temperature to the catalyst bed, Tr. j^e inlet temperature to the
catalyst bed must be calculated to ensure that the inlet temperature is above that necessary to
ignite the combustible organic compounds in the catalyst that was selected for use.
The inlet temperature to the catalyst bed, Tr. shouid ^ such that? ^hen the temperature rise
through the catalyst bed, AT1, is added to it, the resulting temperature is Tf, 900°F. Thus,
*T=Tf,-Tr, (3-21)
The value of A T is determined by an energy balance around the preheater portion of the
combustor. The preheater is required to heat the gases up to the catalyst bed inlet temperature
using auxiliary fuel.1 This energy balance is made with the assumptions made earlier in deriving
Equation 3.18 and further assuming that only auxiliary fuel is combusted in the preheater portion.
The resulting equation is very similar to Equation 3.18 except that (1) the terms with an/
subscript become terms with r, subscripts to denote a catalytic reactor inlet stream rather than a
combustor outlet (flue gas inlet to the primary heat exchanger) and (2) the term for combustion
of the waste gas organics does not appear. The resulting equation is as follows:
3.22
This equation may be rearranged to solve for Tr. explicitly. This produces an equation that
is somewhat complex and non-intuitive.
T . . ,,
i.ic^ (PafQaf + p^a,;
After substituting the example problem parameters into Equation 3.23, we obtain a value for
^r(.of 693 °F. Based on ignition temperatures shown in Table 3.2, this reactor inlet temperature
should be satisfactory. Prior to a more definitive design, the ignition temperatures for the specific
chemicals should be verified.
The temperature rise across the catalyst bed is thus (900 - 693) or 207 °F. These
temperatures are somewhat sensitive to the assumption for energy losses from the combustor.
The assumption for energy losses is perhaps somewhat conservative, i.e., it causes a larger Qaf
to be estimated than would a less conservative assumption, and becomes more conservative as
the combustor size and insulation are increased.
1 At equilibrium, the temperature of the catalyst bed is maintained without requiring auxiliary fuel.
3-33
-------
Step lOc - Calculate the total volumetric flow rate of gas through the incinerator, Qfl The
total volumetric flow rate of gas leaving the incinerator is referred to as the flue gas flow rate, Qfl,
and is the gas rate on which the incinerator sizing and cost correlations are based. The flue gas
flow rate measured at the standard conditions of 77°F and 1 atmosphere, where the increase in
volumetric throughput due to an increase in the number of moles of gas as a result of combustion
is neglected, is the sum of the inlet streams to the incinerator.
0,, = QWo + Qa + Qaf
= 20,000 +0+40
= 20,040 scfin
Step lie - Calculate the volume of catalyst in the catalyst bed If the volumetric flow rate of
gas through the catalyst bed, Op and the nominal residence time (reciprocal space velocity) in the
catalyst bed are known, then the volume of catalyst can be estimated. There exists complex set
of relationships between the catalyst volume and geometry, overall pressure drop across the
catalyst, conversion of the oxidizable components in the gas, gas temperature, and the reaction
rate. These relationships are dependent on the catalyst and the type of compound being oxidized.
It is beyond the scope if this Manual to discuss these relationships, even in an approximate way.
For the purposes of cost estimation, the space velocity, -
cat
where
VCM = Overall bulk volume of the catalyst bed, including interparticle voids (ft3)
By petro-chemical industry convention, the space velocity is computed at the conditions of 60 °F
(not 77 °F) and 1 atm. The volumetric flow rate, Qfi must be corrected to these conditions. The
proper space velocity to achieve a desired level of conversion is based on experimental data for
the system involved. For precious metal monolithic catalysts, the space velocity generally lies
between 10,000 h"1 and 60,000 h1 . (Base metal catalysts operate at lower space velocities,
ranging from 5,000 to 15,000 h-'.)[10]
For the example, using a space velocity of 30,000 h"1 or 500 minl, and using Qfi at 60°F,
3-34
-------
at 60°F = 20,040 60 + 46°
cat
77 + 460
= 19,400 ftVmin
19,400 ft3/min
500 min '
39 ft3
There are a number of catalyst bed parameters, such as catalyst configuration and bed design,
that are not significant for study type cost estimates. Accordingly, design of these factors is not
discussed here.
3.5 Cost Analysis
This section presents procedures and data for estimating capital and annual costs for four types
of incinerators:(l) thermal-recuperative, (2) thermal regenerative, (3) fixed-bed catalytic, and (4)
fluid-bed catalytic.
3.5.1 Estimating Total Capital Investment
Total capital investment, TCI, includes the equipment cost, EC, for the incinerator itself, the cost
of auxiliary equipment (e.g., ductwork), all direct and indirect installation costs, and costs for
buildings, site preparation, offsite facilities, land, and working capital. However, the last five
costs usually do not apply to incinerators. (See Chapter 2 of this Manual for a detailed
description of the elements comprising the TCI)
3.5.1.1 Equipment Costs, EC
As discussed in Section 3.2.3, the equipment costs, EC, given in this chapter apply to packaged
incinerators, except for regenerative incinerators. For regenerative incinerators, the costs apply
to field-erected units. The EC typically includes all flange-to-flange equipment needed to oxidize
the waste gas, including the auxiliary burners, combustion chamber, catalyst, primary heat
exchanger (except for the "zero heat recovery" cases), weathertight housing and insulation, fan,
flow and temperature control systems, a short stack, and structural supports. Smaller units, e.g.,
typically less than 20,000 scfin, are typically preassembled skid-mounted [18]. The various
available incineration systems are presented in four groups delineated according to their similarity
3-35
-------
Table 3.7: Scope of Cost Correlations
Incinerator Type
Thermal - Recuperative
Thermal - Regenerative
Fixed-Bed Catalytic
Fluid-Bed Catalytic
Total (Flue) Gas
Flowrate, scfm
500a-50,000
10,000-100,000
2,000-50,000
2,000-25,000
Figure Number
3.4
3.5
3.6
3.7
• Although Figure 3.4 covers the 1,000 to 50,000 scfin range, the correlation is valid for the 500 to 50,000 scfin range.
of design. These groups are outlined in Table 3.7. With the exception of regenerative thermal
and fluid-bed catalytic incinerators, the maximum size for which costs are given is 50,000 scfin.
Although larger units of each technology can be built, applications are rare at flow rates above
50,000 scfin. Regenerative thermal incinerator costs are provided for flow rates from 10,000 to
100,000 scfrn. Fluid-bed catalytic incinerator costs are provided for flow rates from 2,000 to
25,000 scfin.
The cost curves are least-squares regressions of cost data provided by different vendors. It
must be kept in mind that even for a given incineration technology, design and manufacturing
procedures vary from vendor to vendor, so that costs may vary. As always, once the study
estimate is completed, it is recommended that more than one vendor be solicited for a more
detailed cost estimate.
The additional expense of acid gas clean-up or paniculate control is not treated in this section.
The equipment cost of a gas absorber to remove any acid gases formed in the incinerator can be
quite large, sometimes exceeding the equipment cost of the incinerator itself even for simple
packed tower scrubbers [19]. For more complex absorbers that include venturi scrubbers instead
of, or in addition to, packed beds, the cost of the scrubber alone may be up to 4 times that of the
incinerator [11]. These more complex absorbers are sometimes necessary when particulates, in
addition to acid gases, must be removed from the flue gas. (Note: Chapter 9 of the Manual
provides data and procedures for sizing and costing gas absorbers.)
Thermal Incinerators Among the thermal units, the direct flame (0% energy recovery) and
recuperative systems are treated together because the various levels of energy recovery are
achieved simply by adding heat exchanger surface area. Costs for these units were provided by
several vendors [12,20,21]. The EC of these units are given as a function of total volumetric
throughput, Qton in scfm. "Qor", is the total volume of the gaseous compounds exiting the
combustion chamber; it is identical to the term, "Qf," used in Figures 3.1 and 3.3. This includes
the combustion products, nitrogen, unburned fuel and organics, and other constituents. (See
3-36
-------
Figure 3.4). Note that costs are given free on board (F.O.B.) in April 1988 dollars*. Based on
a least-squares regression analysis, a log-log relationship between throughput and EC was found
for a given level of energy recovery (HR) over the flow rate range from 500 to 50,000 scfrn.
These relationships are as follows:
EC = 1 02940° 2355 HR = 0%
(3 24)
EC = 13149e,°,2609 HR = 35% (3i25)
EC = 17056<2,°,25°2 HR = 50% (3-26)
,,
EC = 213420,!,2500 HR = 70% (3'27)
The regenerative (or excess enthalpy) systems provide up to 95 percent heat recovery at the
expense of higher capital costs. Their unique design [22,23], which combines the heat exchanger
and reactor, is substantially different from traditional thermal units and is therefore treated
separately in Figure 3.5. The ECs of these systems are given as an approximately linear function
of total flow rate over a 10,000 to 100,000 scfm range by the following equation:
EC = 2.204 x 105 + 11.57 Qtot 3.28
Again, the higher capital costs of these units can be substantially offset by the substantial
savings in auxiliary fuel costs.
Catalytic Incinerator The EC for a catalytic incinerator is a function of the type of catalyst
contacting pattern used and the total gas flow rate, Qtot, for a given level of energy recovery.
There are three types of contacting configurations used in catalytic systems: fixed-bed, catalytic
monolith, and fluid-bed. The EC for the first two are generally comparable and are given in
Figure 3.6. The data provided by several vendors [12,20,21,24] exhibited curvilinear
relationships with Qtot for each of the energy recovery rates. Least squares regressions of the
data yielded the following correlations for total flow rates between 2,000 and 50,000 scfrn:
*For information on escalating these and the other incinerator prices to more current dollars, refer to the
EPA report Escalation Indexes for Air Pollution Control Costs and updates thereto, all of which are installed
on the OAQPS Technology Transfer Network (CTC Bulletin Board).
3-37
-------
EC- 11052,°0,5471 HR - 0%
°4189
EC - 3623£,0, HR - 35% (330)
EC - 12152™ HR = 50% (3.31)
£C = 1443&0,5527 HR = 70% (3J2)
Fluid-bed catalytic incinerators aflFord certain advantages over fixed-bed catalyst units in that
they tolerate waste streams with (1) higher heating values, (2) particulate contents, and (3)
chlorinated species. For this enhanced flexibility of feed streams, a higher capital cost is incurred,
as indicated by the EC shown in Figure 3.7. The data shown were provided by vendors [11,19]
and exhibited a linear relationship over the range of flow rates from 2,000 to 25,000 scfm. They
can be approximated by the following equations:
EC = 8.48 x 104 + 13.2gto, HR = 0%
EC = 8.84 x 104 + 14.60to, HR = 35% (334)
EC = 8.66 x 104 + 15.80,0, HR - 50% (3.35)
EC = 8.39 x 104 + 19.20 HR - 70% <3-36)
3-38
-------
en
Q
O
I
(SI
1
o
o
CD
CO
en
cr
a.
03
o
lo-
co
o
o
UJ
Q.
*™1
O
0% ENERGY RECOVERY
3 * 3 • 7 « a 10
10 30 *O 10 60 708090
"LUE GAS VOLUMETRIC FLOW RATE. SCFM
Figure 3.4. Equipment Costs of Thermal Incinerators, Recuperative
3-39
-------
in
10
VI
o
0
00
00
200
2 0
03
c
S
3
a.
o 95 % ENERGY RECOVERY
>r
20 40 60
FLUE GAS VOLUMETRIC PLOW RATE.
60
SCFM (~HCUSA,\3£
•r
Figure 3.5. Equipment Costs of Thermal Incinerators, Regenerative
3-40
-------
1/1
a
z
O
I
1/1
a:
o
a
00
ao
en
-------
A comparison of the thermal, catalytic fixed-bed, and catalytic fluid-bed systems with 50 percent
energy recovery is shown in Figure 3.8.
3.5.1.2 Installation Costs
As explained in Chapter 2, the purchased equipment cost, PEC, is calculated by taking the sum
of the EC and the cost of auxiliary equipment (e.g., ductwork), taxes, freight, and
instrumentation. Average values of direct and indirect installation factors [25] to be applied to
the PEC are given in Table 3.8 for both recuperative thermal and fixed- and fluid-bed catalytic
incinerators.
Table 3.9 shows the itemized installation costs that are obtained when these installation factors
are applied to the PECs for the example incinerators. Depending on the site conditions, the
installation costs for a given incinerator could deviate significantly from costs generated by these
average factors, Vatavuk and Neveril [25] provide some guidelines for adjusting the average
installation factors to account for other-than-average installation conditions. For units handling
total gas flow rates lower than 20,000 scfm the installation costs are minimal, amounting normally
to only utility tie-ins (electrical and, if necessary, combustion or dilution air). The installation
costs for these smaller incinerators would be 20 to 25 % of the PEC. Smaller units may be
installed on the roofs of manufacturing buildings rather than at ground level. In such cases the
installation factors could be as high as (or higher than) the factors shown in Table 3.8, even
though the units would be "packaged".
3.5.2 Estimating Total Annual Cost
The total annual cost (TAC) is the sum of the direct and indirect annual costs. The TAG for both
example systems is given in Table 3.10, alone with suggested factors for calculating them.
3.5.2.1 Direct Annual Costs
Direct annual costs for incinerators include labor (operating and supervisory), maintenance (labor
and materials), fuel, electricity, and (in catalytic units) replacement catalyst. For thermal and
catalytic units, the fuel usage rate is calculated as shown in Sections 3.4.2 and 3.4.3, respectively
where natural gas (methane) is assumed to be the fuel. (Other fuels could be used for thermal
units.)
3-42
-------
10
o
i
CD
CO
a>
03
O
u.
O
U
a.
o
o
tO
o
o
o
o
CM
c
1
.
\
o 0 - % ENERGY RECOVERY
o 35 %
* 50 %
• 70 %
i
mmmm
^
•
j
/
ii ii
u it
ii u
1
!
l
i /
/
'/
j&*
^
r\
. \
.
f
$
i {
/
/
>*rC!
i^^r
i
i
\
1
! LI
i
i
' !
! i
! !
! t
! i
i ;
l : i
j ! ;
|
1 1 ! ! 1
i !
' J
1
i
,
:
0 10 20 30
~_UE GAS VOLUMETRIC "LOW RA"E. $•:
40
Figure 3.7. Equipment Costs of Catalytic Incinerators, Fluid-Bed
3-43
-------
(/I
a
z
in
O
O
O
03
to
en
tz
a.
02"
o
o
u
LU
2
a.
o
O CATALYTIC - FIXED BED
THERMAL - RECUP
50% ENERGY RECOVERY j
20 20 40 SO M 708090
FLUE GAS VOLUMETRIC FLOW RATE, SCFM (THOUSANDS;
Figure 3.8: Equipment Costs Comparison of Incinerator Tj'pes
3-44
-------
Table 3.8. Capital Cost Factors for Thermal and Catalytic Incinerators3
Cost Item Factor
Direct Costs
Purchased equipment costs
Incinerator (EC) + auxiliary equipment* As estimated, A
Instrumentation1" 0.10 A
Sales taxes 0.03 A
Freight 0.05 A
Purchased equipment cost, PEC B = 1.18 A
Direct installation costs
Foundations & supports 0.08 B
Handling & erection 0.14 B
Electrical 0.04 B
Piping 0.02 B
Insulation for ductwork^ 0.01 B
Painting 0.01 B
Direct installation cost 0.30 B
Site preparation As required, SP
Buildings As required, Bldg
Total Direct Cost, DC 1.30 B + SP + Bldg.
Indirect Costs (installation)
Engineering 0.1 OB
Construction and field expenses 0.05 B
Contractor fees 0.10 B
Start-up 0.02 B
Performance test 0.01 B
Contingencies 0.03 B
Total Indirect Cost, 1C 0.31 B
Total Capital Investment = DC + 1C 1.61 B + SP + Bldg.
"Reference [25]
'Ductwork and any other equipment normally not included with unit furnished by incinerator vendor.
Instrumentation and controls often furnished with the incinerator, and those often included in the EC.
''If ductwork dimensions have been established, cost may be estimated based on $10 to $12/ft2 of surface for fluid
application. (Alternatively, refer to Chapter 10 of this Manual. Fan housings and stacks may also be insulated.
3-45
-------
Table 3.9. Capital Costs for Thermal and Catalytic Incinerators
Example Problem
Cost Item
Direct Costa
Purchased equipment costs
Incinerator (EC)
Auxiliary equipment*
Sum = A
Instrumentation, 0.1A
Sales taxes, 0.03A
Freight, 0.05A
Purchased equipment cost, B
Direct installation costs
Foundation and supports, 0.08B
Handling and erection, 0.14B
Electrical, 0.04B
Piping, 0.02B
Insulation (for ductwork), Q.OlB
Painting, 0.01B
Direct installation cost
Site preparation*
Buildings'
Total Direct Cost
Indirect Costs (installation)
Engineering, 0.10B
Construction and field expenses, 0.05B
Contractor fees, 0.1 OB
Start-up, 0.02B
Performance test, O.OlB
Contingencies, 0.03B
Total Indirect Cost
Total Capital Investment (rounded)
Cost,
Thermal-
Recuperative
$254,200
—
1254,200
25,400
7,630
12,700
J300,000
24,000
42,000
12,000
6,000
3,000
3,000
$90,000
—
*"•""
1390,000
30,000
15,000
30,000
6,000
3,000
9,000
593,000
$483,000
5
Fluid- Bed
Catalytic
$468,200
—
$468,200
46,800
14,000
23,400
$552,400
44,200
77,300
22,100
11,000
5,520
5,520
$165,600
—
MM*
$718,000
55,200
27,600
55.200
11,000
5,520
16,600
$171,100
$889.000
•N»B« of thww item* I* rvquirvd.
3-46
-------
Table 3.10: Annual Costs for Thermal and Catalytic Incinerators
Example Problem
Cost Item
Suggested Factor
Unit Cost"
Thermal
Direct Annual Costs*. DC
Operating Labor
Operator
Supervisor
Operating Materials
Maintenance
Labor
Material
Catalyst replacement
Utilities
Natural Gas
Electricity
Total DC
Indirect Annual Costs. 1C
0.5 h/shift
15% of operator
0.5 h/shift
100% of
maint. labor
100% of catalyst
replaced ea. 2 yr.
$12.95/h
$14.25/h
$650/ft3 for
metal oxide
$3.30/kft3
$0.059/kWh
6,480
972
7,130
7,130
264,500
$321,200
Fluid-Bed
Catalyst
6,480
972
7,130
7,130
15,100
Overhead
Administrative charges
Property taxes
Insurance
Capital recovery^
Total Annual Cost (rounded)
60% of sum of
operating, supr., &
maint. labor &
maint. materials.
2% TCI
1%TCI
1%TCI
CRF [TCI- 1.08
(Cat. Cost)]
— 13,000
— 9,650
— 4,830
— 4,830
— 68,600
$101,100
$422,000
13.000
17,800
8,900
8,900
122,700
$171,300
$316,000
"1988 dollars
"Assumes 8.000 h/yr
The capital recovery cost factor, CRF. is a function of the catalyst or equipment life (typically. 2 and 10 years, respectively) and
the opportunity cost of the capital (i.e., interest rate). For example, for a 10-year equipment life and a 7% interest rate. CRF =
0.1424.
3-47
-------
Table 3.11. Typical Pressure Drop Across Selected Equipment
Equipment Type
Thermal Incinerators
Catalytic Fixed-bed Incinerators
Catalytic Fluid-bed Incinerators
Heat Exchangers
M M
?! 11
Energy Recovery, %
0
0
0
35
50
70
A P, in. H
4
6
6-10
4
8
15
,0
The electricity costs are primarily associated with the fan needed to move the gas through the
incinerator. The power (in kilowatts) needed to move a given inlet volumetric flow rate of air
(Q»,per Sections 3.4.2 and 3.4.3) at a total flange-to-flange pressure drop of AP inches of water
and combined motor/fen efficiency, e, is adapted from Equation 2.7, as follows:
1.17 x 1Q-4Q AP
= - -1— (3.37)
Fan efficiencies vary from 40 to 70 percent [15] while motor efficiencies are usually 90 percent.
The total pressure drop across an incinerator system depends on the number and types of
equipment elements included in the system and on design considerations. The estimation of actual
pressure drop requirements involves complex calculations based on the specific system's waste
gas and flue gas conditions and equipment used. For the purposes of this section, however, the
approximate values shown in Table 3.1 1 can be used.
For the example cases, we will assume 8,000 hours per year operation and a 60% efficiency
for the fan and motor together. Using pressure drops of 4 and 8 inches of water, respectively,
for the thermal and fluid-bed catalytic incinerators*, and adding the pressure drop of 15 inches of
water for 70% heat recovery, the fan power requirements can be calculated as follows:
Thermal Incinerator"
= 1-17 x 1(T4 (20,900 acfm) (19 inches water)
owerfan —
= 77.4 kW
Catalytic Incinerator
*A fluid-bed catalytic incinerator is used because the waste gas contains a chlorinated compound which
would poison the catalyst in a fixed-bed incinerator.
"""Computed from inlet waste-gas flow rate (20,000 scfrn) at preheater inlet temperature (100°F).
3-48
-------
1.17 x IP"4(20,900 acfin) (23 inches water)
0.60
= 93.7 kW
The annual electricity costs would be the products of these usages, the annual operating
hours, and the electricity cost ($/kWh), or:
Electricity Cost (Thermal) = 77.4 kW x 8,000 hours/yr x $0.059/kWh
= $36,500 per yr
Electricity Cost (Catalytic) = 93.7 kW x 8,000 hours/yr x $0.059/kWh
= $44,200 per yr
The catalyst replacement costs and scheduling are highly variable and depend on the nature
of the catalyst, the amount of "poisons" and particulates in the gas stream (including the auxiliary
fuel), the temperature history of the catalyst, and the design of the unit. It is impossible to predict
the costs in a general sense. However, noble metal monolith catalysts operating on pure
hydrocarbon gases in air will last longer than fluid-bed base metal catalysts operating on
chlorinated hydrocarbons in air. Noble metal catalysts are also more expensive than base metal
oxide catalysts. The catalyst life for many field units is from 1 to 4 years. The cost, in April 1988
dollars, of the replacement catalyst must be obtained from the vendor, but it may be estimated at
$3,000/ft3 for noble metal catalysts and $650/ft3 for base metal oxide catalysts. For the example
case, the catalyst is a base metal oxide because the waste gas contains a chlorinated compound.
We will assume a two year catalyst life. Knowing that the catalyst volume is 39 ft3 (Section
3.4.3) and using a cost of $650/ft3 and a capital recovery factor of 0.5531 (2-year life at a 7%
interest rate), the annual expense for catalyst replacement is
Annual Catalyst Replacement Cost = 39 ft3 x 650— x 0.5531 x 1.08
ft3
= $15,100 per year
(The " 1.08" factor covers the freight and sales tax for the replacement catalyst.)
To calculate the fuel or electricity annual cost, multiply the fuel usage rate (scfin) or the
electricity usage rate (kW) by the total hours per year of operation (e.g., 333 d/yr x 24 h/d =
8,000 h/yr) and by the appropriate unit cost (e.g., $/scfrn for fuel and $/kWh for electricity).
For the example cases, the fuel costs can be calculated from the fuel usage rates and the
natural gas unit cost of $0.00330/scf. For the thermal incinerator example, the annual fuel cost
is calculated as follows:
Annual Fuel = 0.00330 _L x 167— x 60 -^ x 8,000—
scf min hr yr
Cost, Thermal = $264,500 per year
3-49
-------
For the catalytic incinerator example, the annual fuel cost is found similarly.
Annual Fuel Cost, Catalytic = $63,400 per year
Operating and maintenance labor are estimated as 0.5 hours per 8-hour shift each, supervisory
labor at 15% of operating labor, and maintenance material as 100% of maintenance labor.
3.5.2.2 Indirect Annual Costs
The indirect (fixed) annual costs include capital recovery, overhead, and property taxes,
insurance, and administrative (G&A) charges. The last three of these can be estimated at 1%,
1%, and 2% of the total capital investment, respectively. The system capital recovery cost is
based on an estimated 10-year equipment life. (See Section 2 for a thorough discussion of the
capital recovery cost and the variables that determine it.) The system capital recovery cost is the
product of the system capital recovery factor (CRF) and the total capital investment (TCI) less
the purchased cost of the catalyst (Cca, x 1.08 where the 1.08 is for freight and sales tax). These
values calculated for the example cases are given in Table 3.10.
3.5.3 Cost Comparison for Example Case
The example VOC stream defined in Section 3.4.1 serves to illustrate some typical characteristics
of thermal and catalytic systems. The total annual costs shown in Table 3.10 show that the
catalytic system's auxiliary fuel costs are significantly lower than those of the thermal unit. The
disparity is enough to offset the higher capital costs of the catalytic incinerator over the assumed
10-year lifetime of the units. Two factors that should be noted in the comparison of these two
systems are (1) the 98 percent level of destruction met by the thermal incinerator may be difficult
to reach by the catalytic system (this may be important in some cases), and (2) the example waste
stream is of particularly low heating value (4 Btu/scf) which favors the catalytic system due to
the lower auxiliary fuel requirements.
3-50
-------
3.6 Acknowledgements
The authors gratefully acknowledge the following companies for contributing data to this chapter:
• Peabody Engineering (Stamford, CT)
• Combustion Engineering - Air Preheater, Inc. (Wellsville, NY)
• TEC Systems, Inc. (DePere, WI)
• Air Research, Inc. (ARI) (Palatine, IL)
• Energy Development Associates (EDA) (Itasca, IL)
• Pillar Technologies, Inc. (Hartland, WI)
• Huntington Energy Systems, Inc. (Union, NJ)
• Regenerative Environmental Equipment Co. (REECO) (Morris Plains, NJ)
• Englehard Corp. (Edison, NJ)
3-51
-------
Appendix 3A
Properties of Selected
Compounds
3-52
-------
Table 3.12. Limits of Flammability of Combustible Organic Compounds in Air at
Atmospheric Pressure, Room Temperature*
Compound
Methane
Ethane
Propane
Butane
Pentane
Hexane
Octane
Nonane
Decane
Ethylene
Propylene
Acetylene
Cyclohexane
Benzene
Toluene
Molecular
Weight
16.04
30.07
44.09
58.12
72.15
86.17
114.23
128.25
142.28
28.05
42.08
26.04
84.16
78.11
92.13
LEI/,
vol. %
5.00
3.00
2.12
1.86
1.40
1.18
0.95
0.83
0.77
2.75
2.00
2.50
1.26
1.40
1.27
UELb,
vol. %
15.00
12.50
9.35
8.41
7.80
7.40
28.60
11.10
80.00
7.75
7.10
6.75
*Reference [14]
"Lower Explosive Limit
bUpper Explosive Limit
3-53
-------
Table 3.13: Molar Heat Capabilities of Gases at Zero Pressure*
Compound
Methane
Ethane
Propane
Butane
Pentane
Hexane
Cyclopentane
Cyclohexane
Benzene
Toluene
Nitrogen
Oxygen
Air
Carbon dioxide
a
4.750
1.648
-0.966
0.945
1.618
1.657
-12.957
-15.935
-8.650
-8.213
6.903
6.085
6.713
5.316
Lp = a + bl + d*
rl
pfTJ / 'y
C in calories/ 'g -moles
bxlO2
1.200
4.124
7.279
8.873
10.85
13.19
13.087
16.454
11.578
13.357
-0.03753
0.3631
0.04697
1.4285
+ aij ; 1 m "K
>CpdT
"K Btullb-mole
cxlO6
0.3030
-1.530
-3.755
-4.380
-5.365
-6.844
-7.447
-9.203
-7.540
-8.230
0.1930
-0.1709
0.1147
-0.8362
°R
dxlO9
-2.630
1.740
7.580
8.360
10.10
13.78
16.41
19.27
18.54
19.20
-0.6861
0.3133
-0.4696
1.784
Temperature
Range, °K
273-1500
273-1500
273-1500
273-1500
273-1500
273-1500
273-1500
273-1500
273-1500
273-1500
273-1800
273-1800
273-1800
273-1800
: Reference [26]
3-54
-------
Table 3.14 Heats of Combustion of Selected Gaseous Organic Compounds, -A//c, at 25°C
and constant pressure to form gaseous water and carbon dioxide.*
Compound
Methane
Ethane
Propane
Butane
Pentane
Hexane
Octane
Nonane
Decane
Ethylene
Propylene
Cyclopentane
Cyclohexane
Benzene
Toluene
Molecular
Weight
16.04
30.07
44.09
58.12
72.15
86.17
114.23
128.25
142.28
28.05
42.08
70.13
84.16
78.11
92.13
cal/g.
11,953.6
11,349.6
11,079.2
10,932.3
10,839.7
10,780.0
10,737.2
10,680.0
10,659.7
11,271.7
10,942.3
10,563.1
10,476.7
9,698.4
9,784.7
Btu/lb
21,502
20,416
19,929
19,665
19,499
19,391
19,256
19,211
19,175
20,276
19,683
19,001
18,846
17,446
17,601
*Reference[15]
3-55
-------
Appendix 3B
Design Procedure for Non-Recuperative
Thermal Incinerators
3-56
-------
Not all thermal incinerators are equipped with recuperative heat exchangers to transfer energy
from the flue gas stream to the incoming waste gas stream. These non-recuperative units use
other mechanisms to recovery flue gas energy. One of these types is the regenerative incinerator.
As discussed in Section 3.2.1.3, a regenerative incinerator accomplishes energy recovery by
conveying the flue gas through a ceramic bed which captures a portion of the stream's enthalpy.
After a switching mechanism is engaged, the incoming waste gas passes through this hot bed and
is warmed to its ignition temperature. This process is illustrated in Figure 3.2.
While we can determine the stream inlet and outlet temperatures for a recuperative heat
exchanger fairly accurately, we cannot always do so for a regenerative incinerator bed. For one
thing, these beds do not behave like typical heat exchangers. The bed temperature profiles are
often difficult to predict. More importantly, because regenerative incinerators do not operate at
steady state conditions, the temperatures within the beds and many other parts of the unit vary
with time. For that reason, it is more convenient to view the entire regenerative incinerator as
a "black box" into which waste gas and auxiliary fuel flow and from which flue gas emanates.
Around this black box we may make mass and energy balances. In this way, we need not make
any assumptions about what occurs inside the incinerator regarding temperatures, flowrates, or
other stream parameters.
However, to determine the auxiliary fuel requirement for regenerative incinerators via the
procedure shown in this appendix we have to make two key assumptions, viz.: (1) the
temperatures and flowrates of all streams entering and leaving the incinerator are at steady state
and (2) the combustion temperature (and by inference, the heat loss fraction) are constant as well.
The other assumptions will be addressed in the following design steps:
Steps 1 to 4: These are the same as those for thermal recuperative and catalytic incinerators.
(See Section 3.4.1.)
Step 5t - Establish the incinerator operating temperature: Because their designs are more
resistant to thermal stresses and because they can achieve very high heat recoveries, regenerative
incinerators are usually operated at higher temperatures than recuperative units. Consequently,
higher VOC destruction efficiencies are achieved. Operating temperatures of 1800 to 2000 °F
are typical.
Step 6t - Calculate the waste gas temperature at the exit of the preheater: As explained
above, regenerative incinerators do not employ preheaters. The preheating is done by and within
the ceramic beds. Moreover, because the mass and energy balances are made around the entire
unit, we do not need to know the temperature of the preheated waste gas to calculate the
auxiliary fuel requirement.
Step 7t - Calculate the auxiliary fuel requirement, Q^: Because a regenerative incinerator
recovers nearly all of the energy from the combustion (flue) gas, its auxiliary fuel requirement is
usually lower than that for a recuperative incinerator. However, as discussed above, this fuel
requirement is determined via mass and energy balances taken around the entire unit, not just the
combustion chamber. Consider the following diagram:
3-57
-------
Flue gas
(fo)
Aux. fuel
(af)
Incinerator unit
fany type)
Waste
gas
(wi)
Taking mass and energy balances around the incinerator, we obtain:
Mass balance:
Mass in = Mass out
Mass fuel + Mass waste gas = Mass flue gas
PafQaf + PWQW = PfoQfo (3B-1)
Energy balance:
Next, we take an energy balance around the incinerator unit:
Energy in - Energy out + Energy generated = 0
The terms of the energy balance equation are the inlet waste gas and outlet flue gas enthalpies
(H^ and Hf0, respectively), the energy loss (H^), and the waste gas VOC and fuel (natural gas)
heat contents (Hcwl and H^, in turn):
The variables comprising each of the terms in this energy balance equation are listed in Table
3.6. They are:
"wi ~ PwiQwi^pmwX 1 wi ~ 1 ref)
H =pOC (T-T)
HL = TlPflQfiCpmfl(Tfl - Ttef)
3-58
-------
HCW1 =
where: T\ = energy loss from combustion chamber (fractional)
Tfi = combustion temperature (°F)
We next substitute these variables into eq. (3B-2) and solve for the fuel mass rate
When doing so, we make the following assumptions:
• The streams flowing to and from the incinerator are at steady state conditions.
• The auxiliary air requirements are zero.
• The ambient, reference, and fuel inlet temperatures are equal (77 °F). (This assumption
results in the inlet fuel stream having a zero enthalpy.)
• The heat capacities of the gas streams to and from the unit are approximately the same,
regardless of composition.
• The mean heat capacities of the streams above the reference temperature (77 °F) are
approximately equal, regardless of temperature. Further, the mean heat capacity of the waste
gas/flue gas stream entering/leaving the incinerator is evaluated at the average of the inlet
(T J and combustion (Tfl) temperatures. That is, Cpmwi = C^ = Cpmfo= Cpm.
When we do all this, we get the following expression:
PwQw,Cpm(TM - Tref) - [pfoQfoCpm(Tfo - Tref) + npfiQfiCpm(Tfi - Tref)] +
(Energy in) (Energy out)
[pwlQM(-Ahc J + o^-Ah^)] = 0
(Energy generated)
Substitution for pfoQfo per eq. 3B-1 above yields:
{pw,Qw1Cpm(Tw,-Tref)} - {r,Cpm(paiQaf + pw,QJ(Tfi-Tref) +
Cpm(pafQaf + PwiQJ(Tfo-Tref)} + {pJ^C-AlO + p^-Alw)} = 0
Finally, solving for pafQaf, the auxiliary fuel mass rate (Ib/rnin):
Pw,Qw,{Cpm[Ti(Tfi - Tref) + (Tfo - TJ] - (-
PafQaf = - (3B-3)
- Cpm[Ti(Tfl - Tref) + (Tfo - Tref)}
3-59
-------
Equation (3B-3) provides the auxiliary fuel requirement for any type of thermal incinerator,
as it is independent of any intermediate variables, such as the temperature of the preheated waste
gas. Clearly, this equation can be used with regenerative incinerators, as long as the above-stated
assumptions hold.
The heat loss fraction (T|) will vary according to the incinerator type, how the incinerator
components are configured in the unit, the construction materials, the type and amount of
insulation, and other factors. For instance, for recuperative incinerators, t) is approximately 0.10.
The T) for regenerative incinerators is considerably lower, however. There are two reasons for
this. First, the components of a regenerative incinerator—combustion chamber, ceramic beds,
etc.—are housed in a single enclosure, while in a recuperative incinerator the combustion
chamber, heat exchanger, and interconnecting ductwork are housed separately, thus offering more
heat transfer area. Second, because regenerative units are lined with ceramic, they are better
insulated than recuperative incinerators.
To gain an estimate of this heat loss fraction, we contacted two regenerative incinerator
vendors. [27,28] Based on the heat loss data that they supplied, we calculated T| values ranging
from 0.002 to 0.015 (0.2 to 1.5%). These values varied according to the incinerator
configuration (vertical or horizontal), the waste gas flow rate, the ambient temperature, and the
wind speed.
Step 8t - Verify that the auxiliary fuel requirement is sufficient to stabilize the burner
flame: As explained in Section 3.4.2, only a small amount (< 5% of the total energy input) is
needed to stabilize the burner flame. With recuperative incinerators, the auxiliary fuel
requirement is usually much larger than the burner stabilization requirement, so that this
constraint rarely comes into play. With regenerative incinerators, however, the auxiliary fuel
requirement may be as low or lower than the fuel needed to stabilize the burner. Therefore, it is
important to compare these two requirements. This comparison is made via equations 3.19 and
3.20. If the auxiliary fuel is less, the minimum fuel requirement would be set at 5% of the total
energy input.
Step 9t - Calculate the flue gas volumetric flow rate, Qfl: As with thermal recuperative
incinerators, the regenerative incinerator flue gas flow rate is the rate used to size and cost the
unit. Measured at standard conditions (1 atmosphere and 77° F), Qfl is the sum of the inlet waste
gas (Q^ and fuel (Q^ flow rates. But since Qrf for regenerative units is small compared to Qm,
the waste gas and flue gas flows should be virtually identical.
3-60
-------
References
[1] Prudent Practices for Disposal of Chemicals from Laboratories, National Academy Press,
Washington, D.C., 1983.
[2] Memo fromMascone, D.C., EPA, OAQPS, to Farmer, J. R., OAQPS, EPA, June 11, 1980,
Thermal Incinerator Performance for NSPS.
[3] Memo fromMascone, B.C., EPA, OAQPS, to Farmer, J. R., OAQPS, EPA, July 22, 1980,
Thermal Incinerator Performance for NSPS, Addendum.
[4] Memo from Mascone, D.C., EPA, OAQPS, to Farmer, J. R., OAQPS, EPA, August 22,
1980, Thermal Incinerators and Flares.
[5] Letter from Thomas Schmidt (ARI International, Palatine, IL) to William M. Vatavuk (EPA,
OAQPS, Research Triangle Park, NC), August 16, 1989.
[6] Weldon, J. and S. M. Senkan, Combustion Sci. Technol, 1986, 47.
[7] Manning, P., Hazard Waste, 1984, 1(1).
[8] Pope, D., Walker, D. S., Moss, R. L., Atmos. Environ., 1976, 10.
[9] Musick, J. K., and F. W. Williams, Ind. Eng. Chem. Prod. Res. Dev., 1974, 13(3).
[10] Letter from Robert M. Yarrington (Englehard Corporation, Edison, NJ) to William M.
Vatavuk (EPA, OAQPS, Research Triangle Park, NC), August 14, 1989.
[11] Personal Communication from Bill Shefier (ARI, Inc., Palatine, IL) to Donald R. van der
Vaart (RTI, Research Triangle Park, NC), March 30, 1988.
[12] Personal Communication from Ralph Stettenbenz (Combustion Engineering, Air
Preheater, Inc., Wellsville, NY) to Donald R. van der Vaart (RTI, Research Triangle
Park, NC), March 23, 1988.
3-61
-------
[13] Grelecki, C., Fundamentals of Fire and Explosion Hazards Evaluation, AIChE Today
Series, New York, 1976.
[14] Weast, R.C. (ed.), CRC Handbook of Chemistry and Physics, 49th ed., CRC Press,
Cleveland, Ohio, 1968.
[15] Perry, R. H. and C. H. Chihon(eds.), Chemical Engineers Handbook, 5th ed., McGraw-
Hill, New York, 1973.
[16] Personal Communication from Robert Yarrington (Englehard Corp., Edison, NJ) to
William M. Vatavuk (EPA, OAQPS, Research Triangle Park, NC), June 6, 1989.
[17] Personal Communication from Thomas Schmidt (ARI International, Palatine, IL) to
William M. Vatavuk (EPA, OAQPS, Research Triangle Park, NC), June 7, 1989.
[18] Githens, R. E. and D. M. Sowards, Catalytic Oxidation of Hydrocarbon Fumes, PB-299
132, National Technical Information Service, Springfield, VA.
[19] Personal Communication from Andrew Jones (Energy Development Associates, Itasca,
IL) to Donald R. van der Vaart (RTI, Research Triangle Park, NC), March 4, 1988.
[20] Personal Communication from C. L. Bumford (Peabody Engineering, Stamford, CT) to
Donald R. van der Vaart (RTI, Research Triangle Park, NC), March 28, 1988.
[21] Personal Communication from C. M. Martinson (TEC Systems, DePere, WI) to Donald
R. van der Vaart (RTI, Research Triangle Park, NC), March 28,1988.
[22] Personal Communication from Ronald J. Renko (Huntington Energy Systems, Inc.,
Union, NJ) to Donald R. van der Vaart (RTI, Research Triangle Park, NC), March 16,
1988.
[23] Personal Communication from James H. Mueller (Regenerative Environmental Equipment
Co., Inc., Morris Plains, NJ) to Donald R. van der Vaart (RTI, Research Triangle Park,
NC), January 13, 1988.
[24] Personal Communication from Robert Hablewitz (Pillar Technologies, Hartland, WI) to
Donald R. van der Vaart (RTI, Research Triangle Park, NC), March 20, 1988.
[25] Vatavuk, W. M. and R. Neveril, "Estimating Costs of Air Pollution Control Systems, Part
II: Factors for Estimating Capital and Operating Costs", Chemical Engineering,
November 3, 1980, pp. 157-162.
[26] Kobe, K. A. and associates, "Thermochemistry for the Petrochemical Industry",
Petroleum Refiner, Jan. 1949 through Nov. 1954.
3-62
-------
[27] Letter from William M. Vatavuk (EPA, OAQPS, Research Triangle Park, NC) to Gerald
Schrubba (Salem Engelhard, South Lyon, MI), September 22, 1992.
[28] Letter from William M. Vatavuk (EPA, OAQPS, Research Triangle Park, NC) to Rod
Pennington (REECO/Research Cottrell,
Somerville, NJ), September 22, 1992.
3-63
-------
Chapter 4
CARBON ADSORBERS
William M. Vatavuk
Innovative Strategies and Economics Group, OAQPS
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
William L. Klotz
Chas. T. Main, Inc.
Charlotte, NC 28224
Robert L. Stallings
Ozone Policy and Strategies Group, OAQPS
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
December 1995
4-1
-------
Contents
4.1 Process Description 4-4
4.1.1 Introduction 4-4
4.1.2 Types of Adsorbers 4-4
4.1.2.1 Fixed-bed Units 4-5
4.1.2.2 Cannister Units 4-6
4.1.3 Adsorption Theory 4-8
4.2 Design Procedure 4-14
4.2.1 Sizing Parameters 4-14
4.2.2 Determining Adsorption and Desorption Times 4-15
4.2.3 Estimating Carbon Requirement 4-17
4.2.3.1 Overview of Carbon Estimation Procedures 4-17
4.2.3.2 Carbon Estimation Procedure Used in Manual 4-17
4.3 Estimating Total Capital Investment 4-19
4.3.1 Fixed-Bed Systems 4-19
4.3.1.1 Carbon Cost 4-19
4.3.1.2 Vessel Cost 4-19
4.3.1.3 Total Purchased Cost 4-23
4.3.1.4 Total Capital Investment 4-23
4.3.2 Cannister Systems 4-25
4.4 Estimating Total Annual Cost 4-25
4.4.1 Direct Annual Costs 4-26
4-2
-------
4.4.1.1 Steam 4-26
4.4.1.2 Cooling Water 4-27
4.4.1.3 Electricity 4-27
4.4.1.4 Carbon Replacement 4-30
4.4.1.5 Solid Waste disposal 4-30
4.4.1.6 Operating and Supervisory Labor 4-31
4.4.1.7 Maintenance Labor and Materials 4-31
4.4.2 Indirect Annual Costs 4-31
4.4.3 Recovery Credits 4-32
4.4.4 Total Annual Cost 4-33
4.4.5 Example Problem 4-33
References 4-40
4-3
-------
4.1 Process Description
4.1.1 Introduction
In air pollution control, adsorption is employed to remove volatile organic compounds (VOCs)
from low to medium concentration gas streams, when a stringent outlet concentration must be
met and/or recovery of the VOC is desired. Adsorption itself is a phenomenon where gas
molecules passing through a bed of solid particles are selectively held there by attractive forces
which are weaker and less specific than those of chemical bonds. During adsorption, a gas
molecule migrates from the gas stream to the surface of the solid where it is held by physical
attraction releasing energy—the "heat of adsorption", which approximately equals the heat of
condensation. Adsorptive capacity of the solid for the gas tends to increase with the gas phase
concentration, molecular weight, difiusivity, polarity, and boiling point.
Some gases form actual chemical bonds with the adsorbent surface groups. This phenomenon
is termed "chemisorption".
Most gases ("adsorbates") can be removed ("desorbed") from the adsorbent by heating to a
sufficiently high temperature, usually via steam or (increasingly) hot combustion gases, or by
reducing the pressure to a sufficiently low value (vacuum desorption). The physically adsorbed
species in the smallest pores of the solid and the chemisorbed species may require rather high
temperatures to be removed, and for all practical purposes cannot be desorbed during
regeneration. For example, approximately 3 to 5 percent of organics adsorbed on virgin activated
carbon is either chemisorbed or very strongly physically adsorbed and, for all intents, cannot be
desorbed during regeneration. [1]
Adsorbents in large scale use include activated carbon, silica gel, activated alumina, synthetic
zeolites, fuller's earth, and other clays. This chapter is oriented toward the use of activated
carbon, a commonly used adsorbent for VOCs.
4.1.2 Types of Adsorbers
Five types of adsorption equipment are used in collecting gases: (1) fixed regenerable beds; (2)
disposable/rechargeable canisters; (3) traveling bed adsorbers; (4) fluid bed adsorbers; and (5)
chromatographic baghouses.[2] Of these, the most commonly used in air pollution control are
the fixed-bed and cannister types. This chapter addresses only fixed-bed and cannister units.
4-4
-------
4.1.2.1 Fixed-bed Units
Fixed-bed units can be sized for controlling continuous, VOC-containing streams over a wide
range of flow rates, ranging from several hundred to several hundred thousand cubic feet per
minute (cfin). The VOC concentration of streams that can be treated by fixed-bed adsorbers can
be as low as several parts per billion by volume (ppbv) in the case of some toxic chemicals or as
high as 25% of the VOCs1 lower explosive limit (LEL). (For most VOCs, the LEL ranges from
2500tolO,OOOppmv.[3])
Fixed-bed adsorbers may be operated in either intermittent or continuous modes. In
intermittent operation, the adsorber removes VOC for a specified time (the "adsorption time"),
which corresponds to the time during which the controlled source is emitting VOC. After the
adsorber and the source are shut down (e.g., overnight), the unit begins the desorption cycle
during which the captured VOC is removed from the carbon. This cycle, in turn, consists of three
steps: (1) regeneration of the carbon by heating, generally by blowing steam through the bed in
the direction opposite to the gas flow;1 (2) drying of the bed, with compressed air or a fan; and
(3) cooling the bed to its operating temperature via a fan. (In most designs, the same fan can be
used both for bed drying and cooling.) At the end of the desorption cycle (which usually lasts 1
to 1 Vz hours), the unit sits idle until the source starts up again.
In continuous operation a regenerated carbon bed is always available for adsorption, so that
the controlled source can operate continuously without shut down. For example, two carbon
beds can be provided: while one is adsorbing, the second is desorbing/idled. As each bed must
be large enough to handle the entire gas flow while adsorbing, twice as much carbon must be
provided than an intermittent system handling the same flow. If the desorption cycle is
significantly shorter than the adsorption cycle, it may be more economical to have three, four, or
even more beds operating in the system. This can reduce the amount of extra carbon capacity
needed or provide some additional benefits, relative to maintaining a low VOC content in the
effluent. (See Section 4.2 for a more thorough discussion of this.)
A typical two-bed, continuously operated adsorber system is shown in Figure 4.1. One of the
two beds is adsorbing at all times, while the other is desorbing/idled. As shown here, the VOC-
laden gas enters vessel #1 through valve A, passes through the carbon bed (shown by the shading)
and exits through valve B, from whence it passes to the stack. Meanwhile, vessel #2 is in the
desorption cycle. Steam enters through valve C, flows through the bed and exits through D. The
steam-VOC vapor mixture passes to a condenser, where cooling water condenses the entire
mixture. If part of the VOC is immiscible in water, the condensate next passes to a decanter,
where the VOC and water layers are separated. The VOC layer is conveyed to storage. If
impure, it may receive additional purification by distillation. Depending on its quality (i.e.,
quantity of dissolved organics), the water layer is usually discharged to a wastewater
1 Although steam is the most commonly used regenerant, there are situations where it should not be used.
An example would be a degreasing operation that emits halogenated VOCs. Steaming might cause the VOCs to
decompose.
4-5
-------
treatment facility.
Once steaming is completed, valves C and D are closed and valve E is opened, to allow au-
to enter to dry and cool the bed. After this is done, the bed is placed on standby until vessel #1
reaches the end of its adsorption cycle. At this time, the VOC-laden gas is valved to vessel #2,
while vessel #1 begins its desorption cycle, and the above process is repeated.
In Figure 4.1, the system fan is shown installed ahead of the vessels, though it could also be
placed after them. Further, this figure does not show the pumps needed to bring cooling water
to the condenser. Nor does it depict the solvent pump which conveys the VOC condensate to
storage. Also missing are preconditioning equipment used to cool, dehumidify, or remove
particulate from the inlet gases. Such equipment may or may not be needed, depending on the
condition of the inlet gas. In any case, preconditioning equipment will not be covered in this
chapter.
4.1.2.2 Cannister Units
Cannister-type adsorbers differ from fixed-bed units, in that they are normally limited to
controlling low-volume, (typically 100 ftVmin, maximum) intermittent gas streams, such as those
emitted by storage tank vents, where process economics dictate that either toll regeneration or
throw-away canisters are appropriate. The carbon canisters are not intended for desorption on-
site. However, the carbon may be regenerated at a central facility. Once the carbon reaches a
certain VOC content, the unit is shut down, replaced with another, and disposed of or regenerated
by the central facility. Each cannister unit consists of a vessel, activated carbon, inlet connection
and distributer leading to the carbon bed, and an outlet connection for the purified gas stream. [4]
In one design (Calgon's Ventsorb® ), 150 Ibs of carbon are installed on an 8-inch gravel bed, in
4-6
-------
(Ory1«/CMl
Al
Mutt fas
(Fn»
J_^_ 5onre«)
-£T-
£^*.
Syit« FM
(OrylM/ rvVi
CMtl*f Air) ~~~V«i^-
*t
4 1
Y«tl«l M
y///////////////////
su«— CXH
?
JJ <
Y«u«l 12
-0*-
;•
[•
W*
:
Out !•
«« »•»., T...I
CAM**a«r
„
wx. |
C»«*««Mt«-*— Q*c*nur |
, .
rr*c«atM)
1 r
' (To TrutMiit/
Figure 4.1. Typical-Two-Bed, Continuously Operated Fixed-Bed Carbon Adsorber
System
4-7
-------
a 55-gallon drum. The type of carbon used depends on the nature of the VOC to be treated.
In theory, a cannister unit would remain in service no longer than a regenerable unit would
stay in its adsorption cycle. Doing so would help to insure the allowable outlet concentration
from being exceeded. In reality, however, poor operating practice may result in the cannister
remaining connected until the carbon is near or at saturation. This is because: (1) the carbon (and
often the vessel) will probably be disposed of, so there is the temptation to operate it until the
carbon is saturated; and (2) unlike fixed-bed units, whose outlet VOC concentrations are usually
monitored continuously (via flame ionization detectors, typically), canisters are usually not
monitored. Thus, the user can only guess at the outlet loading, and could tend to leave a unit in
place longer.
4.1.3 Adsorption Theory
At equilibrium, the quantity of gas that is adsorbed on activated carbon is a function of the
adsorption temperature and pressure, the chemical species being adsorbed, and the carbon
characteristics, such as carbon particle size and pore structure. For a given adsorbent-VOC
combination at a given temperature, an adsorption isotherm can be constructed which relates the
mass of adsorbate per unit weight of adsorbent ("equilibrium adsorptivity") to the partial pressure
of the VOC in the gas stream. The adsorptivity increases with increasing VOC partial pressure
and decreases with increasing temperature.
A family of adsorption isotherms having the shape typical of adsorption on activated carbon
is plotted in Figure 4.2. This and other isotherms whose shapes are convex upward throughout,
are designated "Type I" isotherms. The Freundlich isotherm, which can be fit to a portion of a
Type I curve, is commonly used in industrial design. [2]
w = kP (4.1)
e m
where
we = equilibrium adsorptivity (lb adsorbate/lb adsorbent)
P = partial pressure of VOC in gas stream (psia)
k,m = empirical parameters
The treatment of adsorption from gas mixtures is complex and beyond the scope of this chapter.
Except where the VOC in these mixtures have nearly identical adsorption isotherms, one VOC
in a mixture will tend to displace another on the carbon surface. Generally, VOCs with lower
vapor pressures will displace those with higher vapor pressure, resulting in the former displacing
the latter previously adsorbed. Thus, during the course of the adsorption cycle the carbon's
capacity for a higher vapor pressure constituent decreases. This phenomenon should be
considered when sizing the adsorber. To be conservative, one would normally base the
adsorption cycle requirements on the least adsorbable component in a mixture and the desorption
cycle on the most adsorbable component. [1]
4-8
-------
The equilibrium adsorptivity is the maximum amount of adsorbate the carbon can hold at a
given temperature and VOC partial pressure. In actual control systems, however, the entire
carbon bed is never allowed to reach equilibrium. Instead, once the outlet concentration reaches
a preset limit (the "breakthrough concentration"), the adsorber is shut down for desorption or (in
the case of cannister units) replacement and disposal. At the point where the vessel is shut down,
the average bed VOC concentration may only be 50% or less of the equilibrium concentration.
That is, the carbon bed may be at equilibrium ("saturated") at the gas inlet, but contain only a
small quantity of VOC near the outlet.
As Equation 4.1 indicates, the Freundlich isotherm is a power function that plots as a straight
line on log-log paper. Conveniently, for the concentrations/partial pressures normally
encountered in carbon adsorber operation, most VOC-activated carbon adsorption conforms to
Equation 4.1. At very low concentrations, typical of breakthrough concentrations, a linear
approximation (on arithmetic coordinates) to the Freundlich isotherm is adequate. However, the
Freundlich isotherm does not accurately represent the isotherm at high gas concentrations and
thus should be used with care as such concentrations are approached.
Adsorptivity data for selected VOCs were obtained from Calgon Corporation, a vendor of
activated carbon.[5] The vendor presents adsorptivity data in two forms: a set of graphs
displaying equilibrium isotherms [5] and as a modification of the Dubinin-Radushkevich (D-R)
equation, a semi-empirical equation that predicts the adsorptivity of a compound based on its
adsorption potential and polarizability.[6] In this Manual, the modified D-R equation is referred
to as the Calgon fifth-order polynomial. The data displayed in the Calgon graphs [5] has beenfit
to the Freundlich equation. The resulting Freundlich parameters are shown in Table 4.1 for a
limited number of chemicals. The adsorbates listed include aromatics (e.g., benzene, toluene),
chlorinated aliphatics (dichloroethane), and one ketone (acetone). However, the list is far from
all-inclusive.
Notice that a range of partial pressures is listed with each set of parameters, k and m. (Note:
In one case (m-xylene) the isotherm was so curvilinear that it had to be split into two parts, each
with a different set of parameters.) This is the range to which the parameters apply.
Extrapolation beyond this range—especially at the high end—can introduce inaccuracy to the
calculated adsorptivity.
But high-end extrapolation may not be necessary, as the following will show. In most air
pollution control applications, the system pressure is approximately one atmosphere (14.696
psia). The upper end of the partial pressure ranges in Table 4.1 goes from 0.04 to 0.05 psia.
According to Dalton's Law, at a total system pressure of one atmosphere this corresponds to
an adsorbate concentration in the waste gas of 2,720 to 3,400 ppmv. Now, as discussed in
Section 4.1.2, the adsorbate concentration is usually kept at 25% of the lower explosive limit
(LEL)2. For many VOCs, the LEL ranges from 1 to 1.5 volume %, so that 25% of the LEL
Although, Factory Mutual Insurance will reportedly permit operation at up to 50% of the LEL, if proper VOC
monitoring is used.
4-9
-------
would be 0.25 to 0.375% or 2,500 to 3,750 ppmv, which approximates the high end of the
partial pressure ranges in Table 4.1.
Finally, each set of parameters applies to a fixed adsorption temperature, ranging from 77°
to 104° F. These temperatures reflect typical operting conditions, although adsorption can
take place as low as 32°F and even higher than 104°F. As the adsorption temperature
increases to much higher levels, however, the equilibrium adsorptivity decreases to such an
extent that VOC recovery by carbon adsorption may become economically impractical.
The Calgon fifth-order polynomial is somewhat more accurate than the Freundilich
parameters from Table 4.1. The polynomial contains a temperature parameter, and it allows
one to estimate adsorption isotherms for compounds not shown in Table 4.1 if pure
component data are available. The pure component data required are the saturation pressure,
liquid molar volume, and the refractive index. It is, however, somewhat more complex to use
than the Freundlich equation. The Calgon fifth-Oder polynomial is as follows:
Table 4.1: Parameters for Selected Adsorption Isotherms**
Adsorbate Adsorption Temp Isotherm
(°F) Parameters
k m
(1) Benzene
(2) Chlorobenzene
(3) Cyclohexane
(4) Dichloroethane
(5) Phenol
(6) Trichloroethane
(7) Vinyl Chloride
(8) m-Xylene
(9) Acrylonitrile
(10) Acetone
(11) Toluene
77
77
100
77
104
77
100
77
77
100
100
77
0.597
1.05
0.508
0.976
0.855
1.06
0.200
0.708
0.527
0.935
0.412
0.551
0.176
0.188
0.210
0.281
0.153
0.161
0.477
0.113
0.0703
0.424
0.389
0.110
Range of
isotherm*
(psia)
0.0001-0.05
0.0001-0.01
0.0001-0.05
0.0001-0.04
0.0001-0.03
0.0001-0.04
0.0001-0.05
0.0001-0.001
0.001-0.05
0.0001-0.015
0.0001-0.05
0.0001-0.05
' Reference [5].
' Each isotherm is of the form: w, = tP". (See text for definition of terms.) Dala are for adsorption on Calgon type "BPL" carbon.
' Equations should not be extrapolated outside these ranges
4-10
-------
o
1
Adsorhat* Partial Prwsur* (pala)
Figure 4.2. Type I Adsorption Isotherms for Hypothetical Adsorbate
4-11
-------
The mass loading, \ve, is calculated from
(Molecular Wt of Adsorbate)
where
we = mass loading, i.e., equilibrium adsorptivity (g adsorbate per g carbon)*
G = carbon loading at equilibrium (cm3 liquid adsorbate per 100 g carbon)
Vm = liquid molar volume of adsorbate (cm3 per g-mole).
Note that the terms in equation 4.2 are given in metric units, not English. This has been
done because the carbon loading, G, is calculated from a regression equation in which all the
terms are expressed in metric units. This equation for G is the Calgon fifth-order polynomial:
Logw(G) = AQ + A^Y + A2Y2 + A3Y3 + A^Y* + A^Y' (4.3)
where
AO = 1.71
A, = -1.46 x 10'2
A2 = -1.65 x lO'3
A3 = -4.11 x lo-4
A4 = +3.14 xlO-5
A5 = -6.75 x 10'7
and Y is calculated from several equations which follow.
The first step in calculating Y is to calculate X- This can be done by calculating the
adsorption potential, e:
e= RT ln(Ps/P.) (4.4)
where
R = 1.987 (calories per g-mole- °K
T = absolute temperature (°K)
i = aosoiuie temperature ^ j\j
Ps = vapor pressure of adsorbate at the temperature T (kPa)
P, = partial pressure of adsorbate (kPa).
This, of course, is equal to Ib absorbate per Ib carbon.
4-12
-------
The is calculated from:
By substituting for e in the above equation, x can alternatively be calculated from
The next step in calculating Y is to calculate the relative polarizability, F.
r = ®/@0
where
0, = po larizability of component /per unit volume, where component /is the adsorbate
Q0 = polarizability of component o per unit volume, where component o is the reference
component, n-heptane.
For the adsorbate or the reference compound, using the appropriate refractive index of
adsorbate, «, the polarizability is calculated from:
„ n2 - 1
n2 + 2
Once x and F are known, Y can be calculated from:
Y = $ (4-5)
Calgon also has a proprietary, seventh-order form in which two additional coefficients are
added to the Calgon fifth-order polynomial, but the degree of fit reportedly is improved only
modestly. [6] Additional sources of isotherm data include the activated carbon vendors,
handbooks (such as Perry's Chemical Engineer's Handbook), and the literature.
"Alternatively, if the available values for T, P,, P^ end Vm are in English units, they may be substituted into this
equation without conversion. However, to make the result dimensionally consistent with equation 4.3, it would
have to be multiplied by a conversion factor, 34.7.
4-13
-------
4.2 Design Procedure
4.2.1 Sizing Parameters
Data received from adsorber vendors indicate that the size and purchase cost of a fixed-bed or
cannister carbon adsorber system primarily depend on four parameters:
1. The volumetric flow of the VOC laden gas passing through the carbon bed(s);
2. The inlet and outlet VOC mass loadings of the gas stream;
3. The adsorption time (i.e,. the time a carbon bed remains on-line to adsorb VOC before
being taken off-line for desorption of the bed);
4. The working capacity of the activated carbon.
In addition, the cost could also be affected by other stream conditions, such as the
presence/absence of excessive amounts of particulate, moisture, or other substances which
would require the use of extensive pretreatment and/or corrosive-resistant construction
materials.
The purchased cost depends to a large extent on the volumetric flow (usually measured in
actual ftVmin). The flow, in turn, determines the size of the vessels housing the carbon, the
capacities of the fan and motor needed to convey the waste gas through the system, and the
diameter of the internal ducting.
Also important are the VOC inlet and outlet gas stream loadings, the adsorption time, and
the working capacity of the carbon. These variables determine the amount and cost of carbon
charged to the system initially and, in turn, the cost of replacing that carbon after it is
exhausted (typically, five years after startup). Moreover, the amount of the carbon charge
affects the size and cost of the auxiliary equipment (condenser, decanter, bed drying/cooling
fan), because the sizes of these items are tied to the amount of VOC removed by the bed. The
amount of carbon also has a bearing on the size and cost of the vessels.
A carbon adsorber vendor [7] supplied data that illustrate the dependency of the
equipment cost on the amount of the carbon charge. Costs were obtained for fixed-bed
adsorbers sized to handle three gas flow rates ranging from 4,000 to 100,000 scfrn and to
treat inlet VOC (toluene) concentrations of 500 and 5,000 ppm. Each adsorber was assumed
to have an eight-hour adsorption time. As one might expect, the equipment costs for units
4-14
-------
handling higher gas flow rates were higher than those handling lower gas flow rates.
Likewise, at each of the gas flow rates, the units sized to treat the 5,000 ppm VOC streams
had higher equipment costs than those sized to treat the 500 ppm concentration. These cost
differences ranged from 23 to 29% and averaged 27%. These higher costs were partly needed
to pay for the additional carbon required to treat the higher concentration streams. But some
of these higher costs were also needed for enlarging the adsorber vessels to accommodate the
additional carbon and for the added structural steel to support the larger vessels. Also, larger
condensers, decanters, cooling water pumps, etc., were necessary to treat the more
concentrated streams. (See Section 4.3.)
The VOC inlet loading is set by the source parameters, while the outlet loading is set by
the VOC emission limit. (For example, in many states, the average VOC outlet concentration
from adsorbers may not exceed 25 ppm.)
4.2.2 Determining Adsorption and Desorption Times
The relative times for adsorption and desorption and the adsorber bed configuration (i.e.,
whether single or multiple and series or parallel adsorption beds are used) establish the
adsorption/desorption cycle profile. The circle profile is important in determining carbon and
vessel requirements and in establishing desorption auxiliary equipment and utility
requirements. An example will illustrate. In the simplest case, an adsorber would be
controlling a process which emits a relatively small amount of VOC intermittently—say,
during one 8-hour shift per day. During the remaining 16 hours the system would either be
desorbing or on standby. Properly sized, such a system would only require a single bed,
which would contain enough carbon to treat eight hours worth of gas flow at the specified
inlet concentration, temperature, and pressure. Multiple beds, operating in parallel, would be
needed to treat large gas flows (>100,000 actual ftVmin, generally)[7], as there are practical
limits to the sizes to which adsorber vessels can be built. But, regardless of whether a single
bed or multiple beds were used, the system would only be on-line for part of the day.
However, if the process were operating continuously (24 hours), an extra carbon bed
would have to be installed to provide adsorptive capacity during the time the first bed is being
regenerated. The amount of this extra capacity must depend on the number of carbon beds
that would be adsorbing at any one time, the length of the adsorption period relative to the
desorption period, and whether the beds were operating in parallel or in series. If one bed
were adsorbing, a second would be needed to come on-line when the first was shut down for
desorption. In this case, 100% extra capacity would be needed. Similarly, if five beds in
parallel were operating in a staggered adsorption cycle, only one extra bed would be needed
and the extra capacity would be 20% (i.e., 1/5)—provided, of course, that the adsorption time
were at least five times as long as the desorption time. The relationship between adsorption
time, desorption time, and the required extra capacity can be generalized.
4-15
-------
(4.6)
where
Mc M = amounts of carbon required for continuous or intermittent control of a
given source, respectively (Ibs)
/ = extra capacity factor (dimensionless)
This equation shows the relationship between Mc and M . Section 4.2.3 shows how to
ci
calculate these quantities.
The factor,/, is related to the number of beds adsorbing (NA) and desorbing (ND) in a
continuous system as follows:
(Note: Nx is also the number of beds in an intermittent system that would be adsorbing at any
given time. The total number of beds in the system would be Nx + ND.)
It can be shown that the number of desorbing beds required in a continuous system (N0) is
related to the desorption time (0D), adsorption time (QA), and the number of adsorbing beds,
as follows:
(4.8)
(Note: 0D is the total time needed for bed regeneration, drying, and cooling.)
For instance, for an eight-hour adsorption time, in a continuously operated system of seven
beds (six adsorbing, one desorbing) 0D would have to be 1-1/3 hours or less (8 hours/6 beds).
Otherwise, additional beds would have to be added to provide sufficient extra capacity during
desorption.
4-16
-------
4.2.3 Estimating Carbon Requirement
4.2.3.1 Overview of Carbon Estimation Procedures
Obtaining the carbon requirement (Mc or AT ) is not as straightforward as determining the
ci
other adsorber design parameters. When estimating the carbon charge, the sophistication of
the approach used depends on the data and calculational tools available.
One approach for obtaining the carbon requirement is a rigorous one which considers the
unsteady-state energy and mass transfer phenomena occurring in the adsorbent bed. Such a
procedure necessarily involves a number of assumptions in formulating and solving the
problem. Such a procedure is beyond the scope of this Manual at the present time, although
ongoing work in the Agency is addressing this approach.
In preparing this chapter of the Manual, we have adopted a rule-of-thumb procedure for
estimating the carbon requirement. This procedure, while approximate in nature, appears to
have the acceptance of vendors and field personnel. It is sometimes employed by adsorber
vendors to make rough estimates of carbon requirement and is relatively simple and easy to
use. It normally yields results incorporating a safety margin, the size of which depends on the
bed depth (short beds would have less of a safety margin than deep beds), the effectiveness of
regeneration, the particular adsorbate and the presence or absence of impurities in the stream
being treated.
4.2.3.2 Carbon Estimation Procedure Used in Manual
The rule-of-thumb carbon estimation procedure is based on the "working capacity" (W^ Ib
VOC/lb carbon). This is the difference per unit mass of carbon between the amount of VOC
on the carbon at the end of the adsorption cycle and the amount remaining on the carbon at
the end of the desorption cycle. It should not be confused with the "equilibrium capacity"
(W,,) defined above in section 4.1.3. Recall that the equilibrium capacity measures the
capacity of virgin activated carbon when the VOC has been in contact with it (at a constant
temperature and partial pressure) long enough to reach equilibrium. In adsorber design, it
would not be feasible to allow the bed to reach equilibrium. If it were, the outlet
concentration would rapidly increase beyond the allowable outlet (or "breakthrough")
concentration until the outlet concentration reached the inlet concentration. During this
period the adsorber would be violating the emission limit.
The working capacity is some fraction of the equilibrium capacity. Like the equilibrium
adsorptivity, the working capacity depends upon the temperature, the VOC partial pressure,
4-17
-------
and the VOC composition. The working capacity also depends on the flow rate and the
carbon bed parameters.
The working capacity, along with the adsorption time and VOC inlet loading, is used to
compute the carbon requirement for a cannister adsorber or for an intermittently operated
fixed-bed adsorber as follows:
m
-J^-QA (4.9)
w A v '
where m^ = VOC inlet loading (Ib/h)
Combining this with Equations 4.6 and 4.7 yields the general equation for estimating the
system total carbon charge for a continuously operated system:
m
M = -
c
c w
c
Values for wc may be obtained from knowledge of operating units. If no value for wc is
available for the VOC (or VOC mixture) in question, the working capacity may be estimated
at 50% of the equilibrium capacity, as follows:
Wc X °-5"e(max) (4.11)
where M^^ = the equilibrium capacity (Ib VOCAb carbon) taken at the adsorber inlet
(i.e., the point of maximum VOC concentration).
(Note: To be conservative, this 50% figure should be lowered if short desorption cycles, very
high vapor pressure constituents, high moisture contents significant amounts of impurities, or
difficult-to-desorb VOCs are involved. Furthermore, the presence of strongly adsorbed
impurities in the inlet VOC stream may significantly shorten carbon life.)
As Equation 4.10 shows, the carbon requirement is directly proportional to the adsorption
time. This would tend to indicate that a system could be designed with a shorter adsorption
time to minimize the carbon requirement (and equipment cost). There is a trade-off here not
readily apparent from Equation 4.10, however. Certainly, a shorter adsorption time would
require less carbon. But, it would also mean that a carbon bed would have to be desorbed
more frequently. This would mean that the regeneration steam would have to be supplied to
the bed(s) more frequently to remove (in the long run) the same amount of VOC. Further,
4-18
-------
each time the bed is regenerated the steam supplied must heat the vessel and carbon, as well as
drive off the adsorbed VOC. And the bed must be dried and cooled after each desorption,
regardless of the amount of VOC removed. Thus, if the bed is regenerated too frequently, the
bed drying/cooling fan must operate more often, increasing its power consumption. Also,
more frequent regeneration tends to shorten the carbon life. As a rule-of-thumb, the optimum
regeneration frequency for fixed-bed adsorbers treating streams with moderate to high VOC
inlet loadings is once every 8 to 12 hours.fl]
4.3 Estimating Total Capital Investment
Entirely different procedures should be used to estimate the purchased costs of fixed-bed and
cannister-rype adsorbers. Therefore, they will be discussed separately.
4.3.1 Fixed-Bed Systems
As indicated in the previous section, the purchased cost is a function of the volumetric flow
rate, VOC inlet and outlet loadings, the adsorption time, and the working capacity of the
activated carbon. As Figure 4.1 shows, the adsorber system is made up of several different
items. Of these, the adsorber vessels and the carbon comprise from one-half to nearly 90% of
the total equipment cost. (See Section 4.3.1.3.) There is also auxiliary equipment, such as
fans, pumps, condensers, decanters, and internal piping. But because these usually comprise a
small part of the total purchased cost, they may be "factored" from the costs of the carbon and
vessels without introducing significant error. The costs of these major items will be
considered separately.
4.3.1.1 Carbon Cost
This cost (C^S) is simply the product of the initial carbon requirement (Mc) and the current
price of carbon. As adsorber vendors buy carbon in very large quantities (million-pound lots
or larger), their cost is somewhat lower than the list price. A typical vendor cost is $2.00/lb
(fall 1989 dollars).[8] Thus:
Cc - 2.00Mc (4.12)
4.3.1.2 Vessel Cost
The cost of an adsorber vessel is primarily determined by its dimensions which, in turn,
depend upon the amount of carbon it must hold and the superficial gas velocity through the
bed that must be maintained for optimum adsorption. The desired superficial velocity is used
to calculate the cross-sectional area of the bed perpendicular to the gas flow. An acceptable
superficial velocity is established empirically, considering desired removal efficiency, the
carbon particle size and bed porosity, and other factors. For example, one adsorber vendor
4-19
-------
recommends a superficial bed velocity of 85 ft/min[7], while an activated carbon manufacturer
cautions against exceeding 60 ft/min in systems operating at one atmosphere. [5] Another
vendor uses a 65 ft/min superficial face velocity in sizing its adsorber vessels. [8] Lastly, there
are practical limits to vessel dimensions which also influence their sizing. That is, due to
shipping restrictions, vessel diameters rarely exceed 12 feet, while their length is generally
limited to 50 feet. [8]
The cost of a vessel is usually correlated with its weight. However, as the weight is often
difficult to obtain or calculate, the cost may be estimated from the external surface area. This
is true because the vessel material cost—and the cost of fabricating that material—-is directly
proportional to its surface area. The surface area (S, ft2) of a vessel is a function of its length
(L, ft) and diameter (D, ft), which in turn, depend upon the superficial bed face velocity, the
L/D ratio, and other factors.
Most commonly, adsorber vessels are cylindrical in shape and erected horizontally (as in
Figure 4.1). Vessels configured in this manner are generally subjected to the constraint that
the carbon volume occupies no more than 1/3 of the vessel volume [7,8]. It can be shown
that this constraint limits the bed depth to no more than
Maximum bed depth = (4-13)
The vessel length, L, and diameter, D, can be estimated by solving two relationships, namely,
(1) the equation relating carbon volume, and thus vessel volume, to L and D, and (2) the
equation relating volumetric flow rate, superficial velocity, and cross-section normal to flow.
If one assumes that the carbon bulk density is 30 Ib/ft3, then one can show that:
0.127MV
Q'
-" * -
where
D = vessel diameter (ft)
L - vessel length (ft)
VA = bed superficial velocity (ft/min)
Mc ' = carbon requirement per vessel (Ibs)
Q ' = volumetric flow rate per adsorbing vessel (acfrn)
Because the constants in equations 4.14 and 4.15 are not dimensionless, one must be careful
4-20
-------
to use the units specified in these equations.
Although other design considerations can result in different values of L and D, these
equations result in L and D which are acceptable from the standpoint of "study" cost
estimation for horizontal, cylindrical vessels which are larger than 2-3 feet in diameter.
The carbon requirement and flow rate for each adsorber vessel can be calculated as
follows:
M
M' = e-
At gas flow rates (Q ') of less than 9,000 scfin, it is usually more feasible to erect the
adsorber vessels vertically instead of horizontally. [8] If so, the vessel diameter can be
calculated from the volumetric flow rate per adsorbing vessel and the bed superficial velocity
as follows:
4Q'
1/2
The vertical vessel length will depend principally on the carbon bed thickness. Additional
space must be included below the carbon bed for bed support and above and below the bed
for distribution and disengaging of the gas stream and for physical access to the carbon bed.
In smaller diameter vessels, access to both sides of the bed is usually not required. However,
\io\V2 feet must be provided on each side for gas distribution and disengagement, or 2 to 3
feet overall. For longer vessels, 2 to 3 feet at each end of the vessel is typically provided for
access space.
Given the mass of carbon in the bed, the carbon bulk density, and the bed diameter (i.e.,
the cross-sectional area normal to flow), determining the carbon bed thickness is straight
forward using the following equation:
volume of carbon _ Mc/Pb ,» •> -y\
b cross-sectional area nor~?.l to flow Q'/v
4-21
-------
where
pb = carbon bulk density (lb/ft3, assume 30 lb/ft3)
The vessel length is, therefore,
L = tb + ta,9 (4.18)
where
t^g = access / gas distribution allowance
= 2 to 6 feet (depending on vertical vessel diameter)
Finally, use the following equation to calculate the surface area of either a horizontal or
vertical vessel:
S = nD(L + D/2) (4.19)
Similar equations can be developed for other vessel shapes, configurations, etc.
Based on vendor data, we developed a correlation between adsorber vessel cost and
surface area: [8]
Cv = 271S°-778 (4.20)
where
Cv = vessel cost (fall 1989 $), F.O.B. vendor*
and 97 < S <; 2,110 ft2.
These units would be made of 304 stainless steel, which is the most common material used
in fabricating adsorber vessels. [7,8] However, to obtain the cost of a vessel fabricated of
another material, multiply Cv by an adjustment factor (Fm). A few of these factors are listed
below:
For information on escalating these prices to more current dollars, refer to the EPA report
alation Indexes for Air Pollution Control Costs and updates thereto, all of which are
led on the OAQPS Technology Transfer Network (CTC Bulletin Board).
^^^calc
*"'
4-22
-------
Material Fm Factor Reference(s)
Stainless steel, 316
Carpenter 20 CB-3
Monel-400
Nickel-200
Titanium
1.3
1.9
2.3
2
4.j
T7,8,9]
[9]
[7,9]
[9]
[9]
4.3.1.3 Total Purchased Cost
As stated earlier, the costs of such items as the fens, pumps, condenser, decanter,
instrumentation, and internal piping can be factored from the sum of the costs for the carbon
and vessels. Based on four data points derived from costs supplied by an equipment vendor
[8], we found that, depending on the total gas flow rate (Q), the ratio (Rc) of the total
adsorber equipment cost to the cost of the vessels and carbon ranged from 1.14 to 2.24.
These data points spanned a gas flow rate range of approximately 4,000 to 500,000 acfrn.
The following regression formula fit
these four points:
Rc = 5.82Q-0'133 (4.21)
where
4, 000 ± Q (acfrn) < 500,000
Correlation coefficient (r) = 0.872
The total adsorber equipment cost (C^) would be the product of R,, and the sum of the carbon
and vessel costs, or:
C = R [ C + (N + N ) C ] (A 22)
A C L C v A D' vj Ir*""1/
4.3.1.4 Total Capital Investment
As discussed in Chapter 2, in the methodology used in this Manual, the total capita
investment (TCI) is estimated from the total purchased cost via an overall direct/indL ct
installation cost factor. A breakdown of that factor for carbon adsorbers is shown ir, vble
4.2. As Chapter 2 indicates, the TCI also include^1 ^ost* for land, working capital, ar >ff-site
facilities, which are not included in the direct/indirect installation factor. However as >ese
items are rarely required with adsorber systems, they will not be considered here. Fus icr, no
factors have been provided for site preparation (SP) and buildings (Bldg.), as these site-
specific costs depend very little on the purchased equipment cost.
4-23
-------
Table 4.2: Capital Cost Factors for Carbon Adsorbers"
Cost Item Factor
Direct Coats
Purchased equipment costs
Adsorber + auxiliary equipment* As estimated, A
Instrumentation' 0.10 A
Sales taxes 0.03 A
Freight 0.05 A
Purchased equipment cost, PEC B = 1.18 A
Direct installation costs
Foundations & supports
Handling & erection
Electrical
Piping
Insulation
Painting
Direct installation costs
Site preparation As required, SP
Buildings As required, BIdg.
Total Direct Costs, DC 1.30 B + SP -f BIdg.
Indirect Costs (installation)
Engineering 0.10 B
Construction and field expenses 0.05 B
Contractor fees 0.10 B
5
-------
Note that the installation factor is applied to the total purchased equipment cost, which
includes the cost of such auxiliary equipment as the stack and external ductwork and such
costs as freight and sales taxes (if applicable). ("External ductwork" is that ducting needed to
convey the exhaust gas from the source to the adsorber system, and then from the adsorber to
the stack. Costs for ductwork and stacks are shown elsewhere in this Manual) Normally, the
adjustment would also cover the instrumentation cost, but this cost is usually included with
the adsorber equipment cost. Finally, note that these factors reflect "average" installation
conditions and could vary considerably, depending upon the installation circumstances.
4.3.2 Cannister Systems
Once the carbon requirement is estimated using the above procedure, the number of cannisters
is determined. This is done simply by dividing the total carbon requirement (Me) by the
amount of carbon contained by each cannister (typically, 150 Ibs.). This quotient, rounded to
the next highest digit, yields the required number of cannisters to control the vent in question.
Costs for a typical cannister (Calgon's Ventsorb®) are listed in Table 4.3. These costs
include the vessel, carbon, and connections, but do not include taxes, freight, or installation
charges.
Note that the cost per unit decreases as the quantity purchased increases. Each cannister
contains Calgon's "BPL" carbon (4x10 mesh), which is commonly used in industrial
adsorption. However, to treat certain VOCs, more expensive specialty carbons (e.g., "FCA 4
x 10") are needed. These carbons can increase the equipment cost by 60% or more. [4] As is
indicated in the caption of Table 4.3, these prices are in Spring 1986 dollars. Since then,
however, the prices of these cannisters have increased modestly—approximately 10%.[11]
As fewer installation materials and labor are required to install a cannister unit than a
fixed-bed system, the composite installation factor is consequently lower. The only costs
required are those needed to place the cannisters at, and connect them to, the source. This
involves a small amount of piping only; little or no electrical work, painting, foundations, or
the like would be needed. Twenty percent of the sum of the cannister(s) cost, freight charges,
and applicable sales taxes would cover this installation cost.
4.4 Estimating Total Annual Cost
As Chapter 2 of this Manual explains, the total annual cost is comprised of three components:
direct costs, indirect costs, and recovery credits. These will be considered separately.
4-25
-------
Table 4.3: Equipment costs (Spring 1986 $) for a Typical Cannister Adsorber0
Quantity
1-3
4-9
10-29
;>30
Equipmen
tCost
(each)*
$687
659
622
579
" Reference [4].
* These costs are F.O.B., Pittsburgh, PA. They
do not include taxes and freight charges.
4.4.1 Direct Annual Costs
These include the following expenditures: steam, cooling water, electricity, carbon
replacement, operating and supervisor labor, and maintenance labor and materials. Of these,
only electricity and solid waste disposal would apply to the cannister-type adsorbers.
4.4.1.1 Steam
As explained in section 4.1, steam is used during the desorption cycle. The quantity of steam
required will depend on the amount of carbon in the vessel, the vessel dimensions, the type
and amount of VOC adsorbed, and other variables. Experience has shown that the steam
requirement ranges from approximately 3 to 4 Ibs of steam/lb of adsorbed VOC. [7,8] Using
the midpoint of this range, we can develop the following expression for the annual steam cost:
Cs = 3.50 x icr3™ 6 ps (4.23)
where
Cs = steam cost ($/yr)
0j = system operating hours (h/yr)
wvoc = VOC inlet loading (Ibs/h)
p, = steam price ($/thous. Ibs)
If steam price data are unavailable, one can estimate its cost at 120% of the fuel cost. For
example, if the local price of natural gas were $5.00/million BTU, the estimated steam price
4-26
-------
would be $6.00/million BTU which is approximately $6.00/thousand Ibs. (The 20% factor
covers the capital and annual costs of producing the steam.)
4.4.1.2 Cooling Water
Cooling water is consumed by the condenser in which the steam-VOC mixture leaving the
desorbed carbon bed is totally condensed. Most of the condenser duty is comprised of the
latent heat of vaporization (AHJ of the steam and VOC. As the VOC At^ are usually small
compared to the steam AH^ (about 1000 BTU/lb), the VOC AH,, may be ignored. So may
the sensible heat of cooling the water-VOC condensate from the condenser inlet temperature
(about 212°F) to the outlet temperature. Therefore, the cooling water requirement is
essentially a function of the steam usage and the allowable temperature rise in the coolant,
which is typically 30° to 40°F.[7] Using the average temperature rise (35°F), we can write:
(4-24)
where
Cw = cooling water cost ($/yr)
p^ = cooling water price ($/thous. gal.)
If the cooling water price is unavailable, use $0.15 to $0.30/thousand gallons.
4.4.1.3 Electricity
In fixed-bed adsorbers, electricity is consumed by the system fan, bed drying/cooling fan,
cooling water pump, and solvent pump(s). Both the system and bed fans must be sized to
overcome the pressure drop through the carbon beds. But, while the system fan must
continuously convey the total gas flow through the system, the bed cooling fan is only used
during a part of the desorption cycle (one-half hour or less).
For both fans, the horsepower needed depends both on the gas flow and the pressure drop
through the carbon bed. The pressure drop through the bed (AP6) depends on several
variables, such as the adsorption temperature, bed velocity, bed characteristics (e.g., void
fraction), and thickness. But, for a given temperature and carbon, the pressure drop per unit
thickness depends solely on the gas velocity. For instance, for Calgon's "PCB" carbon (4x10
mesh), the following relationship holds:[5]
APb/tfo = 0.03679vb + 1.1.07 xiO'V* (4.25)
4-27
-------
where
APj/tj = pressure drop through bed (inches of water/foot of carbon)
v6 = superficial bed velocity (ft/rnin)
As Equation 4.17 shows, the bed thickness (tA, ft) is the quotient of the bed volume (VA)
and
the bed cross-sectional area (AA). For a 30 lb/ft3 carbon bed density, this becomes:
V. 0.0333M'
*. - • —(4.26)
(For vertically erected vessels, Ab = Q 7vb, while for horizontally erected cylindrical vessels,
A«LD.) Once APA is known, the system fan horsepower requirement (hpsf) can be calculated:
hpsf = 2.50 x 1Q-4 QAPs (4.27)
where
Q = gas volumetric flow through system (acfin)
= total system pressure drop = AP6 + 1
(The extra inch accounts for miscellaneous pressure losses through the external ductwork and
other parts of the system. [7]* However, if extra long duct runs and/or preconditioning
equipment are needed, the miscellaneous losses could be much higher.)
This equation incorporates a fan efficiency of 70% and a motor efficiency of 90%, or 63%
overall.
The horsepower requirement for the bed drying/cooling fan (hpcf) is computed similarly.
While the bed fan pressure drop would still be APA, the gas flow and operating times would be
different. For typical adsorber operating conditions, the drying/cooling air requirement would
be 50 to 150 ftVlb carbon, depending on the bed moisture content, required temperature drop,
and other factors. The operating time (6cf) would be the product of the drying/coating time
per desorption cycle and the number of cycles per year. It can be shown that:
*To obtain a more precise estimate of ductwork pressure drop, refer to Chapter 10 of this Manual.
4-28
-------
(The "0.4" allows for the fact that as a rule-of-thumb, approximately 40% of the desorption
cycle is used for bed drying/cooling.)
The cooling water pump horsepower requirement (hpcwp) would be computed as follows:
2.52 x I0"4g Hs
hPcw = C" (4-29)
where
q^, = cooling water flow (gal/min)
H = required head (nominally 100 feet of water)
s = specific gravity of fluid relative to water at 60 °F
TJ = combined pump-motor efficiency.
The annual operating hours for the cooling water pump (0cwp) would be computed using
Equation 4.28, after substituting "0.6" for 0.4. The 0.6 factor accounts for the fact that the
cooling water pump is only used during the steaming portion of the regeneration, while the
condenser is in operation.
Equation 4.29 may also be used to compute the solvent pump horsepower requirement.
In the latter case, the flow (q,) would be different of course, although the same head—100 ft.
of water—could be used. The specific gravity would depend on the composition and
temperature of the condensed solvent. For example, the specific gravity of toluene at 100°F
would be approximately 0.86 at 70°F. (However, the solvent pump horsepower is usually
very small—usually < 0.1 hp.—so its electricity consumption can usually be neglected.)
Once the various horsepowers are calculated, the electricity usage (in kWh) is calculated,
by multiplying each horsepower value by 0.746 (the factor for converting hp to kilowatts) and
the number of hours each fan or pump operates annually. For the system fan, the hours would
be the annual operating hours for the system (6S). But, as discussed above, the operating
times for the bed drying/cooling fan and cooling water pump would be different.
To obtain the annual electricity cost, simply multiply kWh by the electricity price (in
$/kWh) that applies to the facility being controlled.
For cannister units, use equation 4.27 to calculate the fan horsepower requirement.
However, instead of AP6 use the following to compute the total cannister pressure drop (APe
inches of water): [4]
4-29
-------
APc = 0.0471Qc + 9.29 * 10~4Q* (4.30)
where Qc = flow through the cannister (acfin).
4.4.1.4 Carbon Replacement
As discussed above, the carbon has a different economic life than the rest of the adsorber
system. Therefore, its replacement cost must be calculated separately. Employing the
procedure detailed in Chapter 2, we have:
CRCc = CRFc (1. 08Cc + Ccl) (4.31)
where
CRFC = capital recovery factor for the carbon
1.08 = taxes and freight factor
GC, Cc/ = initial cost of carbon (F.O.B. vendor) and carbon replacement labor cost,
respectively ($)
The replacement labor cost covers the labor cost for removing spent carbon from vessels
and replacing it with virgin or regenerated carbon. The cost would vary with the amount of
carbon being replaced, the labor rates, and other factors. For example, to remove and replace
a 50,000 pound carbon charge would require about 16 person-days, which, at typical wage
rates, is equivalent to approximately $0.05/lb replaced. [12]
A typical life for the carbon Is five years. However, if the inlet contains VOCs that are
very difficult to desorb, tend to polymerize, or react with other constituents, a shorter carbon
lifetime—perhaps as low as two years—would be likely. [1] For a five-year life and 7% interest
rate, CRFC = 0.2439.
4.4.1.5 Solid Waste disposal
Disposal costs are rarely incurred with fixed-bed adsorbers, because the carbon is almost
always regenerated in place, not discarded. In certain cases, the carbon in cannister units is
also regenerated, either off-site or at a central regeneration facility on-site. However, most
cannister adsorbers are disposed of once they become saturated. The entire
cannister—carbon, drum, connections, etc.—is shipped to a secure landfill. The cost of
landfill disposal could vary considerably, depending on the number of cannisters disposed of,
the location of the landfill, etc. Based on data obtained from two large landfills, for instance,
the disposal cost would range from approximately $35 to $65 per cannister excluding
4-30
-------
transportation costs. [ 13,14]
4.4.1.6 Operating and Supervisory Labor
The operating labor for adsorbers is relatively low, as most systems are automated and require
little attention. One-half operator hour per shift is typical. [10] The annual labor cost would
then be the product of this labor requirement and the operating labor wage rate ($/h) which,
naturally, would vary according to the facility location, type of industry, etc. Add to this 15%
to cover supervisory labor, as Chapter 2 suggests.
4.4.1.7 Maintenance Labor and Materials
Use 0.5 hours/shut for maintenance labor [10] and the applicable maintenance wage rate. If
the latter data are unavailable, estimate the maintenance wage rate at 110% of the operating
labor rate, as Chapter 2 suggests. Finally, for maintenance materials, add an amount equal to
the maintenance labor, also per Chapter 2.
4.4.2 Indirect Annual Costs
These include such costs as capital recovery, property, taxes, insurance, overhead, and
administrative costs ("G&A"). The capital recovery cost is based on the equipment lifetime
and the annual interest rate employed. (See Chapter 2 for a thorough discussion of the capital
recovery cost and the variables that determine it.) For adsorbers, the system lifetime is
typically ten years, except for the carbon, which, as stated above, typically needs to be
replaced after five years. Therefore, when figuring the system capital recovery cost, one
should base it on the installed capital cost less the cost of replacing the carbon {i.e., the
carbon cost plus the cost of labor necessary to replace it). Substituting the initial carbon and
replacement labor costs from equation 4.31, we obtain:
CRCs = (TCI - (1.08Cc + Ccl)] CRFs (4.32)
where
CRCS = capital recovery cost for adsorber system ($/yr)
TCI = total capital investment ($)
1.08 = taxes and freight factor
C^C,,, = initial carbon cost (F.O.B. vendor) and carbon replacement cost,
respectively ($)
CRF, = capital recovery factor for adsorber system (defined in Chapter 2).
For a ten-year life and a 7% annual interest rate, the CRFS would be 0.1424.
4-31
-------
As Chapter 2 indicates, the suggested factor to use for property taxes, insurance, and
administrative charges is 4% of the TCI. Finally, the overhead is calculated as 60% of the
sum of operating, supervisory, and maintenance labor, and maintenance materials.
The above procedure applies to cannister units as well, except that, in most cases, the
carbon is not replaced—the entire unit is. Cannisters are generally used in specialized
applications. The piping and ducting cost can usually be considered a capital investment with
a useful life often years. However, whether the cannister itself would be treated as a capital
or an operating expense would depend on the particular application and would need to be
evaluated on a case-by-case basis.
4.4.3 Recovery Credits
These apply to the VOC which is adsorbed, then desorbed, condensed, and separated from the
steam condensate. If the recovered VOC is sufficiently pure, it can be sold. However, if the
VOC layer contains impurities or is a mixture of compounds, it would require further
treatment, such as distillation. Purification and separation costs are beyond the scope of this
chapter. Needless to say, the costs of these operations would offset the revenues generated
by the sale of the VOC. Finally, as an alternative to reselling it, the VOC could be burned as
fuel and valued accordingly. In any case, the following equation can be used to calculate
these credits:
RC = m..._8_P__,_E (4.33)
VOC S VOC
where
RC = recovery credit ($/yr)
m^ = VOC inlet loading (Ibs/h)
0S = system operating hours (h/yr)
Pnc = resale value of the recovered VOC ($/lb)
E = adsorber VOC control efficiency
By definition, the efficiency (E) is the difference between the inlet and outlet VOC mass
loading, divided by the inlet loading. However, during an adsorption cycle the outlet VOC
loading will increase from essentially zero at the start of the cycle to the breakthrough
concentration at the end of the cycle. Because the efficiency is a function of time, it should be
calculated via integration over the length of the absorption cycle. To do this would require
knowledge of the temporal variation of the outlet loading during the adsorption cycle. If this
knowledge is not available to the Manual user, a conservative approximation of the efficiency
may be made by setting the outlet loading equal to the breakthrough concentration.
4-32
-------
4.4.4 Total Annual Cost
Finally, as explained in Chapter 2, the total annual cost (TAG) is the sum of the direct and
indirect annual costs, less any recovery credits, or:
TAG = DC + 1C - RC (4.34)
4.4.5 Example Problem
A source at a printing plant emitting 100 Ib/h of toluene is to be controlled by a carbon
adsorber. The plant proposes to operate the adsorber in a continuous mode for 8,640 h/yr
(360 days). While operating, two carbon beds will be adsorbing, while a third will be
desorbing/on stand by. For its convenience, the plant has selected adsorption and desorption
times of 12 and 5 hours, respectively. The total waste gas flow is 10,000 acfin at the adsorber
inlet conditions (one atmosphere and 77°F). The waste gas contains negligible quantities of
particulate matter and moisture. Further, the applicable VOC regulation requires the adsorber
to achieve a mean removal efficiency of 98% during the entire adsorption cycle. Finally,
assume that the recovered toluene is recycled at the source. Estimate the total capital
investment and total annual cost for the adsorber system.
Carbon Working Capacity: At the stated flow and pollutant loading, the toluene inlet
concentration is 710 ppm. This corresponds to a partial pressure of 0.0104 psia. Substituting
this partial pressure and the toluene isotherm parameters (from Table. 4.1) into equation 4.1,
we obtain an equilibrium capacity of 0.333 Ib/lb. By applying the rule-of-thumb discussed
above (page 4-19), we obtain a working capacity of 0.167 Ib/lb (i.e., 0.333/2).
Carbon Requirement: As stated above, this adsorber would have two beds on-line and a
third off-line. Is this a reasonable assumption? Equation 4.8 can answer this question.
Substitution of the adsorption time and numbers of adsorbing and desorbing beds yields:
Desorption time = 9 < Q(Nn/NJ = 12 h (1/2) = 6 h.
D A D A
Because the stated desorption time (5 hours) is less than 6 hours, the proposed bed
configuration is feasible. Next, calculate the carbon requirement (Mc) from equation 4 0:
4-33
-------
m
M =
w
n
NL
~W.
_ 1001b/h (12h)
1/2)
0.1671b/lb
= 10,800 Ibs
From equation 4.12, the carbon cost is:
Cc = 2.00M, - $21,600.
Adsorber Vessel Dimensions and Cost: Assume that the vessels will be erected horizontally
and select a superficial bed velocity (v6) of 75 ft/min. Next, calculate the vessel diameter (D),
length (L), and surface area (S) from equations 4.14,4.15, and 4.19, respectively. [Note: In
these equations, M 'c = Mc (NA + NJ = 3,600 Ib and Q' = Q/NA = 5,000 acfra]
__ 0.127M vb=0.127(3,600)(75)
Q'
5,000
L =
7.87
M'
7.87 5,000
3,600 I 75
- 9.72 ft
S = nD(L + D/2) = 283 ft'
Because S falls between 97 and 2,110 ft2, equation 4.20 can be used to calculate the cost per
vessel, Cv (assuming 304 stainless steel construction). Thus:
Cv = 2715
0.778 _
= $21,900
Adsorber Equipment Cost: Recall that the adsorber equipment cost is comprised of the
adsorber vessels, carbon, and the condenser, decanter, fan, pumps and other equipment
usually included in the adsorber price. The cost of the latter items are "factored" from the
combined cost of the vessels and carbon. Combining equations 4.21 and 4.22, we have:
CA = 5.82Q
-0.133
[C + (N + N } C ]
^ C ^ A D' vj
4-34
-------
Substitution of the above values yields:
CA = $149,200
Cost of Auxiliary Equipment: Assume that costs for the following auxiliary equipment have
been estimated from data in other parts of the Manual:
Ductwork $16,500
Dampers 7,200
Stack 8,500
Total $32,200
Total Capital Investment: The total capital investment is factored from the sum of the
adsorber unit and auxiliary equipment cost, as displayed in Table 4.4. Note that no line item
cost has been shown for instrumentation, for this cost is typically included in the adsorber
price.
Therefore:
Purchased Equipment Cost = "B" = 1.08 x "A"
= 1.08 x ($149,300 + $32,200) = $196,000
And:
Total Capital Investment (rounded) = 1.61 x "B" = $316,000
Annual Costs: Table 4.5 gives the direct and indirect annual costs for the carbon adsorber
system, as calculated from the factors in Section 4.4. Except for electricity, the calculations in
the table show how these costs were derived. The following discussion will deal wit! he
electricity cost.
First, recall that the electricity includes the power for the system fan, bed drying/c ing
fan, and the cooling water pump. (The solvent pump motor is normaiiy so small that its
power consumption may be neglected.) These consumptions are calculated as follows:
• System fan: From equation 4.27:
4-35
-------
Table 4.4: Capital Costs for Carbon Adsorber System
Example Problem
Tn<:t Item
Direct Costs
Purchased equipment costs
Adsorber vessels and carbon
Auxiliary equipment
Sume = A
Instrumentation, 0.1 Aa
Sales taxes, 0.03 A
Freight, 0.05A
Purchased equipment cost, B
Direct installation costs
Foundation and supports, 0.08B
Handling & erection, 0.14B
Electrical, 0.04B
Piping, 0.02B
Insulation for ductwork, 0.0IB
Painting, 0.01B
Direct installation cost
Site preparation
Facilities and building
Total Direct Cost
Indirect Costs (installation)
Engineering, 0.1 OB
Construction and field expenses, 0.05B
Contractor fees, 0.1 OB
Start-up, 0.02B
Performance test, 0.0IB
Contingencies, 0.03B
Total Indirect Cost
$149,300
32.200
$181,500
5,450
9.080
$196,000
15,680
27,440
7,840
3,920
1,960
1.960
$58,800
$254,800
19,600
9,800
19,600
3,920
1,960
5.880
$ 60,760
Total Capital Investment (rounded)
$316,000
"The cost for this is included in the adsorber equipment cost.
4-36
-------
kWh = 0.746kW/hp x 2.50 * 10"4QAP x
s s
But:
AP (inches water) = AP + 1 = tK(0.03679vK + 1.107 x 10~4vf) + 1
s sob b
(The latter expression was derived from equation 4.25, assuming that the carbon used in this
example system is Calgoris "PCB", 4x10 mesh size.)
By assuming a carbon bed density, of 30 lb/ft3, Equation 4.26 can be used to calculate the bed
thickness (tb):
0.0333 M1 0.0333 M1
Bed thickness = t = = = 1.80 ft
A LD
D
Thus:
APs = 1 + 1.80 (0.03679 x 75 + 1.107 x 10'4 x752) = 7.09 inches
and finally:
kWhsf = 0.746 x 2.50 x 10"4 x 7.09 in. x 10,000 acfm x 8640 h/yr
kWhsf = 114,200 kWh/yr
• Bed drying/cooling fan: During the drying/cooling cycle, the pressure drop through the
bed also equals APb. However, as section 4.4.1.3 indicates, the flow and operating time
are different. For the air flow, take the midpoint of the range given on page 4-30 (100 ft3
air/lb carbon) and divide by 2 hours (the bed drying/cooling time), yielding: 100 ft3/lb x
3,600 Ibs x 1/120 min = 3,000 acfrn. Substituting this into equation 4.27 results in:
2.50 x KT* x 7.09 inches x 3,000 acfrn - 5.32 hp
From equation 4.28, we get:
0cf = (0.4)(5 h)(2)(8,640 h)/12 h = 2,880 h
4-37
-------
Thus:
rf = 0.746 kW/hp x 5.32 hp x 2,880 h = 11,400 kWh/yr
Cooling water pump: The cooling water pump horsepower is calculated from equation
4.29. Here, let r) = 63% and H = 100 ft. The cooling water flow (qcw) is the quotient of
the annual cooling water requirement and the annual pump operating time. From the data
in Table 4.5, we obtain the cooling water requirement: 10,400,000 gal/yr. The pump
annual operating time is obtained from equation 4.28 (substituting 0.6 for 0.4), or 0cwp =
(0.6)(5 h)(2X8,640)/12 = 4,320 h/yr.
Thus:
h = (2-52 x 10"4) (100ft) x 10,400,000 gal/yr = ! 60 h
Pcwp 0.63 X 4,320 h/yr x 60 min/yr
And:
kWhcwp - 0.746 kW/h x 1.60 hp x 4,320 h/yr = 5,160 kWh/yr
Summing the individual power consumptions, we get the value shown in Table
4.5:131,OOOkWh/yr Recovery Credit: As Table 4.5 indicates, a credit for the recovered
toluene has been taken. However, to account for miscellaneous losses and contamination, the
toluene is arbitrarily valued at one-half the November 1989 market price of $0.0533/lb(=
$lll/ton).[15]
Total Annual Cost: The sum of the direct and indirect annual costs, less the toluene
recovery credit, yields a net total annual cost of $76,100. Clearly, this "bottom line" is very
sensitive to the recovery credit and, in turn, the value given the recovered toluene. For
instance, if it had been valued at the full market price ($221/ton), the credit would have
doubled and the total annual cost would have been $29,200. Thus when incorporating
recovery credits, it is imperative to select the value of the recovered product carefully.
4-38
-------
Table 4.5: Annual Costs for Carbon Adsorber System
Example Problem
Cost Item
Calculations
Cost
Direct Annual Costs. DC
Operating Labor
Operator 0.5h/shift x 3 shi/day x 360 days/yr x $ 12/h
Supervisor
15% of operator = .15 x 6,480
Operating materials
Maintenance
Labor 0.5h/shift x 3sh/day x 360days/yr x $ 13.20/hr
Material 100% of maintenance labor
Replacement parts, carbon (5-year life)
Replacement labor
Carbon cost"
Utilities
Electricity
Steam
Cooling water
Total DC
Indirect Annual Costs. 1C
Overhead
Administrative charges
Property tax
Insurance
Capital recovery1'
Total 1C
0.2439 ($0.05/lbxl0,800 Ib)
0.2439 ($21,600x1.08)
$0.06/kWh x 131,000k Wh/yr
3.51b/lbVOCx $6/103lb x lOOlbVOC/h x 8640h/yr
3.43gal/lbsteam x (3.5 x 100 x 8640Mb steam x $0.20/103gal
y
$6,480
970
7,130
7,130
130
5,690
7,860
18,140
2.070
60% of sum of operating, supv., & maint, labor
& maint. materials = 0.6 (6,480 + 970 + 7,130 +
7,130)
2% of Total Capital Investment = 0.02($316,000)
1% of Total Capital Investments 0.01($316,000)
1% of Total Capital Investment = 0.01($316,000)
0.1424 [316,000 - 0.05(10,800) -1.08(21,600)]
Recovery Credit (toluene)
Total Annual Cost (rounded)
$55,600
13,030
6,320
3,160
3,160
41.600
$67,270
(46.8201
$76,100
* The 1.08 factor is for freight and sales taxes.
b The capital recovery cost factor, CRF, is a function of the adsorber or equipment life and the opportunity cost of
the capital (i.e., interest rate). For example, for a 10 year equipment life and a 7% interest rate, CRF = 0.1424.
4-39
-------
References
[1] Correspondence: Robert L. Stallings and William Klotz (Research Triangle Institute,
Research Triangle Park, NC) to William M. Vatavuk (U.S. EPA, OAQPS, Research
Triangle Park, NC), June 24,1986.
[2] Calvert, Seymour and Englund, Harold M. (eds.), Handbook of Air Pollution Control
Technology, John Wiley & Sons, New York, 1984, pp. 135-192.
[3] Handbook of Chemistry and Physics, 54th Edition, The Chemical Rubber Company,
Cleveland, 1973-74, pp. D85-D92.
[4] "Calgon Ventsorb® for Industrial Air Purification" (Bulletin 23-56a), Calgon
Corporation, Pittsburgh, 1986.
[5] Adsorption Handbook, Calgon Corporation, Pittsburgh, 1980.
[6] Rogers, Tony, "Comparison of BED SIZE and Calgon Adsorption Isotherms",
Research Triangle Institute (Research Triangle Park, NC), January 20, 1988.
[7] Correspondence: Richard Selznick (Baron Blakeslee, Inc., Westfield. NJ) to William
M. Vatavuk (U.S. EPA, OAQPS, Research Triangle Park, NC), April 23, 1986.
[8] Correspondence: Denny Clodfelter(M&W Industries, Inc., Rural Hall, NC) to William
M. Vatavuk (U.S. EPA, OAQPS, Research Triangle Park, NC), September 25, 1989.
[9] Matley, Jay (ed.), Modern Cost Engineering, McGraw-Hill Publications Co., New
York, 1984, p. 142.
[ 10] Vatavuk, William M. and Neveril, Robert, "Estimating Costs of Air Pollution Control
Systems, Part II: Factors for Estimating Capital and Operating Costs," Chemical
Engineering, November 3, 1980, pp. 157-162.
[11] Telephone conversation: Robert Bradley (Calgon Corporation, Charlotte, NC) with
William M. Vatavuk (U.S. EPA, OAQPS, Research Triangle Park, NC), December 5,
1989.
[12] Telephone conversation: Robert L. Stallings (Research Triangle Institute, Research
Triangle Park, NC) with William M. Vatavuk (U.S. EPA, OAQPS, Research Triangle
Park, NC), September 11, 1986.
4-40
-------
[13] Correspondence: William Kitto (Chemwaste, Sulphur, LA) to William M. Vatavuk
(U.S. EPA, OAQPS, Research Triangle Park, NC), July 25, 1986.
[14] Correspondence: Jerry Locklear (GSX, Pinewood, SC) to William M. Vatavuk (U.S.
EPA, OAQPS, Research Triangle Park, NC), July 25, 1986.
[15] Chemical Marketing Reporter, December 2, 1989.
4-41
-------
Chapter 5
FABRIC FILTERS
James H. Turner
Andrew S. Viner
Research Triangle Institute
Research Triangle Park, NC 22709
JohnD. McKenna
ETS, Inc.
Roanoke, VA 24018-4394
Richard E. Jenkins
William M. Vatavuk
Innovative Strategies and Economics Group, OAQPS
U.S. Environmental Protection Agency
Research Triangle Park, NC 22711
December 1995
5-1
-------
Contents
5.1 Process Description 5-5
5.1.1 Introduction 5-5
5.1.2 Types of Fabric Filters 5-6
5.1.2.1 Shaker Cleaning 5-6
5.1.2.2 Reverse-air Cleaning 5-7
5.1.2.3 Pulse-jet Cleaning 5-7
5.1.3 Auxiliary Equipment 5-8
5.1.4 Fabric Filtration Theory 5-8
5.1.4.1 Reverse Air/Shake Deflate Baghouses 5-10
5.1.4.2 Pulse-Jet Baghouses 5-12
5.2 Design Procedures 5-14
5.2.1 Gas-to-Cloth Ratio 5-14
5.2.1.1 Gas-to-Cloth Ratio From Similar Applications 5-17
5.2.1.2 Gas-to-Cloth Ratio From Manufacturer's Methods 5-17
5.2.2 Pressure Drop 5-21
5.2.3 Particle Characteristics 5-21
5.2.4 Gas Stream Characteristics 5-21
5.2.4.1 Temperature 5-22
5.2.4.2 Pressure 5-22
5.2.5 Equipment Design Considerations 5-22
5-2
-------
5.2.5.1 Pressure or Suction Housings 5-22
5.2.5.2 Standard or Custom Construction 5-24
5.2.5.3 Filter Media 5-25
5.3 Estimating Total Capital Investment 5-25
5.3.1 Equipment Cost 5-27
5.3.1.1 Bare Baghouse Costs 5-27
5.3.1.2 Bag Costs 5-35
5.3.1.3 Auxiliary Equipment 5-35
5.3.2 Total Purchased Cost 5-35
5.3.3 Total Capital Investment 5-35
5.4 Estimating Total Annual Costs 5-38
5.4.1 Direct Annual Cost 5-38
5.4.1.1 Operating and Supervisory Labor 5-38
5.4.1.2 Operating Materials 5-38
5.4.1.3 Maintenance 5-38
5.4.1.4 Replacement Parts 5-38
5.4.1.5 Electricity 5-39
5.4.1.6 Fuel 5-40
5.4.1.7 Water 5-40
5.4.1.8 Compressed Air 5-40
5.4.1.9 Dust Disposal 5-40
5.4.2 Indirect Annual Cost 5-40
5-3
-------
5.4.3 Recovery Credits 5-41
5.4.4 Total Annual Cost 5-41
5.4.5 Example Problem 5-42
5.5 Acknowledgments 5-47
References 5-48
5-4
-------
5.1 Process Description
5.1.1 Introduction
A fabric filter unit consists of one or more isolated compartments containing rows of fabric
filter bags or tubes. Particle-laden gas passes up (usually) along the surface of the bags then
radially through the fabric. Particles are retained on the upstream face of the bags, while the
cleaned gas stream is vented to the atmosphere. The filter is operated cyclically alternating
between relatively long periods of filtering and short periods of cleaning. During cleaning,
dust that has accumulated on the bags is removed from the fabric surface and deposited in a
hopper for subsequent disposal.
Fabric filters will collect particle sizes ranging from submicron to several hundred microns
in diameter at efficiencies generally in excess of 99 or 99.9 percent. The dust cake collected
on the fabric is primarily responsible for such high efficiency. Gas temperatures up to about
500°F, with surges to about 550°F can be accommodated routinely. Most of the energy used
to operate the system appears as pressure drop across the bags and associated hardware and
ducting. Typical values of pressure drop range from about 5 to 20 inches of water. Fabric
filters are used where high-efficiency particle collection is required. Limitations are imposed
by gas characteristics (temperature and corrosivity) and particle characteristics (primarily
stickiness) that affect the fabric or its operation and that cannot be economically
accommodated.
Important process variables include particle characteristics, gas characteristics, and fabric
properties. The most important design parameter is the air- or gas-to-cloth ratio, and the usual
operating parameter of interest is pressure drop across the filter system. The major operating
feature of fabric filters that distinguishes them from other gas filters is the ability to renew the
filtering surface periodically by cleaning.
Another type of fabric filter currently being developed is the electrostatically enhanced
filter. Pilot plant baghouses employing this new technology have shown substantially lower
pressure drops than conventional filter designs. Further, some cost analyses have shown that
electrostatically enhanced baghouses could have lower lifetime costs than convention
baghouses. The purpose of this chapter, however, is to focus only on currently available
commercial filters. Anyone interested in electrostatically enhanced filtration may consult such
references as Van Osdell et al. [1], Viner et al. [2], or Donovan [3].
In this section, the types of fabric filters and the auxiliary equipment required are
discussed first from a general viewpoint. Then, fabric filtration theory as applied to each type
of filter is discussed to lay a foundation for the sizing procedures outlined in the Section 5.2.
5-5
-------
5.1.2 Types of Fabric Filters
Fabric filters can be categorized by several means, including type of cleaning (shaker, reverse-
air, pulse-jet), direction of gas flow (from inside the bag towards the outside or vice versa),
location of the system fan (suction or pressure), or size (low, medium, or high gas flow
quantity). Of these four approaches, the cleaning method is probably the most distinguishing
feature. Fabric filters are discussed in this section based on the type of cleaning employed.
5.1.2.1 Shaker Cleaning
For any type of cleaning, enough energy, must be imparted to the fabric to overcome the
adhesion forces holding dust to the bag. In shaker cleaning, used with inside to outside gas
flow, this is accomplished by suspending the bag from a motor-driven hook or framework that
oscillates. Motion may be imparted to the bag in several ways, but the general effect is to
create a sine wave along the fabric. As the fabric moves outward, accumulated dust on the
surface moves with the fabric. When the fabric reaches the limit of its extension, the patches
of dust have enough inertia to tear away from the fabric and descend to the hopper.
For small, single-compartment baghouses, a lever attached to the shaker mechanism may
be operated manually at appropriate intervals, typically at the end of a shift. In multi-
compartment baghouses, a timer or a pressure sensor responding to system pressure drop
initiates bag shaking automatically. The compartments operate in sequence so that one
compartment at a time is cleaned. Forward gas flow to the compartment is stopped, dust is
allowed to settle, residual gas flow stops, and the shaker mechanism is switched on for several
seconds to a minute or more. The settling and shaking periods may be repeated, then the
compartment is brought back on-line for filtering. Many large-scale shaker systems employ
a small amount of reverse air during the shaker cycle to assist cleaning by deflating the bags.
Parameters that affect cleaning include the amplitude and frequency of the shaking motion
and the tension of the mounted bag. The first two parameters are part of the baghouse design
and generally are not changed easily. The tension is set when bags are installed. Typical
values are about 4 Hz for frequency and 2 to 3 inches for amplitude (half-stroke). [4] Some
installations allow easy adjustment of bag tension, while others require that the bag be
loosened and reclamped to its attaching thimble.
The vigorous action of shaker systems tends to stress the bags and requires heavier and
more durable fabrics. In the United States, woven fabrics are used almost exclusively for
shaker cleaning.[5] European practice allows the use of felted fabrics at somewhat higher
filtering velocities.
5-6
-------
5.1.2.2 Reverse-air Cleaning
When glass fiber fabrics were introduced, a gentler means of cleaning the bags was needed
to prevent premature degradation. Reverse-air cleaning was developed as a less intensive way
to impart energy to the bags. In this method, gas flow to the bags is stopped in the
compartment being cleaned, and a reverse flow of air is directed through the bags. This
reversal of gas flow gently collapses the bags, and dust is removed from the fabric surface by
shear forces developed between the dust and fabric as the latter changes its contours. Other
differences between reverse-air and shaker cleaning are the installation of metal caps as an
integral part of the bag and sewn-in rings to prevent complete collapse of the bag, which may
be 30 feet long, during cleaning. Without these rings, collected dust tends to choke the bag
as the fabric collapses in on itself during inside-to-outside filtering. As with multi-
compartment shaker baghouses, the same cycle takes place hi reverse-air baghouse of stopping
forward gas flow and allowing dust to settle before cleaning action begins.
The source of reverse air is generally a separate system fan capable of supplying clean,
dry air for one or two compartments at a gas-to-cloth ratio similar to that of the forward gas
flow.
5.1.2.3 Pulse-jet Cleaning
This form of cleaning uses compressed air to force a burst of air down through the bag and
expand it violently. As with shaker baghouses, the fabric reaches its extension limit and the
dust separates from the bag. In pulse jets, however, filtering gas flows are opposite in
direction when compared with shaker or reverse-air baghouses. Bags are mounted on wire
cages to prevent collapse while the dusty gas flows from outside the bag to the inside. Instead
of attaching both ends of the bag to the baghouse structure, the bag and cage assembly
generally is attached only at the top. The bottom end of the assembly tends to move in the
turbulent gas flow and may contact other bags, which accelerates wear.
Although some pulse-jet baghouses are compartmented, most are not. Bags are cleaned
one row at a time when a timer initiates the burst of cleaning air through a quick-opening
valve. A pipe across each row of bags carries the compressed air. The pipe is pierced above
each bag so that cleaning air exits directly downward into the bag. Some systems direct the
air through a short venturi that is intended to entrain additional cleaning air. The pulse
opposes and interrupts forward gas flow only for a few tenths of a second. However, the
quick resumption of forward flow redeposits most of the dust back on the clean bag or on
adjacent bags. An advantage of pulse-jet cleaning is the reduction in baghouse size allowed
by not having to build an extra compartment for off-line cleaning.
5-7
-------
5.1.3 Auxiliary Equipment
The typical auxiliary equipment associated with fabric filter systems is shown in Figure 5.1.
Along with the fabric filter itself, a control system typically includes the following auxiliary
equipment: a capture device (i.e., hood or direct exhaust connection); ductwork; dust removal
equipment (screw conveyor, etc.); fans, motors, and starters; and a stack. In addition, spray
chambers, mechanical collectors, and dilution air ports may be needed to precondition the gas
before it reaches the fabric filter. Capture devices are usually hoods that exhaust pollutants
into the ductwork or direct exhaust couplings attached to a process vessel. Hoods are more
common, yet poorly designed hoods will allow pollutants to escape. Direct exhaust couplings
are less common, requiring sweep air to be drawn through the process vessel, and may not
be feasible in some processes. Ductwork provides a means of moving the exhaust stream to
the control device. Spray chambers and dilution air ports are used to decrease the temperature
of the pollutant stream to protect the filter fabric from excessive temperatures. When a
substantial portion of the pollutant loading consists of relatively large particles, mechanical
collectors such as cyclones are used to reduce the load on the fabric filter itself. The fans
provide motive power for air movement and can be mounted before (pressure baghouse) or
after (suction baghouse) the filter. A stack, when used, vents the cleaned stream to the
atmosphere. Screw conveyors are often used to remove captured dust from the bottom of the
hoppers. Air conveying (pneumatic) systems and direct dumping into containers are also used.
Figure 5.1: Typical alternative auxiliary equipment items used with fabric filter control
systems.
5-8
-------
5.1.4 Fabric Filtration Theory
The key to designing a baghouse is to determine the face velocity that produces the optimum
balance between pressure drop (operating cost) and baghouse size (capital cost). Major factors
that affect design face velocity (or gas-to-cloth ratio), discussed in Section 5.2, include particle
and fabric characteristics and gas temperature.
Although collection efficiency is another important measure of baghouse performance, it
is generally assumed that a properly designed and well run baghouse will be highly efficient.
Therefore, the design process focuses on the pressure drop.
There are several contributions to the pressure drop across a baghouse compartment,
including the pressure drop from the flow through the inlet and outlet ducts, from flow
through the hopper regions, and from flow through the bags. The pressure drop through the
baghouse compartment (excluding the pressure drop across the bags) depends largely on the
baghouse design and ranges from 1 to 2 inches of H2O[3] in conventional designs and up to
about 3 inches of H2O in designs having complicated gas flow paths. This loss can be kept
to a minimum (i.e., 1 inch of H2O or less) by investing in a flow modeling study of the
proposed design. A study of this sort would cost on the order of $50,000 (in 1986). The
pressure drop across the bags (also called the tube-sheet pressure drop) can be as high as 10
inches of H2O or more. The tube-sheet pressure drop is a complex function of the physical
properties of the dust and the fabric and the manner in which the baghouse is designed and
operated. The duct and hopper losses are constant and can be minimized effectively through
proper design based on a knowledge of the flow through the baghouse.1
Fabric filtration is inherently a batch process that has been adapted to continuous operation
through clever engineering. One requirement for a continuously operating baghouse is that the
dust collected on the bags must be removed periodically. Shaker and reverse-air baghouses
are similar in the sense that they both normally use woven fabric bags, run at relatively low
face velocities, and the filtration mechanism is cake filtration. That is, the fabric merely
serves as a substrate for the formation of a dust cake that is the actual filtration medium.
Pulse-jet baghouses generally use felt fabric and run with a high gas-to-cloth ratio (about
double that of shaker or reverse-air baghouses). The felt fabric may play a much more active
role in the filtration process. This distinction between cake filtration and fabric filtration has
important implications for the rate of pressure loss across the filter bags. The theoretical
description of cake filtration is quite different from that for fabric filtration, and the design
processes are quite different. Accurate fabric selection is aided by bench-scale filtration tests
at less than one-tenth the cost of flow modeling. These tests can be used to investigate fabric
effects on pressure drop, cake release during cleaning, and collection efficiency. Electrical
properties of the fabric, such as resistivity and triboelectric order, may be measured to aid in
'A procedure for estimating duct pressure losses is given in chapter 10 ("Hoods, Ductwork, and Stacks")
of this Manual.
5-9
-------
fabric selection. Although their effects are generally poorly understood, electrical/electrostatic
effects influence cake porosity and particle adhesion.[6,7,8]
The general equations used to size a baghouse follow beginning with the reverse air/shake
deflate type of baghouse.
5.1.4.1 Reverse Air/Shake Deflate Baghouses
The construction of a baghouse begins with a set of specifications including average pressure
drop, total gas flow, and other requirements; a maximum pressure drop may also be specified.
Given these specifications, the designer must determine the maximum face velocity that can
meet these requirements. The standard way to relate baghouse pressure drop to face velocity
is given by the relation:
&P(0) = Ssys(0)Vavg (5.1)
where
AP(#) = the pressure drop across the filter, a function of time, 6 (in. H2O)
ASsys(#) = system drag, a function of time [in. H2O/(ft/min)]
Vavg = average (i.e., design) face velocity, constant (ft/min)
For a multi-compartment baghouse, the system drag, which accounts for most of the drag from
the inlet flange to the outlet flange of the baghouse, is determined as a combination of
resistances representative of several compartments. For the typical case where the pressure
drop through each compartment is the same, it can be shown that:[16]
(5.2)
1 M 1
i y i
M /e, S,(8)
1
, M
1 V
M,e, '
M
i M 1
1 jp 1
where
M = number of compartments in the baghouse
S,(#) = drag across compartment i
The compartment drag is a function of the amount of dust collected on the bags in that
compartment. In general, the dust will be distributed is a very nonuniform manner. That is,
there will be a variation of dust load from one bag to the next and within a given bag there
will also be a variation of dust load from one area to another. For a sufficiently small area
j within compartment /, it can be assumed that the drag is a linear function of dust load:
5-10
-------
= Se + K2W,j(0) (5.3)
where
Se = drag of a dust-free filter bag [in. H2O/(ft/min)]
K2 = dust cake flow resistance {[in. H2O/(ft/min)]/(lb/ft2)}
W, j(#) = dust mass per unit area of area j in compartment i, "areal density" (lb/ft2)
If there are N different areas of equal size within compartment i, each with a different drag
S,;, then the total drag for compartment i can be computed in a manner analogous to Equation
5.2:
(5.4)
The constants Se and Kg depend upon the fabric and the nature and size of the dust. The
relationships between these constants and the dust and fabric properties are not understood
well enough to permit accurate predictions and so must be determined empirically, either from
prior experience with the dust/fabric combination or from laboratory measurements. The dust
mass as a function of time is defined as:
Cin
(5.5)
where
Wr = dust mass per unit area remaining on a "clean" bag (lb/ft2)
Cm = dust concentration in the inlet gas (lb/ft3)
VtJ( 6) = face velocity through area j of compartment / (ft/min)
It is assumed that the inlet dust concentration and the filter area are constant. The face
velocity, (gas-to-cloth ratio) through each filter area j and compartment i changes with time,
starting at a maximum value just after clearing and steadily decreasing as dust builds up on
the bags. The individual compartment face velocities are related to the average face velocity
by the expression:
V
avg
E,
i
£
E, ;
^
c,
/0) A,J
>j
5-11
-------
(for M compartments with
equal area)
Equations 5.1 through 5.6 reveal that there is no explicit relationship between the design face
velocity and the tube-sheet pressure drop. On the contrary, the pressure drop that results from
a given design can only be determined by the simultaneous solution of Equations 5.1 through
5.5, with Equation 5.6 as a constraint on that solution. This conclusion has several
implications for the design process. The design requires an iterative procedure: one must
begin with a known target for the average pressure drop, propose a baghouse design (number
of compartments, length of filtration period, etc.), assume a face velocity that will yield that
pressure drop, and solve the system of Equations 5.1 through 5.6 to verify that the calculated
pressure drop equals the target pressure drop. This procedure is repeated until the specified
face velocity yields an average pressure drop (and maximum pressure drop, if applicable) that
is sufficiently close to the design specification.
5.1.4.2 Pulse-Jet Baghouses
The distinction between pulse-jet baghouses and reverse-air and shaker baghouses is basically
the difference between cake filtration and composite dust/fabric filtration (noncake filtration).
This distinction is more a matter of convenience than physics. In reality, pulse-jet baghouses
have been designed to operate in a variety of modes. Some pulse jets remain on-line at all
times and are cleaned frequently. Others are taken off-line for cleaning at relatively long
intervals. Obviously, if a compartment remains on-line long enough without being cleaned,
then the filtration mechanism becomes that of cake filtration. A complete model of pulse-jet
filtration therefore must account for the depth filtration occurring on a relatively clean pulse-
jet filter, the cake filtration that inevitably results from prolonged periods on-line, and the
transition period between the two regimes.
Besides the question of filtration mechanism, there is also the question of cleaning method.
If a compartment is taken off-line for cleaning, then the dust that is removed from the bags
will fall into the dust hopper before forward gas flow resumes. If a compartment is cleaned
while on-line, then only a small fraction of the dust removed from the bag will fall into the
hopper. The remainder of the dislodged dust will be redeposited (i.e., "recycled") on the bag
by the forward gas flow. The redeposited dust layer has different pressure drop characteristics
than the freshly deposited dust. The modeling work that has been done to date focuses on the
on-line cleaning method. Dennis and Klemm[9] proposed the following model of drag across
a pulse-jet filter:
S = Se + (K2)CWC + K2W0 (5.7)
5-12
-------
where
S = drag across the filter
Se = drag of a just-cleaned filter
(K2)c = specific dust resistance of the recycling dust
Wc = areal density of the recycling dust
K2 = specific dust resistance of the freshly deposited dust
W0 = areal density of the freshly deposited dust
This model has the advantage that it can easily account for all three regimes of filtration in a
pulse-jet baghouse. As hi Equations 5.1 to 5.6, the drag, filtration velocity and areal densities
are functions of time, 6. For given operating conditions, however, the values of Se, (K^, and
Wc may be assumed to be constant, so that they can be grouped together:
AP - (PE)Aw + K2W0V (5.8)
where
AP = pressure drop (in. H2O)
V = filtration velocity (ft/min)
(PE)Aw = (Se +(K2)CWJ V
Equation 5.8 describes the pressure drop behavior of an individual bag. To extend this single
bag result to a multiple-bag compartment, Equation 5.7 would be used to determine the
individual bag drag and the total baghouse drag would then be computed as the sum of the
parallel resistances. Pressure drop would then be calculated as in Equation 5.1. It seems
reasonable to extend this analysis to the case when the dust is distributed unevenly on the bag
and then apply Equation 5.7 to each area on the bag, followed by an equation analogous to 5.4
to compute the overall bag drag. The difficulty in doing this is that one must assume values
for W? for each different area to be modeled.
The disadvantage of the model represented by Equations 5.7 and 5.8 is that the constants,
S,, K2, and W,, cannot be predicted at this time. Consequently, correlations of laboratory data
must be used to determine the value of (PE)Aw. For the fabric-dust combination of Dacron felt
and coal fly ash, Dennis and Klemm[9] developed an empirical relationship between (PE)Aw,
the face velocity, and the cleaning pulse pressure. This relationship (converted from metric
to English units) is as follows:
(PE)Aw = 6.0SVfPj065 (5.9)
where
Vf = face velocity, (ft/min)
P, = pressure of the cleaning pulse
(usually 60 to 100 psig; see Section 5.4.1)
5-13
-------
This equation is essentially a regression fit to a limited amount of laboratory data and
should not be applied to other dust/fabric combinations. The power law form of Equation 5.9
may not be valid for other dusts or fabrics. Consequently, more data should be collected and
analyzed before the model represented by Equation 5.9 can be used for rigorous sizing
purposes.
Another model that shows promise in the prediction of noncake filtration pressure drop is
that of Leith and Ellenbecker[10] as modified by Koehler and Leith.[ll] In this model, the
tube-sheet pressure drop is a function of the clean fabric drag, the system hardware, and the
cleaning energy. Specifically:
AP = A |P + Kiyf ~ ^/(P - KlVfY - 4WK2/K3\ + KV2f (5.8)
where
Ps = maximum static pressure achieved in the bag during cleaning
K! = clean fabric resistance
Vf = face velocity
K2 = dust deposit flow resistance
K3 = bag cleaning efficiency coefficient
Kv = loss coefficient for the venturi at the inlet to the bag
Comparisons of laboratory data with pressure drops computed from Equation 5.10 [10,11] are
in close agreement for a variety of dust/fabric combinations. The disadvantage of Equation
5.10 is that the constants Kt, K2, and K3 must be determined from laboratory measurements.
The most difficult one to determine is the K3 value, which can only be found by making
measurements in a pilot-scale pulse-jet baghouse. A limitation of laboratory measurements
is that actual filtration conditions cannot always be adequately simulated. For example, a
redispersed dust may not have the same size distribution as the original dust, thereby yielding
different values of Kl5 K2, and K3 than would be measured in an operating baghouse.
5.2 Design Procedures
5.2.1 Gas-to-Cloth Ratio
The gas-to-cloth ratio is difficult to estimate from first principles. However, shortcut methods
of varying complexity allow rapid estimation. Descriptions of three methods of increasing
difficulty follow. For shaker and reverse-air baghouses, the third method is best performed
with publicly available computer model programs.
5-14
-------
The methods outlined below pertain to conventional baghouses. Use of electrostatic
stimulation (an emerging technology) allow a higher gas-to-cloth ratio at a given pressure
drop; thus a smaller baghouse structure and fewer bags are needed. This reduces total annual
cost by up to 30%. Viner and Locke[13] discuss cost and performance models for
electrostatically stimulated fabric filters.
5-15
-------
Table 5.1: Gas-to-Cloth Ratios*'3
(actual ft3/min)(ft2 of net cloth area)
Shaker/Woven Puke Jet/Felt
Dust Reverse-Air/Woven Reverse-Air Felt
Alumina
Asbestos
Bauxite
Carbon Black
Coal
Cocoa, Chocolate
Clay
Cement
Cosmetics
Enamel Frit
Feeds, Grain
Feldspar
Fertilizer
Flour
Fly Ash
Graphite
Gypsum
Iron Ore
Iron Oxide
Iron Sulfate
Lead Oxide
Leather Dust
Lime
Limestone
Mica
Paint Pigments
Paper
Plastics
Quartz
Rock Dust
Sand
Sawdust (Wood)
Silica
Slate
Soap, Detergents
Spices
Starch
Sugar
Talc
Tobacco
Zinc Oxide
2.5
3.0
2.5
1.5
2.5
2.8
2.5
2.0
1.5
2.5
3.5
2.2
3.0
3.0
2.5
2.0
2.0
3.0
2.5
2.0
2.0
3.5
2.5
2.7
2.7
2.5
3.5
2.5
2.8
3.0
2.5
3.5
2.5
3.5
2.0
2.7
3.0
2.0
2.5
3.5
2.0
8
10
8
5
8
12
9
8
10
9
14
9
8
12
5
5
10
11
7
6
6
12
10
8
9
7
10
7
9
9
10
12
7
12
5
10
8
13
5
*Reference [12]
••Generally safe design values; application requires condsideration of particle size and grain loading.
5-16
-------
5.2.1.1 Gas-to-Cloth Ratio From Similar Applications
After a fabric has been selected, the gas-to-cloth ratio can be determined using Table 5.1.
Column 1 shows the type of dust; column 2 shows the gas-to-cloth ratios for woven fabric;
and column 3 shows gas-to-cloth ratios for felted fabrics. Notice that these values are all net
gas-to-cloth ratios. The net gas-to-cloth ratio is equal to the total actual volumetric flow rate
in cubic feet per minute divided by the net cloth area in square feet. This ratio, in units of feet
per minute, affects pressure drop and bag life. The net cloth area is determined by dividing
the gas-to-cloth ratio into the actual cubic feet per minute flow of the exhaust gas stream. For
an intermittent-type baghouse that is shut down for cleaning, this is the total, or gross, cloth
area. However, for continuously operated filters, the area must be increased to allow the
shutting down of one or more compartments for cleaning. Table 5.2 provides a guide for
adjusting the net area to the gross area, which determines the size of a continuously cleaned
filter.
5.2.1.2 Gas-to-Cloth Ratio From Manufacturer's Methods
Manufacturers have developed nomographs and charts that allow rapid estimation of the gas-
to-cloth ratio. Two examples are given below, one for shaker-cleaned baghouses and the other
for pulse-jet cleaned baghouses.
For shaker baghouses, Table 5.3 gives a factor method for estimating the ratio. Ratios for
several materials in different operations are presented, but are modified by factors for particle
size and dust load. Directions and an example are included. Gas-to-cloth ratios for reverse-
air baghouses would be about the same or a little more conservative compared to the Table
5.3 values.
For pulse-jet baghouses, which normally operate at two or more times the gas-to-cloth
ratio of reverse-air baghouses, another factor method[15] has been modified with equations
to represent temperature, particle size, and dust load:
V = 2.878,4 B r°2335L-006021 (0.7471 + 0.0853 InD) (5.11)
where
V = gas-to-cloth ratio (ft/min)
A = material factor, from Table 5.4
B = application factor, from Table 5.4
T = temperature, (°F, between 50 and 275)
L = inlet dust loading (gr/ft3, between 0.05 and 100)
D = mass mean diameter of particle (/im, between 3 and 100)
5-17
-------
Table 5.2: Approximate Guide to Estimate Gross
Cloth Area*
Net Cloth Area
(ft2)
1-4,000
4,001-12,000
12,001-24,000
24,001-36,000
36,001-48,000
48,001-60,000
60,001-72,000
72,001-84,000
84,001-96,000
96,001-108,000
108,001-132,000
132,001-180,000
above 180,001
Gross Cloth Area
(ft2)
Multiply by 2
1.5
1.25
1.17
1.125
1.11
1.10
1.09
1.08
1.07
1.06
1.05
1.04
*Reference [14]
For temperatures below 50°F, use T = 50 but expect decreased accuracy; for temperatures
above 275°F, use T = 275. For particle mass mean diameters less than 3 /*m, the value of
D is 0.8, and for diameters greater than 100 /xm, D is 1.2. For dust loading less than 0.05
gr/ft3, use L = 0.05; for dust loading above 100 gr/ft3, use L = 100.
Pulse-jet baghouses The overall process of designing a pulse jet baghouse is actually simpler
than that required for a reverse-air or shaker baghouse if the baghouse remains on-line for
cleaning. The first step is to specify what the desired average tube-sheet pressure drop should
be. Second, the operating characteristics of the baghouse must be established (e.g., on-line
time, cleaning energy). Third, the designer must obtain values for the coefficients in either
Equation 5.9 or Equation 5.10 from field, pilot plant, or laboratory measurements. Fourth,
a value hi estimated for the face velocity and the appropriate equation (Equation 5.8 or 5.10)
is solved for the pressure drop as a function of time for the duration of the filtration cycle.
This information is used to calculate the cycle average pressure drop. If the calculated
pressure drop matches the specified pressure drop, then the procedure is finished. If not, then
the designer must adjust the face velocity and repeat the procedure.
5-18
-------
Table 5.3: Manufacturer's Factor Method for Estimating Gas-to-Cloth
Ratios for Shaker Baehouses
A
•!'! RATIO
I 5/1 RATIO
MATLRIAL
MAT! RIAL
OPERATION
MATERIAL
OPERATION
Cardboard
Feeds
Flour
Gram
Leather Dust
Tobacco
Supply Air
Wood Dusi
Chips
2 3 4.5 6.7
: 3 4 5 6 7
2 3 4 5 6.7
1.7 8
I 4 6.7
13
1 67
Asbcsios
Aluminum Dusi
Fibrous Mai I
Cellulose Mafl
Gypsum
Lime
(Hydraledl
Pcrlite
Rubber Chcm
Sail
SanO •
Iron Scale
SiKia Ash
Talc
MaLluntni:
Operand!
i 7 »
I 7 X
1478
1478
I 1 5,6 7
2467
2 -1 S 6
J S h 1 8
: 3 4 5 6 7
J 5 I. 7 V 15
Alumina
Carbon Black
Cement
Coke
Ceramic I'igm
Clay & Brick Dust
Cial
kaolin
Limestone
Rock Ore Dusi
Silica
Su|:ar
23456
4 5.6 7
34567
2 3.5.6
4 5 6.7
2 4 6.12
2 3.67 12
457
234567
234567
234567
34567
Ammonium
Phosphate Fcrl
Diatomaceous
Earth
Dry Peirochem
Dyes
Ply Ash
Metal Powders
Plastics
Resms
Silicates
Starch
Soaps
234567
4.5 67
2.3 4.5 6.7 14
2.34 5 6.7
10
2.345 67 14
2.3 4 5 6 7 14
2 3 4 5 6 7 14
2 3 4 5 6 7 U
67
34507
Activated Charcoal
Carbon Black
Detergents
Metal Fumes.
Oxides and other
Solid Dispersed
Products
2.4.5.6.7
11.14
2.4.5.67
10,11
( I [TING
CRUSHING
PULVERIZING
MIXING
SCRhl.NING
STORAGE
< t)s\ n ING
GRINDING
SHAM OUT
I IIRNACL I I'ML
Rl ACTION R'MF.
DUMPING
INI \kl ( 1.1 \MNG - H
PRO< I Ss 14
Kl.ASllMi IS
-UNtJNt.S6
MICRON SI7.F.
> 100
50- 1(10
10- 50 •
3 - 10
1 - 3
I-AC IUK
1 •XCTOR
1 2
1 1
1 0
s>
8
This information constitutes a guide for commonly encountered situations and should not be
considered a "hard-and-fast" rule. Air-to-cloth ratios are dependent on dust loading, size
distribution, particle shape and "cohesiveness" of the deposited dust. These conditions must
be evaluated for each application. The longer the interval between bag cleaning, the lower
the air-to-cloth ratio must be. Finely-divided, uniformly sized particles generally form more
dense filter cakes and require lower air-to-cloth ratios than when larger particles are
interspersed with the lines. Sticky, oily particles, regardless of shape or size, form dense
filter cakes and require lower air-to-cloth ratios.
DUST LOAD FACTOR
IOADING GR (( IT
4-8
EXAMPLE: Foundry shakeout unit handling 26000 CFM and collecting 3500 Ib / hr. of
sand The particle distribution shows 90% greater than 10 microns. The air is to exhaust to
room in winter, to atmosphere in summer.
3500 Ib _ 60 min _ 26000 cu ft
hr hr mm
iL =
Ib
Sr
cu.ft.
* Chart A - 3/1 ratio. Chart B - Factor 1.0. Chart C — .95: 3 x 1 x .95 - 2.9 air to
cloth ratio. 26000^ 2.9 = 9,000 sq. ft.
Reprinted with permission from Buffalo Forge Company Bulletin AHD-29.
-------
Table 5.4: Factors for Pulse-Jet Gas-to-Cloth Ratios*
A. Material Factor
15"
Cake mix
Cardboard
dust
Cocoa
Feeds
Flour
Grain
Leather
dust
Sawdust
Tobacco
12
10
9.0
B. Application Factor
Nuisance Venting
Relief of transfer points,
conveyors, packing stations, etc.
Product Collection
Air conveoying-venting, mills,
flash driers, classifiers, etc.
Process Gas Filtration
Spray driers, kilns, reactors, etc.
6.0b
Asbestos
Buffing dust
Fiborous and
cellulosic
material
Foundary
shakeout
Gypsum
Lime
(hydrated)
Perlite
Rubber
chemicals
Salt
Sand
Sandblast
dust
Soda ash
Talc
Alumina
Aspirin
Carbon black
(finished)
Cement
Ceramic
pigments
Clay and brick
dusts
Coal
Fluorspar
Gum, natural
Kaolin
Limestone
Perchlorates
Rock dust, ores
and minerals
Silica
Sorbic acid
Sugar
Ammonium
phosphate-
fertilizer
Cake
Diatomaceous
earth
Dry petro-
chemicals
Dyes
Fly ash
Metal powder
Metal oxides
Pigments
metallic and
synthetic
Plastics
Resins
Silicates
Starch
Stearates
Tannic acid
Activated
carbon
Carbon black
(molecular)
Detergents
Fumes and
other
dispersed
products direct
from reactions
Powdered milk
Soap
1.0
0.9
0.8
*Reference [15]
aln general, physically and chemically stable material.
bAlso includes those solids that are unstable in their physical or chemical state due to
hygroscopic nature, sublimation, and/or polymerization
5-20
-------
5.2.2 Pressure Drop
Pressure drop for the bags can be calculated rigorously from the equations given in the
preceding section if values for the various parameters are known. Frequently they are not
known. For quick estimation, a maximum pressure drop of 5 to 10 in. H2O across the
baghouse and 10 to 20 in. H2O across the entire system can be assumed if it contains much
ductwork.
A comparable form of Equations 5.1 and 3.3 that may be used for pressure drop across
the fabric in a shaker or reverse-air baghouse is:
AP = SeV + K2C,V26 (5.12)
where
AP = pressure drop (in. H2O)
Se = effective residual drag of the fabric [in. H2O/(ft/min)]
V = superficial face velocity or gas-to-cloth ratio (ft/min)
K2 = specific resistance coefficient of the dust
{[in. H2O/(ft/min)]/(lb /ft2)}
C, = inlet dust concentration (lb/ft3)
6 = filtration time (min)
Although there is much variability, values for Se may range from about 0.2 to 2 in.
H20/(ft/min) and for K2 from 1.2 to 30-40 [in. H20/(ft/min)]/(lb/ft2). Typical values for coal
fly ash are about 1 to 4. Inlet concentrations vary from less than 0.05 gr/ft3 to more than 100
gr/ft3, but a more nearly typical range in from about 0.5 to 10 gr/ft . Filtration times may
range from about 20 to 90 minutes for continuous duty baghouses, but 30 to 60 minutes is
more frequently found. For pulse-jet baghouses, use Equations 5.8 and 5.9 to estimate AP,
after substituting C,V0for W0 and (PE)Aw for S,V.
5.2.3 Particle Characteristics
Particle size distribution and adhesiveness are the most important particle properties that affect
design procedures. Smaller particle sizes can form a denser cake, which increases pressure
drop. As shown in Table 5.3 and Equation 5.11, the effect of decreasing average particle size
is a lower applicable gas-to-cloth ratio.
Adhering particles, such as oily residues or electrostatically active plastics, may require
installing equipment that injects a precoating material onto the bag surface, which acts as a
buffer that traps the particles and prevents them from blinding or permanently plugging the
fabric pores. Informed fabric selection may eliminate electrostatic problems.
5-21
-------
5.2.4 Gas Stream Characteristics
Moisture and corrosives content are the major gas stream characteristics requiring design
consideration. The baghouse and associated ductwork should be insulated and possibly heated
if condensation may occur. Both the structural and fabric components must be considered,
as either may be damaged. Where structural corrosion is likely, stainless steel substitution for
mild steel may be required, provided that chlorides are not present. (Most austenitic stainless
steels are susceptible to chloride corrosion.)
5.2.4.1 Temperature
The temperature of the pollutant stream to be cleaned must be above and remain above the
dew point of any condensables in the stream. If the temperature is high and it can be lowered
without approaching the dew point, spray coolers or dilution air can be used to drop the
temperature so that temperature limits of the fabric will not be exceeded. The additional cost
of a precooler will have to be weighed against the higher cost of bags with greater temperature
resistance. The use of dilution air to cool the stream also constitutes a tradeoff between a less
expensive fabric and a larger filter necessary to accommodate the additional volume of the
dilution air. Generally, precooling would not be necessary if fabric that will handle the
temperature and the chemical action of the pollutant stream is available. (Costs for spray
chambers, quenchers, and other precoolers are found in the "Precoolers" section of the
Manual) Table 5.5 lists several of the fabrics in current use and provides information on
temperature limits and chemical resistance. The column labeled "Flex Abrasion" indicates the
fabric's suitability for cleaning by mechanical shakers.
5.2.4.2 Pressure
Standard fabric filters can be used in pressure or vacuum service but only within the range of
about + 25 inches of water. Because of the sheet metal construction of the house, they are
not generally suited for more severe service. However, for special applications, high-pressure
shells can be built.
5.2.5 Equipment Design Considerations
5.2.5.1 Pressure or Suction Housings
The location of the baghouse with respect to the fan in the gas stream affects the capital cost.
A suction-type baghouse, with the fan located on the downstream side of the unit, must
withstand high negative pressures and therefore must be more heavily constructed and
reinforced than a baghouse located downstream of the fan (pressure baghouse). The negative
5-22
-------
pressure in the suction baghouse can result in outside air infiltration, which can result in
condensation, corrosion, or even explosions if combustible gases are being handled. In the
case of toxic gases, this inward leakage can have an advantage over the pressure-type
baghouse, where leakage is outward. The main advantage of the suction baghouse is that the
fan handling the process stream is located at the clean-gas side of the baghouse. This reduces
the wear and abrasion on the fan and permits the use of more efficient fans (backward-curved
blade design). However, because for some designs the exhaust gases from each compartment
are combined in the outlet manifold to the fan, locating compartments with leaking bags may
be difficult and adds to maintenance costs. Pressure-type baghouses are generally less ex-
pensive because the housing must only withstand the differential pressure across the fabric.
In some designs the baghouse has no external housing. Maintenance also is reduced because
the compartments can be entered and leaking bags can be observed while the compartment is
in service. With a pressure baghouse, the housing acts as the stack to contain the fumes with
the subsequent discharge at the roof of the structure, which makes it easier to locate leaking
bags. The main disadvantage of the pressure-type baghouse in that the fan is exposed to the
dirty gases where abrasion and wear on the fan blades may become a problem.
5.2.5.2 Standard or Custom Construction
The design and construction of baghouses are separated into two groups, standard and
custom,[14] which are further separated into low, medium, and high capacity. Standard
baghouses are predesigned and factory built as complete off-the-shelf units that are shop-
assembled and bagged for low-capacity units (hundreds to thousands of acfm throughput.
Medium-capacity units (thousands to less than 100,000 acfm) have standard designs, are shop-
assembled, mar or may not be bagged, and have separate bag compartment and hopper
sections. High-capacity baghouses (larger than 50,000 or 100,000 acfm) can be designed as
shippable modules requiring only moderate field assembly. These modules may have bags
installed and
can be shipped by truck or rail. Upon arrival, they can be operated singly or combined to
form units for larger-capacity applications. Because they are preassembled, field labor for
installation is less costly.
The custom baghouse, also high capacity, is designed for a specific application and is
usually built to the specifications prescribed by the customer. Generally, these units are much
larger than standard baghouses. For example, many, are used on power plants. The cost of
the custom baghouse is much higher per square foot of fabric because it is not an off-the-shelf
item and requires special setups for manufacture and expensive field labor for assembly upon
arrival. The advantages of the custom baghouse are many and are usually directed towards
ease of maintenance, accessibility, and other customer preferences. In some standard
baghouses, a complete set of bags must be replaced in a compartment at one time because of
the difficulty in locating and replacing single leaking bags, whereas in custom baghouses,
single bags are accessible and can be replaced one at a time as leaks develop.
5-23
-------
pressure in the suction baghouse can result in outside air infiltration, which can result in
condensation, corrosion, or even explosions if combustible gases are being handled. In the
case of toxic gases, this inward leakage can have an advantage over the pressure-type
baghouse, where leakage is outward. The main advantage of the suction baghouse is that the
fan handling the process stream is located at the clean-gas side of the baghouse. This reduces
the wear and abrasion on the fan and permits the use of more efficient fans (backward-curved
blade design). However, because for some designs the exhaust gases from each compartment
are combined in the outlet manifold to the fan, locating compartments with leaking bags may
be difficult and adds to maintenance costs. Pressure-type baghouses are generally less ex-
pensive because the housing must only withstand the differential pressure across the fabric.
In some designs the baghouse has no external housing. Maintenance also is reduced because
the compartments can be entered and leaking bags can be observed while the compartment is
in service. With a pressure baghouse, the housing acts as the stack to contain the fumes with
the subsequent discharge at the roof of the structure, which makes it easier to locate leaking
bags. The main disadvantage of the pressure-type baghouse in that the fan is exposed to the
dirty gases where abrasion and wear on the fan blades may become a problem.
5.2.5.2 Standard or Custom Construction
The design and construction of baghouses are separated into two groups, standard and
custom,[14] which are further separated into low, medium, and high capacity. Standard
baghouses are predesigned and factory built as complete off-the-shelf units that are shop-
assembled and bagged for low-capacity units (hundreds to thousands of acfm throughput.
Medium-capacity units (thousands to less than 100,000 acfm) have standard designs, are shop-
assembled, mar or may not be bagged, and have separate bag compartment and hopper
sections. High-capacity baghouses (larger than 50,000 or 100,000 acfm) can be designed as
shippable modules requiring only moderate field assembly. These modules may have bags
installed and
can be shipped by truck or rail. Upon arrival, they can be operated singly or combined to
form units for larger-capacity applications. Because they are preassembled, field labor for
installation is less costly.
The custom baghouse, also high capacity, is designed for a specific application and is
usually built to the specifications prescribed by the customer. Generally, these units are much
larger than standard baghouses. For example, many, are used on power plants. The cost of
the custom baghouse is much higher per square foot of fabric because it is not an off-the-shelf
item and requires special setups for manufacture and expensive field labor for assembly upon
arrival. The advantages of the custom baghouse are many and are usually directed towards
ease of maintenance, accessibility, and other customer preferences. In some standard
baghouses, a complete set of bags must be replaced in a compartment at one time because of
the difficulty in locating and replacing single leaking bags, whereas in custom baghouses,
single bags are accessible and can be replaced one at a time as leaks develop.
5-24
-------
5.2.5.3 Filter Media
The type of filter material used in baghouses is dependent on the specific application in terms
of chemical composition of the gas, operating temperature, dust loading, and the physical and
chemical characteristics of the particulate. A variety of fabrics, either felted or woven, is
available and the selection of a specific material, weave, finish, or weight is based primarily
on past experience. The type of yarn (filament, spun, or staple), the yarn diameter, and twist
are also factors in the selection of suitable fabrics for a specific application. For some difficult
applications, Gore-Tex, a polytetrafluoroethylene (PTFE) membrane laminated to a fabric (felt
or woven) may be used. Because of the violent agitation of mechanical shakers, spun or heavy
weight staple yarn fabrics are commonly used with this type of cleaning, while lighter weight
filament yarn fabrics are used with reverse air cleaning.
The type of material will limit the maximum operating gas temperature for the baghouse.
Cotton fabric has the least resistance to high temperatures (about 180°F), while fiberglass
hasthe most (about 500 °F). The temperature of the exhaust-gas stream must be well above
the dew point of any, of its contained condensables, as liquid particles will usually plug the
fabric pores quickly. However, the temperature must be below the maximum limit of the
fabric in the bags. These maximum limits are given in Table 5.5.
5.3 Estimating Total Capital Investment
Total capital investment includes costs for the baghouse structure, the initial complement of
bags, auxiliary equipment, and the usual direct and indirect costs associated with installing or
erecting new structures. These costs are described below. (Costs for improving baghouse
performance with electrical enhancement are not discussed in this section, but are mentioned
in the example problem.)
5-25
-------
Table 5.5 Properties of Leading Fabric Materials*
Fabric
Cotton
Creslan"
Dacrorf
Temp
opa
180
250
275
Acid
Resistance
Poor
Good in mineral acids
Good in most mineral
Alkali
Resistance
Very good
Good in weak alkali
Good in weak alkali;
Flex
Abrasion
Very good
Good to very
good
Very good
Dynelc
Fiberglasd
Filtron'
Gore-Texf
Nomexe
Nylonc
Orlonc
Polypropylene
Teflon0
Wool
acids; dissolves
partially in
concentrated H2SO4
160 Little effect even at
high concentration
500 Fair to good
270 Good to excellent
Depends Depends on backing
on
backing
375 Fair
200 Fair
260 Good to excellent in
mineral acids
200 Excellent
450 Inert except to fluorine
200 Very good
fair in strong alkali
Little effect even in
high concentration
Fair to good
Good
Fair to good
Fair
Good to very
good
Depends on backing Fair
Excellent at low
temperature
Excellent
Fair to good in weak
alkali
Excellent
Inert except to
trifluoride, chlorine,
and molten alkaline
metals
Poor
Excellent
Excellent
Good
Excellent
Fair
Fair to good
^Reference [20]
aMaximum continuous operating temperatures recommended by the Institute of Clean Air Companies.
""American Cyanamid registered trademark.
cDu Pont registered trademark.
dOwens-Corning Fiberglas registered trademark.
eW. W. Criswell Div. of Wheelabrator-Fry, Inc. trade name.
fW. L. Gote and Co., registered trademark
5-26
-------
5.3.1 Equipment Cost
5.3.1.1 Bare Baghouse Costs
Cost correlations for six types of baghouses are presented. These six types, five of which are
preassembled and one, field-assembled, are outlined in Table 5.6.
Table 5.6: Scope of Cost Correlations
Intermittent
Continuous
Continuous
Continuous
Continuous
Baghouse Type
Preassembled Units
Shaker
Shaker
Pulse-jet (common housing)
Pulse-jet (modular)
Reverse-air
Figure No.
5.2
5.3
5.4
5.5
5.6
Continuous
Field-assembled units
Any method
5.7
Each figure gives costs for the filter without bag and additional costs for stainless steel
construction and for insulation.* All curves are based on a number of actual quotes. A least
squares line has been fitted to the quotes and the line's equation is given. In most cases these
lines should not be extrapolated in either direction. The reader should not be surprised if he
obtains quotes that differ from these curves by as much as + 25%. Significant savings can
be obtained by soliciting multiple quotes. All units include inlet and exhaust manifold
supports, platforms, handrails, and hopper discharge devices. The indicated prices are flange
to flange. The scales on both axes change from one figure to another to accommodate the
differing gas flow ranges over which the various types of baghouses operate.
*Cost in Figures 5.2 to 5.7 are in third quarter 1986 dollars. For information on escalating these prices to more current
dollars, refer to the EPA report Escalation Indexes for Air Pollution Control Costs and updates thereto, all of which are
installed on the OAQPS Technology Transfer Network (CTC Bulletin Board).
5-27
-------
The 304 stainless steel add-on cost is used when such construction is necessary to prevent
the exhaust gas stream from corroding the interior of the baghouse. Stainless steel is
substituted for all metal surfaces that are in contact with the exhaust gas stream.
Insulation costs are for 3 inches of shop-installed glass fiber encased in a metal skin. One
exception is the custom baghouse, which has field-installed insulation. Costs for insulation
include only the flange-to-flange baghouse structure on the outside of all areas in contact with
the exhaust gas stream. Insulation for ductwork, fan casings, and stacks must be calculated
separately as discussed later.
The first baghouse type is the intermittent service baghouse cleaned by a mechanical
shaker. This baghouse is shut down and cleaned at convenient times, such as the end of the
shift or end of the day. Although few units are sold, they are applicable for operations that
require infrequent cleaning. Figure 5.2 presents the unit cost with price in dollars plotted
against the gross square feet of cloth required. [21] Because intermittent service baghouses do
not require an extra compartment for cleaning, gross and net fabric areas are the same. The
plot is linear because baghouses are made up of modular compartments and thus have little
economy of scale. Because of the modular construction, the price line should not be
extrapolated downward.
Figure 5.3 presents similar costs for a continuously operated baghouse cleaned by
mechanical shaker.[21,22] Again, price is plotted against the gross cloth area in square feet.
As in Figure 5.2, the units are modular in construction. Costs for these units, on a square foot
basis, are higher because of increased complexity and generally heavier construction.
The third and fourth types are common-housing pulse jets and modular pulse jets. The latter
are constructed of separate modules that may be arranged for off-line cleaning, and the former
have all bags within one housing. The costs for these units are shown in Figures 5.4 and 5.5,
respectively.[21] Note that in the single-unit (common-housing) pulse jet, for the range shown,
the height and width of the unit are constant and the length increases; thus, for a different
reason than that for the modular units discussed above, the cost increases linearly with size.
Because the common housing is relatively inexpensive, the stainless steel add-on is proportion-
ately higher than for modular units. Added material costs and setup and labor charges
associated with the less workable stainless steel account for most of the added expense.
5-28
-------
UJ
Caution: Do not extrapolate.
Cost without bagsH
S - 1J75 + 0.450 B2
4 6 8 10 12 14
Goss Cloth ATM (1000 n 2)
16
1E
Figure 5.2: Equipment costs for intermittent shaker filters.
5-29
-------
Caution: Do not extrapolate.
Slainl*M st*«l add on
20 30 40 50 60
Gross Ootn ATM (1000 ft2)
Sourca: ETS. Inc.; Futior Co.
70
viO
Figure 5.3: Equipment costs for continuous shaker filters.
5-30
-------
Caution: Do not extrapolate.
a a w t2
Grass Ootn ATM (1000ft2)
16
Soirca: ETS, Inc.
Figure 5.4: Equipment costs for pulse-jet filters (common housing).
5-31
-------
Caution: Do not extrapolate.
IU
4 6 8 10 12
Grass CkXh ATM (itXDB2)
14
16
Source: ETS, Inc.
Figure 5.5: Equipment costs for pulse-jet filters (modular).
5-32
-------
Figure 5.6 shows the costs for the reverse-air baghouse.[21] Again, the construction is
modular. The final type is the custom baghouse which, because of its large size, must be field
assembled. It is often used on power plants, steel mills, or other applications too large for the
factory-assembled baghouses. Prices for custom units are shown in Figure 5.7.[21]
Caution: Do not extrapolate.
SUInitu «*•< add on
0 100
Sourca: ETS, Inc.
200 300
&088Oo
-------
Table 5.7: Bag Prices
Type of
Pulse jet
Pulse jet
Shaker,
Shaker,
Reverse
Reverse
Cleaning
, TRb
,BBR
Strap top
Loop top
air with rings
air w/o rings
Bag Diameter
(inches)
4-1/2
6
4-1/2
6
to 5-1/8
to 8
to 5-1/8
to 8
5
5
8
11-1/2
8
11-1/2
PE
0.
0,
0
0
0
0
0
0
0
0
.59
.43
.37
.32
.45
.43
.46
.47
.32
.32
PP
0.61
0.44
0.40
0.33
0.48
0.45
NA
NA
NA
NA
NO
1.88
1.56
1.37
1.18
1.28
1.17
1.72
1.69
1.20
1.16
Type
HA
0.92
0.71
0.66
0.58
0.75
0.66
NA
NA
NA
NA
of Material2
FG
1.29
1.08
1.24
0.95
NA
NA
0.99
0.76
0.69
0.53
CO
NA
NA
NA
NA
0.44
0.39
NA
NA
NA
NA
TF
9.05
6.80
8.78
6.71
NA
NA
NA
NA
NA
NA
NA = Not applicable.
•"Materials:
PE = 16-oz polyester FG = 16-oz fiberglass with 10% Teflon
PP = 16-oz polypropylene CO = 9-oz cotton
NO = 14-oz nomex TF = 22-oz Teflon felt
HA = 16-oz homopolymer acrylic
hBag removal methods:
TR = Top bag removal (snap in)
BBR = Bottom bag removal
NOTE:For pulse-jet baghouses, all bags are felts except for the fiberglass, which is woven.For bottom access pulse
jets, the mild steel cage price for one cage can be calculated from the single-bag fabric area using:
$ = 4.941 + 0.163 ft2 in 50 cage lots
$ = 4.441 + 0.163 ft2 in 100 cage lots
$ = 3.941 + 0.163 ft2 in 500 cage lots
These costs apply to 4-1/2 inch or 5-5/8 inch diameter, 8-foot and 10-foot cages made of 11 gauge mild steel and
having 10 vertical wires and "Roll Band" tops. For flanged tops, add $1 per cage.If flow control Venturis are used
(as they are in about half of the pulse-jet manufacturers' designs) add $5 per cage.For stainless steel cages use-
$ = 23.335 + 0.280 ft2 in 50 cage lots
$ = 21.791 + 0.263 ft2 in 100 cage lots
$ = 20.564 + 0.248 ft2 in 500 cage lots
For shakers and reverse air baghouses, all bags are woven.All prices are for finished bags, and prices can vary
from one supplier to another. For Gore-Tex bag prices, multiply base fabric price by factors of 3 to 4.5.
Source: ETS Inc. [21]
5-34
-------
5.3.1.2 Bag Costs
Table 5.7 gives the price per square foot of bags by type of fabric and by type of cleaning
system used. Actual quoted prices may vary by ± 10% from the values in the table. In
calculating thecost, the gross area as determined from Table 5.2 should be used. Gore-Tex
fabric costs are a combination of the base fabric cost and a premium for the PTFE laminate
and its application. As fiber market conditions change, the costs of fabrics relative to each
other also change. The bag prices are based on typical fabric weights, in ounces/square yard,
for the fabric being priced. Sewn-in snap rings are included in the price, but other mounting
hardware, such as clamps or cages, is an added cost.
5.3.1.3 Auxiliary Equipment
The auxiliary equipment depicted in Figure 5.1 is discussed elsewhere in the Manual.
Because hoods, ductwork, precoolers, cyclones, fans, motors, dust removal equipment and
stacks are common to many pollution control systems, they are (or will be) given extended
treatment in separate chapters.
5.3.2 Total Purchased Cost
The total purchased cost of the fabric filter system is the sum of the costs of the baghouse,
bags, auxiliary equipment, instruments and controls; and of taxes and freight. The last three
items generally are taken as percentages of the estimated total cost of the first three items.
Typical values, from Chapter 2, are 10% for instruments and controls, 3% for taxes, and 5%
for freight.
Bag costs can vary, from less than 15% to more than 100% of bare baghouse cost,
depending on type of fabric required. This situation makes it inadvisable to estimate total
purchased cost without considering both costs, and prevents effective use of factors to estimate
a single cost for the baghouse and bags.
5.3.3 Total Capital Investment
Using Chapter 2 methodology, the total capital investment (TCI) for most baghouses is
estimated from a series of factors applied to the purchased equipment cost to obtain direct and
indirect installation costs. The TCI is the sum of these three costs (i.e., purchased equipment
and direct and indirect installation costs). The required factors are given in Table 5.8.
Because bag costs can have such a large effect on the total purchased equipment cost, the
factors may cause overestimation of total capital investment when expensive bags are used.
Using stainless steel components may also cause overestimation. Because baghouses may vary
from small units installed within existing buildings to large, separate structures, specific
5-35
-------
factors for site preparation or for buildings are not given. However, costs for buildings may
be obtained from such references as Means Square Foot Costs 1986.[23] Land, working
capital, and off-site facilities are excluded from the table, as they are not normally required.
For very large installations, however, they may be needed and would be estimated on an as-
needed basis.
5-36
-------
Table 5.8 Capital Cost Factors for Fabric Filters3
Cost Item Factor
Direct costs
Purchased equipment costs
Fabric filter (EC) + bags + auxiliary equipment As estimated, A
Instrumentation 0.10 A
Sales taxes 0.03 A
Freight 0.05 A
Purchased Equipment Cost, PEC B = 1.18 A
Direct installation costs
Foundations & supports
Handling & erection
Electrical
Piping
Insulation for ductworkb
Painting"
Direct installation cost
Site preparation As required. SP
Buildings As required, Bldg.
Total Direct Cost 1.72 B + SP + Bldg.
Indirect Costs (installation)
Engineering 0.10 B
Construction and field expense 0.20 B
Contractor fees 0.1 OB
Start-up 0.01 B
Performance test 0.01 B
Contingencies 0.03 B
Total Indirect Cost, 1C 0.45 B
Total Capital Investment = DC + 1C 2.17 B + SP + Bldg.
"Reference [24]
blf ductwork dimensions have been established, cost may be estimated based on $10 to $12/ft (fourth quarter
1986) of surface for field application (Alternatively, refer to Chapter 10 of this Manual.) Fan housings and slocks
may also be insulated. [21]
The increased use of special coatings may increase this factor to 0.06B or higher. [The factors given in Table
5.8 are for average installation conditions. Considerable variation may be seen with other-than-average
installation circumstances. Moreover, the Table 5.8 factors may be too large for "packaged" fabric filters—those
pre-assembled baghouses that consist of the compartments, bags, waste gas fan and motor, and instruments and
controls. Because these packaged units require very little installation, their installation costs would be lower
(20-25% of the purchased equipment cost).
5-37
-------
5.4 Estimating Total Annual Costs
5.4.1 Direct Annual Cost
Direct annual costs include operating and supervisory labor, operating materials, replacement
bags, maintenance (labor and materials), utilities, and dust disposal. Most of these costs are
discussed individually below. They vary considerably with location and time, and, for this
reason, should be obtained to suit the specific baghouse system being coated. For example,
current labor rates may be found in such publications as the Monthly Labor Review, published
by the U.S. Department of Labor, Bureau of Labor Statistics.
5.4.1.1 Operating and Supervisory Labor
Typical operating labor requirements are 2 to 4 hours per shift for a wide range of filter
sizes.[24] Small or well-performing units may require less time, while very large or
troublesome units may require more. Supervisory labor is taken as 15% of operating labor.
5.4.1.2 Operating Materials
Operating materials are generally not required for baghouses. An exception is the use of
precoat materials injected on the inlet side of the baghouse to provide a protective dust layer
on the bags when sticky or corrosive particles might harm them. Adsorbents may be similarly
injected when the baghouse is used for simultaneous particle and gas removal. Costs for these
materials should be included on a dollars-per-mass basis (e.g., dollars per ton).
5.4.1.3 Maintenance
Maintenance labor varies from 1 to 2 hours per shift. [24] As with operating labor, these
values may be reduced or exceeded depending on the size and operating difficulty of a
particular unit. Maintenance materials costs are assumed to be equal to maintenance labor
costs. [24]
5.4.1.4 Replacement Parts
The major replacement part items are filter bags, which have a typical operating life of about
2 years. The following formula is used for computing the bag replacement cost:
CRCB - (CB + CL) * CRFB (5.13)
5-38
-------
where
CRCB = bag capital recovery cost ($/year)
CB = initial bag cost including taxes and freight ($)
CL = bag replacement labor ($)
CRFB = capital recovery factor whose value is a function of the annual interest rate
and the useful life of the bags (For instance, for a 7% interest rate and a 2-
year life, CRFB = 0.5531.)
The bag replacement labor cost (Q) will depend on such factors as the number, size, and
type of the bags; their accessibility; how they are connected to the baghouse tube-sheet; etc.
For example, in a reverse-air baghouse it would probably take from 10 to 20 man-minutes to
change an 8-inch by 24-foot bag that is clamped in place. This bag has a filtering surface area
of approximately 50 ft2. If the replacement labor rate were $21.12/h (including overhead),
CL would be from $0.07 to $0.14/ft2 of bag area. As Table 5.7 shows, for some bags (e.g.,
cotton), this range of CL would constitute a significant fraction of the purchased cost. For
pulse jets, replacement time would be about 5 to 10 man-minutes for a 5 inch by 10 foot bag
in a top-access baghouse. These bag replacement times are based on changing a minimum of
an entire module and on having typical baghouse designs. Times would be significantly longer
if only a few bags were being replaced or if the design for bag attachment or access were
atypical.
This method treats the bags as an investment that is amortized over the useful life of the
bags, while the rest of the control system is amortized over its useful life (typically 20 years;
see Subsection 5.4.2). Values of CRFB for bag lives different from 2 years can be calculated
from Equation 2.3.
5.4.1.5 Electricity
Power is required to operate system fans and cleaning equipment. Fan power for primary gas
movement can be calculated from Equation 2.7. After substituting into this equation a
combined fan-motor efficiency of 0.65 and a specific gravity of 1.000, we obtain:[26]
Power/fln = 0.0001810(AP)<9 (5.14)
where
Power/fln = fan power requirement (kWh/yr)
Q — system flow rate (acfm)
AP = system pressure drop (in. H2O)
6 = operating time (h/yr)
Cleaning energy for reverse-air systems can be calculated from the number of
compartments to be cleaned at one time (usually one, sometimes two), and the reverse gas-to-
cloth ratio (from about one to two times the forward gas-to-cloth ratio). Reverse-air pressure
drop varies up to 6 or 7 in. H2O depending on location of the fan pickup (before or after the
main system fan).[27] The reverse-air fan generally runs continuously.
5-39
-------
Typical energy consumption in kWh/yr for a shaker system operated 8,760 h/yr can be
calculated from: [5]
P = 0.053,4 (5.15)
where
A = gross fabric area (ft2)
5.4.1.6 Fuel
If the baghouse or associated ductwork is heated to prevent condensation, fuel costs should
be calculated as required. These costs can be significant, but may be difficult to predict. For
methods of calculating heat transfer requirements, see Perry.[26]
5.4.1.7 Water
Cooling process gases to acceptable temperatures for fabrics being used can be done by
dilution with air, evaporation with water, or heat exchange with normal equipment. The last
two cases require consumption of plant water, although costs are not usually significant.
Section 4.4 of Chapter 4 provides information on estimating cooling-water costs.
5.4.1.8 Compressed Air
Pulse-jet filters use compressed air at pressures of about 60 to 100 psig. Typical consumption
is about 2 scfm/1,000 cfm of gas filtered. [5] For example, a unit filtering 20,000 cfm of gas
uses about 40 scfm of compressed air for each minute the filter is operated.
5.4.1.9 Dust Disposal
If collected dust cannot be recycled or sold, it must be landfilled or disposed of in some other
manner. Disposal costs are site-specific, but they may typically run $20 or $30 per ton
exclusive of transportation (see Section 2.4, Chapter 2).
5.4.2 Indirect Annual Cost
These include such costs as capital recovery, property taxes, insurance, administrative costs
("G&A"), and overhead. The capital recovery cost is based on the equipment lifetime and the
annual interest rate employed. (See Chapter 2 for a thorough discussion of the capital recovery
cost and the variables that determine it.) For fabric filters, the system lifetime varies from 5
to 40 years, with 20 years being typical.[24] However, this does not apply to the bags, which
usually have much shorter lives. (See Section 5.4.1.) Therefore, as Chapter 2 suggests, when
figuring the system capital recovery cost, one should base it on the installed capital cost less
5-40
-------
the cost of replacing the bags (i.e., the purchased cost of the bags plus the cost of labor
necessary to replace them). In other words:
CRCS = [TCI - CB - CJ CRFS (5.16)
where
CRCS = capital recovery cost for fabric filter system ($/yr)
TCI = total capital investment ($)
CB = initial cost of bags including taxes and freight ($)*
CL = labor cost for replacing bags ($)
CRFS = capital recovery, factor for fabric filter system (defined in Chapter 2).
For example, for a 20-year system life and a 7% annual interest rate, the CRFS would be
0.09439.
As Chapter 2 suggests, the suggested factor to use for property taxes, insurance, and
administrative charges is 4% of the TCI. Finally, the overhead is calculated as 60% of the
sum of operating, supervisory, and maintenance labor, and maintenance materials.
5.4.3 Recovery Credits
For processes that can reuse the dust collected in the baghouse or that can sell the dust in a
local market, such as fly ash sold as an extender for paving mixes, a credit should be taken.
As used below, this credit (RC) appears as a negative cost.
5.4.4 Total Annual Cost
Total annual cost for owning and operating a fabric filter system is the sum of the components
listed in Sections 5.4.1 through 5.4.3, i.e.:
TAC = DC + IC-RC (5.17)
where
TAG = total annual cost ($)
DC = direct annual cost ($)
1C = indirect annual cost ($)
RC = recovery credits (annual) ($)
Typically, 8% of the bag initial cost.
5-41
-------
5.4.5 Example Problem
Assume a baghouse is required for controlling fly ash emissions from a coal-fired boiler. The
flue gas stream is 50,000 acfm at 325 °F and has an ash loading of 4 gr/ft3. Analysis of the
ash shows a mass median diameter of 7 jum. Assume the baghouse operates for 8,640 h/yr
(360 d).
The gas-to-cloth ratio (G/C) can be taken from Table 5.1 as 2.5, for woven fabrics in
shaker or reverse-air baghouses, or 5, for felts used in pulse-jet baghouses. If a factor method
were used for estimating G/C, Table 5.3 for shakers would yield the following values: A =
2, B = 0.9, and C = 1.0. The gas-to-cloth ratio would be:
2x0.9x 1.0 = 1.8.
This value could also be used for reverse-air cleaning. For a pulse-jet unit, Table 5.4 gives
a value of 9.0 for factor A and 0.8 for factor B. Equation 5.11 becomes:
V = 2.878 x 9.0 x 0.8(275)"°2335(4)"006021 (0.7471 + 0.0853 In 7) ,, ...
= 4.69 ( }
Because this value is so much greater than the shaker/reverse-air G/C, we conclude that
the pulse-jet baghouse would be the least costly design.2 Assume the use of on-line cleaning
in a common housing structure and, due to the high operating temperature, the use of glass
filter bags (see Table 5.5). At a gas-to-cloth ratio of 4.69, the fabric required is3
50,000 acfm/4.69 fpm = 10,661 ft2.
From Figure 5.4, the cost of the baghouse ("common housing" design) is:
Cost = 9,688 + 5.552(10,661) = $68,878 (5.18)
2This conclusion is based on the inference that a much bigger G/C would yield lower capital and, in turn, annual costs. However,
to make a more rigorous selection, we would need to calculate and compare the total annual costs of all three baghouse designs
(assuming all three are technically acceptable). The reader is invited to make this comparison. Further discussion of the effects of G/C
increases, and accompanying pressure drop increases, on overall annual costs will be found in Reference 27.
3This is the total (gross) bag area required. No bag area adjustment factor has been applied here, because this is a common housing
pulse jet unit that is cleaned continuously during operation. Thus, no extra bag compartment is needed, and the gross and net bag are.
are equal.
5-42
-------
Insulation is required. The insulation add-on cost from Figure 5.4 is:
Cost = 1,428 + 0.931(10,661) = $11,353 (5.19)
From Table 5.7, bag costs are $1.24/ft2 for 5-1/8-inch diameter glass fiber, bottom
removal bags. Total bag cost is
10,661 ft2 x $1.24/ft2 = $13,220.
For 10 ft long cages,
fabric area per cage = 5]/g in./12 in./ft x n; x 10 ft
= 13.42 ft2.
The number of cages = 10,661 ft2 /13.42 ft2
= 795 cages (rounded up to next integer)
From Table 5.7, individual cage cost is
3.941 + 0.163(13.42ft2) = $6.128.
Total cage cost is
795 cages x $6.128/cage = $4,872.
Assume the following auxiliary costs have been estimated from data in other parts of the
Manual:
Ductwork $14,000
Fan 14,000
Motor 7,000
Starter 3,500
Dampers 7,200
Compressor 6,000
Screw conveyor 4,000
Stack 7.000
Total $62,700
Direct costs for the fabric filter system, based on the factors in Table 5.8, are given in
Table 5.9. (Again, we assume site preparation and buildings costs to be negligible.) Total
capital investment is $412,000. Table 5.10 gives the direct and indirect annual costs, as
calculated from the factors given in Section 5.4.1. For bag replacement labor, assume 10 min
5-43
-------
per bag for each of the 795 bags. At a maintenance labor rate of $321.12 (including
overhead), the labor cost is $2,809 for 133 h. The bags are assumed to be replaced every 2
yr. The replacement cost is calculated using Equation 5.13.
Pressure drop (for energy costs) can be calculated from Equations 5.8 and 3.9, with the
following assumed values:
in. H2O/l(ft/min)
fL. = 15
2 lb/ft2
Pj = 100 psig
cleaning interval = 10 min
We further assume that a G/C of 4.69 ft/min is a good estimate of the mean face velocity over
the duration of the filtering cycle.
Wo - CVQ
, gr 1 Ib , .,„ ft in-
= 4_&_ x x 4.69 x 10 mm
ft3 7,000gr min
= 0.0268 lb/ft2
5-44
-------
Table 5.9 Capital Costs for Fabric Filter System
Example Problem
Cost Item
Cost
Direct Costs
Purchased equipment costs
Fabric filter (with insulation)(EC) $80,231
Bags and cages 18,092
Auxiliary equipment 62.700
Sum = A $161,023
Instrumentation, 0.1 A 16,102
Sales taxes, 0.03A 4,831
Freight, 0.05A 8.051
Purchased equipment cost, B $190,007
Direct installation costs
Foundation and supports, 0.04B 7,600
Handling and erection, 0.50B 95,004
Electrical, 0.08B 15,201
Piping, 0.01B 1,900
Insulation for ductwork, 0.07B 13,300
Painting, 0.02B 3.800
Direct installation cost $136,805
Site preparation
Facilities and buildings
Total Direct Cost $326,812
Indirect Costs (installation)
Engineering, 0.1 OB 19,001
Construction and field expenses, 0.20B 38,001
Contractor fees, 0.10B 19,001
Start-up, 0.0IB 1,900
Performance test, 0.01B 1,900
Contingencies, 0.03B 5.700
Total Indirect Cost $85,503
Total Capital Investment (rounded) $412,000
5-45
-------
Table 5.10 Annual Costs for Fabric Filter System
Example Problem
Cost Item
Direct Annual Costs. DC
Operating labor
Operator
Supervisor
Operating materials
Maintenance
Labor
Material
Replacement parts, bags
Utilities
Electricity
Compressed air
(dried and filtered)
Waste disposal
Total DC
(rounded)
Calculations
6 h v 360 days , $12
day yr h
15% of operator = 0.15 x 25,920
3 h v 360 days v $13.20
day yr h
1 00% of maintenance labor
[2,809 + (13,220 x 1.08a)] x 0.5531
0000181 v 50000 acfm v 103 in HO v ^-(^ ^ ^ $0-06
"' '" '" "2~ yr kWh
2 scfm v <00()0 acfin v $0.16 v 60 min v 8,640
1,000 acfm ' 1,000 scf h yr
at $20/ton on-site for essentially 100% collection
cc • 4 gr 1 Ib _„ „„„ £.3 60 min
ft3 7.000 gr h
8,640 h 1 ton v $20
Cost
$25,920
3,888
—
14,256
14,256
9,451
48,322
8,294
148.114
272,500
2,000 Ib ton
Indirect Annual Costs. 1C
Overhead
Administrative charges
Property Tax
Insurance
Capital recovery*1
Total 1C (rounded)
Total Annual Cost
(rounded)
60% of sum of operating, supv., & maint. labor & maint.
materials = 0.6(25,920+3,888+14,256+14,256)
2% of Total Capital Investment = 0.02
1% of Total Capital Investment = 0.01
1% of Total Capital Investment = 0.01
0.09439 (412,315 - 2,809 - 13,220 x 1
($412,315)
($412.315)
($412,315)
.08)
34,992
8.246
4,123
4.123
37.306
88,800
361,000
aThe 1.08 factor is for freight and sales taxes.
bThe capital recovery cost factor, CRF, is a function of the fabric filter or equipment life and the opportunity
cost of the capital (i.e., interest rate). For example, for a 20-year equipment life and a 7% interest rate. CRF
= 0.09439.
5-46
-------
ft .„ in. FLO/(ft/min) ,
AP = 6.08 x 4.69— x (100 psig)'065 + 15 ?— -/ft2
min Ib
x 0.0268— x 4.69 —
ft2 min
= 3.32 in. H2O across the fabric (when fully loaded).
We will assume that the baghouse structure and the ductwork contribute an additional 3 in. H2O
and 4 in. H2O, respectively. The total pressure drop is, therefore, 10.3 inches.
The total annual cost is $361,000, nearly half of which is for ash disposal. If a market for the
fly ash could be found, the total annual cost would be greatly reduced. For example, if $2/ton
were received for the ash, the total annual cost would drop to $198,000 ($361,000 - $148,000
- $14,800), or 55% of the cost when no market exists. Clearly, the total annual cost is extremely
sensitive to the value chosen for the dust disposal cost in this case. In this and in similar cases,
this value should be selected with care.
Finally, as discussed hi the Design Procedures Section (Section 5.2), an electrostatically
enhanced baghouse (an emerging technology) may have up to 30% lower total annual costs than
the baghouse estimated in the example problem.
5.5 Acknowledgments
We gratefully acknowledge the following companies for contributing data to this section:
• Aget Manufacturing Company (Adrian, Michigan)
• BACT Engineering, Inc. (Arlington Heights, Illinois)
• The BHA Group (Kansas City, Missouri)
• Dustex Corporation (Charlotte, North Carolina)
• Fuller Company (Bethlehem, Pennsylvania)
• W. L. Gore and Associates, Inc. (Elkton, Maryland)
• Griffin Environmental Company, Inc. (Syracuse, New York)
• W. W. Sly Manufacturing Company (Cleveland, Ohio)
• Zurn Industries, Inc. (Birmingham, Alabama)
5-47
-------
References
[1] Van Osdell, D. W., M. B. Ranade, G. P. Greiner, and D. F. Furlong, Electrostatic
Augmentation of Fabric Filtration: Pulse-jet Pilot Unit Experience, November 1982 (EPA-
600/7-82-062).
[2] Viner, A. S., G. P. Greiner, D. F. Furlong, and R. G. Hurst, Pilot-Scale Evaluation of Top-
Inlet and Advanced Electrostatic Filtration, October 1986 (EPA-600/7-86-042).
[3] Donovan, R. P., Fabric Filtration For Combustion Sources, Marcel Dekker, Inc., New York,
1985.
[4] Turner, J. H., "Bag Filtration," in Handbook of Multiphase Systems, ed. by G. Hetsroni,
Hemisphere, New York, 1982.
[5] Turner, J. E., and J. D. McKenna, "Control of Particles by Filters," in Handbook of Air
Pollution Technology, ed. by S. Calvert and E. Englund, John Wiley & Sons, New York,
1984.
[6] Penny, C. W., Electrostatic Effects in Fabric Filtration: Volume 1. Fields, Fabrics, and
Particles (Annotated Data), September 1978 (EPA-600/7-78-142A[NTIS PB 288576]).
[7] Frederick, E. R., Electrostatic Effects in Fabric Filtration: Volume 11. Triboelectric
Measurements and Bag Performance, July 1978 (EPA600/7-78.1A2B[NTIS PB 287207]).
[8] Frederick, E. R., Electrical Effects in Paniculate Matter Processes, Filter Media
Specification, Pittsburgh, 1987.
[9] Dennis, R., and H. A. Klemm, "Modeling Concepts for Pulse Jet Filtration." JAPCA, 30(1),
January 1980.
[10] Leith, D. and M. J. Ellenbecker, "Theory for Pressure Drop in a Pulse-Jet Cleaned Fabric
Filter." Atm. Environment, 14, 1980, pp. 845-852.
[11] Koehler, J. L. and D. Leith, "Model Calibration for Pressure Drop in a Pulse-Jet Cleaned
Fabric Filter," Atm. Environment, 17(10), 1983, pp. 1909-1913.
[12] Northrop Services, Inc. Fabric Filter Workshop Reference Materials, 1977 Workshop,
Air Pollution Training Institute.
[13] Viner, A. S., and B. R. Locke, Cost and Performance Models for Electrostatically
Stimulated Fabric Filters, April 1984 (EPA 600/8-84-016).
5-48
-------
[ 14] Vatavuk, W. M., and R. B. Neveril, "Estimating Costs of Air-Pollution Control Systems,
Part XI: Estimate the Size and Cost of Baghouses," Chemical Engineering, March 22,
1982, pp. 153-158.
[15] Frey, R. F., and T. V. Reinauer, "New Filter Rate Guide," Air Engineering, 30 April
1964.
[16] Dennis, R., et al.., Filtration Model for Coal Fly Ash with Glass Fabrics, August 1977
(EPA-600/7-77-084 [NTIS PB 276489]).
[17] Owen, M. K. and A. S. Viner, Microcomputer Programs for Paniculate Control, June
1985 (EPA-600/M5-025a).
[18] Dennis, R. and H. A. Klemm, Fabric Filter Model Change: Vol. 1, Detailed Technical
Report, February 1979 (EPA-600/7-79-043a [NTISS PB 293551]).
[19] Viner, A. S., et al, "Comparison of Baghouse Test Results with the GCA/EPA Design
Model", JAPCA, 34(8), August 1984.
[20] Reigel, S. A. and R. P. Bundy. "Why the Swing to Baghouses?", Power, 121-1, January
1977, pp.68-73.
[21] ETS, Inc., Roanoke, VA.
[22] Fuller Company, Bethlehem, PA.
[23] R. S. Means Company, Inc., Means Square Foot Costs 1986, Kingston, MA.
[24] Vatavuk, W. M., and R. B. Neveril, "Estimating Costs of Air-Pollution Control S\ stems.
Part II: Factors for Estimating Capital and Operating Costs," Chemical Engineering,
November 3, 1980, pp. 157-162.
[25] Personal Communication from Frank Smith, Griffin Environmental, to Jim Turner,
Research Triangle Institute, November 8, 1988.
[26] Perry, Robert H., et al., Perry's Chemical Engineers' Handbook (Fourth Edition).
McGraw-Hill, New York, 1963, p. 6-20.
[27] Personal communication from Gary Greiner, ETS, Inc., to Jim Turner, Research Triangle
Institute, October 24, 1986.
[28] Perry, Robert H., et al., Perry's Chemical Engineers' Handbook (Sixth Edition),
McGraw-Hill, New York, 1984.
5-49
-------
[29] McKenna, J. D., J. H. Turner, D. Furlong, and D. S. Beachler, Fabric Filters-Baghouses,
I, Theory, Design and Selection, in preparation, ETS, Inc., Roanoke, VA.
5-50
-------
Chapter 6
ELECTROSTATIC
PRECIPITATORS
James H. Turner
Phil A. Lawless
Toshiaki Yamamoto
David W. Coy
Research Triangle Institute
Research Triangle Park, NC 27709
Gary P. Greiner
John D. McKenna
ETS. Inc.
Roanoke. VA 24018-4394
William M. Vatavuk
Innovative Strategies and Economics Group, OAQPS
U.S. Environmental Protection Agency
Research Trianele Park, NC 27711
December, 1995
6-1
-------
Contents
6.1 Process Description 6-5
6.1.1 Introduction 6-5
6.1.2 Types of ESPs 6-5
6.1.2.1 Plate-Wire Precipitators 6-5
6.1.2.2 Flat Plate Precipitators 6-7
6.1.2.3 Tubular Precipitators 6-10
6.1.2.4 Wet Precipitators 6-10
6.1.2.5 Two-Stage Precipitators 6-10
6.1.3 Auxiliary Equipment 6-11
6.1.4 Electrostatic Precipitation Theory 6-12
6.1.4.1 Electrical Operating Point 6-13
6.1.4.2 Particle Charging 6-15
6,1.4.3 Particle Collection 6-18
6.1.4.4 Sneakage and Rapping Reentrainment 6-20
6.2 ESP Design Procedure 6-21
6.2.1 Specific Collecting Area 6-21
6.2.1.1 SCA Procedure with Known Migration Velocity 6-22
6.2.1.2 Full SCA Procedure 6-24
6.2.1.3 Specific Collecting Area for Tubular Precipitators 6-32
6.2.2 Flow Velocity 6-32
6.2.3 Pressure Drop Calculations 6-33
6-2
-------
6.2.4 Particle Characteristics 6-36
6.2.5 Gas Characteristics 6-37
6.2.6 Cleaning 6-37
6.2.7 Construction Features 6-38
6.3 Estimating Total Capital Investment 6-39
6.3.1 Equipment Cost 6-39
6.3.1.1 ESP Costs 6-39
6.3.1.2 Retrofit Cost Factor 6-42
6.3.1.3 Auxiliary Equipment 6-43
6.3.1.4 Costs for Two-Stage Precipitators 6-45
6.3.2 Total Purchased Cost 6-45
6.3.3 Total Capital Investment (TCI) 6-46
6.4 Estimating Total Annual Costs 6-46 "
6.4.1 Direct Annual Costs 6-46
6.4.1.1 Operating and Supervisory Labor 6-46
6.4.1.2 Operating Materials 6-47
6.4.1.3 Maintenance 6-49
6.4.1.4 Electricity 6-49
6.4.1.5 Fuel 6-50
6.4.1.6 Water 6-51
6.4.1.7 Compressed Air 6-51
6.4.1.8 Dust Disposal 6-51
6.4.1.9 Wastewater Treatment 6-51
6-3
-------
6.4.1.10 Conditioning Costs 6-52
6.4.2 Indirect Annual Costs 6-52
6.4.3 Recovery Credits 6-52
6.4.4 Total Annual Cost 6-52
6.4.5 Example Problem 6-53
6.4.5.1 Design SCA 6-53
6.4.5.2 ESP Cost 6-57
6.4.5.3 Costs of Auxiliaries 6-57
6.4.5.4 Total Capital Investment 6-57
6.4.5.5 Annual Costs-Pressure Drop 6-58
6.4.5.8 Total Annual Cost 6-61
6.5 Acknowledgments 6-61
Appendix 6A 6-62
References 6-67
6-4
-------
6.1 Process Description
6.1.1 Introduction
An electrostatic precipitator (ESP) is a particle control device that uses electrical forces to move
the particles out of the flowing gas stream and onto collector plates. The particles are given an
electrical charge by forcing them to pass through a corona, a region in which gaseous ions flow.
The electrical field that forces the charged particles to the walls comes from electrodes
maintained at high voltage in the center of the flow lane.
Once the particles are collected on the plates, they must be removed from the plates without
reentraining them into the gas stream. This is usually accomplished by knocking them loose
from the plates, allowing the collected layer of particles to slide down into a hopper from which
they are evacuated. Some precipitators remove the particles by intermittent or continuous
washing with water.
6.1.2 Types of ESPs
ESPs are configured in several ways. Some of these configurations have been developed for
special control action, and others have evolved for economic reasons. The types that will be
described here are (1) the plate-wire precipitator, the most common variety; (2) the flat plate
precipitator, (3) the tubular precipitator: (4) the wet precipitator, which may have any of the
previous mechanical configurations; and (5) the two-stage precipitator.
6.1.2.1 Plate-Wire Precipitators
Plate-wire ESPs are used in a wide variety of industrial applications, including coal-fired boilers.
cement kilns, solid waste incinerators, paper mill recovery boilers, petroleum refining catalytic
cracking units, sinter plants, basic oxygen furnaces, open hearth furnaces, electric arc furnaces,
coke oven batteries, and glass furnaces.
In a plate-wire ESP, gas flows between parallel plates of sheet metal and high-voltage
electrodes. These electrodes are long wires weighted and hanging between the plates or are
supported there by mast-like structures (rigid frames). Within each flow path, gas flow must
pass each wire in sequence as flows through the unit.
The plate-wire ESP allows many flow lanes to operate in parallel, and each lane can be quite
tall. As a result, this type of precipitator is well suited for handling large volumes of gas. The
need for rapping the plates to dislodge the collected material has caused the plat to be divided
into sections, often three or four in series with one another, which can be rapped independent.
The power supplies are often sectionalized in the same way to obtain higher operating voltages,
and further electrical sectionalization may be used for increased reliability. Dust also deposits
on the discharge electrode wires and must be periodically removed similarly to the collector
plate.
6-5
-------
The power supplies for the ESP convert the industrial ac voltage (220 to 480 V) to pulsating
dc voltage in the range of 20,000 to 100,000 V as needed. The supply consists of a step-up
transformer, high-voltage rectifiers, and sometimes filter capacitors. The unit may supply either
half-wave or full-wave rectified dc voltage. There are auxiliary components and controls to
allow the voltage to be adjusted to the highest level possible without excessive sparking and to
protect the supply and electrodes in the event a heavy arc or short-circuit occurs.
The voltage applied to the electrodes causes the air between the electrodes to break down
electrically, an action known as a "corona". The electrodes usually are given a negative polarity
because a negative corona supports a higher voltage than a positive corona before sparking
occurs. The ions generated in the corona follow electric field lines from the wires to the
collecting plates. Therefore, each wire establishes a charging zone through which the particles
must pass.
Particles passing through the charging zone intercept some of the ions, which become
attached. Small aerosol particles (<1 /urn diameter) can absorb tens of ions before their total
charge becomes large enough to repel further ions, and large particles (>10 ^m diameter) can
absorb tens of thousands. The electrical forces are therefore much stronger on the large
particles.
As the particles pass each successive wire, they are driven closer and closer to the collecting
walls. The turbulence in the gas, however, tends to keep them uniformly mixed with the gas.
The collection process is therefore a competition between the electrical and dispersive forces.
Eventually, the particles approach close enough to the walls so that the turbulence drops to low
levels and the particles are collected.
If the collected particles could be dislodged into the hopper without losses, the ESP would
be extremely efficient. The rapping that dislodges the accumulated layer also projects some of
the particles (typically 12 percent for coal fly ash) back into the gas stream. These reentrained
particles are then processed again by later sections, but the particles reentrained in the last
section of the ESP have no chance to be recaptured and so escape the unit.
Practical considerations of passing the high voltage into the space between the lanes and
allowing for some clearance above the hoppers to support and align electrodes leave room for
part of the gas to flow around the charging zones. This is called "sneakage" and amounts to 5
to 10 percent of the total flow. Antisneakage baffles usually are placed to force the sneakage
flow to mix with the main gas stream for collection in later sections. But, again, the sneakage
flow around the last section has no opportunity to be collected.
These losses play a significant role in the overall performance of an ESP. Another major
factor is the resistivity of the collected material. Because the particles form a continuous layer
on the ESP plates, all the ion current must pass through the layer to reach the ground-plates.
This current creates an electric field in the layer, and it can become large enough to cause local
electrical breakdown. When this occurs, new ions of the wrong polarity are injected into the
wire-plate gap where they reduce the charge on the particles and may cause sparking. This
breakdown condition is called "back corona"
6-6
-------
Back corona is prevalent when the resistivity of the layer is high, usually above 2 x 10n
ohm-cm. For lower resistivities, the operation of the ESP is not impaired by back coronas, but
resistivities much higher than 2 x 10" ohm-cm considerably reduce the collection ability of the
unit because the severe back corona causes difficulties in charging the particles. At resistivities
below 108 ohm-cm, the particles are held on the plates so loosely that rapping and nonrapping
reentrainment become much more severe. Care must be taken in measuring or estimating
resistivity because it is strongly affected by variables such as temperature, moisture, gas
composition, particle composition, and surface characteristics.
6.1.2.2 Flat Plate Precipitators
A significant number of smaller precipitators (100,000 to 200,000 acfm) use flat plates instead
of wires for the high-voltage electrodes. The flat plates (United McGill Corporation patents)
increase the average electric field that can be used to collect the particles, and they provide an
increased surface area for the collection of particles. Corona cannot be generated on flat plates
by themselves, so corona-generating electrodes are placed ahead of and sometimes behind the
flat plate collecting zones. These electrodes may be sharp-pointed needles attached to the edges
of the plates or independent corona wires. Unlike place-wire or tubular ESPs. this design
operates equally well with either negative or positive polarity. The manufacturer has chosen to
use positive polarity to reduce ozone generation.
A flat plate ESP operates with little or no corona current flowing through the collected dust.
except directly under the corona needles or wires. This has two consequences. The first is that
the unit is somewhat less susceptible to back corona than conventional units are because no back
corona is generated in the collected dust, and particles charged with both polarities of ions have
large collection surfaces available. The second consequence is that the lack of current in the
collected layer causes an electrical force that tends to remove the layer from the collecting
surface: this can lead to high rapping losses.
Flat plate ESPs seem to have wide application for high-resistivity particles with small (1 to
2 \uri) mass median diameters (MMDs). These applications especially emphasize the strengths
of the design because the electrical dislodging forces are weaker for small particles than for large
ones. Fly ash has been successfully collected with this type of ESP. but low-flow velocity
appears to be critical for avoiding high rapping losses.
6-7
-------
10
11
15
12
1. Weather Enclosure 9.
2. Transf oner/Rectifier 10.
3. High Voltage Bus. Duct 11.
4. High Voltage Support Regulator 12,
5. Roof Door 13.
6. Discharge Electrode Rapper 14.
(Side Location Available) 15.
7. Gas Passage 16.
8. Sidewall
Sidewall Door
Collecting Surface
Collecting Surface Rapper
Dischare Electrode
High Voltage Support Frane
Slide Bearing
Gas Distribution Device
Inlet Nozzle
Figure 6.1: Electrostatic Precipitator Coiponents
(Courtesy of the Institute of Clean Air Companies)
6-8
-------
Collector Plate
54 in.
Discharge Plate
36 in.
^1
T6 in!'
Plan view of McGill E.P. electrodes with
typical dinensions
Collector Plate
--•1
.Discha/geliire,
Plan view of conventional E. P. electrodes with typical diiensions
Figure 6.2: Flat-plate and Plate-wire ESP Configurations
(Courtesy of United McGill Corporation)
6-9
-------
6.1.2.3 Tubular Precipitators
The original ESPs were tubular like the smokestacks they were placed on, with the high-voltage
electrode running along the axis of the tube. Tubular precipitators have typical applications in
sulfuric add plants, coke oven by-product gas cleaning (tar removal), and, recently, iron and
steel sinter plants. Such tubular units are still used for some applications, with many tubes
operating in parallel to handle increased gas flows. The tubes may be formed as a circular,
square, or hexagonal honeycomb with gas flowing upwards or downwards. The length of the
tubes can be selected to fit conditions. A tubular ESP can be tightly scaled to prevent leaks of
material, especially valuable or hazardous material.
A tubular ESP is essentially a one-stage unit and is unique in having all the gas pass through
the electrode region. The high-voltage electrode operates at one voltage for the entire length of
the tube, and the current varies along the length as the particles are removed from the system.
No sneakage paths are around the collecting region, but corona nonuniformities may allow some
particles to avoid charging for a considerable fraction of the tube length.
Tubular ESPs comprise only a small portion of the ESP population and are most commonly
applied where the paniculate is either wet or sticky. These ESPs, usually cleaned with water.
have reentrainment losses of a lower magnitude than do the dry paniculate precipiiators.
6.1.2.4 Wet Precipitators
Any of the precipitator configurations discussed above may be operated with wet walls instead
of dry. The water flow may be applied intermittently or continuously to wash the collected
particles into a sump for disposal. The advantage of the wet wall precipitator is that it has no
problems with rapping reentrainment or with back coronas. The disadvantage is the increased
complexity of the wash and the fact that the collected slurry must be handled more carefully than
a dry product, adding to the expense of disposal.
6.1.2.5 Two-Stage Precipitators
The previously described precipitators are all parallel in nature, i.e., the discharge and collecting
electrodes are side by side. The two-stage precipitator invented by Penney is a series device
with the discharge electrode, or ionizer, preceding the collector electrodes. For indoor
applications, the unit is operated with positive polarity to limit ozone generation.
Advantages of this configuration include more time for particle charging, less propensity for
back corona, and economical construction for small sizes. This type of precipitator is generally
used for gas flow volumes of 50,000 acfm and less and is applied to submicrometer sources
6-10
-------
Spray Cooler
Hood
Ductwork
Direct Exhaust
Mechanical CoUector
Screw Conveyor
Figure 6.3: Control Device and Typical Auxiliary Equipment
emitting oil mists, smokes, fumes, or other sticky particulates because there is little electrical
force to hold the collected particulates on the plates. Modules consisting of a mechanical
prefilter, ionizer, collecting-plate cell, after-filter, and power pack may be placed in parallel or
series-parallel arrangements. Preconditioning of gases is normally part of the system. Cleaning
may be by water wash of modules removed from the system up to automatic, in-place detergent
spraying of the collector followed by air-blow drying.
Two-stage precipitators are considered to be separate and distinct types of devices compared
to large, high-gas-volume, single-stage ESPs. The smaller devices are usually sold as pre-
engineered, package systems.
6.1.3 Auxiliary Equipment
Typical auxiliary equipment associated with an ESP system is shown schematically in Figure
6.3. Along with the ESP itself, a control system usually includes the following auxiliary
equipment: a capture device (i.e., hood or direct exhaust connection); ductwork; dust removal
equipment (screw conveyor, etc.); fans, motors, and starters; and a stack. In addition, spray
coolers and mechanical collectors may, be needed to precondition the gas before it reaches the
ESP. Capture devices are usually hoods that exhaust pollutants into the ductwork or are direct
exhaust couplings attached to a combustor or process equipment. These devices are usually
refractory lined, water cooled, or simply fabricated from carbon steel, depending on the gas-
stream temperatures. Refractory or water-cooled capture devices are used where the wall
temperatures exceed 800°F; carbon steel is used for lower temperatures. The ducting, like the
capture device, should be water cooled, refractory, or stainless steel for hot processes and carbon
steel for gas temperatures below approximately 1,150°F (duct wall temperatures <800°F). The
ducts should be sized for a gas velocity of approximately 4,000 ft/min for the average case to
prevent particle deposition in the ducts. Large or dense particles might require higher velocities,
but rarely would lower velocities be used. Spray chambers may be required for processes where
6-11
-------
the addition of moisture, or decreased temperature or gas volume, will improve precipitation or
protect the ESP from warpage. For combustion processes with exhaust gas temperatures below
approximately 700°F, cooling would not be required, and the exhaust gases can be deli- J
directly to the precipitator.
When much of the pollutant loading consists of relatively large particles, mecha; ,1
collectors, such as cyclones, may be used to reduce the load on the ESP, especially at high ;t
concentrations. The fans provide the motive power for air movement and can be mot :d
before or after the ESP. A stack, normally used, vents the cleaned stream to the atmo.c~ re.
Screw conveyors or pneumatic systems are often used to remove captured dust from the :om
of the hoppers.
Wet ESPs require a source of wash water to be injected or sprayed near the to )f the
collector plates either continuously or at timed intervals. The water flows with the collected
particles into a sump from which the fluid is pumped. A portion of the fluid may be recycled
to reduce the total amount of water required. The remainder is pumped directly to a settling
pond or passed through a dewatering stage, with subsequent disposal of the sludge.
Gas conditioning equipment to improve ESP performance by changing dust resistivity is
occasionally used as part of the original design, but more frequently it is used to upgrade
existing ESPs. The equipment injects an agent into the gas stream ahead of the ESP. Usually,
the agent mixes with the particles and alters their resistivity to promote higher migration
velocity, and thus higher collection efficiency. However, electrical properties of the gas may
change, rather than dust resistivity. For instance, cooling the gas will allow more voltage to be
applied before sparking occurs. Significant conditioning agents that are used include SO3,
F^SO^ sodium compounds, ammonia, and water, but the major conditioning agent by usage is
SO3. A typical dose rate for any of the gaseous agents is 10 to 30 ppm by volume.
The equipment required for conditioning depends on the agent being used. A typical S03
conditioner requires a supply of molten sulfur. It is stored in a heated vessel and supplied to a
burner, where it is oxidized to SO2. The SO2 gas is passed over a catalyst for further oxidation
to S03. The S03 gas is then injected into the flue gas stream through a multi-outlet set of probes
that breach a duct. In place of a sulfur burner to provide S02, liquid S02 may be vaporized from
a storage tank. Although their total annual costs are higher, the liquid SO2 systems have lower
capital costs and are easier to operate than the molten sulfur based systems.
Water or ammonia injection requires a set of spray nozzles in the duct, along with pumping
and control equipment.
Sodium conditioning is often done by coating the coal on a conveyor with a powder
compound or a water solution of the desired compound. A hopper or storage tank is of ton
positioned over the conveyor for this purpose.
6.1.4 Electrostatic Precipitation Theory
6-12
-------
The theory of ESP operation requires many scientific disciplines to describe it thoroughly.
The ESP is basically an electrical machine. The principal actions are the charging of particles
and forcing them to the collector plates. The amount of charged paniculate matter affects the
electrical operating point of the ESP. The transport of the particles is affected by the level of
turbulence in the gas. The losses mentioned earlier, sneakage and rapping reentrainment, are
major influences on the total performance of the system. The particle properties also leave a
major effect on the operation of the unit.
The following subsections will explain the theory behind (1) electrical operating points in
the ESP, (2) particle charging, (3) particle collection, and (4) sneakage and rapping
reentrainment. General references for these topics are White [1] or Lawless and Sparks [2].
6.1.4.1 Electrical Operating Point
The electrical operating point of an ESP section is the value of voltage and current at which the
section operates. As will become apparent, the best collection occurs when the highest electric
field is present, which roughly corresponds to the highest voltage on the electrodes. In this work,
the term "section" represents one set of plates and electrodes in the direction of flow. This unit
is commonly called a "field", and a "section" or "bus section" represents a subdivision of a
"field" perpendicular to the direction of flow. In an ESP model and in sizing applications, the
two terms "section" and "field" are used equivalently because the subdivision into bus sections
should have no effect on the model. This terminology has probably arisen because of the
frequent use of the word "field" to refer to the electric field.
The lowest acceptable voltage is the voltage required for the formation of a corona, the
electrical discharge that produces ions for charging particles. The (negative) corona is produced
when an occasional free electron near the high-voltage electrode, produced by a cosmic ray.
gains enough energy from the electric field to ionize the gas and produce more free electrons.
The electric field for which this process is self-sustained has been determined experimentally.
For round wires, the field at the surface of the wire is given by:
E^, = 3 .126 x 106dr [1 + 0.0301 (dr/rj0 5] (6.1)
where
Ee = corona onset field at the wire surface (V/m)
dr - relative gas density, referred to 1 atm pressure and 20 °C (dimensionless)
rw = radius of the wire, meters (m)
This is the field required to produce "glow" corona, the form usually seen in the laboratory
on smooth, clean wires. The glow appears as a uniform, rapidly moving diffuse light around
the electrode. After a period of operation, the movement concentrates into small spots on the
6-13
-------
wire surface, and the corona assumes a tuft-like appearance. The field required to produce "tuft"
corona, the form found in full-scale ESPs, is 0.6 times the value of Ec.
The voltage that must be applied to the wire to obtain this value of field, Vc, is found by
integrating the electric field from the wire to the plate. The field follows a simple "1/r"
dependence in cylindrical geometry. This leads to a logarithmic dependence of voltage on
electrode dimensions. In the plate-wire geometry, the field dependence is somewhat more
complex, but the voltage still shows the logarithmic dependence. Vc is given by:
(6.2)
where
Vc = corona onset voltage (V)
d = outer cylinder radius for tubular ESP (m)
4/Ti x (wire-plate separation) for plate-wire ESP (m)
No current will flow until the voltage reaches this value, but the amount of current will
increase steeply for voltages above this value. The maximum current density (amperes/square
meter) on the plate or cylinder directly under the wire is given by:
V2
3 = ue— (6.3)
L
where
j = maximum current density (A/m2)
/j = ion mobility rrr/Vs) (meter/volt second)
e = free space permittivity (8.845 x K)"12 F/m)(Farad/meter)
V = applied voltage (V)
L = shortest distance from wire to collecting surface (m)
For tuft corona, the current density is zero until the corona onset voltage is reached, when
it jumps almost to this value of j within a few hundred volts, directly under a tuft.
The region near the wire is strongly influenced by the presence of ions there, and the corona
onset voltage magnitude shows strong spatial variations. Outside the corona region, it is quite
uniform.
The electric field is strongest along the line from wire to plate and is approximated very
well, except near the wire, by:
6-14
-------
(6.4)
where
= maximum field strength (V/m)
When the electric field throughout the gap between the wire and the plate becomes strong
enough, a spark will occur, and the voltage cannot be increased without severe sparking
occurring. The field at which sparking occurs is not sharply defined, but a reasonable value is
given by:
1.65
(6.5)
where
Es = sparking field strength (V/m)
T = absolute temperature (K)
P = gas pressure (atm)
This field would be reached at a voltage of, for example, 35,000 V for a plate-wire spacing of
11.4 cm (4.5 in.) at a temperature of 149°C (300°F). The ESP will generally operate near this
voltage in the absence of back corona. Emfu will be equal to or less than Es.
Instead of sparking, back corona may occur if the electric field in the dust layer, resulting
from the current flow in the layer, reaches a critical value of about 1 x 106 V/m. Depending on
conditions, the back corona, may enhance sparking or may generate so much current that the
voltage cannot be raised any higher. The field in the layer is given by:
E, = IP
where
E, = electric field in dust layer (V/m)
p = resistivity of the collected material (ohm-m)
6.1.4.2 Particle Charging
Charging of particles takes place when ions bombard the surface of a particle. Once an ion is
close to the particle, it is tightly bound because of the image charge within the particle. The
"image charge" is a representation of the charge distortion that occurs when a real charge
6-15
-------
approaches a conducting surface. The distortion is equivalent to a charge of opposite magnitude
to the real charge, located as far below the surface as the real charge is above it. The notion of
the fictitious charge is similar to the notion of an image in a mirror, hence the name. As more
ions accumulate on a particle, the total charge tends to prevent further ionic bombardment.
There are two principal charging mechanisms: diffusion charging and field charging.
Diffusion charging results from the thermal kinetic energy of the ions overcoming the repulsion
of the ions already on the particle. Field charging occurs when ions follow electric field lines
until they terminate on a particle. In general, both mechanisms are operative for all sizes of
particles. Field charging, however, adds a larger percentage of charge on particles greater than
about 2/jm in diameter, and diffusion charging adds a greater percentage on particles smaller
than about 0.5/mi.
Diffusion charging, as derived by White [1], produces a logarithmically increasing level of
charge on particles, given by:
£^lln(l+r)
where
q(t) = particle charge (C) as function of time, t, in seconds
r = particle radius (m)
k = Boltzmann's constant (j/K)
T - absolute temperature (K)
e = electron charge (1.67 x 10"19O
T = dimensionless time given by:
T —
nrv Ne2G
kT
(6.8)
where
v = mean thermal speed of the ions (m/s)
N = ion number concentration near the particle (No./m3) 6 = real time (exposure
6 = real time (exposure time in the charging zone) (s)
Diffusion charging never reaches a limit, but it becomes very slow after about three
dimensionless time units. For fixed exposure times, the charge on a particle is proportional to
its radius.
Field charging also exhibits a characteristic time-dependence, given by:
6-16
-------
= gs6/ (6 + T') (6.9)
where
qs = saturation charge, charge at infinite time (C)
6 = real time (s)
t' = another dimensionless time unit
The saturation charge is given by:
qs = 12ner2 E (6.10)
where
e = free space permittivity (F/m)
E = external electric field applied to the particle (V/m)
The saturation charge is proportional to the square of the radius, which explains why field
charging is the dominant mechanism for larger particles. The field charging time constant is
given by:
r' = 4e/A/eM (6.11)
where
/j - ion mobility
Strictly speaking, both diffusion and field charging mechanisms operate at the same time on
all particles, and neither mechanism is sufficient to explain the charges measured on the
particles. It has been found empirically that a very good approximation to the measured charge
is given by the sum of the charges predicted by equations 6.7 and 6.9 independently of one
another:
(6.12)
where
particle charge due to both mechanisms
particle charge due to diffusion charging
6-17
-------
- particle charge due to field charging
6.1.4.3 Particle Collection
The electric field in the collecting zone produces a force on a particle proportional to the
magnitude of the field and to the charge:
Fe =
(6.13)
where
Fe - force due to electric field (N)
q = charge on particle (C)
E = electric field (V/m)
Because the field charging mechanism gives an ultimate charge proportional to the electric
field, the force on large particles is proportional to the square of the field, which shows the
advantage for maintaining as high a field as possible.
The motion of the particles under the influence of the electric field is opposed by the viscous
drag of the gas. By equating the electric force and the drag force component due to the electric
field (according to Stokes' law), we can obtain the particle velocity:
q(E, r) x E x C(r)
v(q, E, r) = 2 : (6.14)
6nrir v
where
v(q,E,r) - particle velocity (m/s)
q(E,r) = particle charge (C)
C(r) = Cunningham correction to Stokes' law (dimensionless)
r) = gas viscosity (kg/ms)
The particle velocity, is the rate at which the particle moves along the electric field lines, i.e.,
toward the walls.
For a given electric field, this velocity is usually at a minimum for particles of about 0.5
diameter. Smaller particles move faster because the charge does not decrease very much, but
the Cunningham factor increases rapidly as radius decreases. Larger particles have a charge
increasing as r2 and a viscous drag only increasing as r]\ the velocity then increases as r.
6-18
-------
Equation 6.14 gives the particle velocity with respect to still air. In the ESP, the flow is
usually very turbulent, with instantaneous gas velocities of the same magnitude as the particles
velocities, but in random directions. The motion of particles toward the collecting plates is
therefore a statistical process, with an average component imparted by the electric field and a
fluctuating component from the gas turbulence.
This statistical motion leads to an exponential collection equation, given by:
N(r) = Nn(r) x exp ( - v (r) / v.
(6.15)
where
N(r) - particle concentration of size r at the exit of the collecting zone (No./m3
N0(>'} - particle concentration of size r at the entrance of the
zone (No./nr)
v(r) = size-dependent particle velocity (m/s)
\0 = characteristic velocity of the ESP (m/s), given by:
vn = Q/A = 1/SCA
(6.16)
where
Q = volume flow rate of the gas (m3/s)
A = plate area for the ESP collecting zone (m2)
SCA = specific collection area (A/Q) (s/m)
When this collection equation is averaged over all the particle sizes and weighted according
to the concentration of each size, the Deutsch equation results, with the penetration (fraction of
particles escaping) given by:
p = exp (-w x SCA)
(6.17)
where
p - penetration (fraction)
we = effective migration velocity for the particle ensemble (m/s)
The efficiency is given by:
Eff(%) = 100 (1 - p)
(6.18)
and is the number most often used to describe the performance of an ESP.
6-19
-------
6.1.4.4 Sneakage and Rapping Reentrainment
Sneakage and rapping reentrainment are best considered on the basis of the sections within an
ESP. Sneakage occurs when a part of the gas flow bypasses the collection zone of a section.
Generally, the portion of gas that bypasses the zone is thoroughly mixed with the gas that passes
through the zone before all the gas enters the next section. This mixing cannot always be
assumed, and when sneakage paths exist around several sections, the performance of the whole
ESP is seriously elected. To describe the effects of sneakage and rapping reentrainment
mathematically we first consider sneakage by itself and then consider the effects of rapping as
an average over many rapping cycles.
On the assumption that the gas is well mixed between sections, the penetration for each
section can be expressed as:
PS = SN + t(1 ~ SN} X PC^'}] (6.19)
where
p, = section's fractional penetration
SN - fraction of gas bypassing the section (sneakage)
PC(Q ) = fraction of particles penetrating the collection zone, which is functionally
dependent on Q', the gas volume flow in the collection zone.reduced by the
sneakage (mVs)
The penetration of the entire ESP is the product of the section penetrations. The sneakage
sets a lower limit on the penetration of particles through the section.
To calculate the effects of rapping, we first calculate the amount of material captured on the
plates of the section. The fraction of material that was caught is given by:
m/m. = l-p=l-S- [ (1 - SN) x p^(Q') } (6.20)
where
m/m0 = mass fraction collected from the gas stream
This material accumulates until the plates are rapped, whereupon most of the material falls
into the hopper for disposal, but a fraction of it is reentrained and leaves the section.
Experimental measurements have been conducted on fly ash ESPs to evaluate the fraction
reentrained, which averages about 12 percent.
The average penetration for a section including sneakage and rapping reentrainments. is:
6-20
-------
[(!-£„) x Pc(Q'n + RR (1 - SN) [1 - pc (6.21)
where
RR - fraction reentrained
This can be written in a more compact form as:
ps = LF + [ (1 - LF) x pc (Q1) ] (6 22)
by substituting LF (loss factor) for SN + RR(\ - $ ). These formulas can allow for variable
amounts of sneakage and rapping reentrainment for each section, but there is no experimental
evidence to suggest that it is necessary.
Fly ash precipitators analyzed in this way have an average SN of 0.07 and an RR of 0.12.
These values are the best available at this time, but some wet ESPs, which presumably have no
rapping losses, have shown SN- values of 0.05 or less. These values offer a means for estimating
the performance of ESPs whose actual characteristics are not known, but about which general
statements can be made. For instance, wet ESPs would be expected to have RR = 0, as would
ESPs collecting wet or sticky particles. Paniculate materials with a much smaller mass mean
diameter, MMD, than fly ash would be expected to have a lower RR factor because they are held
more tightly to the plates and each other. Sneakage factors are harder to account for: unless
special efforts have been made in the design to control sneakage. the 0.07 value should be used.
6.2 ESP Design Procedure
6.2.1 Specific Collecting Area
Specific collecting area (SCA) is a parameter used to compare ESPs and roughly estimate their
collection efficiency. SCA is the total collector plate, area divided by gas volume flow rate and
has the units of s/m or s/ft. Since SCA is the ratio of A/Q, it is often expressed as m2/(m3/s) or
ftVkacfm, where kacfm is thousand acfm. SCA is also one of the most important factors in
determining the capital and several of the annual costs (for example, maintenance and dust
disposal costs) of the ESP because it determines the size of the unit. Because of the various
ways in which SCA can be expressed, Table 6.1 gives equivalent SCAs in the different units for
what would be considered a small, medium, and large SCA.
6-21
-------
Table 6.1: Small, Medium, and Large SCAs as Expressed by Various Units
Units Small Medium Large
ftVkacfm
s/m
s/ft
100
19.7
6
400
78.8
24
900
177
54
5.080 tf/kacfin = 1 (s/m).
The design procedure is based on the loss factor approach of Lawless and Sparks [2] and
considers a number of process parameters. It can be calculated by hand, but it is most
conveniently used with a spreadsheet program. For many uses, tables of effective migration
velocities can be used to obtain the SCA required for a given efficiency. In the following
subsection, tables have been calculated using the design procedure for a number of different
particle sources and for differing levels of efficiency. If a situation is encountered that is not
covered in these tables, then the full procedure that appears in the subsequent subsection should
be used.
6.2.1.1 SCA Procedure with Known Migration Velocity
If the migration velocity is known, then equation 6.17 can be rearranged to give the SCA:
SCA = -ln(p) /we (6.23)
6-22
-------
A graphical solution to equation 6.23 is given in Figure 6.4. The migration velocities have
been calculated for three main precipitator types: plate-wire, flat plate, and wet wall ESPs of
the plate-wire type. The following three tables, keyed to design efficiency as an easily
quantified variable, summarize the migration velocities under various conditions:
• In Table 6.2, the migration velocities are given for a plate-wire ESP with conditions of
no back corona and severe back corona; temperatures appropriate for each process have
been assumed.
• In Table 6.3, the migration velocities calculated for a wet wall ESP of the plate-wire
type assume no back corona and no rapping reentrainment.
• In Table 6.4, the flat plate ESP migration velocities are given only for no back corona
conditions because they appear to be less affected by high-resistivity dusts than the plate-
wire types.
It is generally expected from experience that the migration velocity will decrease with
increasing efficiency. In Tables 6.2 through 6.4, however, the migration velocities show some
fluctuations. This is because the number of sections must be increased as the efficiency
increases, and the changing sectionalization affects the overall migration velocity. This effect
is particularly noticeable, for example, in Table 6.4 for glass plants. When the migration
velocities in the tables are used to obtain SCAs for the different efficiencies in the tables , the
SCAs will increase as the efficiency increases.
•»§
Figure 6.4: Chart for Finding SCA
6-23
-------
6.2.1.2 Full SCA Procedure
The full procedure for determining the SCA for large plate-wire, flat plate, and (with
restrictions) tubular dry ESPs is given here. This procedure does not apply to the smaller, two-
stage precipitators because these are packaged modules generally sized and sold on the basis of
the waste gas volumetric flow rate. Nor does this procedure apply to determining the SCA for
wet ESPs. The full procedure consists of the 15 steps given below:
Step 1 - Determine the design efficiency, Eff (%). Efficiency is the most commonly used term
in the industry and is the reference value for guarantees however, if it has not been specified,
it can be computed as follows:
Eff(%) = 100 x (1 - outlet load/inlet load)
Step 2 - Compute design penetration, p:
p=l- (Eff/100)
Step 3 - Compute or obtain the operating temperature, T^, K. Temperature in Kelvin is required
in the calculations which follow.
6-24
-------
Table 6.2: Plate-wire ESP Migration Velocities
(cm/s)a
Design Efficiency, %
Particle Source
Bituminous coal fly ash6
Sub-bituminous coal fly ash in
tangential- tired' 6oi'lcr>
Other coalfc
Cement kilne
Glass plant1*
Iron/steel sinter plant dust with
mechanical precollector1'
Kraft-paper recovery boiler*
Incinerator fly ash*
Copper reverberatory furnace'
Copper converter'
Copper roaster*
Coke plant combustion stack1
(no BC)
(BC)
(no BC)
(BCJ
(no BC)
(BC)
(no BC)
(BC)
(no BC)
(no BC)
(BC)
(no BC)
(no BC)
(no BC)
(no BC)
(no BC)
(no BC)
95
12.6
3.1
17.0
4.9
9.7
2.9
1.5
0.6
1.6
U.5
6.8
2.2
2.6
15.3
6.2
5.5
6.2
1.2'
99
10.1
2.5
11.8
3.1
7.9
2.2
1.5
0.6
1.6
J.5
6.2
1.8
2.5
11.4
4.2
4.4
5.5
—
99.5
9.3
2.4
10.3
2.6
7.9
2.1
1.8
0.5
1.5
U.5
6.6
1.8
3,1
10.6
3.7
4.1
5.:
—
99.9
8.2
2.1
8.8
2.2
7.2
1.9
1.8
0.5
1.5
0.5
6.3
1.7
2.9
9.4
2.9
3.6
4.S
—
BC a Back eoron*.
*To teoeft em/i !• ft/*, multiply on/i kry 0.0338- CvapalaiioMl procedure UM« SI unit*, to con-
cert an/* l» m/*, multiply OB/( by 0.01. AMOIB>» tmmt p*rticl« mm u (ivea in full comput*tion*i
,uiiu<.»>iui».
'At 30O*F. Depending oa iaatritiuBl fomaev/boilv eoadiiiocw, efacaural natur* of lh« fly wru and
of DAinrolly ociiuiiuf evooltieainf ««cau («-»^ IBMBIUT* in the gw itrcun), migr»tion
•**Tf cotutdvmUy £ra«B th«>« -»«lu«». Liloiy **iua« «r» in ttM rmnfv from b*ek corona
«A« WO'T.
-------
Table 6.3: Wet Wall Plate-wire ESP Migration Velocities
(No back corona, cm/s)a
Design Efficiency, %
Particle Source* 95 99 99.5 99.9
Situminous coal' tfy asri 31.4 33.G 33.$ 24.3
Sub-bituminous coal fly ash in 40.0 42.7 44.1 31.4
tangential-fired boiler
Other coal 21.1 21.4 21.5 17.0
Cement kiln 6.4 5.6 5.0 5.7
Glass plant 4.5 4.5 4.3 3.S
Iron/steel sinter plant dust with 14.0 13.7 13.3 11.5
mechanical precollector
"To convert cm/a to ft/a, multiply eat/a by 0.0328. Computational procedure
uaea SI nrnta; to convert cm/a to m/a, multiply cm/a by 0.01. Aaaumea same
particle six u given in full computational procedure.
6-26
-------
Table 6.4: Flat Plate ESP Migration Velocities0
(No back corona, cm/s)fc
Design Efficiency, T<
Particle Source
95
99 99.5 99.9
Bituminous coal fly ashe
13.2 15.1 18.6 16.0
Sub-bituminous coal fly ash in tangential- 28.6 18.2 21.2
fired boiler*
Cement kiln1'
Glass plant*
IJ.J
2.4
2.3
1.8 1.9 2.6
2.6
Iron/steel sinter plant dust with mechanical 13.4 12.1 13.1 12.4
precollector*
Kraft-paper recovery boiler' 5.0 4.T 6.1 5.5
Incinerator fly ash'
25.2 16.9 21.
is.:
'Assumes same particle use u given in full computational procedure. Taeae vaiues
give the grounded collector plate SCA, from which the collector plate area :s derived
In Sat plate ESPs, the discharge or high-voltage plate area is typically 40 oercint
of the ground-plate area. The fiat-plate manufacturer usual]? counu ul liie pia:e
area (collector piatea plos discaaxge plates) in tneeting an SCA specification. -»7i:c.-.
means that the velocities tabulated »bove must be divided by 1.4 to be used on the
manufacturer'; basis.
convert cm/s to ft/a, multiply CTO/S by 0.0378. Computational procedure uses SI
v Ja- &fir*ee* far/'f v\y ar/'f, jrn'A^'f car/ 'fbrf
-------
Step 4 - Determine whether severe back corona is present. Severe back corona usually occurs
for dust resistivities above 2 x 1011 ohm-cm. Its presence will greatly increase the size of the
ESP required to achieve a certain efficiency.
Step 5 - Determine the MMD of the inlet particle distribution MMD, (urn). If this is not known,
assume a value from the following table:
6-28
-------
Source MMD) (urn)
Bituminous coal 16
Subbituminous coal,
tangential boiler 21
Sub-bituminous coal,
other boiler types 10 to 15
Cement kiln 2 to 5
Glass plant 1
Wood burning boiler 5
Sinter plant, 50
with mechanical precollector 6
2
Kraft Process Recovery 15 to 30
Incinerators 1
Copper reverberatory furnace 1
Copper converter 1
Coke plant combustion stack 1
Unknown
Step 6 - Assume value for sneakage. SN, and rapping reentrainment, RR. from the following
tables:
ESP Type Sv
Plate-wire 0.07
Wet wall 0.05
Flat plate 0.10
ESP / Ash Type RR
Coal fly ash, or not known 0.14
Wet Wall 0.0
Flat plate with gas velocity
> 1.5 m/s (not glass or
cement) 0.15
Glass or cement 0.10
Step 7 - Assume values for the most penetrating size. MMD , and rapping puff size. MMDr:
6-29
-------
= 2 urn
MMDr = 5 pin for ash with MMD, > 5 [im
MMDr = 3 urn for ash with MMD, < 5
where
MMDp = the MMD of the size distribution emerging from a very efficient collecting
zone
MMDp = the MMD of the size distribution of rapped/reentrained material.
Step 8 - Use or compute the following factors for pure air:
e0 = 8.845 x 10"12 free space permittivity (F/m)
77 = 1.72 x lO'5 (Tk/273)071 gas viscosity (kg/m.s)
Ebd = 630,000 (273/Tk)165 electric field at sparking (Vim)
LF = SN + RR(1-SN) loss factor (dimensionless)
For plate-wire ESPs:
EMg = EbJ/l.l5 average field with no back corona
Eavg = 0.7 x Ebdl\.15 average field with severe back corona
For Hat plate ESPs:
£a,s = Ebd x 5/6.3 average field, no back corona, positive polarity
£m,? = 0.7 x Ehd\ 5/6.3 average field, severe back corona, positive polarity
Step 9 - Assume the smallest number of sections for the ESP, n, such that LF" < p. Suggested
values of n are:
Eff(%) n
<96.5 2
<99 3
<99.8 4
<99.9 5
>99.9 6
These values are for an LF of 0.185, corresponding to a coal fly ash precipitator. The values
are approximate, but the best results are for the smallest allowable n.
Step 10 - Compute the average section penetration, pv: v
6-30 ™
-------
=P
1/72
Step 11 - Compute the section collection penetration, pc:
p - PS ~ LF
1 - LF
If the value of n is too small, then this value will be negative and n must be increased.
Step 12 - Compute the particle size change factors, D and MMD^, which are constants used for
computing the change of particle size from section to section:
D = Ps = SN + Pc(l - SN) + RR (1 ~ SN) (1 - pj
= MMDr - RR (I ~ SN) (1 - pJMMDr/D
Step 13 - Compute a table of particle sizes for sections 1 through n:
Section
MMD
! = MMD;
2 MMD, = {MMD! x
xpe}/D + MMD^
3 MMD3 = (MMD,x
xpe}/D +
n MMDn =
MMDn.,]
((l-pe)x MMD,~
[(1-p.) x MMDF-».
[(1 - pe) x MMD? -
6-31
-------
Step 14 - Calculate the SCA for sections 1 through n, using MMDn, r\, e, Eavg. and/?e:
SCA, = -(n/e) x (1 - SN) x In (pc)/(E2avg x MMD, x 1CT6)
SCAn = -(n/e) x (1 - SN] x In (pc)/(E2avg x MMDn x 1CT6)
where the factor 10~6 converts micrometers to meters. Note that the only quantity changing in
these expressions is MMD,; therefore, the following relation can be used:
^; = SCAn x MMD/MMDn+;
Step 15 - Calculate the total SCA and the English SCA, ESC A:
n
SCA (s/m) = £ SCAi
i = 1
ESCA (ft2/kacfm) = 5.080 x SCA (s/m)
This sizing procedure works best for pc values less than the value of LF, which means the
smallest value of n. Any ESP model is sensitive to the values of particle diameter and electric
field. This one shows the same sensitivity, but the expressions for electric field are based on
theoretical and experimental values. The SCA should not be strongly affected by the number
of sections chosen; if more sections are used, the SCA of each section is reduced.
6.2.1.3 Specific Collecting Area for Tubular Precipitators
The procedure given above is suitable for large plate-wire or flat plate ESPs. but must be used
with restrictions for tubular ESPs. Values of Sv = 0.015 and RR = 0 are assumed, and only one
section is used.
Table 6.5 gives migration velocities that can be used with equation 6.23 to calculate SCAs
for several tubular ESP applications.
6.2.2 Flow Velocity
A precipitator collecting a dry paniculate material runs a risk of nonrapping (continuous)
reentrainment if the gas velocity becomes too large. This effect is independent of SCA and has
been learned through experience. For fly, ash applications, the maximum acceptable velocity
is about 1.5 m/s (5 ft/s) for plate-wire ESPs and about 1 m/s (3 ft/s) for flat plate ESPs. For low
resistivity applications, design velocities of 3 ft/s or less are common to avoid nonrapping
reentrainment. The frontal area of the ESP (W x H), i. e., the area normal to the direction of gas
flow, must be chosen to keep gas velocity low and to accommodate electrical requirements (e.g..
6-32
-------
wire-to-plate spacing) while also ensuring that total plate area requirements are met. This area
can be configured in a variety of ways. The plates can be short in height, long in the direction
of flow, with several in parallel (making the width narrow). Or, the plates can be tall in height,
short in the direction of flow, with many in parallel making the width large). After selecting a
configuration, the gas velocity can be obtained by dividing the volume flow rate, Q, by the
frontal area of the ESP:
(6-24)
WH
where:
vgas = Sas velocity (m/s)
W = width of ESP entrance (m)
H = height of ESP entrance (m)
When meeting the above restrictions, this value of velocity also ensures that turbulence is
not strongly developed, thereby assisting in the capture of particles.
6.2.3 Pressure Drop Calculations
The pressure drop in an ESP is due to four main factors:
• Diffuser plate (usually present)—(perforated plate at the inlet)
• Transitions at the ESP inlet and outlet
• Collection plate baffles (stiffeners) or corrugations
• Drag of the flat collection plate
The total pressure drop is the sum of the individual pressure drops, but any one of these
sources may dominate all other contributions to the pressure drop. Usually, the pressure drop
is not a design-driving factor, but it needs to be maintained at an acceptably low value. Table
6.6 gives typical pressure drops for the four factors. The ESP pressure drop, usually less than
about 0.5 in. H2O, is much lower than for the associated collection system and ductwork. With
the conveying velocities used for dust collected in ESPs, generally 4,000 ft/min or greater,
system pressure drops are usually in the range of 2 to 10 in. H2O, depending upon the ductwork
length
6-33
-------
Table 6.5: Tubular ESP Migration Velocities
(cm/s)6
Design Efficiency, %
Particle Source
Cement kiln
Glass plant
Kraft- paper
recorery boiler
Incinerator
15 /xm MMD
Wet, at 200° F
MMD (/im)
I
2
5
10
20
(no BC)
(BC)
(no BC)
CHCJ
(no BC)
(no BC)
90
2.2-5.4
1.1-2.7
1.4
u.r
4.7
40.8
3.2
6.4
16.1
32.2
64.5
95
2.1-5.
1.0-2.
1.3
tf.r
4.4
39.
J.I
6.2
15.4
30.8
61.6
1
6
BC = Back corona
"These rates were calculated on the basis ofc
S/«r = 0.015, RB, = 0, one section only.
These ate in abetment with operating tabular ESPi; exten-
sion of reaoits to more than one section is not retro mm ended.
*To conTert cm/s to fl/s, moltipiy cm/a by 0.0328.
6-34
-------
Table 6.7: Components of ESP Pressure Drop
Typical Pressure Drop (in. HjO)
Component
Diffuse:
Inlet transition
Outlet transition
Baffles
Collection plates
Total
Low
0.010
0.07
0.007
0.0006
0.0003
0.09
High
0.09
0.14
MLS
0.123
0.008
0.38
and configuration as well as the type(s) of preconditioning devices(s) used upstream.
The four main factors contributing to pressure drop are described briefly below.
The diffuser plate is used to equalize the gas flow across the face of the ESP. It typically
consists of a flat plate covered with round holes of 5 to 7 cm diameter (2 to 2.5 in.) having an
open area of 50 to 65 percent of the total. Pressure drop is strongly dependent on the percent
open area, but is almost independent of hole size.
The pressure drop due to gradual enlargement at the inlet is caused by the combined effects
of flow separation and wall friction and is dependent on the shape of the enlargement. At the
ESP exit, the pressure drop caused by a short, well-streamlined gradual contraction is small.
Baffles are installed on collection plates to shield the collected dust from the gas flow and
to provide a stiffening effect to keep the plates aligned parallel to one another. The pressure
drop due to the baffles depends on the number of baffles, their protrusion into the gas stream
with respect to electrode-to-plate distance, and the gas velocity in the ESP.
The pressure drop of the flat collection plates is due to friction of the gas dragging along the
flat surfaces and is so small compared to other factors that it may usually be neglected in
engineering problems.
6-35
-------
6.2.4 Particle Characteristics
Several particle characteristics are important for particle collection. It is generally assumed that
the particles are spherical or spherical enough to be described by some equivalent spherical
diameter. Highly irregular or elongated particles may not behave in ways that can be easily
described.
The first important characteristic is the mass of particles in the gas stream, /. e., the particle
loading. This quantity usually is determined by placing a filter in the gas stream, collecting a
known volume of gas, and determining the weight gain of the filter. Because the ESP operates
over a wide range of loadings as a constant efficiency device, the inlet loading will determine
the outlet loading directly. If the loading becomes too high, the operation of the ESP will be
altered, usually for the worse.
The second characteristic is the size distribution of the particles, often expressed as the
cumulative mass less than a given particle size. The size distribution describes how many
particles of a given size there are, which is important because ESP efficiency varies with particle
size. In practical terms, an ESP will collect all particles larger than 1.0 um in diameter better
than ones smaller than 10 um. Only if most of the mass in the particles is concentrated above
10 (am would the actual size distribution above 10 um be needed.
In lieu of cumulative mass distributions, the size distribution is often described by log-
normal parameters. That is, the size distribution appears as a probabilistic normal curve if the
logarithm of particle size used is the abscissa. The two parameters needed to describe a log-
normal distribution are the mass median (or mean) diameter and the geometric standard
deviation.
The MMD is the diameter for which one-half of the paniculate mass consists of smaller
particles and one-half is larger (see the Procedure, Step 5, of Subsection 6.2.1.2). If the MMD
of a distribution is larger than about 3 um, the ESP will collect all particles larger than the MMD
at least as well as a 3 um particle, representing one-half the mass in the inlet size distribution.
The geometric standard deviation is the equivalent of the standard deviation of the normal
distribution: It describes how broad the size distribution is. The geometric standard deviation
is computed as the ratio of the diameter corresponding to 84 percent of the total cumulative mass
to the MMD; it is always a number greater than 1. A distribution with particles of all the same
size (monodisperse) has a geometric standard deviation of 1. Geometric standard deviations less
than 2 represent rather narrow distributions. For combustion sources, the geometric standard
deviations range from 3 to 5 and are commonly in the 3.5 to 4.5 range.
A geometric standard deviation of 4 to 5, coupled with an MMD of less than 5 um, means
that there is a substantial amount of submicrometer material. This situation may change the
electrical conditions in an ESP by the phenomenon known as "space charge quenching", which
results in high operating voltages but low currents. It is a sign of inadequate charging and
reduces the theoretical efficiency of the ESP. This condition must be evaluated carefully to be
sure of adequate design margins.
6-36
-------
6.2.5 Gas Characteristics
The gas characteristics most needed for ESP design are the gas volume flow and the gas
temperature. The volume flow, multiplied by the design SCA, gives the total plate area required
for the ESP. If the volume flow is known at one temperature, it may be estimated at other
temperatures by applying the ideal gas law. Temperature and volume uncertainties will
outweigh inaccuracies of applying the ideal gas law.
The temperature of the gas directly affects the gas viscosity, which increases with
temperature. Gas viscosity is affected to a lesser degree by the gas composition, particularly the
water vapor content. In lieu of viscosity values for a particular gas composition, the viscosity
for air may be used. Viscosity enters the calculation of SCA directly, as seen in Step 14 of the
design procedure (page 6-33).
The gas temperature and composition can have a strong effect on the resistivity of the
collected particulate material. Specifically, moisture and acid gas components may be
chemisorbed on the particles in a sufficient amount to lower the intrimic resistivity dramatically
(orders of magnitude). For other types of materials, there is almost no effect. Although it is not
possible to treat resistivity here, the designer should be aware of the potential sensitivity of the
size of the ESP to resistivity and the factors influencing it.
The choice of power supplies' size (current capacity and voltage) to be used with a particular
application may be influenced by the gas characteristics. Certain applications produce gas
whose density may vary significantly from typical combustion sources (density variation may
result from temperature, pressure, and composition). Gas density affects corona starting
voltages and voltages at which sparking will occur.
6.2.6 Cleaning
Cleaning the collected materials from the plates often is accomplished intermittently or
continuously by rapping the plates severely with automatic hammers or pistons, usually along
their top edges, except in the case of wet ESPs that use water. Rapping dislodges the material.
which then falls down the length of the plate until it lands in a dust hopper. The dust
characteristics, rapping intensity, and rapping frequency determine how much of the material
is reentrained and how much reaches the hopper permanently.
For wet ESPs, consideration must be given to handling waste waters. For simple systems
with innocuous dusts, water with particles collected by the ESP may be discharged from the ESP
system to a solids-removing clarifier (either dedicated to the ESP or part of the plant wastewater
treatment system) and then to final disposal. More complex systems may require skimming and
sludge removal, clarification in dedicated equipment, pH adjustment, and/or treatment to remove
dissolved-solids. Spray water from the ESP preconditioner may be treated separately from the
water used to flood the ESP collecting plates, so that the cleaner of the two treated waters may
be returned to the ESP. Recirculation of treated water to the ESP may approach 100 percent.
6-37
-------
The hopper should be designed so that all the material in it slides to the very bottom, where
it can be evacuated periodically, as the hopper becomes full. Dust is removed through a valve
into a dust-handling system, such as a pneumatic conveyor. Hoppers often are supplied with
auxiliary heat to prevent the formation of lumps or cakes and the subsequent blockage of the
dust handling system.
6.2.7 Construction Features
The use of the term "plate-wire geometry" may be somewhat misleading. It could refer to three
different types of discharge electrodes: weighted wires hung from a support structure at the top
of the ESP, wire frames in which wires are strung tautly in a rigid support frame, or rigid
electrodes constructed from a single piece of fabricated metal. In recent years, there has been
a trend toward using wire frames or rigid discharge electrodes in place of weighted wire
discharge electrodes (particularly in coal-fired boiler applications). This trend has been
stimulated by the user's desire for increased ESP reliability. The wire frames and rigid
electrodes are less prone to failure by breakage and are readily cleaned by impulse-type cleaning
equipment.
Other differences in construction result from the choice of gas passage (flow lane) width or
discharge electrode to collecting electrode spacing. Typically, discharge to collecting electrode
spacing varies from 11 to 19 cm (4.3 to 7.5 in.). Having a large spacing between discharge and
collecting electrodes allows higher electric fields to be used, which tends to improve dust
collection. To generate larger electric fields, however, power supplies must produce higher
operating voltages. Therefore, it is necessary to balance the cost savings achieved with larger
electrode spacing against the higher cost of power supplies that produce higher operating
voltages.
Most ESPs are constructed of mild steel. ESP shells are constructed typically of 3/16 or 1 /4
in. mild steel, plate. Collecting electrodes are generally fabricated from lighter gauge mild steel.
A thickness of 18 gauge is common, but it will vary with size and severity of application.
Wire discharge electrodes come in varied shapes from round to square or barbed. A
diameter of 2.5 mm (0.1 in.) is common for weighted wires, but other shapes used have much
larger effective diameters, e.g., 64 mm (0.25 in.) square electrodes.
Stainless steel may be used for corrosive applications, but it is uncommon except in wet
ESPs. Stainless steel discharge electrodes have been found to be prone to fatigue failure in dry
ESPs with impact-type electrode cleaning systems.[3]
Precipitators used to collect sulfuric acid mist in sulfuric acid plants are constructed of steel,
but the surfaces in contact with the acid mist are lead-lined. Precipitators used on paper mil]
black liquor recovery boilers are steam-jacketed. Of these two, recovery boilers have by far the
larger number of ESP applications.
6-38
-------
Table 6.7: Standard Options for Basic Equipment
Option
Cost adder (%)
1 - Inlet and outlet nozzles and diffuser plates
2 - Hopper auxiliaries /heaters, level detectors
3 • Weather enclosure and stair access
4 - Structural supports
5 - Insulation
Total options 1 to 5
8 to 10
8 to 10
8 to 1C
5
8 to 10
1.37 to 1.45xBase
6.3 Estimating Total Capital Investment
Total capital investment (TCI) for an ESP system includes costs for the ESP structure, the
internals, rappers, power supply, auxiliary equipment, and the usual direct and indirect costs
associated with installing or erecting new control equipment. These costs, in second-quarter
1987 dollars, are described in the following subsections.*
6.3.1 Equipment Cost
6.3.1.1 ESP Costs
Five types of ESPs are considered: plate-wire, flat plate, wet. tubular, and two-stage. Basic
For information on escalating these prices to more current dollars, refer to the EPA report Escalation Indexes for Air Pollution Control Costs
and updates thereto, all of which are installed on the OAQPS Technology Transfer Network (CTC Bulletin Board).
6-39
-------
1 4 5 < 7«
4561851
Rigid Electrode Design
with All Standard Options
(see Table 6-7) _
Rigid Electrode Basic
Flange-to-Flange Hardware
Field Erected
10,000
100,000
Plate Area - ft3
Figure 6.5: Dry-type ESP flange-to-flange purchase price vs Plate area.
6-40
-------
costs for the first two are taken from Figure 6.5, which gives the flange-to-flange, field-erected
price based on required plate area and a rigid electrode design. This plate area is calculated from
the sizing information given previously for the four types. Adjustments are made for standard
options listed in Table 6.7. Costs for wet/tubular ESPs are discussed under Recent Trends,
below, and costs for two-stage precipitators are given in a later subsection.
The costs are based on a number of actual quotes. Least squares lines have been fitted to the
quotes, one for sizes between 50,000 and 1,000,000 ft2 and a second for sizes between 10,000
and 50,000 ft2. An equation is given for each line. Extrapolation below 10,000 or above
1,000,000 ft2 should not be used. The reader should not be surprised if quotes are obtained that
differ from these curves by as much as ±25 percent. Significant savings can be had by soliciting
multiple quotes. All Units include the ESP casing, pyramidal hoppers, rigid electrodes and
internal collecting plates, transformer rectifier (TR) sets and microprocessor controls, rappers,
and stub supports (legs) for 4 feet clearance below the hopper discharges. The lower curve is
the basic unit without the standard options. The upper curve includes all of the standard options
(see Table 6.7) that are normally utilized in a modern system. These options add approximately
45 percent to the basic cost of the flange-to-flange hardware. Insulation costs are for 3 in. of
field-installed glass fiber encased in a metal skin and applied on the outside of all areas in
contact with the exhaust gas stream. Insulation for ductwork, fan casings, and stacks must be
calculated separately.
Impact of alternative electrode designs All three designs—rigid electrode, weighted wire, and
rigid frame—can be employed in most applications. Any cost differential between designs will
depend on the combination of vendor experience and site-specific factors that dictate equipment
size factors. The rigid frame design will cost up to 25 percent more if the mast or plate height
is restricted to the same used in other designs. Several vendors can now provide rigid frame
collectors with longer plates, and thus the cost differential can approach zero.
The weighted wire design uses narrower plate spacings and more internal discharge
electrodes. This design is being employed less; therefore, its cost is increasing and currently is
approximately the same as that for the rigid electrode collector. Below about 15.000 ft2 of plate
area. ESPs are of different design and are not normally field erected, and the costs will be
significantly different from values extrapolated from Figure 6.5.
Impact of materials of construction Metal thickness and stainless steel Corrosive or
other adverse operating conditions may suggest the specification of thicker metal sections in the
precipitator. Reasonable increases in metal sections result in minimal cost increases. For
example, collecting plates are typically constructed of 18 gauge mild steel. Most ESP
manufacturers can increase the section thickness by 25 percent without significant design
changes or increases in manufacturing costs of more than a few percent.
Changes in type of material can increase purchase cost of the ESP from about 30 to 50
percent for type 304 stainless steel collector plates and precipitator walls, and up to several
hundred percent for more expensive materials used for all elements of the ESP. Based on the
type 304 stainless steel cost, the approximate factors given below can be used for other
materials:
6-41
-------
Material Factor Reference(s)
Stainless steel, 316 1.3 [4,5,6]
Carpenter 20 CB-3 1.9 [6]
Monel-400 2.3 [4,6]
Nickel-200 3.2 [6]
Titanium 4.5 [6]
Appendix 6A provides more detail on the effects of material thickness and type.
Recent trends most of today's market (1987) is in the 50,000 to 200,000 ft2 plate area size
range. ESP selling prices have increased very little over the past 10 years because of more
effective designs, increased competition from European suppliers, and a shrinking utility market.
Design improvements have allowed wider plate spacings that reduce the number of internal
components and higher plates and masts that provide additional plate area at a low cost.
Microprocessor controls and energy management systems have lowered operating costs.
Few. if any, hot-side ESPs (those used upstream from an air preheater on a combustion
source) are being specified for purchase. Recognition that low sodium coals tend to build
resistive ash layers on the collection plates, thus reducing ESP efficiency, has almost eliminated
sales of these units. Of about 150 existing units, about 75 are candidates for conversion to cold-
side units over the next 10 years.
Specific industry application has little impact on either ESP design or cost, with three
exceptions: paper mills and sulfuric acid manufacturing plants, and coke by-product plants.
Paper mill ESPs use drag conveyor hoppers that add approximately 10 percent to the base
flange-to-flange equipment cost. For emissions control in sulfuric acid plants and coke by-
product ovens, wet ESPs are used. In sulfuric acid manufacture, wet ESPs are used to collect
acid mist. These precipitators usually are small, and they use lead for all interior surfaces;
hence, they normally cost $65 to $95/ft2 of collecting area installed (mid-1987 dollars) and up
to $120/ft2 in special situations. In addition, a wet circular ESP is used to control emissions
from a coke oven off-gas detarring operation. These precipitators are made using high-alloy
stainless steels and typically cost $90 to $120/ft2, installed. Because of the small number of
sales, small size of units sold, and dependency on site-specific factors, more definitive costs are
not available.
6.3.1.2 Retrofit Cost Factor
Retrofit installations increase the costs of an ESP because of the common need to remove
something to make way for the new ESP. Also, the ducting usually is much more expensive.
The ducting path is often constrained by existing structures, additional supports are required, and
the confined areas make erection more labor intensive and lengthy. Costs are site-specific:
however, for estimating purposes, a retrofit multiplier of 1.3 to 1.5 applied to the total capital
6-42
-------
investment can be used. The multiplier should be selected within this range based on the relative
difficulty of the installation.
A special case is conversion of hot-to-cold side ESPs for coal-fired boiler applications. The
magnitude of the conversion is very site-specific, but most projects will contain the following
elements:
• Relocating the air preheater and the ducting to it
• Resizing the ESP inlet and outlet duct to the new air volume and rerouting it
• Upgrading the ID (induced draft) fan size or motor to accommodate the higher static
pressure and horsepower requirements
• Adding or modifying foundations for fan and duct supports
• Assessing the required SCA and either increasing the collecting area or installing an SO3
gas-conditioning system
• Adding hopper heaters
• Upgrading the analog electrical controls to microprocessor-type controls
• Increasing the number of collecting plate rappers and perhaps the location of rapping
In some installations, it may be cost-effective to gut the existing collector totally, utilize only
the existing casing and hoppers, and upgrade to modern internals.
The cost of conversion is a multimillion dollar project typically running at least 25 to 35
percent of the total capital investment of a new unit.
6.3.1.3 Auxiliary Equipment
The auxiliary equipment depicted in Figure 6.2 is discussed elsewhere in the Manual. Because
dust-removal equipment (e.g., screw conveyers), hoods, precoolers, cyclones, fans, motors, and
stacks are common to many pollution control systems, they are (or will be) given extended
treatment in separate chapters.
6-43
-------
100
System Without
Precooler. Installed
Cell Washer, or Fan
t 10
Flow Rated.000 acfin)
Figure 6.6: Purchase Costs for Two-stage, Two-cell Precipitators (7)
6-44
-------
6.3.1.4 Costs for Two-Stage Precipitators
Purchase costs for two-stage precipitators, which should be considered separately from large-
scale, single-stage ESPs, are given in Figure 6.6.[7] To be consistent with industry practice,
costs are given as a function of flow rate through the system. The lower cost curve is for a two-
cell unit without precooler, an installed cell washer, or a fan. The upper curve is for an
engineered, package system with the following components: inlet diffuser plenum, prefilter,
cooling coils with coating, coil plenums with access, water flow controls, triple pass
configuration, system exhaust fan with accessories, outlet plenum, and in-place foam cleaning
system with semiautomatic controls and programmable controller. All equipment is fully
assembled mechanically and electrically, and it is mounted on a steel structural skid.
Table 6.8: Items That Increase ESP Costs
Item
Factor
Applied to
Rigid frame electrode with
restricted plate height
Type 304 stainless steel collector
plates and precipitator walls*
All stainless steel construction*
ESP with drag conveyor hoppers
(paper mill)
Retrofit installations
1.0 to 1.25
1.3 to 1.5
2 to 3
1.1
1.3 to 1.5
ESP Base cost
ESP total capital investment
(new facility installation)
Wet ESP
Sulfuric acid mist
Sulfuric acid mist
(special installation)
Coke oven off gas
See 6.3.1.1.
See 6.3.1.1.
See 6.3.1.1.
*See table on page 6-42 for other materials' cost factors.
6.3.2 Total Purchased Cost
The total purchased cost of an ESP system is the sum of the costs of the ESP, options, auxiliary
equipment, instruments and controls, taxes, and freight. The last three items generally are taken
as percentages of the estimated total cost of the first three items. Typical values, from Chapter
2 of the Manual, are 10 percent for instruments and controls, 3 percent for taxes, and 5 percent
for freight.
6-45
-------
Costs of standard and other options can vary from 0 to more than 150 percent of bare ESP
cost, depending on site and application requirements. Other factors that can increase ESP costs
are given in Table 6.8.
6.3.3 Total Capital Investment (TCI)
Using the Chapter 2 methodology, TCI is estimated from a series of factors applied to the
purchased equipment cost to obtain direct and indirect costs for installation. The TCI is the sum
of these three costs. The required factors are given in Table 6.9. Because ESPs may vary from
small units attached to existing buildings to large, separate structures, specific factors for site
preparation or for buildings are not given. However, costs for buildings may be obtained from
such references as Means Square Foot Costs 1987 [10]. Land, working capital, and off-site
facilities are excluded from the table because they are not normally required. For very large
installations, however, they may be needed and could be estimated on an as-needed basis.
Note that the factors given in Table 6.9 are for average installation conditions, e.g., no unusual
problems with site earthwork, access, shipping, or interfering structures. Considerable variation
may be seen with other-than-average installation circumstances. For two-stage precipitators
purchased as packaged systems, several of the costs in Table 6.9 would be greatly reduced or
eliminated. These include instruments and controls, foundations and supports, erection and
handling, painting, and model studies. An installation factor of 0.20 B to 0.25 B would be more
nearly appropriate for the two-stage ESPs.
6.4 Estimating Total Annual Costs
6.4.1 Direct Annual Costs
Direct annual costs include operating and supervisory labor, operating materials, replacement
rappers and electrodes, maintenance (labor and materials), utilities, dust disposal, and
wastewater treatment for wet ESPs. Most of these costs are discussed individually below. They
vary considerably with location and time and, for this reason, should be obtained to suit the
specific ESP system being costed. For example, current labor rates may be found in such
publications as the Monthly Labor Review, published by the U.S. Department of Labor, Bureau
of Labor Statistics.
6.4.1.1 Operating and Supervisory Labor
Proper operation of the ESP is necessary both to meet applicable particulate emission regulations
and to ensure minimum costs. An ESP is an expensive piece of equipment. Even well-designed
equipment will deteriorate rapidly if improperly maintained and will have to be replaced long
before it should be necessary. Not only can proper operation and maintenance save the operator
6-46
-------
money, such an operation and maintenance program can also contribute to good relations with
the governing pollution control agency by showing good faith in efforts to comply with air
regulations.
Although each plant has its own methods for conducting an operation and maintenance
program, experience has shown that plants that assign one individual the responsibility of
coordinating all the pieces of the program operate better than those where different departments
look after only a certain portion of the program. The separate departments have little knowledge
of how their portion impacts the overall program. In other words, a plant needs one individual
to coordinate the operation, maintenance, and troubleshooting components of its ESP program
if it expects to have a relatively trouble-free operation. The coordinator typically is an engineer
who reports to plant management and interfaces with the maintenance and plant process
supervisors, the laboratory, and the purchasing department. For companies with more than one
plant, he would be responsible for all ESPs. The portion of his total time that this individual
spends an the ESP then becomes an operating expense for the ESP. This can be expressed as:
AC = X(LCC) (6.25)
where
AC - annual coordination cost ($/yr)
X = fraction of total time spent on ESP
LCC - individual annual labor cost for ESP coordinator ($/yr)
In addition to coordination costs, typical operating labor requirements are 1/2 to 2 hours per shift
for a wide range of ESP sizes.[8] Small or well-performing units may require less time, and very
large or troublesome units may require more time. Supervisory labor is taken as 15 percent of
operating labor.
6.4.1.2 Operating Materials
Operating materials are generally not required for ESPs. An exception is the use of gas-
preconditioning agents for dust resistivity control.
6-47
-------
Table 6.9 Capital Cost Factors for ESPsa
Cost Item Factor
Direct Cost
Purchase equipment costs
ESP + auxiliary equipment As estimated, A
Instrumentation 0.10 A
Sales taxes 0.03 A
Freight 0.05 A
Purchased equipment cost, PEC B = 1.18 A
Direct installation costs
Foundations & supports 0.04 B
Handling & erection 0.50 B
Electrical 0.08 B
Piping 0.01 B
Insulation for ductwork* 0.02 B
Painting 0.02 B
Direct installation costs 0.67 B
Site preparation As required, SP
Buildings As required, Bldg.
Total Direct Cost, DC 1.67 B + SP - Bide.
Indirect costs (installation)
Engineering 0.20 B
Construction and field expenses 0.20 B
Contractor fees 0.10B
Start-up 0.01 B
Performance test 0.01 B
Model study 0.02 B
Contingencies 0.03 B
Total Indirect Costs, 1C 0.57 B
Total Capital Investment = DC + 1C 2.24 B + SP + Bldg.
•Reference (8 )
blf ductwork dimensions have been established, cost may be estimated based on $10 to $12/ft2
(fourth quarter 1986) of surface for field application. (Alternatively, refer to Chapter 10 of this manual.) Fan housings and stacks may also be
insulated. (9)
'For two-stage precipitators. total installation direct costs are more nearly 0.20 to 0.25B + SP + Bldg.
6-48
-------
6.4.1.3 Maintenance
The reader should obtain Publication No. EPA/625/1-85/017, Operating and Maintenance Manual for
ESPs,[\ 1] for suggested maintenance practices. Routine ESP maintenance labor costs can be estimated using
data provided by manufacturers. If such data are unavailable, the following procedure can be used. Based on
data for a 100,000 ft2 collector, maintenance labor is estimated to require 15 h/wk, 44 wk/yr. At a direct labor
cost of $12.50/h (mid-1987 costs), an estimated annual maintenance labor cost of $8,250 or $0.0825/ft2 of
collector area is established. This relationship can be assumed to be linear above a 50,000 ft2 collector-size and
constant at $4,125 below this size. To the maintenance labor cost must be added the cost of maintenance
materials. Based on an analysis of vendor information, annual maintenance materials are estimated as 1 percent
of the flange-to-flange precipitator purchase cost:
MC = 0.01 (FCC) + labor cost (6.26)
where
MC = annual maintenance cost ($/yr)
FCC = ESP flange-to-flange purchase cost ($)
labor cost = $4,125 if A < 50,000 ft2
0.825 A if > 50,000ft2
where A = ESP plate area (ft2)
6.4.1.4 Electricity
Power is required to operate system fans, transformer-rectifier (TR) sets, and cleaning equipment. Fan power
for primary gas movement can be calculated from Equation 2.7 of the Manual. After substituting into this
equation a combined fan-motor efficiency of 0.65 and a specific gravity of 1.0. we obtain:
FP = 0. 000181£(AP) (6') (6.27)
where
FP = fan power requirement (kWh/yr)
<2 = system flow rate (acfm)
AP = system pressure drop (in. H2O)
0' = annual operating time (h/yr)
Pump power for wet ESPs can be calculated from [8]:
PP - (0.746£?1Z5ff6/) / (3,960n) (6.28)
6-49
-------
where
PP = pump power requirement (kWh/yr)
Q = water flow rate (gal/min)
Z = fluid head (ft)
Sg ~ specific gravity of water being pumped compared to water at 70°F and 29.92 in. Hg
& = annual operating time (h/yr)
t) = pump-motor efficiency (fractional)
Energy for TR sets and motor-driven or electromagnetic rapper systems is the sum of the energy
consumption for operating both items. Manufacturers' averaged data indicate that the following relationship
can be used:
OP = 1.94 x lOAG (6.29)
where
OP = annual ESP operating power (kWh/yr)
A = ESP plate area (ft2)
6' = annual operating time (h/yr)
For installations requiring hopper heaters, hopper heater power can be
similarly estimated: ^\
HH = 2 (HN)Q' (6.30)
where
HH = annual hopper heater power consumption (kWh/yr)
HN = number of hoppers
0' = annual operating time (h/yr)
For two-stage precipitators, power consumption ranges from 25 to 100 W/kacfm, with 40 W/kacfm being
typical.
6.4.1.5 Fuel
If the ESP or associated ductwork is heated to prevent condensation, fuel costs should be calculated as required.
These costs can be significant, but they may be difficult to predict. For methods of calculating heat transfer
requirements, see Perry [12].
6-50
-------
6.4.1.6 Water
Cooling process gases for preconditioning can be done by dilution with air, evaporation with water, or heat
exchange with normal equipment. Spray cooling requires consumption of plant water (heat exchangers may
also require water), although costs are not usually significant. Section 4.4 of the Manual provides information
on estimating cooling water costs. Water consumption in wet ESPs is estimated at 5 gal/min kacfrn [13] for
large single-stage units and 16 gal/min-kacfm for two-stage precipitators [14].
6.4.1.7 Compressed Air
ESPs may use compressed air at pressures of about 60 to 100 psig for operating rappers. Equivalent power
cost is included in Equation 6.29 for operating power.
6.4.1.8 Dust Disposal
If collected dust cannot be recycled or sold, it must be landfilled or disposed of in some other manner.
Costs may typically run $20/ton or $30/ton for nonhazardous wastes exclusive of transportation (see Section
2.4 of the Manual}. Landfilling of hazardous wastes may cost 10 times as much. The disposal costs are highly
site-specific and depend on transportation distance to the landfill, handling rates, and disposal unloading
(tipping) fees. If these factors are known, they lead to the relationship:
DD = 4.29 x 10"6Ge'0[T + (TAT) D] (7.6)
where
DD = annual dust disposal cost ($/yr)
G = ESP inlet grain loading or dust concentration (gr /ft?)
6' - annual operating time (h/yr)
<2 = gas flow rate through ESP (acfm)
T = tipping fee ($/ton)
TM = mileage rate ($/ton-mile)
D = dust hauling distance (miles)
6.4.1.9 Wastewater Treatment
As indicated above, the water usage for wet ESPs is about 5 gal/min kacfm [13]. Treatment cost of the
resulting wastewater may vary from about $1.30 to $2.15/1,000 gal [15] depending on the complexity of the
treatment system. More precise costs can be obtained from Gumerman et al. [16].
6-51
-------
6.4.1.10 Conditioning Costs
Adaptation of information on utility boilers [17] suggests that S03 conditioning for a large ESP (2.6 x 106
acfm) costs from about $1.60/106 fl3 of gas processed for a sulfur burner providing 5 ppm SO3 to about
S2.30/106 ft3 (in first-quarter 1987 dollars) for a liquid SO2 system providing 20 ppm of S03.
6.4.2 Indirect Annual Costs
Capital recovery, property taxes, insurance, administrative costs ("G&A"), and overhead are examples of
indirect annual costs. The capital recovery cost is based on the equipment lifetime and the annual interest rate
employed. (See Chapter 2 for a thorough discussion of the capital recovery cost and the variables that
determine it.) For ESPs, the system lifetime varies from 5 to 40 years, with 20 years being typical. Therefore,
as Chapter 2 of the Manual suggests, when figuring the system capital recovery cost, one should base it on the
total capital investment. In other words:
CRCs = TCI x CRFs (6J2)
where
CRCS = capital recovery cost for ESP system ($/yr)
TCI = total capital investment ($)
CRF5 = capital recovery factor for ESP system (defined in Chapter 2)
For example, for a 20-year system life and a 7 percent annual interest rate, the CRFs would be 0.09439.
The suggested factor to use for property taxes, insurance, and administrative charges is 4 percent of the TCI.
Overhead is calculated as 60 percent of the sum of operating, supervisory, coordination, and maintenance labor.
as well as maintenance materials.
6.4.3 Recovery Credits
For processes that can reuse the dust collected in the ESP or that can sell the dust in a local market, such as fly
ash sold as an extender for paving mixes, a credit should be taken. As used below, this credit (RC) appears as
a negative cost.
6.4.4 Total Annual Cost
Total annual cost for owning and operating an ESP system is the sum of the components listed in Subsections
6.4.1 through 6.4.3, i.e.:
6-52
-------
TAC = DC + 1C - RC (6.33)
where
TAC = total annual cost ($)
DC = direct annual cost ($)
1C = indirect annual cost ($)
RC = recovery credits (annual) ($)
6.4.5 Example Problem
Assume an ESP is required for controlling fly ash emissions from a coal-fired boiler burning bituminous coal.
The flue gas stream is 50 kacfm at 325 °F and has an inlet ash loading of 4 gr/ft3. Analysis of the ash shows
of 7 urn and a resistivity of less than 2 x 10n ohm-cm. Assume that the ESP operates for 8,640 h/yr (360 d)
and that an efficiency of 99.9 percent is required.
6.4.5.1 Design SCA
The SCA can be calculated from Equation 6.23. Assuming that a flat plate
ESP design is chosen, the fly ash migration velocity is 16.0 cm/ s (see Table 6.4). Then:
SCA = -ln(l - 0.999)716.0 = 0.432 s/cm = 43.2 s/m
Converting to English units (see page 6-33, Step 15, in the procedure):
ESCA = 5.080 x 43.2 = 219 ftVkacfm
Total collector plate area is then:
219 ftVkacfm x 50 kacfm = 10.950 ft2
To obtain a more rigorous answer, we can follow the steps of the procedure given in Subsection 6.2.1:
Step 1 - Design efficiency is required as 99.9.
Step 2 - Design penetration:
1 - (99.9/100) = 0.001
6-53
-------
Step 3 - Operating temperature in Kelvin:
(325°F - 32°F) x 5/9 + 273°C = 436°^
Step 4 - Because dust resistivity is less than 2 x 1011 ohm-cm (see page 6-30, Step 4), no severe back corona
is expected and back corona = 0.
Step 5 - The MMD of the fly ash is given as 7 pm.
Step 6 - Values for sneakage and rapping reentrainment (from the table
presented in Step 6, page 6-24) are:
SN = 0.10
RR = 0.124 (assuming gas velocity <1-5 m/s)
Step 7 - The most penetrating particle size, from Step 7 of the procedure on page 6-31, is:
MMDp =
The rapping puff size is:
MMDr = 5 vim
Step 8 - From the procedure (Subsection 6.2.1):
ec = 8.845 x 10'12
ri = 1.72 x 10'5(436/273)071 = 2.40 x 10'5
EM = 6.3 x 105(273/436)'65 = 2.91 x 105 V/m
Emg = Ebdx 5/6.3 = 2.31 x 10s
LF = SN + RR(1-8^=0.1+0.124(1-0.1) = 0.212
Step 9 - Choose the number of sections for LFn < p, p = 0.001. Try four sections:
LF" = 0.2124 = 0.002
6-54
-------
This value is larger than p. Try five sections:
LF" = 0.2125 = 0.000428
This value is smaller than p and is acceptable.
Step 10 - Average section penetration is:
ps = pi/" = 0.0011/5 = 0.251
Step 11 - Section collection penetration is:
pc = (ps - LF) I (1 -LF) = (0.251 - 0.212) / (1 - 0.212) = 0.0495
Step 12 - Particle size change factors are:
D = Ps = SN + Pc(l - Sv) + RR(1 - SN)(1 - Pc)
= 0.10 + 0.0495(1 - 0.) + 0.214(1 - 0.1)(1 - 0.0495)
= 0.251
MMD.J, = RR(1 - SN)(1 - PC)MMD/D
= 0.124(1-1Q.1)(1 - 0.0495)(5)/0.251
= 2.11
6-55
-------
Step 13 - Particle sizes for each section are:
Section
MMD (/im)
1 MMD,
2 MMD,
3 MMD3
4 MMD,
MMDi = 7
x Sy + [(1 - Pe) x MMDF + Prx
x?e}/D + MMDn
{7 x 0.1 + [(1 . 0.0495) x 2 r 0.0495 x 7!
x 0.0495}/0.251 + 2.11
5.34
{5,34x0.1+[(l-0.0495)x2-0.0495x5.34]x
0.0495 }/0.251 + 2.11
4.67
{4.67x0.1 + [(l-0.0495) x 2-0.0495 x 4.67! /
0.0495 J/0.251 + 2.11
4.39
0.0495 }/0.251 - 2.11
4.28
Step 14 - SCAs for each section are:
Section
SCA (a/m)
3
4
5
SCA, =
=
=
SCA3 =
=
SCA-, =
=
SCA, =
-(2.40 x 10-'/8.845 x L
ln(0.0495)/I"(2.31 x 10')rCT x
19.65
SCA^x MMDA/MMD,
19.65 (7/5.34)
25.76
25.76 (5.34/4.67)
29.46
29.46 (4.67/4.39)
31.34
31.34 (4.39/4.28)
32.15
- 0.1;
6-56
•
-------
Step 15 - Calculate the total SCA.
Total SCA = 19.65 + 25.76 + 29.46 + 31.34 + 32.15 = 138.36 s/m
English SCA = 5.080 x 138.36 = 702.87 ftVkacfm
Note that the more rigorous procedure calls for an SCA that is considerably higher than the value found by
using Equation 6.23. This discrepancy is caused by the considerably smaller particle size used in the example
problem than is assumed for Table 6.4. In this case, the shorter method would lead to an unacceptably low cost
estimate.
Total collector plate area is:
702.87 ft2/kacfm x 50 kacfm = 35,144 ff
6.4.5.2 ESP Cost
From Figure 6.5, the basic flange-to-flange cost of the rigid electrode ESP is $438,060 (mid-1987 dollars).
Assuming all standard options are purchased, the ESP cost rises to $635,189 (mid-1987 dollars).
6.4.5.3 Costs of Auxiliaries
Assume the following auxiliary costs have been estimated from data in other parts of the Manual:
Ductwork $16.000
Fan 16,000
Motor 7,500
Starter 4,000
Dampers 7,200
Pneumatic conveyor 4,000
Stack 8.000
Total $62,700
6.4.5.4 Total Capital Investment Direct costs for the ESP system, based on the factors in Table 6.9, are given
in Table 6.10. (Again, we assume site preparation and building costs to be negligible.) TCI is $1,840,000
(rounded, mid-1987 dollars).
6-57
-------
6.4.5.5 Annual Costs-Pressure Drop
Table 6.11 gives the direct and indirect annual costs, as calculated from the factors given in Section 6.4.
Pressure drop (for energy costs) can be taken from Table 6.6 in Subsection 6.2.2. Using the higher values from
the table, pressure drop for the inlet diffuser plate, inlet and outlet transitions, baffles, and plates is:
AP = 0.09 + 0.14 + 0.015 + 0.123 + 0.008 = 0.38 in. H2O
Assume the ductwork contributes an additional 4.1 in. H£)* The total pressure drop is, therefore, 4.48 in. H2O.
As is typical, the ductwork pressure drop overwhelms the ESP pressure drop.
*For ductwork pressure drop data, refer to Chapter 10 ("Hoods, Ductwork, and
Stacks") of the Manual.
6-58
-------
Table 6.10: Capital Costs for ESP System
Example Problem
Cost Item Cost
Direct Costs
Purchased equipment costs
Adsorber vessels and carbon $635,189
Auxiliary equipment _ 62,700
Sum = A $697,889
Instrumentation, 0.1A 69,789
Sales taxes, 0.03A 20,937
Freight, 0.05A 34,894
Purchased equipment cost, B $823,509
Direct installation costs
Foundation and supports, 0.04B 32,940
Handling & erection, 0.50B 411,755
Electrical, 0.08B 65,881
Piping, 0.01B 8,235
Insulation for ductwork, 0.02B 16,470
Painting, 0.02B
_
Direct installation cost $551,751
Site preparation —
Facilities and buildings —
Total Direct Cost $1,375,260
Indirect Costs (installation)
Engineering, 0.20B 164,702
Construction and field expenses, 0.20B 164,702
Contractor fees, 0.10B 82,351
Start-up, 0.01B 8,235
Performance test, 0.01B 8,235
Model study, 0.02B 16,470
Contingencies, 0.03B 24,705
Total Indirect Cost $469,400
Total Capital Investment (rounded) $1,840,000
6-59
-------
Table 6.11: Annual Costs for ESP System
Example Problem
Cost Item
Calculations
Cost
Direct Annual Costa, DC
Operating labor
Operator
Supervisor
Coordinator
Operating materials
Maintenance
Labor
Material
Utilities
Electricity-fan
Electricity-operating
Waste disposal
Total DC
15% of operator = .15 x 12,960
1/3 of operator = 1/3 x 12,960
$ 4,125 for collector area < 50,000 ft3
1% of Purchased equipment cost = 0.01 x 823,509
0.000181x50,000 acfm x 4.48 in. H2O x 8'6yr° h y
$0.06
kWh
1.94 x 10-3 x 35,144 ft1 x 8,640 h x $0.06/VWh
at $20/ton tipping fee at 2 miles and $0.50/tun-
mile for essentially 100% collection efficiency: 1.29 y
10-8 x 1^ x 6'6yr° h x 50,000 acfm x (20 + 0.50 y
2)lon
312,960
1,944
4,320
4,125
8,235
35,344
155,HTG
"$243,622
Indirect Annual Costs, 1C
Overhead
Administrative charges
Property tax
Insurance
Capital recovery"
Total 1C
60% of sum of operating, supv., coord., fc maint.
labor k maint. materials = 0.6(12,960 -r 1,944 -
4,320 + 4,125 + 8,235)
2% of Total Capital Investment = 0.02($1,844,660)
1% of Total Capital Investment = 0.01(81,844,660)
1% of Total Capital Investment = 0 01($1.844,660)
0.1175 (1,844,360)
Total Annual Cost (rounded)
18,950
36.593
18.447
18 -447
216,7i*
$309,485
$553,000
" The capital recovery cost factor, CRF, is a function of the fabric filter or equipment life
and the opportunity cost of the capital (i r. , interest rate) For example, for a 20 yrar
equipment life and a 10% interest rate, CRF = 0.1175.
6-60
-------
6.4.5.8 Total Annual Cost The total annual cost, calculated in Table 6.11, is $511,000 (rounded). Had the
particle sizes being captured been larger, the ESP cost would have been considerably less. Also, for a much
larger gas flow rate, the $/acfm treated cost would have been more favorable. Reviewing components of the
TAG, dust disposal is the largest single item. Care should be taken in determining this cost and the unit disposal
cost ($/ton). Finding a market for the dust, for example, as an extender in asphalt or a dressing for fields, even
at giveaway prices, would reduce TAG dramatically.
6.5 Acknowledgments
We gratefully acknowledge C. G. Noll, United McGill Corp. (Columbus, OH), for extensive review and the
following companies for contributing data to this chapter:
• Research-Cottrell
• Joy Industrial Equipment Co., Western Precipitation Division (Los Angeles, CA)
• Environmental Elements Corp. (Baltimore, MD)
6-61
-------
Appendix 6A
Effects of Material Thickness
and Type on ESP Costs
The impact of material thickness and composition of collecting plates and the ESP casing can be estimated
using the following equations and Figure 6.5:
Plates:
J =
—- x FS) -0.90
M + SP
(6.34)
SP
Casing:
where
I =
W, =
FS =
M =
SP =
10
x FS) - 0. 58
M + SP
(6.35)
SP
incremental increase of flange-to-flange selling price ($/ft2)
weight of steel (lb/ft2)
fabricated steel selling price ($/lb) (normally assume approximately 2 times material cost)
manufacturer's markup factor of fabricated cost (direct labor, wages, and material cost before
genera] and administrative expense and profit) to selling price (normally 2 to 3)
flange-to-flange selling price from Figure 6.5 ($/ft2)
Most vendors can produce ESPs with collecting plate material thicknesses from 16 to 20 gauge and casing
material thicknesses from 1/8 through 1/4 in. without affecting the 2 times material cost = fabricated cost
relationship. Thus, the impact of increasing the collecting plates from 18 to 16 gauge and the casing from 3/16
to 1/4 in. plate on a 72,000 ft2 collector having a selling price of $10/ft2 and assuming a markup factor of 2 is
as follows:
6-62
-------
Plates:
Casing:
I =
2
.5
2
X
0.
90
- 0.
90
2 -
* 10
10
= 1.045 =4.5 percent increase
I =
10.21
10
x 0.76 - 0.58
2+10
10
= 1.039 =3.9 percent increase
Equations 6.34 and 6.35 were developed using the following assumptions:
Material selling price increase + Standard ESP selling price
Standard ESP selling price
6-63
-------
Figure 6.7: ESP Dimensions
Because figure 6.5 identifies the standard ESP selling price /ft2 of collecting area, the material selling price
increase = (New material cost - Standard material cost)M. Then it follows that:
It) steel
Material selling price = x Fabricated cost in S/lb x M
ft2 collecting area
The ESP dimensions given in Figure 6.7 include:
• Casing area = 30 ft 30 ft x 8 = 7,200 ft2 (assume 4 walls, 1 top, 2 hopper sides, 2 triangular hopper ends = 8
equivalent sides)
6-64
-------
Collecting plate area =
30 ft x 30 ft x 2 sides/plate x 30 ft plates
s
54,000 2 ft2 for s = 0>75ft
where i = plate spacing (ft)
Thus, there are:
• 7.50/s ft2 of collecting area per 1 ft2 of casing and
• 2 ft2 of collecting area per 1 ft2 of collecting plate
Material cost per ft2 collecting area is:
Plates = — ~^^, ^ x $/lb
Ib steel/ft2
_ . Ib steel/ ft2 . ..,
Casing = x $ / Ib
7.50/s
For the standard ESPs described by Figure 6.5,18 gauge collecting plates and 3/16 in. plate casing were specified. Assum
Material cost for 18 gauge mild steel = $0.45/lb
Material cost for 3/16 in. plate mild steel = $0.38/lb
Material cost to fabricated cost factor = 2
These costs yield fabricated material costs of:
Plates:
2 lb/£t x $0.45/lb x 2 = $0.90/ft2 of collecting area
Casing:
7.50/s
x $0.38 x 2 = $0.78 s/ft of collecting area
6-65
-------
At a typical 9 in. plate spacing the casing cost would be $0.58/ft2 of collecting area
Selling
price =
impact
Cost of _ Cost of j Original overall
new material old material/ + selling price
Original overall selling price
which gives us equations 6.34 and 6.35. Note that the value 0.58 will change significantly if a plate spacing other than 9 in.
is chosen.
Thus, for a less than 5 percent increase in flange-to-flange cost, all the precipitator exposed wall sections can be increased
by more than 25 percent to provide increased life under corrosive conditions. Section thickness increases that are greater than
those just discussed would probably result in significant cost increases because of both increased material costs and necessary
engineering design changes.
The impact of changing from mild steel to 304 stainless steel assuming material costs of $1.63/lb for 18 gauge collecting
plates, $1.38/lb for the 3/16 in. casing, and a markup factor of 3 is as follows:
Plates:
I =
2
~2
x 1.
63
/
- 0
9
3
* 10
10
=21.9 percent increase
J =
7.66
10
x 1
3.)
o
58
3
+ 10
10
= 14.3 percent increase
To these material costs must be added extra fabrication labor and procurement costs that will increase the ESP flange-to-
flange cost by a factor of 2 to 3. Note that a totally stainless steel collector would be much more expensive because the
discharge electrodes, rappers, hangers, etc., would be also converted to stainless. The preceding equations can be used for
other grades of stainless steel or other materials of construction by inserting material costs obtained from local vendors on a
$/lb basis.
6-66
-------
References
[1] White, H. J., Industrial Electrostatic Precipitation, Addison-Wesley, Reading, MA, 1963.
[2] Lawless, P. A., and L. E. Sparks, "A Review of Mathematical Models for ESPs and Comparison of Their Successes,"
Proceedings of Second International Conference on Electrostatic Precipitation, S. Masuda, ed., Kyoto, 1984, pp. 513-
522.
[3] Bump. R. L. (Research Cottrell, Inc.), "Evolution and Design of Electrostatic Precipitator Discharge Electrodes," paper
presented at the APCA Annual Meeting, New Orleans, LA, June 1982.
[4] Correspondence: Richard Selznick (Baron Blakeslee, Inc., Westfield. NJ) to William M. Vatavuk. April 23, 1986.
[5] Correspondence: James Jessup (M&W Industries, Inc., Rural Hall, NC) to William M. Vatavuk, May 16, 1986.
[6] Matley, Jay (ed.), Modern Cost Engineering, McGraw-Hill, New York, 1984, p. 142.
[7] Personal communication: Robert Snipe, Jr. (American Air Filter Co., Louisville. KY). and S. A. Sauerland (United Air
Specialists, Inc.. Cincinnati, OH), to Roger Ellefson (JACA Corp., Fort Washington. PA). June 1987.
[8] Vatavuk, W. M., and R. B. Neveril, "Estimating Costs of Air Pollution Control Systems, Part 11: Factors for Estimating
Capital and Operating Costs." Chemical Engineering. November 3, 1980, pp. 157-162.
[9] Telecon: Gary Gremer (ETS. Inc., Roanoke. VA) to James H. Turner. October 1986.
[10] R.S. Means Company, Inc., Means Square Fool Costs 1987, Kingston. MA.
[11] PEDCo Environmental, Inc.. Operating and Maintenance Manual for ESPs. Publication No. EPA/625/1-85/017.
Office of Research and Development. Air and Energy Engineering Research Lab, Research Triangle Park, NC.
September 1985.
[12] Perry,, R. H., el al, Perry's Chemical Engineers Handbook (Sixth Edition), McGraw-Hill, New York, 1984.
[13] Bakke, E., "Wet Electrostatic Precipitators for Control of Submicron Particles," Proceedings of the Symposium on
Electrostatic Precipitators for the Contra! of Fine Particles, Pensacola, FL, September 30 to October 2, 1974,
Publication No. EPA 650/2-75-016, 1975.
[14] Beltran Associates, Inc., "Poly-Stage Precipitator for Stack and Duct Emissions," November 1978.
6-67
-------
[15] Vatavuk, W.M., and R.B. Neveril, "Estimating Costs of Air-Pollution Control Systems, Part XVII: Particle
Emissions Control," Chemical Engineering (adapted), April 2,1984, pp. 97-99.
[16] Gumerman, R.C., B.E. Burris, and S.P. Hansen, Estimation of Small System Water Treatment Costs, Publication
No. EPAy600/2-84/184a, NTIS No. PB85-161644, 1984.
[17] Gooch, J.P., A Manual on the Use of Flue Gas Conditioning for ESP Performance Enhancement, Electric Power
Research Institute Report No. CS-4145,1985.
6-68
-------
Chapter 7
FLARES
Diana K. Stone
Susan K. Lynch
Richard F. Pandullo
Radian Corporation
Research Triangle Park, NC 27709
Leslie B. Evans, Organic Chemicals Group
William M. Vatavuk, Innovative Strategies and Economics Group
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
December 1995
Contents
7.1 Introduction 7-4
7.1.1 Flare Types 7-4
7.1.1.1 Steam-Assisted Flares 7-5
7.1.1.2 Air-Assisted Flares 7-5
7.1.1.3 Non-Assisted Flares 7-5
7.1.1.4 Pressure-Assisted Flares 7-5
7-1
-------
7.1.1.5 Enclosed Ground Flares 7-6
7.1.2 Applicability 7-6
7.1.3 Performance 7-7
7.1.3.1 Factors Affecting Efficiency 7-7
7.1.3.2 Flare Specifications 7-7
7.2 Process Description 7-8
7.2.1 Gas Transport Piping 7-10
7.2.2 Knock-out Drum 7-10
7.2.3 Liquid Seal 7-10
7.2.4 Flare Stack 7-15
7.2.5 Gas Seal 7-15
7.2.6 Burner Tip 7-15
7.2.7 Pilot Burners 7-16
7.2.8 Steam Jets 7-16
7.2.9 Controls 7-17
7.3 Design Procedures 7-17
7.3.1 Auxiliary Fuel Requirement 7-17
7.3.2 Flare Tip Diameter 7-18
7.3.3 Flare Height 7-19
7.3.4 Purge Gas Requirement 7-21
7.3.5 Pilot Gas Requirement 7-22
7.3.6 Steam Requirement 7-23
7.3.7 Knock-out Drum 7-23
7-2
-------
7.3.8 Gas Mover System 7-25
7.4 Estimating Total Capital Investment 7-26
7.4.1 Equipment Costs 7-27
7.4.2 Installation Costs 7-28
7.5 Estimating Total Annual Costs 7-29
7.5.1 Direct Annual Costs 7-29
7.5.2 Indirect Annual Costs 7-30
7.6 Example Problem 7-32
7.6.1 Required Information for Design 7-32
7.6.2 Capital Equipment 7-35
7.6.2.1 Equipment Design 7-35
7.6.2.2 Equipment Costs 7-37
7.6.3 Operating Requirements 7-40
7.6.4 Total Annual Costs 7-40
7.7 Acknowledgments 7-42
References 7-43
7-3
-------
7.1 Introduction
Flaring is a volatile organic compound (VOC) combustion control process in which the VOCs
are piped to a remote, usually elevated, location and burned in an open flame in the open air
using a specially designed burner tip, auxiliary fuel, and steam or air to promote mixing for
nearly complete (> 98%) VOC destruction. Completeness of combustion in a flare is governed
by flame temperature, residence time in the combustion zone, turbulent mixing of the
components to complete the oxidation reaction, and available oxygen for free radical formation.
Combustion is complete if all VOCs are converted to carbon dioxide and water. Incomplete
combustion results in some of the VOC being unaltered or converted to other organic
compounds such as aldehydes or acids.
The flaring process can produce some undesirable by-products including noise, smoke, heat
radiation, light, SOX, NQ, CO, and an additional source of ignition where not desired.
However, by proper design these can be minimized.
7.1.1 Flare Types
Flares are generally categorized in two ways: (1) by the height of the flare tip (i.e., ground or
elevated), and (2) by the method of enhancing mixing at the flare tip (i.e., steam-assisted, air-
assisted, pressure-assisted, or non-assisted). Elevating the flare can prevent potentially
dangerous conditions at ground level where the open flame (i.e., an ignition source) is located
near a process unit. Further, the products of combustion can be dispersed above working areas
to reduce the effects of noise, heat, smoke, and objectionable odors.
In most flares, combustion occurs by means of a diffusion flame. A diffusion flame is one
in which air diffuses across the boundary of the fuel/combustion product stream toward the
center of the fuel flow, forming the envelope of a combustible gas mixture around a core of fuel
gas. This mixture, on ignition, establishes a stable flame zone around the gas core above the
burner tip. This inner gas core is heated by diffusion of hot combustion products from the flame
zone.
Cracking can occur with the formation of small hot particles of carbon that give the flame
its characteristic luminosity. If there is an oxygen deficiency and if the carbon particles are
cooled to below their ignition temperature, smoking occurs. In large diffusion flames,
combustion product vortices can form around burning portions of the gas and shut off the supply
of oxygen. This localized instability causes flame flickering, which can be accompanied by soot
formation.
As in all combustion processes, an adequate air supply and good mixing are required to
complete combustion and minimize smoke. The various flare designs differ primarily in their
accomplishment of mixing.
7-4
-------
7.1.1.1 Steam-Assisted Flares
Steam-assisted flares are single burner tips, elevated above ground level for safety reasons, that
burn the vented gas in essentially a diffusion flame. They reportedly account for the majority
of the flares installed and are the predominant flare type found in refineries and chemical
plants. [1,2]
To ensure an adequate air supply and good mixing, this type of flare system injects steam
into the combustion zone to promote turbulence for mixing and to induce air into the flame.
Steam-assisted flares are the focus of the chapter and will be discussed in greater detail in
Sections 7.2 through 7.4.
7.1.1.2 Air-Assisted Flares
Some flares use forced air to provide the combustion air and the mixing required for smokeless
operation. These flares are built with a spider-shaped burner (with many small gas orifices)
located inside but near the top of a steel cylinder two feet or more in diameter. Combustion air
is provided by a fan in the bottom of the cylinder. The amount of combustion air can be varied
by varying the fan speed. The principal advantage of the air-assisted flares is that they can be
used where steam is not available. Although air assist is not usually used on large flares
(because it is generally not economical when the gas volume is large[3]) the number of large air-
assisted flares being built is increasing.[4]
7.1.1.3 Non-Assisted Flares
The non-assisted flare is just a flare tip without any auxiliary provision for enhancing the mixing
of air into its flame. Its use is limited essentially to gas streams that have a low heat content and
a low carbon/hydrogen ratio that burn readily without producing smoke. [5] These streams
require less air for complete combustion, have lower combustion temperatures that minimize
cracking reactions, and are more resistant to cracking.
7.1.1.4 Pressure-Assisted Flares
Pressure-assisted flares use the vent stream pressure to promote mixing at the burner tip.
Several vendors now market proprietary, high pressure drop burner tip designs. If sufficient
vent stream pressure is available, these flares can be applied to streams previously requiring
steam or air assist for smokeless operation. Pressure-assisted flares generally (but not
necessarily) have the burner arrangement at ground level, and consequently, must be located in
a remote area of the plant where there is plenty of space available. They have multiple burner
heads that are staged to operate based on the quantity of gas being released. The size, design,
number, and group arrangement of the burner heads depend on the vent gas characteristics.
7-5
-------
7.1.1.5 Enclosed Ground Flares
An enclosed flare's burner heads are inside a shell that is internally insulated. This shell reduces
noise, luminosity, and heat radiation and provides wind protection. A high nozzle pressure drop
is usually adequate to provide the mixing necessary for smokeless operation and air or steam
assist is not required. In this context, enclosed flares can be considered a special class of
pressure-assisted or non-assisted flares. The height must be adequate for creating enough draft
to supply sufficient air for smokeless combustion and for dispersion of the thermal plume.
These flares are always at ground level.
Enclosed flares generally have less capacity than open flares and are used to combust
continuous, constant flow vent streams, although reliable and efficient operation can be attained
over a wide range of design capacity. Stable combustion can be obtained with lower Btu content
vent gases than is possible with open flare designs (50 to 60 Btu/scf has been reported)[2],
probably due to their isolation from wind effects. Enclosed flares are typically found at
landfills.
7.1.2 Applicability
Flares can be used to control almost any VOC stream, and can handle fluctuations in VOC
concentration, flow rate, heating value, and inerts content. Flaring is appropriate for continuous,
batch, and variable flow vent stream applications. The majority of chemical plants and
refineries have existing flare systems designed to relieve emergency process upsets that require
release of large volumes of gas. These large diameter flares designed to handle emergency
releases, can also be used to control vent streams from various process operations.
Consideration of vent stream flow rate and available pressure must be given for retrofit
applications. Normally, emergency relief flare systems are operated at a small percentage of
capacity and at negligible pressure. To consider the effect of controlling an additional vent
stream, the maximum gas velocity, system pressure, and ground level heat radiation during an
emergency release must be evaluated. Further, if the vent stream pressure is not sufficient to
overcome the flare system pressure, then the economics of a gas mover system must be
evaluated, If adding the vent stream causes the maximum velocity limits or ground level heat
radiation limits to be exceeded, then a retrofit application is not viable.
Many flare systems are currently operated in conjunction with baseload gas recovery
systems. These systems recover and compress the waste VOC for use as a feedstock in other
processes or as fuel. When baseload gas recovery systems are applied, the flare is used in a
backup capacity and for emergency releases. Depending on the quantity of usable VOC that can
be recovered, there can be a considerable economic advantage over operation of a flare alone.
Streams containing high concentrations of halogenated or sulfur containing compounds are
not usually flared due to corrosion of the flare tip or formation of secondary pollutants (such as
SOj). If these vent types are to be controlled by combustion, thermal incineration, followed by
scrubbing to remove the acid gases, is the preferred method.[3]
7-6
-------
7.1.3 Performance
This section discusses the parameters that affect flare VOC destruction efficiency and presents
the specifications that must be followed when flares are used to comply with EPA air emission
standards.
7.1.3.1 Factors Affecting Efficiency
The major factors affecting flare combustion efficiency are vent gas flammability, auto-ignition
temperature, heating value (Btu/scf), density, and flame zone mixing.
The flammability limits of the flared gases influence ignition stability and flame extinction.
The flammability limits are defined as the stoichiometric composition limits (maximum and
minimum) of an oxygen-fuel mixture that will burn indefinitely at given conditions of
temperature and pressure without further ignition. In other words, gases must be within their
flammability limits to burn. When flammability limits are narrow, the interior of the flame may
have insufficient air for the mixture to burn. Fuels, such as hydrogen, with wide limits of
flammability are therefore easier to combust.
For most vent streams, the heating value also affects flame stability, emissions, and flame
structure. A lower heating value produces a cooler flame that does not favor combustion
kinetics and is also more easily extinguished. The lower flame temperature also reduces buoyant
forces, which reduces mixing.
The density of the vent stream also affects the structure and stability of the flame through
the effect on buoyancy and mixing. By design, the velocity in many flares is very low;
therefore, most of the flame structure is developed through buoyant forces as a result of
combustion. Lighter gases therefore tend to burn better. In addition to burner tip design, the
density also directly affects the minimum purge gas required to prevent flashback, with lighter
gases requiring more purge.[5]
Poor mixing at the flare tip is the primary cause of flare smoking when burning a given
material. Streams with high carbon-to-hydrogen mole ratio (greater than 0.35) have a greater
tendency to smoke and require better mixing for smokeless flaring. [3] For this reason one
generic steam-to-vent gas ratio is not necessarily appropriate for all vent streams. The required
steam rate is dependent on the carbon to hydrogen ratio of the gas being flared. A high ratio
requires more steam to prevent a smoking flare.
7.1.3.2 Flare Specifications
At too high an exit velocity, the flame can lift off the tip and flame out, while at too low a
velocity, it can burn back into the tip or down the sides of the stack.
7-7
-------
The EPA requirements for flares used to comply with EPA air emission standards are
specified in 40 CFR Section 60.18. The requirements are for steam-assisted, air-assisted, and
non-assisted flares. Requirements for steam-assisted, elevated flares state that the flare shall be
designed for and operated with:
an exit velocity at the flare tip of less than 60 ft/sec for 300 Btu/scf gas streams and less
than 400 ft/sec for >1,000 Btu/scf gas streams. For gas streams between 300-1,000
Btu/scf the maximum permitted velocity (Vmax, in ft/sec) is determined by the following
equation:
B + 1214
where Bv is the net heating value in Btu/scf.
no visible emissions. A five-minute exception period is allowed during any two
consecutive hours.
a flame present at all times when emissions may be vented. The presence of a pilot
flame shall be monitored using a thermocouple or equivalent device.
the net heating value of the gas being combusted being 300 Btu/scf or greater.
In addition, owners or operators must monitor to ensure that flares are operated and
maintained in conformance with their design.
7.2 Process Description
The elements of an elevated steam-assisted flare generally consist of gas vent collection piping,
utilities (fuel, steam, and air), piping from the base up, knock-out drum, liquid seal, flare stack,
gas seal, burner tip, pilot burners, steam jets, ignition system, and controls. Figure 7.1
7-8
-------
(6)
H*p» Pr«v«nt Flash Back
Fl«r» Slack
(5)
Gai Collvetlon HMdar
(D
Knockout
Drum -*
(21
Purg. j
_1_ Qt» 1 i
^U, w ^rfT
V
Oram
SMI
(3)
Figure 7.1: Steam-Assisted Elevated Flare System
7-9
-------
is a diagram of a steam-assisted elevated smokeless flare system showing the usual components
that are included.
7.2.1 Gas Transport Piping
Process vent streams are sent from the facility release point to the flare location through the gas
collection header. The piping (generally schedule 40 carbon steel) is designed to minimize
pressure drop. Ducting is not used as it is more prone to air leaks. Valving should be kept to
an absolute minimum and should be "car-sealed" (sealed) open. Pipe layout is designed to avoid
any potential dead legs and liquid traps. The piping is equipped for purging so that explosive
mixtures do not occur in the flare system either on start-up or during operation.
7.2.2 Knock-out Drum
Liquids that may be in the vent stream gas or that may condense out in the collection header and
transfer lines are removed by a knock-out drum. (See Figure 7.2.) The knock-out or
disentrainment drum is typically either a horizontal or vertical vessel located at or close to the
base of the flare, or a vertical vessel located inside the base of the flare stack. Liquid in the vent
stream can extinguish the flame or cause irregular combustion and smoking. In addition, flaring
liquids can generate a spray of burning chemicals that could reach ground level and create a
safety hazard. For a flare system designed to handle emergency process upsets this drum must
be sized for worst-case conditions (e.g., loss of cooling water or total unit depressuring) and is
usually quite large. For a flare system devoted only to vent stream VOC control, the sizing of
the drum is based primarily on vent gas flow rate with consideration given to liquid entrainment.
7.2.3 Liquid Seal
Process vent streams are usually passed through a liquid seal before going to the flare stack. The
liquid seal can be downstream of the knockout drum or incorporated into the same vessel. This
prevents possible flame flashbacks, caused when air is inadvertently introduced into the flare
system and the flame front pulls down into the stack. The liquid seal also serves to maintain a
positive pressure on the upstream system and acts as a mechanical damper on any explosive
shock wave in the flare stack.(51 Other devices, such as flame arresters and check valves, may
sometimes replace a liquid seal or be used in conjunction with it. Purge gas (as discussed in
Section 7.3.4) also helps to prevent flashback in the flare stack caused by low vent gas flow.
7-10
-------
ConawiMd/Entrmirwd
• To Storage
Figure 7.2: Typical Vertical Knock-out Drum
7-11
-------
if
I
/ \\\
I
Figure 7.3: Self-Supported Elevated Flare
7-12
-------
Figure 7.4: Derrick-Supported Elevated Flare
7-13
-------
Figure 7.5: Guy-Supported Elevated Flare
7-14
-------
7.2.4 Flare Stack
For safety reasons a stack is used to elevate the flare. The flare must he located so that it does
not present a hazard to surrounding personnel and facilities. Elevated flares can be self-
supported (free-standing), guyed, or structurally supported by a derrick. Examples of these three
types of elevated flares are shown in Figures 7.3, 7.4, and 7.5 for self-supported, derrick -
supported, and guy-supported flares, respectively. Self-supporting flares are generally used for
lower flare tower heights (30-100 feet) but can be designed for up to 250 feet. Guy towers are
designed for over 300 feet, while derrick towers are designed for above 200 feet.[4, 6, 7, 8, 9,
10]
Free-standing flares provide ideal structural support. However, for very high units the costs
increase rapidly. In addition, the foundation required and nature of the soil must be considered.
Derrick-supported flares can be built as high as required since the system load is spread over
the derrick structure. This design provides for differential expansion between the stack, piping,
and derrick. Derrick-supported flares are the most expensive design for a given flare height.
The guy-supported flare is the simplest of all the support methods. However, a considerable
amount of land is required since the guy wires are widely spread apart. A rule of thumb for
space required to erect a guy-supported flare is a circle on the ground with a radius equal to the
height of the flare stack.[6]
7.2.5 Gas Seal
Air may tend to flow back into a flare stack due to wind or the thermal contraction of stack gases
and create an explosion potential. To prevent this, a gas seal is typically installed in the flare
stack. One type of gas seal (also referred to as a flare seal, stack seal, labyrinth seal, or gas
barrier) is located below the flare tip to impede the flow of air back into the flare gas network.
There are also "seals" which act as orifices in the top of the stack to reduce the purge gas volume
for a given velocity and also interfere with the passage of air down the stack from the upper rim.
These are known by the names "internal gas seal, fluidic-seal, and arrestor seal" .[5] These seals
are usually proprietary in design, and their presence reduces the operating purge gas
requirements.
7.2.6 Burner Tip
The burner tip, or flare tip, is designed to give environmentally acceptable combustion of the
vent gas over the flare system's capacity range. The burner tips are normally proprietary in
design. Consideration is given to flame stability, ignition reliability, and noise suppression. The
maximum and minimum capacity of a flare to burn a flared gas with a stable flame (not
necessarily smokeless) is a function of tip design. Flame stability can be enhanced by flame
7-15
-------
holder retention devices incorporated in the flare tip inner circumference. Burner tips with
modern flame holder designs can have a stable flame over a flare gas exit velocity range of 1
to 600 ft/sec. [2] The actual maximum capacity of a flare tip is usually limited by the vent stream
pressure available to overcome the system pressure drop. Elevated flares diameters are normally
sized to provide vapor velocities at maximum throughput of about 50 percent of the sonic
velocity of the gas subject to the constraints of CFR 60.18.[1]
7.2.7 Pilot Burners
EPA regulations require the presence of a continuous flame. Reliable ignition is obtained by
continuous pilot burners designed for stability and positioned around the outer perimeter of the
flare tip. The pilot burners are ignited by an ignition source system, which can be designed for
either manual or automatic actuation. Automatic systems are generally activated by a flame
detection device using either a thermocouple, an infra-red sensor or, more rarely, (for ground
flare applications) an ultra-violet sensor.[4]
7.2.8 Steam Jets
A diffusion flame receives its combustion oxygen by diffusion of air into the flame from the
surrounding atmosphere. The high volume of fuel flow in a flare may require more combustion
air at a faster rate than simple gas diffusion can supply. High velocity steam injection nozzles,
positioned around the outer perimeter of the flare tip, increase gas turbulence in the flame
boundary zones, drawing in more combustion air and improving combustion efficiency. For the
larger flares, steam can also be injected concentrically into the flare tip.
The injection of steam into a flare flame can produce other results in addition to air
entrainment and turbulence. Three mechanisms in which steam reduces smoke formation have
been presented.[l] Briefly, one theory suggests that steam separates the hydrocarbon molecule,
thereby minimizing polymerization, and forms oxygen compounds that burn at a reduced rate
and temperature not conducive to cracking and polymerization. Another theory claims that
water vapor reacts with the carbon particles to form CO, CO2, and H2, thereby removing the
carbon before it cools and forms smoke. An additional effect of the steam is to reduce the
temperature in the core of the flame and suppress thermal cracking.[5] The physical limitation
on the quantity of steam that can be delivered and injected into the flare flame determines the
smokeless capacity of the flare. Smokeless capacity refers to the volume of gas that can be
combusted in a flare without smoke generation. The smokeless capacity is usually less than the
stable flame capacity of the burner tip.
Significant disadvantages of steam usage are the increased noise and cost. Steam aggravates
the flare noise problem by producing high-frequency jet noise. The jet noise can be reduced by
the use of small multiple steam jets and, if necessary, by acoustical shrouding. Steam injection
is usually controlled manually with the operator observing the flare (either directly or on a
television monitor) and adding steam as required to maintain smokeless operation. To optimize
7-16
-------
steam usage infrared sensors are available that sense flare flame characteristics and adjust the
steam flow rate automatically to maintain smokeless operation. Automatic control, based on
flare gas flow and flame radiation, gives a faster response to the need for steam and a better
adjustment of the quantity required. If a manual system is used, steam metering should be
installed to significantly increase operator awareness and reduce steam consumption.
7.2.9 Controls
Flare system control can be completely automated or completely manual. Components of a flare
system which can be controlled automatically include the auxiliary gas, steam injection, and the
ignition system. Fuel gas consumption can be minimized by continuously measuring the vent
gas flow rate and heat content (Btu/scf) and automatically adjusting the amount of auxiliary fuel
to maintain the required minimum of 300 Btu/scf for steam-assisted flares. Steam consumption
can likewise be minimized by controlling flow based on vent gas flow rate. Steam flow can also
be controlled using visual smoke monitors. Automatic ignition panels sense the presence of a
flame with either visual or thermal sensors and reignite the pilots when flameouts occur.
7.3 Design Procedures
Flare design is influenced by several factors, including the availability of space, the
characteristics of the flare gas (namely composition, quantity, and pressure level) and
occupational concerns. The sizing of flares requires determination of the required flare tip
diameter and height. The emphasis of this section will be to size a steam-assisted elevated flare
for a given application.
7.3.1 Auxiliary Fuel Requirement
The flare tip diameter is a function of the vent gas flow rate plus the auxiliary fuel and purge gas
flow rate. The purge gas flow rate is very small relative to the vent gas and fuel flow rates, so
it may be ignored when determining the tip diameter. The flow rate of the auxiliary fuel, if
required, is significant, and must be calculated before the tip diameter can be computed.
Some flares are provided with auxiliary fuel to combust hydrocarbon vapors when a lean
flare gas stream falls below the flammability range or heating value necessary to sustain a stable
flame. The amount of fuel required, F, is calculated based on maintaining the vent gas stream
net heating value at the minimum of 300 Btu/scf required by rules defined in the Federal
Register (see next section):
Q Bv + F Bf = (Q + F) (300 Btu / scf) (72)
7-17
-------
where
Q = the vent stream flow rate, scfm
Bv and Bf are the Btu/scf of the vent stream and fuel, respectively.
Rearranging gives:
F(sc£m) -
300 - B
(7.3)
The annual auxiliary fuel requirement, Fa, is calculated by:
Fa(Mscf/yr) = (F scfm)(60min/hr)(8760hr/yr) = 526F
(7.4)
Typical natural gas has a net heating value of about 1,000 Btn/scf. Automatic control of the
auxiliary fuel is ideal for processes with large fluctuations in VOC compositions. These flares
are used for the disposal of such streams as sulfur tail gases and ammonia waste gases, as well
as any low Btu vent streams.[2]
A 7.3.2 Flare Tip Diameter
Flare tip diameter is generally sized on a velocity basis, although pressure drop must also be
checked. Flare tip sizing for flares used to comply with EPA air emission standards is governed
by rules defined in the Federal Register (see 40 CFR 60.18). To comply with these requirements,
the maximum velocity of a steam-assisted elevated flare is determined as follows:
Net Heating Value of
Vent Stream
Bv (Btu/ scf)
Joo
300 - 1,000
>1,000
Maximum Velocity
V' (^/See)
Iog
10
60
400
1,214)/852
By determining the maximum allowed velocity, Vmax (ft/sec), and knowing the total
volumetric flow rate, Qtot (acfm), including vent stream and auxiliary fuel gas, a minimum flare
tip diameter, D,^ (in), can be calculated. It is standard practice to size the flare so that the
design velocity of flow rate Qtot, is 80 percent of Vmax, i.e.:
7-18
-------
4 Q
10 n 60 (
^ 0.
Otot
sec/min)
max
max
(7.5)
where
Qtot = Q + F (measured at stream temperature and pressure)
The flare tip diameter, D, is the calculated diameter, D = Dmm, rounded up to the next
commercially available size. The minimum flare size is 1 inch; larger sizes are available in 2-
inch increments from 2 to 24 inches and in 6-inch increments above 24 inches. The maximum
size commercially available is 90 inches.[5]
A pressure drop calculation is required at this point to ensure that the vent stream has
sufficient pressure to overcome the pressure drop occurring through the flare system at
maximum flow conditions. The pressure drop calculation is site specific but must take into
account losses through the collection header and piping, the knock-out drum, the liquid seal, the
flare stack, the gas seal, and finally the flare tip. Piping size should be assumed equal to the
flare tip diameter. Schedule 40 carbon steel pipe is typically used. If sufficient pressure is not
available, the economics of either a larger flare system (pressure drop is inversely proportional
to the pipe diameter) or a mover such as a fan or compressor must be weighed. (Refer to Section
7.3.8 for typical pressure drop relationships.)
7.3.3 Flare Height
The height of a flare is determined based on the ground level limitations of thermal radiation
intensity, luminosity, noise, height of surrounding structures, and the dispersion of the exhaust
gases. In addition, consideration must also be given for plume dispersion in case of possible
emission ignition failure. Industrial flares are normally sized for a maximum heat intensity of
1,500-2,000 Btu/hr-ft2 when flaring at their maximum design rates.[1,2] At this heat intensity
level, workers can remain in the area of the flare for a limited period only. If, however.
operating personnel are required to remain in the unit area performing their duties, the
recommended design flare radiation level excluding solar radiation is 500 Btu/hr-ft2. [1] The
intensity of solar radiation is in the range of 250-330 Btu/hr-ft2. [1] Flare height may also be
determined by the need to safely disperse the vent gas in case of flameout. The height in these
cases would be based on dispersion modeling for the particular installation conditions and is not
addressed here. The minimum flare height normally used is 30 feet.[5] Equation (7.6) by Hajek
and Ludwig may be used to determine the minimum distance, L, required from the center of the
flare flame and a point of exposure where thermal radiation must be limited.[l]
7-19
-------
where
T = fraction of heat intensity transmitted
/ = fraction of heat radiated
R = net heat release (Btu/hr)
K = allowable radiation (500 Btu/hr-ft2)
The conservative design approach used here ignores wind effects and calculates the distance
assuming the center of radiation is at the base of the flame (at the flare tip), not in the center.
It is also assumed that the location where thermal radiation must be limited is at the base of the
flare. Therefore, the distance, L, is equal to the required flare stack height (which is a minimum
of 30 feet). The/factor allows for the fact that not all the heat released in a flame can be
released as radiation. Heat transfer is propagated through three mechanisms: conduction,
convection, and radiation. Thermal radiation may be either absorbed, reflected, or transmitted.
Since the atmosphere is not a perfect vacuum, a fraction of the heat radiated is not transmitted
due to atmospheric absorption (humidity, particulate matter). For estimating purposes, however,
assume all of the heat radiated is transmitted (i.e., r = 1). The following is a summary of heat
radiated from various gaseous diffusion flames: [1]
7-20
-------
Gas
Hydrogen
Butane
Methane
Natural Gas
Flare Tip Diameter (in)
<1
1.6
3.3
8.0
16.0
<1
1.6
3.3
8.0
16.0
<1
1.6
3.3
8.0
16.0
Fraction of Heat Radiated (/)
.10
.11
1.6
1.5
1.7
.29
.29
.29
.28
.30
.16
.16
.15
.19
.23
In general, the fraction of heat radiated increases as the stack diameter increases. If stream-
specific data are not available, a design basis of/= 0.2 will give conservative results. [4] The
heat release, R, is calculated from the flare gas flow rate, W, and the net heating value, Bv, as
follows:
R (Btu/hr) = (fVlb/hr) (BvBtu/lb)
(7.8)
7.3.4 Purge Gas Requirement
The total volumetric flow to the flame must be carefully controlled to prevent low flow
flashback problems and to avoid flame instability. Purge gas, typically natural gas, N2, or C()2.
is used to maintain a minimum required positive flow through the system. If there is a
possibility of air in the flare manifold, N2, another inert gas, or a flammable gas must be used
to prevent the formation of an explosive mixture in the flare system. To ensure a positive flow
through all flare components, purge gas injection should be at the farthest upstream point in the
flare transport piping.
The minimum continuous purge gas required is determined by the design of the stack seals,
which are usually proprietary devices. Modern labyrinth and internal gas seals are stated to
require a gas velocity of 0.001 to 0.04 ft/sec (at standard conditions).[6, 7, 8, 9, 10] Using the
7-21
-------
conservative value of 0.04 ft/sec and knowing the flare diameter (in), the annual purge gas
volume, Fpu, can be calculated:
Fm (Mscf/yr) = (0.04 ft/sec)
pu
144
(3,600 sec/hr) (8,760 hr/yr)
(7.9)
= 6.88D2 (Mscf/yr)
There is another minimum flare tip velocity for operation without burn lock or instability. This
minimum velocity is dependent on both gas composition and diameter and can range from
insignificant amounts on small flares to 0.5 ft/sec on greater than 60-inch diameter units. [5]
Purge gas is also required to clear the system of air before startup, and to prevent a vacuum
from pulling air back into the system after a hot gas discharge is flared. (The cooling of gases
within the flare system can create a vacuum.) The purge gas consumption from these uses is
assumed to be minor.
7.3.5 Pilot Gas Requirement
The number of pilot burners required depends on flare size and, possibly, on flare gas
composition and wind conditions. Pilot gas usage is a function of the number of pilot burners
required to ensure positive ignition of the flared gas, of the design of the pilots, and of the mode
of operation. The average pilot gas consumption based on an energy-efficient model is 70 scf/hr
(of typical 1000 Btu per scf gas) per pilot burner.[6, 7, 8, 9, 10] The number of pilot burners,
N, based on flare size is:[6, 7, 8, 9, 10]
Flare Tip Diameter (in)
Number of Pilot Burners (N)
1-10
12-24
30-60
>60
1
2
3
4
The annual pilot gas consumption, F is calculated by:
7-22
-------
Fpi (Mscf/yr) = (70 scf/hr)(AO (8,760 hr/yr)
= 613 N
7.3.6 Steam Requirement
The steam requirement depends on the composition of the vent gas being flared, the steam
velocity from the injection nozzle, and the flare tip diameter. Although some gases can be flared
smokelessly without any steam, typically 0.01 to 0.6 pound of steam per pound of flare gas is
required.[6, 7, 8,9,10] The ratio is usually estimated from the molecular weight of the gas, the
carbon-to-hydrogen ratio of the gas, or whether the gas is saturated or unsaturated. For example,
olefins, such as propylene, require higher steam ratios than would paraffin hydrocarbons to burn
smokelessly. [2]
In any event, if a proprietary smokeless flare is purchased, the manufacturer should be
consulted about the minimum necessary steam rate. A small diameter flare tip (less than 24
inches) can use steam more effectively than a large diameter tip to mix air into the flame and
promote turbulence.[2] For a typical refinery, the average steam requirement is typically 0.25
Ib/lb, with this number increasing to 0.5 Ib/lb in chemical plants where large quantities of
unsaturated hydrocarbons are flared.[10]
For general consideration, the quantity of steam required, S, can be assumed to be 0.4
pounds of steam per pound of flare gas, W. Using a 0.4 ratio, the amount of steam required is:
S (Ibs/yr) = 0.4 (w Ib/yr) (8,760 hr/yr)
= 3,500 (W Ibs/hr) t'-11'
Operating a flare at too high a steam-to-gas ratio is not only costly, but also results in a lower
combustion efficiency and a noise nuisance. The capacity of a steam-assisted flare to burn
smokelessly may be limited by the quantity of steam that is available.
7.3.7 Knock-out Drum
As explained previously, the knock-out drum is used to remove any liquids that may be in the
vent stream. Two types of drums are used: horizontal and vertical. The economics of vessel
design influences the choice between a horizontal and a vertical drum. When a large liquid
storage vessel is required and the vapor flow is high, a horizontal drum is usually more
economical. Vertical separators are used when there is small liquid load, limited plot space, or
where ease of level control is desired. It is assumed here that the drum is not sized for
emergency releases and that liquid flow is minimal. Flares designed to control continuous vent
streams generally have vertical knockout drums, whereas emergency flares typically have
7-23
-------
horizontal vessels. The procedure described below applies to vertical drums exclusively. A
typical vertical knock-out drum is presented in Figure 7.2.
Liquid particles will separate when the residence time of the vapor is greater than the time
required to travel the available vertical height at the dropout velocity of the liquid particles, i.e.,
the velocity is less than the dropout velocity. In addition, the vertical gas velocity must be
sufficiently low to permit the liquid droplets to fall. Since flares are designed to handle small-
sized liquid droplets, the allowable vertical velocity is based on separating droplets from 300
to 600 micrometers in diameter.[l] The dropout velocity, U, of a particle in a stream, or the
maximum design vapor velocity, is calculated as follows:[11]
U (ft/sec) = G
(7.12)
where
G = design vapor velocity factor
/?, and pv = liquid and vapor densities, lb/ft3
Note that in most cases,
Pv
The design vapor velocity factor, G, ranges from 0.15 to 0.25 for vertical gravity separators at
85%offlooding.[ll]
Once the maximum design vapor velocity has been determined the minimum vessel cross-
sectional area, A, can be calculated by:
0 ftVmin
A (ft2) = * (7.12)
(60 sec/min) (U ft/sec)
where Q, is the vent stream flow in actual ft3/min, or Q adjusted to the vent stream temperature
and pressure.
The vessel diameter, d^, is then calculated by:
7-24
-------
(7.13)
In accordance with standard head sizes, drum diameters in 6-inch increments are assumed so:
d = dmin rounded to the next largest size (7 14)
Some vertical knockout drums are sized as cyclones and utilize a tangential inlet to generate
horizontal separating velocities. Vertical vessels sized exclusively on settling velocity (as in the
paragraph above) will be larger than those sized as cyclones.[5]
The vessel thickness, /, is determined based on the following:[13]
Diameter, d (inches) Thickness, t (inches)
d<36
36 144
0.25
0.37
50.5
0.75
1.0
Proper vessel height, h, is usually determined based on required liquid surge volume. The
calculated height is then checked to verify that the height-to-diameter ratio is within the
economic range of 3 to 5.[11] For small volumes of liquid, as in the case of continuous VOC
vent control, it is necessary to provide more liquid surge than is necessary to satisfy the h/d>3
condition. So for purposes of flare knock-out drum sizing:
h (in) = 3d (7.15)
7.3.8 Gas Mover System
The total system pressure drop is a function of the available pressure of the vent stream, the
design of the various system components, and the flare gas flow rate. The estimation of actual
pressure drop requirements involves complex calculations based on the specific system's vent
gas properties and equipment used. For the purposes of this section, however, approximate
7-25
-------
values can be used. The design pressure drop through the flare tip can range from ~ 0.1 to 2 psi
with the following approximate pressure drop relationships:[5]
Gas seal: 1 to 3 times flare tip pressure drop
Stack: 0.25 to 2 times flare tip pressure drop
Liquid seal and Knock- 1 to 1.5 times flare tip pressure drop plus
out drum: pressure drop due to liquid depth in the seal, which is
normally 0.2 to 1.5 psi.
Gas collection system: calculated based on diameter, length, and flow. System
is sized by designer to utilize the pressure drop
available and still leave a pressure at the stack base of
between 2 and 10 psi.
Typical total system pressure drop ranges from about 1 to 25 psi.[5]
7.4 Estimating Total Capital Investment
The capital costs of a flare system are presented in this section and are based on the
design/sizing procedures discussed in Section 7.3. The costs presented are in March 1990
dollars.*
Total capital investment, TCI, includes the equipment costs, EC, for the flare itself, the cost
of auxiliary equipment, the cost of taxes, freight, and instrumentation, and all direct and indirect
installation costs.
The capital cost of flares depends on the degree of sophistication desired (i.e., manual vs
automatic control) and the number of appurtenances selected, such as knock-out drums, seals,
controls, ladders, and platforms. The basic support structure of the flare, the size and height,
and the auxiliary equipment are the controlling factors in the cost of the flare. The capital
investment will also depend on the availability of utilities such as steam, natural gas, and
instrument air.
The total capital investment is a battery limit cost estimate and does not include the
provisions for bringing utilities, services, or roads to the site, the backup facilities, the land, the
research and development required, or the process piping and instrumentation interconnections
that may be required in the process generating the waste gas. These costs are based on a new
plant installation; no retrofit cost considerations such as demolition, crowded construction
working conditions, scheduling construction with production activities, and long interconnecting
*For information on escalating these prices to more current dollars, refer to
the EPA report Escalation Indexes for Air Pollution Control Costs and updates
thereto, all of which are installed on the OAQPS Technology Transfer Network
(CTC Bulletin Board).
7-26
-------
piping are included. These factors are so site-specific that no attempt has been made to provide
their costs.
7.4.1 Equipment Costs
Flare vendors were asked to provide budget estimates for the spectrum of commercial flare sizes.
These quotes [6, 7, 8, 9, 10] were used to develop the equipment cost correlations for flare units,
while the cost equations for the auxiliary equipment were based on references [12] and [13]
(knock-out drums) and [14] and [15] (piping). The expected accuracy of these costs is ± 30%
(i.e., "study" estimates). Keeping in mind the height restrictions discussed in Section 7.2.4,
these cost correlations apply to flare tip diameters ranging from 1 to 60 inches and stack heights
ranging from 30 to 500 feet. The standard construction material is carbon steel except when it
is standard practice to use other materials, as is the case with burner tips.
The flare costs, CFpresented in Equations 7.16 through 7.18 are calculated as a function of
stack height, L (ft) (30 ft minimum), and tip diameter, D (in), and are based on support type as
follows:
Self Support Group:
Cp ($) = (78.0 + 9.14D + 0.749L)2 (7.16)
Guv Support Group:
Cp ($) = (103 + 8.68D + 0.470L)2 (7.17)
Derrick Support Group:
CF ($) = (76.4 + 2.72D + 1.64L)2 (7.18)
The equations are least-squares regression of cost data provided by different vendors. It
must be kept in mind that even for a given flare technology (i.e., elevated, steam-assisted),
design and manufacturing procedures vary from vendor to vendor, so that costs may vary. Once
a study estimate is completed, it is recommended that several vendors be solicited for more
detailed cost estimates.
Each of these costs includes the flare tower (stack) and support, burner tip, pilots, utility
(steam, natural gas) piping from base, utility metering and control, liquid seal, gas seal, and
galvanized caged ladders and platforms as required. Costs are based on carbon steel
construction, except for the upper four feet and burner tip, which are based on 310 stainless
steel.
The gas collection header and transfer line requirements are very site specific and depend
on the process facility where the emission is generated and on where the flare is located. For
7-27
-------
the purposes of estimating capital cost it is assumed that the transfer line will be the same
diameter as the flare tip[6] and will be 100 feet long. Most installations will require much more
extensive piping, so 100 feet is considered a minimum.
The costs for vent stream piping, Cp, are presented separately in Equation 7.19 or 7.20 and
are a function of pipe, or flare, diameter Z).[15]
C ($) = 127.D1-21 (where 17/ < D < 24") (7_19)
C ($) - 139D1'07 (where 307/< D< 607/) (7.20)
The costs, Cp, include straight, Schedule 40, carbon steel pipe only, are based on 100 feet of
piping, and are directly proportional to the distance required.
The costs for a knock-out drum, CK, are presented separately in Equation 7.21 and are a
function of drum diameter, d (in), and height, h (in).[12, 13]
CR ($) = 14.2 [dt (h + 0.812d) ]°'737 (7.21)
where t is the vessel thickness, in inches, determined based on the diameter.
Flare system equipment cost, EC, is the total of the calculated flare, knock-out drum, and
piping costs.
EC ($) = Cp + CK + Cp (7.22)
Purchased equipment costs, PEC, is equal to equipment cost, EC, plus factors for ancillary
instrumentation (i.e., control room instruments) (.10), sales taxes (0.03), and freight (0.05) or,
PEC ($) = EC (1 + 0.10 + 0.03 + 0.05) = 1.18 EC (7.23)
7.4.2 Installation Costs
The total capital investment, TCI, is obtained by multiplying the purchased equipment cost,
PEC, by an installation factor of 1.92.
7-28
-------
TCI ($) - 1.92 PEC (7.24)
These costs were determined based on the factors in Table 7.1. The bases used in calculating
annual cost factors are given in Table 7.2. These factors encompass direct and indirect
installation costs. Direct installation costs cover foundations and supports, equipment handling
and erection, piping, insulation, painting, and electrical. Indirect installation costs cover
engineering, construction and field expenses, contractor fees, start-up, performance testing, and
contingencies. Depending on the site conditions, the installation costs for a given flare could
deviate significantly from costs generated by these average factors. Vatavuk and Neveril provide
some guidelines for adjusting the average installation factors to account for other-than-average
installation conditions .[16]
7.5 Estimating Total Annual Costs
The total annual cost, TAG, is the sum of the direct and indirect annual costs. The bases used
in calculating annual cost factors are given in Table 7.2
7.5.1 Direct Annual Costs
Direct annual costs include labor (operating and supervisory), maintenance (labor and materials),
natural gas, steam, and electricity. Unless the flare is to be dedicated to one vent stream and
specific on-line operating factors are known, costs should be calculated based on a continuous
operation of 8,760 hr/yr and expressed on an annual basis. Flares serving multiple process units
typically run continuously for several years between maintenance shutdowns.
Operating labor is estimated at 630 hours annually.[3] A completely manual system could
easily require 1,000 hours. A standard supervision ratio of 0.15 should be assumed.
T Maintenance labor is estimated at 0.5 hours per 8-hour shift. Maintenance materials costs
are assumed to equal maintenance labor costs. Flare utility costs include natural gas, steam, and
electricity.
Flare systems can use natural gas in three ways: in pilot burners that fire natural gas, in
combusting low Btu vent streams that require natural gas as auxiliary fuel, and as purge gas.
The total natural gas cost, Cf to operate a flare system includes pilot, Cpl, auxiliary fuel, Ca, and
purge costs, Cpu:
Cf ($/yr) = Cpi + Ca + Cpu (7.25)
where, Cp, is equal to the annual volume of pilot gas, Fpj, multiplied by the cost per scf, i.e.:
Cpi ($/yr) = (Fp. scf/yr) ($/scf) (7.26)
7-29
-------
Ca and C are similarly calculated.
Steam cost (Cs) to eliminate smoking is equal to the annual steam consumption 8,760 S
multiplied by the cost per Ib, i.e.:
Cs ($/yr) = (8,760 hr/yr) (S Ib/hr) ($/lb) (7 27)
The use of steam as a smoke suppressant can represent as much as 90% or more of the total
direct annual costs.
7.5.2 Indirect Annual Costs
The indirect (fixed) annual costs include overhead, capital recovery, administrative (G & A)
charges, property taxes, and insurance. Suggested indirect annual cost factors are presented in
Table 7.2.
Overhead is calculated as 60% of the total labor (operating, maintenance, and supervisory)
and maintenance material costs. Overhead cost is discussed in Chapter 2 of this Manual.
7-30
-------
Table 7.1: Capital Cost Factors for Flare Systems
Cost Item Factor
Direct Costs
Purchased equipment costs
Rare system, EC As estimated, A
Instrumentation 0.10 A
Sales taxes 0.03 A
Freight 0.05 A
Purchased equipment cost, PEC B = 1.18 A
Direct installation costs
Foundations & supports 0.12 B
Handling & erection 0.40 B
Electrical 0.01 B
Piping 0.02 B
Insulation 0.01 B
Painting 0.01 B
Direct installation costs 0.57 B
Site preparation As required, SP
Buildings As required, Bldg.
Total Direct Costs, DC 1.57 B + SP + Bldg.
Indirect Annual Costs, DC
Engineering 0.1 OB
Construction and Field expenses 0.10 B
Contractor fees 0.10 B
Start-up 0.01 B
Performance test 0.01 B
Contingencies 0.03 B
Total Indirect Costs, 1C 0.35 B
Total Capital Investment = DC + 1C 1.92 B + SP + Bldg.
The system capital recovery cost, CRC, is based on an estimated 15-year equipment life.
(See Chapter 2 of this Manual for a thorough discussion of the capital recovery cost and the
variables that determine it.) For a 15-year life and an interest rate of 7%, the capital
7-31
-------
recovery factor is 0.1098. The system capital recovery cost is the product of the system
capital recovery factor, CRF, and the total capital investment, TCI, or:
CRC ($/yr) = CRF x TCI = 0.1098 x TCI (7.28)
As shown in Table 7.2, G & A, taxes, and insurance can be estimated at 2%, 1%, and 1% of
the total capital investment, TCI, respectively.
7.6 Example Problem
The example problem described in this section shows how to apply the flare sizing and
costing procedures to the control of a vent stream associated with the distillation
manufacturing of methanol.
7.6.1 Required Information for Design
The first step in the design procedure is to determine the specifications of the vent gas to be
processed. The minimum information required to size a flare system for estimating costs are
the vent stream:
Volumetric or mass flow rate
Heating value or chemical composition
Temperature
System pressure
Vapor and liquid densities
In addition the following are needed to calculate direct annual costs.
Labor costs
Fuel costs
Steam costs
Vent stream parameters and cost data to be used in this example problem are listed in Table
7.3.
7-32
-------
Table 7.2: Suggested Annual Cost Factors for Flare Systems
Cost Item
Factor
Direct Annual Costs. DC
Operating labor{ 3}
Operator
Supervisor
Operating materials
Maintenance
Labor
Material
630 man-hours/year
15% of operator
Vi hour per shift
100% of maintenance labor
Utilities
Electricity
Purge gas
Pilot gas
Auxiliary fuel
Steam
All utilities equal to:
(Consumption rate) x
(Hours/yr) x (unit cost)
Indirect Annual Costs. 1C
Overhead
Administrative charges
Property tax
Insurance
Capital recovery3
Total Annual Cost
60% of total labor and material costs
2% of Total Capital Investment
1 % of Total Capital Investment
1 % of Total Capital Investment
0.1315 x Total Capital Investment
Sum of Direct and Indirect Annual Costs
"See Chapter 2.
7-33
-------
Table 7.3: Example Problem Data
Vent Stream Parameters
How rate 63.4 acfma
399.3 Ib/hr
Heat content 449 Btu/sci*
System pressure 10 psigc
Temperature 90 °F
Liquid density[ 17] 49.60 Ib/ft3d
Vapor density[ 17] 0.08446 Ib/ft3d
Cost Data (March 1990)[18.191
Operating hours 8,760 hrs/yr
Natural gas 3.03 $71000 scf
Steam 4.65 $71000 Ibs
Operating labor 15.64 $/hr
Maintenance labor 17.21 $/hr
"Measured at flare tip. Flow rate has been adjusted to account for drop in pressure
from 10 psig at source to 1 psig at flare tip.
"Standard conditions: 77°F, 1 atmosphere.
Tressure at source (gas collection point). Pressure at flare tip is lower: 1 psig.
dMeasured at standard conditions.
7-34
-------
7.6.2 Capital Equipment
The first objective is to properly size a steam-assisted flare system to effectively destroy 98%
of the VOC (methanol) in the vent gas stream. Using the vent stream parameters and the
design procedures outlined in Section 7.3, flare and knock-out drum heights and diameters
can be determined. Once equipment has been specified, the capital costs can be determined
from equations presented in Section 7.4.1.
7.6.2.1 Equipment Design
The first step in flare sizing is determining the appropriate flare tip diameter. Knowing the
net (lower) heating value of the vent stream, the maximum allowed velocity can be
calculated from the Federal Register requirements. Since the heating value is in the range of
300 to 1,000 Btu/scf, the maximum velocity, Vmax, is determined by Equation 7.1.
449 Btu/scf + 1,214
= 1.95
so,
Because the stream heating value is above 300 Btu/scf, no auxiliary fuel is required. Hence,
Qtot equals the vent stream flow rate. Based on Qtot and Vmax, the flare tip diameter can be
calculated using Equation 7.5.
min
= 1.
= 1.
= 1.
95
95.
64
Qtot
max
63.
89.5
m
4 acfm
ft/ sec
The next largest commercially available standard size of 2 inches should he selected for D.
The next parameter to determine is the required height of the flare stack. The heat
release from the flare is calculated using Equation 7.7.
R (Btu/hr) = (Wlb/hr) (B^Btu/lb)
7-35
-------
First the heat of combustion, or heating value, must be converted from Btu/scf to Btu/lb.
The vapor density of the vent stream at standard temperature and pressure is 0.08446 Ib/scf.
So,
= 449 Btu/scf = 531g
0.08446 Ib/scf
and,
R = (399.3 lb/hr)(5,316 Btu/lb) = 2,123,000 Btu/hr
Substituting R and appropriate values for other variables into Equation 7.6:
(1) (0.2) (2,123,000 Btu/hr)
4n (500 Btu/hr-ft2;
= 68 ft2
gives a height of L = 8.2 ft. The smallest commercially available flare is 30 feet, so L = 30
ft.
Next the knock-out drum must be sized. Assuming a design vapor velocity factor, G, of
0.20, and substituting the vapor and liquid densities of methanol into Equation 7.11 yields a
maximum velocity of:
U = G
\
P, - P,
•, ft/sec
| 49.60 - 0.08446
— (J . ^ (J . I
\ 0.08446
= 4.84 ft/sec
Given a vent gas flow rate of 63.4 scfm, the minimum vessel cross-sectional, diameter is
calculated by Equation 7.12:
7-36
-------
Q acfm
A =
'60 sec/min) (U ft/sec!
63.4
(60) (4.84)
= 0.218 ft2
This results in a minimum vessel diameter of:
d - 13 -
= 13.5^0.218
=6.3 inches
The selected diameter, d, rounded to the next largest 6 inches is 12 inches. Using the rule of
the height to diameter ratio of three gives a vessel height of 36 inches, or 3 feet.
7.6.2.2 Equipment Costs
Once the required flare tip diameter and stack height have been determined the equipment
costs can be calculated. Since the height is 30 feet, the flare will be self-supporting. The
costs are determined from Equation 7.16.
Cp = (78.0 + 9.14D + 0.749L)2
= [78.0 + 9. 14 (2 inches) +0.749 (30 ft)]2
= $14,100
Knock-out drum costs are determined using Equation 7.21, where t is determined from the
ranges Presented in Section 7.3.7. Substituting 0.25 for t:
C, = 14.2[dt (h + 0.812d)]
K
0.737
= 14.2H12) (0.25) (36+ 0 . 812 (12 ) ) ] °'737
= $530
Transport piping costs are determined using Equation 7.19.
C = 127D1'21
p
= 127 (2)1'21
= $290
The total auxiliary equipment cost is the sum of the knock-out drum and transport piping
costs, or $530 + $290 = $820.
7-37
-------
The total capital investment is calculated using the factors given in Table 7.1. The
calculations are shown in Table 7.4. Therefore:
Purchased Equipment Cost = "B" = 1.18 x A
=1.18 x (14,920) = $17,610
And:
Total Capital Investment (rounded) = 1. 92 x B
= 1.92 x (17,610) = $33,800.
7-38
-------
Table 7.4: Capital Costs for Flare Systems
Example Problem
Cost Item Cost
Direct Costs
Purchased equipment costs
Flare (self supporting) $14,100
Auxiliary equipment" 820
Sum = A $14,920
Instrumentation, 0.1A 1,490
Sales taxes, 0.03A 450
Freight, 0.05A 750
Purchased equipment cost, B $17,610
Direct installation costs
Foundation and supports, 0.12B 2,110
Handling & erection, 0.40B 7,040
Electrical, 0.01B 180
Piping, 0.02B 350
Insulation, 0.01B 180
Painting, O.OlB 180
Direct installation cost $10,040
Site preparation —
Facilities and buildings —
Total Direct Cost $27,650
Indirect Costs (installation)
Engineering, 0.10B 1,760
Construction and field expenses, 0.10B 1,760
Contractor fees, 0.10B 1,760
Start-up, O.OlB 180
Performance test, O.OlB 180
Contingencies, 0.03B 530
Total Indirect Cost 56,170
Total Capital Investment (rounded) S33.800
"Includes costs for knock-out drum and transport piping.
7-39
-------
7.6.3 Operating Requirements
Operating labor is estimated at 630 hours annually with supervisory labor at 15% of this
amount. Maintenance labor is estimated at 1/2 hour per shift. Maintenance material costs
are assumed to be equal to maintenance labor costs.
As stated above, since, the heat content of the example stream is above 300 Btu/scf (i.e.,
449 Btu/scf) no auxiliary fuel is needed. Natural gas is required, however, for purge and
pilot gas. Purge gas requirements are calculated from Equation 7.8.
F u = 6.88D2 = 6.88(2)2 = 27.5 Mscf / yr
Since the flare tip diameter is less than 10 inches, pilot gas requirements are based on one
pilot burner, (see Section 7.3.5) and are calculated by Equation 7.9.
F . = 613N
WhenN =
F .. = 613 Mscf / yr
Steam requirements are calculated from Equation 7.10:
S (Ib/yr) - 3,500 W
Inserting the methanol mass flow rate of 399.3 Ib/hr yields:
S = (3, 500) (399.3 Ib/hri
= 1,400 Mlb/yr
7.6.4 Total Annual Costs
The sum of the direct and indirect annual costs yields a total annual cost of $61,800. Table
7.5 shows the calculations of the direct and indirect annual costs for the flare system as
calculated from the factors in Table 7.2. Direct costs include labor, materials, and utilities.
Indirect costs are the fixed costs allocated to the project, including capital recovery costs and
such costs as overhead, insurance, taxes, and administrative charges.
Electrical costs of a mover system (fan, blower, compressor) would have to be included
if the vent stream pressure was not sufficient to overcome the flare system pressure drop. In
this example case, the pressure is assumed to be adequate.
7-40
-------
Table 7.5: Annual Costs for Flare System
Example problem
Cost Item
Calculations
Cost
Direct Annual Costs. DC
Operating Labor
Operator
Supervisor
Operating materials
Maintenance
Labor
Material
Utilities
Electricity
Purge gas
Pilot gas
Steam
Total DC (rounded)
Indirect Annual Costs, 1C
Overhead
630 hx $15.64
year h
15% of operator = 0.15 x 9,850
0.5hxshiftx8.760hx$17.21
shift 8 h yr h
100% of maintenance labor
27.5MscfxS3.03
yr Mscf
613 Mscfx $3.03
yr Mscf
1.400x 103lbx$4.65
yr 10%
60% of total labor and material costs
= 0.6(9,850 + 1,480 + 9,420 + 9,420)
Administrative charges 2% of Total Capital Investment = 0.02 ($33,800)
1% of Total Capital Investment = 0.01 ($33,800)
1% of Total Capital Investment = 0.01 ($33.800)
0.1098 x $33,800
Property tax
Insurance
Capital recovery"
Total 1C (rounded)
$ 9,850
1,480
9,420
9,420
80
1,860
6.510
$38,600
18,100
680
340
340
3.710
23,200
Total Annual Cost (rounded)
$61,800
aThe capital recovery cost factor, CRF, is a function of the flare equipment life and the opportunity cost of the
capital (i.e. interest rate). For example, for a 15 year equipment life and 7% interest rate, CRF = 0.1098.
7-41
-------
7.7 Acknowledgments
The authors gratefully acknowledge the following companies for contributing data to this
chapter:
• Flaregas Corporation (Spring Valley, NY)
• John Zink Company (Tulsa, OK)
• Kaldair Incorporated (Houston, TX)
• NAO Incorporated (Philadelphia, PA)
• Peabody Engineering Corporation (Stamford, CT)
• Piedmont HUB, Incorporated (Raleigh, NC)
7-42
-------
References
[1] Guide for Pressure-Relieving and Depressurizing Systems, Refining Department, API
Recommended Practice 521, Second Edition, September 1982.
[2] Kalcevic, V. (IT Enviroscience), "Control Device Evaluation Flares and the Use of
Emissions as Fuels", Organic Chemical Manufacturing Volume 4; Combustion
Control Devices, U.S. Environmental Protection Agency, Research Triangle Park,
NC, Publication no. EPA-450/3-80-026, December 1980, Report 4.
[3] Reactor Processes in Synthetic Organic Chemical Manufacturing Industry-
Background Information for Proposed Standards, U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC,
Preliminary Draft, EPA 450/3-90-016a, June 1990.
[4] Letter from J. Keith McCartney (John Zink Co., Tulsa, OK) to William M. Vatavuk
(U.S. Environmental Protection Agency, Research Triangle Park, NC), November 19,
1990.
[5] Letter from David Shore (Flaregas Corp., Spring Valley, NY) to William M. Vatavuk
(U.S. Environmental Protection Agency, Research Triangle Park, NC), October 3,
1990.
[6] Letter from Pete Tkatschenko (NAO, Inc., Philadelphia, PA) to Diana Stone (Radian,
Research Triangle Park, NC), May 2, 1990.
[7] Letter to Gary Tyler (Kaldair, Inc., Houston, TX) to Diana Stone (Radian, Research
Triangle Park, NC), April 10, 1990.
[8] Letter from Zahir Bozai (Peabody Engineering Corp., Stamford, CT) to Diana Stone
(Radian, Research Triangle Park, NC), May 7, 1990.
[9] Letter from James Parker (John Zink Co., Tulsa, OK) to Diana Stone (Radian,
Research Triangle Park, NC), April 17, 1990.
[10] Letter from Nick Sanderson (Flaregas Corp., Spring Valley, NY) to Diana Stone
(Radian, Research Triangle Park, NC), May 2, 1990.
[11] Wu, F.H., "Drum Separator Design, A New Approach," Chemical Engineering, April
2, 1984, pp. 74-81.
[12] Mulet, A., "Estimate Costs of Pressure Vessels Via Correlations," Chemical
Engineering, October 5, 1981, pp. 145-150.
7-43
-------
[13] Process Plant Construction Estimating Standards, Richardson Engineering Services,
Inc., Volume 4, 1988 Edition.
[14] Peters, Max S. and Klaus D. Timmerhaus, Plant Design and Economics for Chemical
Engineers, Third Edition, McGraw-Hill, 1980.
[15] Cost information from Piedmont HUB, Incorporated, Raleigh, NC, August 1990.
[16] Vatavuk, W.M., and R. Neveril, "Estimating Costs of Air Pollution Control Systems,
Part II: Factors for Estimating Capital and Operating Costs," Chemical Engineering,
November 3,1980, pp. 157-162.
[17] Handbook of Chemistry and Physics, 55th Edition, CRC Press, 1974-1975.
[18] Green, G.P. and Epstein, R.K., Employment and Earnings, Department of Labor,
Bureau of Labor Statistics, Volume 37, No. 4, April 1990.
[19] Monthly Energy Review, Energy Information Administration, Office of Energy
Markets and End Use, U.S. Department of Energy, DOE-EIA-0035(90/12), February
1990.
7-44
-------
Chapter 8
REFRIGERATED
CONDENSERS
Gunseli Sagun Shareef
Wiley J. Barbour
Susan K. Lynch
W. Richard Pelt
Radian Corporation
Research Triangle Park, NC 27709
William M. Vatavuk
Innovative Strategies and Economics Group, OAQPS
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
December 1995
8-1
-------
Contents
8.1 Introduction 8-4
8.1.1 System Efficiencies and Performance 8-4
8.2 Process Description 8-4
8.2.1 VOC Condensers 8-6
8.2.2 Refrigeration Unit 8-7
8.2.3 Auxiliary Equipment 8-8
8.3 Design Procedures 8-9
8.3.1 Estimating Condensation Temperature 8-10
8.3.2 VOC Condenser Heat Load 8-12
8.3.3 Condenser Size 8-14
8.3.4 Coolant Flow Rate 8-15
8.3.5 Refrigeration Capacity 8-15
8.3.6 Recovered VOC 8-16
8.3.7 Auxiliary Equipment 8-16
8.3.8 Alternate Design Procedure 8-16
8.4 Estimating Total Capital Investment 8-17
8.4.1 Equipment Costs for Packaged Solvent Vapor Recovery Systems 8-18
8.4.2 Equipment Costs for Nonpackaged (Custom) Solvent Vapor Recovery
Systems 8-22
8.4.3 Equipment Costs for Gasoline Vapor Recovery Systems 8-23
8.4.4 Installation Costs 8-23
8-2
-------
8.5 Estimating Total Annual Cost 8-28
8.5.1 Direct Annual Costs 8-28
8.5.2 Indirect Annual Costs 8-29
8.5.3 Recovery Credit 8-29
8.5.4 Total Annual Cost 8-30
8.6 Example Problem #1 8-30
8.6.1 Required Information for Design 8-30
8.6.2 Equipment Sizing 8-30
8.6.3 Equipment Costs 8-34
8.6.4 Total Annual Cost 8-35
8.7 Example Problem #2 8-35
8.7.1 Required Information for Design 8-35
8.8 Acknowledgments 8-37
Appendix 8A 8-38
Appendix 8B 8-41
References 8-45
8-3
-------
8.1 Introduction
Condensers in use today may fall in either of two categories: refrigerated or non-refrigerated. Non-
refrigerated condensers are widely used as raw material and/or product recovery devices in chemical
process industries. They are frequently used prior to control devices (e.g., incinerators or
absorbers). Refrigerated condensers are used as air pollution control devices for treating emission
streams with high VOC, concentrations (usually > 5,000 ppmv) in applications involving gasoline
bulk terminals, storage, etc.
Condensation is a separation technique in which one or more volatile components of a vapor
mixture are separated from the remaining vapors through saturation followed by a phase change.
The phase change from gas to liquid can be achieved in two ways: (a) the system pressure can be
increased at a given temperature, or (b) the temperature may be lowered at a constant pressure. In
a two-component system where one of the components is noncondensible (e.g., air), condensation
occurs at dew point (saturation) when the partial pressure of the volatile compound is equal to its
vapor pressure. The more volatile a compound (i.e., the lower the normal boiling point), the larger
the amount that can remain as vapor at a given temperature; hence the lower the temperature
required for saturation (condensation). Refrigeration is often employed to obtain the low
temperatures required for acceptable removal efficiencies. This chapter is limited to the evaluation
of refrigerated condensation at constant (atmospheric) pressure.
8.1.1 System Efficiencies and Performance
The removal efficiency of a condenser is dependent on the emission stream characteristics including
the nature of the VOC in question (vapor pressure/temperature relationship), VOC concentration,
and the type of coolant used. Any component of any vapor mixture can be condensed if brought
to a low enough temperature and allowed to conic to equilibrium. Figure 8.1 shows the vapor
pressure dependence on temperature for selected compounds.[l] A condenser cannot lower the inlet
concentration to levels, below the saturation concentration at the coolant temperature. Removal
efficiencies above 90 percent can be achieved with coolants such as chilled water, brine solutions,
ammonia, or chlorofluorocarbons, depending on the VOC composition and concentration level of
the emission stream.
8.2 Process Description
Figure 8.2 depicts a typical configuration for a refrigerated surface condenser system as an emission
control device. The basic equipment required for a refrigerated condenser system includes a VOC
condenser, a, refrigeration unit(s). and auxiliary, equipment (e.g., precooler, recovery/storage tank,
pump/blower, and piping).
8-4
-------
1000.0
400 —
100.0 —
40
s
B
10.0
1 0 —
0.1
001
441
261
552 -9
Temp«rBiur» f F)
-59
-99
-131.7
Figure 8.1: Vapor Pressures of Selected compounds vs. Temperature[l]
8-5
-------
Alr/VOC
Vnpcx in
Precootor ConO»n*a*
WMW/VOC
Figure 8.2: Schematic Diagram for a Refrigerated Condenser System
8.2.1 VOC Condensers
The two most common types of condensers used are surface and contact condensers. [21 In surface
condensers, the coolant does not contact tile gas stream. Most surface condensers in refrigerated
systems are the shell and tube type (see Figure 8.3).[3] Shell and tube condensers circulate the
coolant through tubes. The VOCs condense on the outside of the tubes (i. e., within the shell). Plate
and frame type heat exchangers are also used as condensers in refrigerated systems. In these
condensers, the coolant and the vapor flow separately over thin plates. In either design, the
condensed vapor forms a film on the cooled surface and drains away to a collection tank for storage,
reuse, or disposal.
In contrast, to surface condensers where the coolant does not contact either the vapors or the
condensate, contact condensers cool tile vapor stream by spraying either a liquid at ambient
temperature or a chilled liquid directly into the gas stream.
Spent coolant containing the VOCs from contact condensers usually can not be reused directly and
can be a waste disposal problem. Additionally, VOCs in the spent coolant can not be directly
recovered without further processing. Since the coolant from surface condensers does not contact
the vapor stream, it is not contaminated and can be recycled in a closed loop. Surface condensers
also allow for direct recovery of VOCs from tile gas Stream. This chapter addresses the design and
costing of refrigerated surface condenser systems only.
8-6
-------
Coolant
Inlet
Vapor
Outlet
Vapor
Inlet
Condensed
VOC
Coolant
Outlet
Figure 8.3: Schematic Diagram for a Refrigerated Condenser System
8.2.2 Refrigeration Unit
The commonly used mechanical vapor compression cycle to produce refrigeration consists of four
stages: evaporation, compression, condensation, and expansion (see Figure 8.4).[4] Tile cycle which
is used for single-stage vapor compression involves two pressures, high and low, to enable a
continuous process to produce a cooling effect. Heat absorbed from tile gas stream evaporates the
liquid coolant (refrigerant). Next, the, refrigerant (now in vapor phase) is compressed to a higher
temperature and pressure by tile system compressor. Then, the superheated refrigerant vapor is
condensed, rejecting its sensible and latent heat in the condenser. Subsequently, the liquid
refrigerant flows from the condenser through the expansion valve, where pressure and temperature
are reduced to those in the evaporator, thus completing the cycle.
The capacity of a refrigeration unit is the rate at which heat is removed, expressed in tons of
refrigeration. One ton of refrigeration is the refrigeration produced by melting one ton of ice at
32°F in 24 hours. It is the rate of removing heat equivalent to 12,000 Btu/h or 200 Btu/min. For
more details on refrigeration principles, see References [5] and [6].
Applications requiring low temperatures (below about -30 °F), multistage refrigeration systems
are frequently employed.[4] Multistage systems are designed and marketed in two different types-
compound and cascade. In compound systems, only one refrigerant is used. In a cascade system,
two or more separate refrigeration systems are interconnected in such a manner that one provides
a means of heat rejection for the other. Cascade systems are desirable for applications requiring
temperatures between -50 and -150°F and allow the use of different refrigerants in each cycle, [4]
Theoretically, any number of cascaded stages are possible, each stage requiring an additional
condenser and an additional stage of compression.
In refrigerated condenser systems, two kinds of refrigerants are used, primary and secondary.
Primary refrigerants are those that undergo a phase change from liquid to gas after absorbing heat.
Examples are ammonia (R-717), and chlorofluorocarbons such as chlorodifluoromethane (R-22) or
8-7
-------
Condenser
High Pressure Side
Expansion
Valve
Low Pressure Side
Compressor
Evaporator
Figure 8.4: Basic Refrigeration Cycle [4]
dichlorodifluorormethane (R-12). Recent concerns about the latter causing depletion of the ozone
layer is prompting development of substitute refrigerants.
Secondary refrigerants such as brine solutions act only as heat carriers and remain in liquid
phase. Conventional systems use a closed primary refrigerant loop that cools the secondary loop
through the heat transfer medium in the evaporator. The secondary heat transfer fluid is then
pumped to a VOC vapor condenser where it is used to cool the, air/VOC vapor stream. In some
applications, however, the primary refrigeration fluid is directly used to cool the vapor stream.
8.2.3 Auxiliary Equipment
As shown in Figure 8.2, some applications may require auxiliary equipment such as precoolers,
recovery/storage tanks, pumps/blowers, and piping.
If water vapor is present in the treated gas stream or if the VOC has a high freezing point (e.g.,
benzene), ice or frozen hydrocarbons may form on the condenser tubes or plates. This will reduce
the heat transfer efficiency of the condenser and thereby reduce the removal efficiency. Formation
of ice will also increase the pressure drop across the condenser. In such cases, a precooler may be
needed to condense the moisture prior to the VOC condenser. This precooler would bring the
temperature of the stream down to approximately 35 to 40 °F, effectively removing the moisture
from the gas. Alternatively, an intermittent heating cycle can be used to melt away ice build-up.
This may be accomplished by circulating ambient temperature brine through the condenser or by
the use of radiant heating coils. If a system is not operated continuously, the ice can also be
removed by circulating ambient air.
8-8
-------
A VOC recovery tank for temporary storage of condensed VOC prior to reuse, reprocessing, or
transfer to a larger storage tank may be necessary in some cases. Pumps and blowers are typically
used to transfer liquid (e.g., coolant or recovered VOC) and gas streams, respectively, within the
system.
8.3 Design Procedures
In this section are presented two procedures for designing (sizing) refrigerated surface condenser
systems to remove VOC from air/VOC mixtures. With the first procedure presented, one calculates
the condenser exit temperature needed to obtain a given VOC recovery efficiency. In the second
procedure, which is the inverse of the first, the exit temperature is given and the recover efficiency
corresponding to it is calculated.
The first procedure depends on knowledge of the following parameters:
1. Volumetric flow rate of the VOC-containing gas stream;
2. Inlet temperature of the gas stream;
3. Concentration and composition of the VOC in the gas stream;
4. Required removal efficiency of the VOC;
5. Moisture content of the emission stream; and
6. Properties of the VOC (assuming the VOC is a pure compound):
• Heat of condensation,
• Heat capacity, and
• Vapor pressure.
The design of a refrigerated condenser system requires determination of the VOC condenser size
and the capacity of the refrigeration unit. For a given VOC removal efficiency, the condensation
temperature and the heat load need to be calculated to determine these parameters. The data
necessary to perform the sizing procedures below as well as the variable names and their respective
units are listed in Table 8.1.
8-9
-------
Table 8.1:Required Input Data
Data Variable Name Units
Inlet Stream Flow Rate Qm scfm (77 °F; 1 atm)
°F
Inlet Stream Temperature Tin volume fraction
VOC Inlet Volume Fraction ymc m fractional (volume)
Required VOC Removal Efficiency 77
Antoine Equation Constants8 A,B,C Btu/lb-mole
Heat of Condensation of the VOCa AHVOC Btu/lb-mole- °F
Heat Capacity of the VOCa CpiWX Btu/lb- ° F
Specific Heat of the Coolant CpiCOOl Btu/lb-mole- °F
Heat Capacity of Air Cp_air
"See Appendix 8A for these properties of selected organic compounds.
The steps outlined below for estimating condensation temperature and the heat load apply to a
two-component mixture (VOC/air) in which one of the two components is considered to be
noncondensible (air). The VOC component is assumed to consist of a single compound. Also, the
emission stream is assumed to be free of moisture. The calculations are based on the assumptions
of ideal gas and ideal solution to simplify the sizing procedures. For a more rigorous analysis, See
Reference [5].
8.3.1 Estimating Condensation Temperature
The temperature necessary to condense the required amount of VOC must, be estimated to
determine the heat load. The first step is to determine the VOC concentration at the outlet of the
condenser for a given removal efficiency. This is calculated by first determining the partial pressure
of the VOC at the outlet assuming that the ideal gas law applies:
VOC partial pressure (outlet) = 760 Moles VOC in outlet stream ,8 i ^
Moles inlet stream - Moles VOC removed v"-1;
However:
Moles VOC in outlet stream = (Moles VOC in inlet stream)(l - r]) (8.2)
Moles VOC in inlet stream = (Moles in inlet stream) yvocm (8.3)
Moles VOC removed = (Moles VOC in inlet stream)r) (8.4)
8-10
-------
where
r) = removal efficiency of the condenser system (fractional)
= Moles VOC removed/Moles VOC in inlet
yVOc,in - Volume fraction of VOC in inlet stream
After substituting these variables in Equation 8.1, we obtain:
£
where
Pvoc = Partial pressure of the VOC in the exit stream (mm Hg).
The condenser is assumed to operate at a constant pressure of one atmosphere (760 mm Hg).
At the condenser outlet, the VOC In the, gas stream is assumed to be at equilibrium with the
VOC condensate. At equilibrium, the partial pressure of the VOC in the gas stream is equal to its
vapor pressure at that temperature assuming the condensate is pure VOC (i.e., vapor pressure Pvoc).
Therefore, by determining the temperature at which this condition occurs, the condensation
temperature can be specified. This calculation is based on the Antoine equation that defines the
relationship between vapor pressure and temperature for a particular compound:
P
log (vapor pressure) = logPvoc = A - (8.6)
con
where Tcon is the condensation temperature (°C). Note that T,on is in degrees Centigrade in this
equation. In Equation 8.6, A, B, and C are VOC-specific constants pertaining to temperature
expressed in °C and pressure in mm Hg (see Appendix 8A). Solving for Tcon and converting to
degrees Fahrenheit:
T = - - - - C 1.8 + 32
r P
10 fvoc i
con a- locr P
iog0 f
The calculation methods for a gas stream containing multiple VOCs are complex, particularly
when there are significant departures from the ideal behavior of gases and liquids. However, the
temperature necessary for condensation of a mixture of VOCs can be estimated by the weighted
average of the temperatures necessary to condense each VOC in the gas stream at a concentration
equal to the total VOC concentration.[1]
8-11
-------
8.3.2 VOC Condenser Heat Load
Condenser heat load is the amount of heat that must be removed from the inlet stream to attain the
specified removal efficiency. It is determined from an energy balance, taking into account the
enthalpy change due to the temperature change of the VOC, the enthalpy change due to the
condensation of the VOC, and the enthalpy change due to the temperature change of the air.
Enthalpy change due to the presence of moisture in the inlet gas stream is neglected in the following
analysis.
For the purpose of this estimation, it is assumed that, the total heat load on the system is equal
to the VOC Condenser heat load. Realistically, when calculating refrigeration capacity requirements
for low temperature cooling units, careful consideration should be given to the, process line, losses
and heat input of the process pumps. Refrigeration unit capacities are typically rated in terms of net
output and do not reflect any losses through process pumps or process lines.
First, the number of Ib-moles of VOC per hour in the inlet stream must be calculated by the
following expression:
Q .
voc, in ~ 292 Vvoc>in) (o-o)
where Mvocm is the molar flow rate of VOC in the inlet stream. The factor 392 is the volume (ft3)
occupied by one Ib-mole of inlet gas stream at standard conditions (77 °F and 1 atm). The number
of Ib-moles of VOC per hour in the outlet gas stream is calculated as follows:
MvocoH; = Mvocin(l-r|) (8.9)
where M^^is the molar flow rate of VOC in the exit stream. Finally, the number of Ib-moles of
VOC per hour that are condensed is calculated as follows:
Mvoc,con = MVOCiin - MVOC,OM( (8.10)
where Mvoc con is the flow rate of the VOC that is condensed.
The condenser heat load is then calculated by the following equation:
Hload = A//COT + &Huncon + kHnoncon (8.11)
where
VOC (Btu/hr)
— condenser heat load (Btu/hr)
= enthalpy change associated with the condensed
VOC (Btu/hr)
AHimcan = enthalpy change associated with the uncondensed
VOC (Btu/lir)
A//noncon = enthalpy change associated with the noncondensible
8-12
-------
air (Btu/hr).
Tile change in enthalpy of the condensed VOC is calculated as follows:
= MVOC,con [AHUOC + Cp,VOC ( T in ~ Tcon>] (8.12)
where A//voc is the molar heat of condensation and £>voc is the molar heat capacity of the VOC.
Each parameter varies as function of temperature. In Equation 8.12, AHVOC and Cpvoc are evaluated
at the mean temperature, i.e., (Tin + Tcon)/2.
The heat of condensation at a specific temperature, T2, (°R), can be calculated from the heat of
condensation at a reference temperature, Tj (°R), using the Watson Equation:[7]
1 - TJT '°'38
*tTi>|-; =T7^| (8-13)
where Tc (°R) is the VOC critical temperature.
The heat capacity can also be calculated for a specific temperature, T2, if heat capacity constants
(a, b, c, and d) are known for the particular compound. The heat capacity equation is:
CpiVOC = a + bT2 + cT22 + dT23 (8.14)
However, to simplify the heat load analysis, Cp,voc can be assumed to remain constant over the
temperature range of operation (i.e., Tm - Tcon) without much loss of accuracy in the heat load
calculations, as the sensible heat change in Equation 8.12 is relatively small compared to the
enthalpy change due to condensation.
Heat of condensation and heat capacity data are provided in Appendix 8A. The heat of
condensation for each compound is reported at its boiling point, while its heat capacity is given at
77°F. To estimate the heat of condensation at another temperature, use Equation 8.13. However,
the Appendix 8A heat capacity data may be used to approximate CftVOC at other temperatures, since
sensible heat changes are usually small, compared to condensation enthalpy changes.
The enthalpy change associated with the uncondensed VOC is calculated by the following
expression:
Cp,VOc(Tm ' Tcon) (8.15)
8-13
-------
Finally, the enthalpy change of the noncondensible air is calculated as follows:
AH
TOO
~> -J ^
- M
.
VOC, in
C . (T. - T )
~>, air in con
(8.16)
where Cpair is the specific heat of air. In both Equations 8.15 and 8.16, the Cp's are evaluated at the
mean temperature, (Tin + Tcon)/2.
8.3.3 Condenser Size
Condensers are sized based on the heat load, the logarithmic mean temperature difference between
the emission and coolant streams, and the overall heat transfer coefficient. The overall heat transfer
coefficient, U, can be estimated from individual heat transfer coefficients of the gas stream and the
coolant. The overall heat transfer coefficients for tubular heat exchangers where organic solvent
vapors in noncondensible gas are condensed on the shell side and water/brine is circulated on the
tube side typically range from 20 to 60 Btu/hr-ft2-°F according to Perry's Chemical Engineers'
Handbook[4]. To simplify the calculations, a single "U" value may be used to size these condensers.
This approximation is acceptable for purposes of making study cost estimates.
Accordingly, an estimate of 20 Btu/hr-ft2-°F can be used to obtain a conservative estimate of
condenser size. The following equation is used to determine the required heat transfer area:
H
r/AT
1m
where
Acon = condenser surface area (ft2)
U = overall heat transfer coefficient (Btu/hr-ft2-°F)
A7"/m = logarithmic mean temperature difference (°F).
The logarithmic mean temperature difference is calculated by the following equation, which is based
on the use of a countercurrent flow condenser:
AT, =
1m
(T. - T , ) -
in cool, out
(T - T , . )
con cool, in
T - T
in cool, out
T ~ T
con cool,in
(8.18)
8-14
-------
where
Tcooliin = coolant inlet temperature (°F)
Tcooiom ~ coolant outlet temperature (°F).
The temperature difference ("approach") at the condenser exit can be assumed to be 15 °F. In
other words, the coolant inlet temperature, Tcoo^, will be 15°F less than the calculated condensation
temperature, Tcon. Also, the temperature rise of the coolant is specified as 25 °F. (These two
temperatures- the condenser approach and the coolant temperature rise-reflect good design practice
that, if used, will result in an acceptable condenser size.) Therefore, the following equations can
be applied to determine the coolant inlet and outlet temperature:
Tcoal,m = Tcon-15 (8.19)
Tcool,out = Tcool.in + 25 (8.20)
8.3.4 Coolant Flow Rate
The heat removed from the emission stream is transferred to the coolant. By a simple energy
balance, the flow rate of the coolant can be calculated as follows:
H, .
W . = ^ (821)
cool f-i i rp — T } \"-^'±/
p, cool cool,out cool,in
where Wcool is the coolant flow rate (Ib/hr), and CpiCOOl is the coolant specific heat (Btu/lb- °F). Cp coo,
will vary according to the type of coolant used. For a 50-50 (volume %) mixture of ethylene glycol
and water, CptCOOi is approximately 0.65 Btu/lb- °F. The specific heat of brine
(salt water), another commonly used coolant, is approximately 1.0 Btu/lb-°F.
8.3.5 Refrigeration Capacity
The refrigeration unit is assumed to supply the coolant at the required temperature to the condenser.
The required refrigeration capacity is expressed in terms of refrigeration tons as follows:
H, ,
R = i£f£_ (8.22)
12,000
Again, as explained In section 8.3.2, Hhad does not include any heat losses.
8-15
-------
8.3.6 Recovered VOC
The mass of VOC recovered in the condenser can be calculated using the following expression:
Wvoc_cm=MVOCtConxMWmc (8.23)
where
WVOCiCon = mass of VOC recovered (or condensed) (Ib/hr)
MWmc = molecular weight of the VOC (Ib/lb-mole).
8.3.7 Auxiliary Equipment
The auxiliary equipment for a refrigerated surface condenser system may include:
• precooler,
• recovered VOC storage tank,
• pumps/blowers, and
• piping/ductwork.
If water vapor is present in the treated gas stream, a precooler may be needed to remove
moisture to prevent ice from forming in the VOC condenser. Sizing of a precooler is influenced by
the moisture concentration and the temperature of the emission stream. As discussed in Section
8.2.3, a precooler may not be necessary for intermittently operated refrigerated surface condenser
systems where the ice will have time to melt between successive operating periods.
If a precooler is required, a typical operating temperature is 35 to 40°F. At this temperature,
almost all of the water vapor present will be condensed without danger of freezing. These
condensation temperatures roughly correspond to a removal efficiency range of 70 to 80 percent if
the inlet stream is saturated with water vapor at 77 °F. The design procedure outlined in the
previous sections for a VOC condenser can be used to size a precooler, based on the psychometric
chart for the air-water vapor system (see Reference [4]).
Storage/recovery tanks are used to store the condensed VOC when direct recycling is not a
suitable option. The size of these tanks is determined from the amount of VOC condensate to be
collected and the amount of time necessary before unloading. Sizing of pumps and blowers is based
on the liquid and gas flow rates, respectively, as well as the system pressure drop between the inlet
and outlet. Sizing of the piping and ductwork (length and diameter) primarily depends upon the
stream flow rate, duct/pipe velocity, available space, and system layout.
8.3.8 Alternate Design Procedure
In some applications, it may be desirable to size and cost a refrigerated condenser system that will
use a specific coolant and provide a particular condensation temperature. The design procedure to
8-16
-------
be implemented in such a case would essentially be the same as the one presented in this section
except that instead of calculating the condenser exit temperature needed to obtain a specified VOC
recovery efficiency, the exit temperature is given and the corresponding recovery efficiency is
calculated.
The initial calculation would be to estimate the partial (=vapor) pressure of the VOC at the given
condenser exit temperature, Tcon, front Equation 8.6. Next, calculate r\ using Equation 8.24, which
is Equation 8.5 rearranged:
~ PVOC
8'24
(760 - P
voc,in{15° Pvoc
Finally, substitute the calculated Pwc into this equation to obtain r\. In the remainder of the
calculations to estimate condenser heat load, refrigeration capacity, coolant flow rate, etc., follow
the procedure presented in Sections 8.3.2 through 8.3.7.
8.4 Estimating Total Capital Investment
This section presents the procedures and data necessary for estimating capital costs for refrigerated
surface condenser systems in solvent vapor recovery and gasoline vapor recovery applications.
Costs for packaged and nonpackaged solvent vapor recovery systems are presented in Sections 8.4.1
and 8.4.2, respectively. Costs for packaged gasoline vapor recovery systems are described in
Section 8.4.3. Costs are calculated based on the design/sizing procedures discussed in Section 8.3.
Total capital investment, TCI, includes equipment cost, EC, for the entire refrigerated condenser
unit, auxiliary equipment costs, taxes, freight charges, instrumentation, and direct, and indirect
installation costs. All costs in this chapter are presented 3rd quarter 1990 dollars.*
For these control systems, the total capital investment is a battery limit cost estimate and does
not include the provisions for bringing utilities, services, or roads to the site; the, backup facilities;
the land; the working capital, the research and development required; or the process piping and
instrumentation interconnections that may be required in the process generating tile waste gas.
These costs are based on new plant installations; no retrofit cost considerations are included. The
retrofit cost factors are so site specific that no attempt has been made to provide them.
The expected accuracy of the cost estimates presented in this chapter is ±30 percent (/. e., "study"
* For information on escalating these prices to more current dollars, refer to
the EPA report Escalation Indexes for Air Pollution Control Costs and updates
thereto, all of which are installed on the OAQPS Technology Transfer Network
(CTC Bulletin Board).
8-17
-------
estimates). It must be kept in mind that even for a given application, design and manufacturing
procedures vary from vendor to vendor, so costs may vary.
In the next two sections, equipment costs are presented for packaged and nonpackaged (custom)
solvent vapor recovery systems, respectively. With the packaged systems the equipment cost is
factored from the refrigeration the custom systems, the equipment cost is determined as the sum of
the costs of the individual system components. Finally, equipment costs for packaged gasoline,
vapor recovery systems are given in Section 8.4.3.
8.4.1 Equipment Costs for Packaged Solvent Vapor Recovery Systems
Vendors were asked to provide refrigerated unit cost estimates for a wide range of applications. The
equations shown below for refrigeration unit equipment costs, ECr, are multivariable regressions of
data provided by two vendors and are only valid for the ranges listed in Table 8.2.[8,9] In this table,
the capacity range of refrigeration units for which cost data were available are shown as a function
of temperature.
Single Stage Refrigeration Units (less than 10 tons)
Ect = exp (9.83 - 0.0147^ + 0.340 In R) (8.25)
8-18
-------
Table 8.2: Applicability Ranges for the Refrigeration Unit Cost Equations
(Equations 8.25 to 8.27)
Temperature
Tcon(°F)a
40
30
20
10
Oto-5
-10
-20 to -25
-30
-40
-45 to -50
-55 to -60
-70
-75 to -80
-90
-100
Minimum Size Available
.R(tons)
Single Stage
0.85
0.63
0.71
0.44
0.32
0.21
0.13
NA
NA
NA
NA
NA
NA
NA
NA
Multistage
NAb
NA
NA
NA
NA
3.50
2.92
2.42
1.92
1.58
1.25
1.33
1.08
0.83
0.67
Maximum Size Available
/?(tons)
Single Stage
174
170
880
200
133
6.6
200
NA
NA
100C
100°
NA
NA
NA
NA
Multistage
NA
NA
NA
NA
NA
81
68
85
68
55
100
42
150
28
22
Tor condensation temperatures that lie between the levels shown, round off to the nearest level (e.g., if Tcm = 16°F, use 20°F)to determine minimum and
maximum available size.
''NA = System not available based on vendor data collected in this study
cOnly one data point available.
8-19
-------
Single Stage Refrigeration Units (greater than or equal to 10 tons)
ECt = exp (9.26 - 0.007rcon + 0.627 In R)
Multistage Refrigeration Units
ECt = exp (9.73 - 0.012Tcon + 0.584 In R)
NOTE: exp(a) = ea * 2.718a
(8.26)
(8.27)
Equations 8.25 and 8.26 provide costs for refrigeration units based on single stage designs, while
Equation 8.27 gives costs for multistage units. Equation 8.27 covers both types of multistage units,
"cascade" and "compound". Data provided by a vendor show that the costs of cascade and
compound units compare well, generally differing by less than 30%.[8] Thus, only one cost
equation is provided. Equation 8.25 applies to single stage refrigeration units smaller than 10 tons
and Equation 8.26 applies to single stage refrigeration units as large or larger than 10 tons. Single
stage units typically achieve temperatures between 40 and -20°F, although there are units that are
capable of achieving -60°F in a single stage.[8, 10] Multistage units are capable of lower
KWOOO
160.000
e
~o
0
8
o>
Jj 120,000
«
a
•a
a.
D
BO.OOO
40.000
20*F
40T
20T
20
60
Capacity (tons)
Figure 8.5: Refrigeration Unit Equipment Cost (Single Stage)[8,9]
8-20
-------
temperature operation between -10 and -100°F,
Single stage refrigeration unit costs are depicted graphically for selected temperatures in Figure
8.5. Figure 8.6 shows the equipment cost curves for multistage refrigeration units.
(NOTE: In Figure 8.5, the discontinuities in the curves at the 10 ton capacity are a result of the two
regression equations used. Equation 8.25 is used for capacities of less than 10 tons; Equation 8.26
is used for capacities greater than or equal to 10 tons.)
These costs are for outdoor models that are skid-mounted on steel beams and consist of the
following components: walk-in weatherproof enclosure, air-cooled low temperature refrigeration
machinery with dual pump design, storage reservoir, control panel and instrumentation, vapor
condenser, and necessary piping. AR refrigeration units have two pumps: a system pump and a
bypass pump to short-circuit the vapor condenser during no-load conditions. Costs for heat transfer
fluids (brine) are not included.
VOC condenser, recovery tank, the necessary connections, piping, and additional
instrumentation. Thus,
The equipment cost of packaged solvent vapor recovery systems (ECp) is estimated to be 25 percent
greater than the cost of the refrigeration unit alone [9]. The additional cost includes the
700.000
600,000
e
JO 500.000
3
TJ
400.000
8 300,000
o
a.
200.000
100.000
-80-F
•100-F .
•40-F
-30-F
•20-F
20
40 60
Capac«y(tona)
80
Figure 8.6: Refrigeration Unit Equipment Cost (Multistage)[9]
8-21
-------
EC=1.25EC (8.28)
p
Purchased equipment cost, PEC, includes the packaged equipment cost, ECp, and factors for
sales taxes (0.03) and freight (0.05). Instrumentation and controls are included with the packaged
units. Thus,
PECP = EcD (1 + 0.03 + 0.05) = 1.08ECP (8.29)
8.4.2 Equipment Costs for Nonpackaged (Custom) Solvent Vapor Recovery
Systems
To develop cost estimates for nonpackaged or custom refrigerated systems, information was
solicited from vendors on costs of refrigeration units, VOC condensers, and VOC storage/recovery
tanks [9, 1 1, 12]. Quotes from the vendors were used to develop the estimated costs for each type
of equipment. Only one set of vendor data was available for each type of equipment.
Equations 8.25, 8.26, and 8.27 shown above are applicable for estimating the costs for the
refrigeration units.
Equation 8.30 shows the equation developed for the VOC condenser cost estimates[l 1]:
ECcon = 34Acm + 3,775 (8.30)
This equation is valid for the range of 38 to 800 ft2 and represents costs for shell arid tube type heat
exchangers with 304 Stainless Steel tubes.
The following equation represents the storage/recovery tank cost data obtained from one
vendor[12]:
EC^= 2.72V,^ + 1,960 (8.31)
These costs are applicable for the range of 50 to 5,000 gallons and pertain to 316 stainless steel
vertical tanks.
Costing procedures for a precooler (ECpre) that includes a separate condenser/refrigeration unit
and a recovery tank are similar to that for a custom refrigerated condenser system. Hence,
Equations 8.25 through 8.31 would be applicable, with the exception of Equation 8.27, which
represents multistage systems. Multistage systems operate at much lower temperatures than that
required by a precooler.
Costs for auxiliary equipment such as ductwork, piping, fans, or pumps are designated as ECauz,
These items should be costed separately using methods described elsewhere in this Manual.
8-22
-------
The total equipment cost for custom systems, ECC is then expressed as:
ECC = ECr + ECcon + EC,mk + ECpre + EC^ (8.32)
The purchased equipment cost including ECC and factors for sales taxes (0.03), freight (0.05),
and instrumentation and controls (0.10) is given below:
PECC = ECC(1 + 0.03 + 0.05 + 0.10) = 1.18ECC (8.33)
8.4.3 Equipment Costs for Gasoline Vapor Recovery Systems
Separate quotes were obtained for packaged gasoline vapor recovery systems because these systems
are specially designed for controlling gasoline vapor emissions from such sources as storage tanks,
gasoline bulk terminals, and marine vessel loading and unloading operations. Systems that control
marine vessel gasoline loading and unloading operations also must meet U.S. Coast Guard safety
requirements.
Quotes obtained from one vendor were used to develop equipment cost estimates for these
packaged systems (see Figure 8.7. The cost equation shown below is a least squares regression of
these cost data and is valid for tile range 20 to 140 tons. [91
ECp = 4,91QK + 212,000 (8.34)
The vendor data in process flow capacity (gal/min) vs cost ($) were transformed into Equation 8.34
after applying the design procedures in Section 8.3. Details of the data transformation are given in
Appendix 8B.
The cost estimates apply to skid-mounted refrigerated VOC condenser systems for hydrocarbon
vapor recovery primarily at gasoline loading/storage facilities. The systems are intermittently
operated at -80 to -120°F allowing 30 to 60 minutes per day for defrosting by circulation of warm
brine. Multistage systems are employed to achieve these lower temperatures. Tile achievable VOC
removal efficiencies for these systems are in the range of 90 to 95 percent.
The packaged systems include the refrigeration unit with the necessary pumps, compressors,
condensers/evaporators, coolant reservoirs, the VOC condenser unit and VOC recovery tank,
precooler, instrumentation and controls, and piping. Costs for heat transfer fluids (brines) are not
included. The purchased equipment cost for these systems includes sales tax and freight and is
calculated using Equation 8.29.
8.4.4 Installation Costs
The total capital investment, TCI, for packaged systems is obtained by multiplying the purchased
equipment cost, PECp by the total installation factor:[13]
8-23
-------
o
Cg
8
CO
en
o
CO
O
O
•P
0)
1O
CT
LLJ
s
CD
•P
co
>,
en
o
o
CD
oc
o
Q.
CD
0)
c
CD
CD
r^
CO
CD
(H
O)
•H
o —' o
8
^
O
0661
PJC '
'o
a
O
8-24
-------
TCI = 1.15PECp (8.35)
For nonpackaged (custom) systems, the total installation factor is 1.74:
8-25
-------
Table 8.3: Capital Cost Factors for Nonpackaged (Custom) Refrigerated Condenser Systems
Cost Item Factor
Direct Costs
Purchased equipment costs
Refrigerated condenser system, EC As estimated, A
Instrumentation 0.10 A
Sales taxes 0.03 A
Freight 0.05 A
Purchased equipment cost, PEC B = 1.18 Aa
Direct installation costs
Foundations & supports 0.08 B
Handling & erection 0.14 B
Electrical 0.08 B
Piping 0.02 B
Insulation 0.1 OB
Painting 0.01 B
Direct installation costs 0.43 B
Site preparation As required, SP
Buildings As required, Bldg.
Total Direct Costs, DC 1.43 B + SP + Bldg.
Indirect Costs (installation)
Engineering
Construction and field expenses
Contractor fees
Start-up
Performance test
Contingencies
Total Indirect Costs, 1C
Total Capital Investment = DC + 1C 1.74 B + SP + Bldg.b
"Purchased equipment cost factor for packaged systems is 1.08 with instrumentation included.
bFor packaged systems, total capital investment = 1.15PEC .
8-26
-------
Table 8.4: Suggested Annual Cost Factors for Refrigerated Condenser Systems
Cost Item
Factor
Direct Annual Costs. DC
Operating Labor
Operator
Supervisor
Operating materials
Maintenance
Labor
Material
Electricity[9]
at 40°F
at 20°F
at -20°F
at -50°F
at-100°F
Indirect Annual Costs, 1C
Overhead
Administrative charges
Property tax
Insurance
Capital recovery3
Recovery Credits. RC
Recovered VOC
Total Annual Cost
1/2 hour per shift
15% of operator
1/2 hour per shift
100% of maintenance labor
1.3kW/ton
2.2 kW/ton
4.7 kW/ton
5.0 kW/ton
11.7kW/ton
60% of total labor and
maintenance material costs
2% of Total Capital Investment
1 % of Total Capital Investment
1% of Total Capital Investment
0.1098 x Total Capital Investment
Quantity recovered x operating hours
DC + 1C - RC
"Assuming a 15 year life at 7%[13]. See Chapter 2.
8-27
-------
TCI=1.74PECC (8.36)
An itemization of tile total installation factor for nonpackaged systems is shown in Table 8.3.
Depending on the site conditions., the installation costs for a given system could deviate
significantly, from costs generated by there average factors. Guidelines are available for adjusting
these average installation factors.[14]
8.5 Estimating Total Annual Cost
The total annual cost (TAG) is the sum of the direct and indirect annual costs. The bases used in
calculating annual cost factors are given in Table 8.4
8.5.1 Direct Annual Costs
Direct annual costs, DC, include-labor (operating and supervisory), maintenance (labor and
materials), and electricity.
Operating labor is estimated at 1/2-hour per 8-hour shift. The supervisory labor cost is estimated
at 15% of the operating labor cost. Maintenance labor is estimated at 1/2-hour per 8-hour shift.
Maintenance materials costs are assumed to equal maintenance labor costs.
Utility costs for refrigerated condenser systems include electricity requirements for the
refrigeration unit and any pumps/blowers. The power required by the pumps/blowers is negligible
when compared with the refrigeration unit power requirements. Electricity requirements for
refrigerated condenser systems are summarized below:
Electricity (E. kW/ton) Temperature (°F)
1.3 40
2.2 20
4.7 -20
5.0 -50
11.7 -100
These estimates were developed from product literature obtained from one vendor. [9] The
electricity cost, Ce, can then be calculated from the following expression:
C = - E6 p (9
e 085 s \o.
8-28
-------
where
0S = system operating hours (hr/yr)
pe = electricity cost
The factor 0.85 accounts for the mechanical efficiency of the compress or. [1]
8.5.2 Indirect Annual Costs
Indirect annual costs, 1C, are calculated as the sum of capital recovery costs plus general and
administrative (G&A), overhead, property tax, and insurance costs. Overhead is assumed to be
equal to 60 percent of the sum of operating, supervisory, and maintenance labor, and maintenance
materials. Overhead cost is discussed in Chapter 2 of this Manual.
The system capital recovery cost, CRC, is based on an estimated 15-year equipment life. [13]
(See Chapter 2 of the Manual for a discussion of the capital recovery cost.) For a 15-year life and
an interest rate of 7 percent, the capital recovery factor is 0.1098. The system capital recovery cost
is then estimated by:
CRC = 0.1098 TCI (8.38)
G&A costs, property tax, and insurance are factored from total capital investment, typically at
2 percent, 1 percent, and 1 percent, respectively.
8.5.3 Recovery Credit
If the condensed VOC can be directly reused or sold without further treatment, then the credit from
this operation is incorporated in the total annual cost estimates. The following equation can be used
to estimate the VOC, recovery credit, RC:
RC = Wvoc,con 0sPvoc (8.39)
where
pvoc = resale value of recovered VOC ($/lb)
Wvoccon = quantity of VOC recovered (Ib/hr).
8-29
-------
8.5.4 Total Annual Cost
The total annual cost (TAG) is calculated as the sum of the direct and indirect annual costs, minus the
recovery credit:
TAG = DC + 1C - RC (8.40)
8.6 Example Problem #1
The example problem described in this section shows how to apply the refrigerated condenser system
sizing and costing procedures to the control of a vent stream consisting of acetone, air, and a
negligible amount of moisture. This example problem assumes a required removal efficiency and
calculates the temperature needed to achieve this level of control.
8.6.1 Required Information for Design
The first step in the design procedure is to specific the gas stream to be processed. Gas stream
parameters to be used in this example problem are listed in Table 8.5. The values for the Antoine
equation constants, heat, of condensation, and heat capacity of acetone are obtained from Appendix
8A. Specific heat of the coolant is obtained from Perry's Chemical Engineers' Handbook[4].
8.6.2 Equipment Sizing
The first step in refrigerated condenser sizing is determining the partial pressure of the VOC at the
outlet of the condenser for a given removal efficiency. Given the stream flow rate, inlet VOC
concentration, and the required removal efficiency, the partial pressure of the VOC at the outlet can
be calculated using Equation 8.5.
= 760 0.375(1-0.90) = 4
voc (1 - 0.375(0.90)
Next, the temperature necessary to condense the required amount of VOC must be determined
ing Equation 8.7:
using Equation
1210-595
- 229.664 1.8 + 32 = 16°F
7.117 - Iog1043
8-30
-------
Table 8.5: Example Problem Data
Vent Stream Parameters
Inlet Stream Row Rate 100 scfrna
Inlet Stream Temperature 86°F
VOC to be Condensed Acetone
VOC Inlet Volume Fraction 0.0375
Required VOC Removal Efficiency .90
Antoine Equation Constants for Acetone:
A 7.117
B 1210.595
C 229.664
Heat of Condensation of Acetoneb 12,510 Btu/lb-mole
Heat Capacity of Acetone0 17.90 Btu/lb-mole-°F
Specific Heat of Coolant0 (ethylene glycol) 0.65 Btu/lb-°F
Heat Capacity of Airc 6.95 Btu/lb-mole-°F
Annual Cost Data Assumed
Operating Labor $15.64/hr
Maintenance Labor $17,2/hr
Electricity $0.0461/kWh
Acetone Resale Value $0.10/lb
'Standard conditions: 77°F and 1 atmosphere.
Evaluated at the acetone boiling point (134°F).
These properties were evaluated at 77°F.
8-31
-------
The next step is to estimate the VOC condenser heat load. Calculate: (1) the VOC flow rate
for the inlet/outlet emission streams, (2) the flow rate of the condensed VOC, and (3) the
condenser heat balance. The flow rate of VOC in the inlet stream is calculated from Equation 8.8.
MVQc ,n = -122 (0.375) 60 = 5.74 lb-moles/hr
The flow rate of VOC in the outlet stream is calculated using Equation 8.9 as follows:
Mmc,ml = 5.74 (1 - 0.90) = 0.574 lb-moles/hr
Finally, the flow rate of condensed VOC is calculated with Equation 8.10:
MVOCtCon = 5.74 - 0.574 = 5.166 lb-moles/hr
Next, the condenser heat balance is conducted. As indicated in Table 8.5, the acetone heat of
condensation is evaluated at its boiling point, 134°F. However, it is assumed (for simplicity) that
all of the acetone condenses at the condensation temperature, Tcon = 16°F. To estimate the heat of
condensation at 16°F, use the Watson equation (Equation 8.13) with the following inputs: Tc =
918°R(Appendix 8A); T, = 134 + 460 = 594°R; T2 = 16 + 460 = 476°R. Upon substitution, we
obtain:
(AHwcatl6°F) =12
voc I - 594/918
= 14,080 Btu/lb-mole
As Table 8.5 shows, the heat capacities of acetone and air and the specific heat of the, coolant
were all evaluated at, 77 °F. This temperature is fairly close to the condenser mean operating
temperature, i.e., (86 + 16)/2 = 51°F. Consequently, using tile 77 °F values would not add
significant additional error to the heat load calculation.
The change in enthalpy of the condensed VOC is calculated using Equation 8.12:
AHcon = 5.166 [14,080 + 17.90(86 - 16)] = 79,210 Btu/hr
The enthalpy change associated with the uncondensed VOC is calculated from Equation 8.15:
8-32
-------
mcon = (0.574)(17.90)(86 - 16) = 719 Btu/hr
Finally, the enthalpy change of the noncondensible air is estimated from Equation 8.16:
100
AH
noncon
392
•60 - 5.74
6.95(86 - 16) = 4,654 Btu/hr
The condenser heat load is then calculated by substituting tJlcon, &Huncon, and tHnoncon in Equation
8.11:
Hload = 79,210 + 719 + 4,654 = 84,583 Btu/hr
The next step is estimation of the VOC condenser size. The logarithmic mean temperature
difference is calculated using Equation 8.18. In this calculation, TcooUn =16- 15 = 1 °F and Tcoolou,
1 + 25 = 26 °F from Equations 8.19 and 8.20, respectively:
= (86-26) - (16-l)s32t5.p
n 86-26
In -
16-1
The condenser surface area can then be calculated using Equation 8.17.
A - 84'583 =130
20(32.5)
In this equation, a conservative value of 20 Btu//hr-ft2-°F is used as the overall heat transfer
coefficient.
The coolant flow rate can be calculated using Equation 8.21.
= 84'583 = 5,205 Ib/hr
0001 0.65(26 ~ 1)
The refrigeration capacity can be estimated from Equation 8.22 as follows:
84,583
12,000
R= 84'583 =7.05 tons
Finally, the quantity of recovered VOC can be estimated using Equation 8.23:
Wvoc,con = 5.166 x 58.08 = 300 Ib/hr
8-33
-------
where the molecular weight of acetone is obtained from Appendix 8A.
Note that in this example case, the partial pressure of acetone at the condenser exit is relatively
high (43 mm Hg). In applications where much lower outlet concentrations are desired, a second
control device (e.g., incinerator, adsorber) to operate in series with the condenser may need to be
considered.
8.6.3 Equipment Costs
Once the system sizing parameters have been determined, the equipment costs can be calculated.
For the purpose of this example, a custom refrigerated condenser system, including a refrigeration
unit, a VOC condenser, and a recovery tank will be costed.
From Table 8.2, a single stage refrigeration unit appears to be suitable for the example problem
with an estimated condensation temperature of 16°F and capacity of 7.05 tons. Hence Equation
8.25, which is applicable to units less than 10 tons, is selected for estimating costs. Application of
this equation results in the following value for the refrigeration unit cost:
ECT = exp [9.83 - 0.014(16) + 0.3401n(7.05)] = $28,855
VOC condenser cost is computed using Equation 8.30 as follows:
ECcon = 34(130) + 3,775 = $8,195
Recovery tank cost can be calculated from Equation 8.31. For this case, Wvoccon = 300 Ib/hr, which
is equivalent to 45.5 gal/hr (density of acetone is about 6.6 Ib/gal). Assuming an 8-hour daily
operation, the interim storage capacity requirement would be 364 gallons. Application of Equation
8.31 leads to the following:
ECtmk = 2.72(364) + 1,960 = $2,950
Assuming there are no additional costs due to precooler or other auxiliary equipment, the total
equipment cost is calculated from Equation 8.32:
ECC = 28,855 + 8,195 + 2,950 + 0 + 0 = $40,000
The purchased equipment cost including instrumentation, controls, taxes, and freight is estimated
using Equation 8.33:
PECC = 1.18(40,000) = $47,200
The total capital investment is calculated using Equation 8.36:
TCI = 1.74(47,200) = $82,128 » $82,100
8-34
-------
8.6.4 Total Annual Cost
Table 8.6 summarizes the estimated annual costs for the, example problem. The cost calculations
are shown in the table.
Direct annual costs for refrigerated systems include labor, materials, and utilities. Labor costs
are based on 8-hr/day, 5-day/week operation. Supervisory labor is computed at 15 percent of
operating labor, and operating and maintenance labor are each based on 1/2 hr per 8-hr shift. The
electricity cost is based on a requirement of 2.2 kW/ton (see page 8-30), because the condensation
temperature (16°F) is close to the 20°F temperature given for this value.
Indirect annual costs include overhead, capital recovery, administrative charges, property tax,
and insurance.
Total annual cost is estimated using Equation 8.40. For this example case, application of
refrigerated condensation as a control measure results in an annual savings of $37,500. As Table
8.6 shows, the acetone recovery credit is over twice the direct and Indirect costs combined.
Clearly, this credit has more influence on the total annual cost than any other Component.
Although the credit depends on three parameters-the acetone recovery rate, the annual operating
hours, and the acetone salvage value ($0.10/lb)-the last parameter is often the most difficult to
estimate. This is mainly because the salvage value varies according to the facility location as well
as the current state of the chemical market.
8.7 Example Problem #2
In this example problem, the alternate design procedure described in Section 8.3.8 is illustrated. The
temperature of condensation is given, and the resultant removal efficiency is calculated. The example
stream inlet parameters are identical to Example Problem #1 with the exception that removal efficiency
is not specified and the required temperature of condensation is assumed to be 16°F.
8.7.1 Required Information for Design
The first step is to calculate the partial pressure of the VOC at the specified temperature (16°F) using
Equation 8.6 to solve for Pvoc'-
log pimn = A - B
VOC
T + C
con
Remember to convert Tcon to degrees Centigrade, i.e., 16°F = -8.9°C.
8-35
-------
Table 8.6: Annual Cost for Refrigerated Condenser System
Example Problem
Cost Item
Calculations
Cost
Direct Annual Costs. DC
Operating Labor
Operator
Supervisor
Operating materials
Maintenance
Labor
Material
Utilities
Electricity
Total DC
Indirect Annual Costs. JC
Overhead
Administrative charges
Property tax
Insurance
Capital recovery*
Total 1C
Recovery Credits. RC
Recovered Acetone
Total Annual Cost (rounded)
0.5hxshiftx2.080hxS15.64
shift 8h yr h
15% of operator = 0.15 x 2,030
O.Shx shift x2.080hxS17.21
shift 8h yr h
100% maintenance labor
7.05tons x 2.2kw x 2.080h x $0.0461
0.85 Ton yr kwh
60% of total labor and maintenance material
= 0.6 (2,030 + 305 + 2,240 + 2,240)
2% of Total Capital Investment =0.02($82,100)
1% of Total Capital Investment =0.01($82,100)
1% of Total Capital Investment =0.01($82,100)
0.1098x882,100
3001bx2.080hx$0.10
h yr Ib
$2,030
300
2,240
2,240
1.750
$8,560
4,090
1,640
820
820
9.010
$16,380
($62,400)
($37,500)
(Savings)
The capital recovery cost factor, CRF, is function of the refrigerated condenser equipment life and the opportunity cost
of the capital (i.e., interest rate). For example, for a 15 year equipment life and a 7% interest rate, CRF = 0.1098.
8-36
-------
Substituting the values for the Antoine equation constants lor acetone as listed in Table 8.5:
=7 117 1210.595
voc ~ -8.9 f 229.664
/>voc = 43 mm Hg.
Using Equation 8.24, the removal efficiency is:
_ [760(0.375)] -43 =0_9Q
0.375(760 - 43)
The remainder of the calculations in this problem are identical to those in Example Problem #1.
8.8 Acknowledgments
The authors gratefully acknowledge the following companies for contributing data to this chapter:
• Edwards Engineering (Pompton Plains, NJ)
• Piedmont Engineering (Charlotte, NC)
• Universal Industrial Refrigeration (Gonzales, LA)
• FIT Standard (Atlanta, GA)
• XChanger (Hopkins, MN)
• Buffalo Tank Co. (Jacksonville, FL)
8-37
-------
Appendix 8A
Properties of-Selected
Compounds
8-38
-------
Table 8.7: Properties of Selected Compounds
Compound
Acetone
Acetylene
Acrylonitrile
Aniline
Benzene
Benzonitrile
Butane
Chloroethane
Chloroform
Chloromethane
Cyclobutane
Cyclohexane
Cyclopentane
Cyclopropane
Diethyl ether
Dimethylamine
Ethylbenzene
Ethylene oxide
Heptane
Hexane
Methanol
Octane
Pentane
Toluene
o - Xylene
m - Xylene
p - Xylene
Critical
Temp."
(°R)
918
555
-
1259
1012
1259
766
829
966
750
-
997
921
716
840
788
1111
845
973
914
923
1024
846
1065
1135
1111
1109
Boiling
Point
(°F)
134
-119
171
364
176
376
31
54
143
-12
55
177
121
-27
94
44
277
51
209
156
148
258
97
231
292
282
281
Molecular
Weight
(Ib/lb-mole)
58.08
26.02
53.06
93.13
78.11
103.12
58.12
64.52
119.39
50.49
56.10
84.16
70.13
42.08
74.12
45.09
106.17
44.05
100.12
86.18
32.04
114.23
72.15
92.14
106.17
106.17
106.17
Heat of
Condensation6
(Btu/lb-mole)
12,510
7,290
14,040
19,160
13,230
19,800
9,630
10,610
12,740
9,260
10,410
12,890
11,740
8,630
11,480
11,390
15,300
10,980
13,640
12,410
14,830
14,810
11,090
14,270
15,840
15,640
15.480
Heat
Capacit/
( B
\ lb-mc
30.22
17.90
10.50
15.24
45.90
25.91
19.52
26.07
23.29
14.97
15.63
9.74
17.26
37.4
25.40
30.80
19.84
13.37
40.8
26.89
16.50
30.69
11.54
53.76
39.67
45.2
34.20
19.40
10.49
45.14
28.73
37.58
24.77
44.9
31.85
43.8
30.49
30.32
State
Liquid
Gas
Gas
Gas
Liquid
Gas
Liquid
Gas
Gas
Gas
Gas
Gas
Gas
Liquid
Gas
Liquid
Gas
Gas
Liquid
Gas
Gas
Gas
Gas
Liquid
Gas
Liquid
Gas
Liquid
Gas
Gas
Gas
Liquid
Gas
Liquid
Gas
Liquid
Gas
Gas
"Reprinted with permission from Lange's Handbook of Chemistry (12th edition), Table 9-7. [15]
'Reprinted with permission from Lange's Handbook of Chemistry (12th edition), Table 9-4.[15]
(Measured at boiling point.)
'Reprinted with permission fromLanqe's Handbook of Chemistry (12th edition), Table 9-2 [15]
(Measured at77°F)
8-39
-------
Table 8.8: Antoine Equation Constants for Selected Compounds3
Compound
Acetone
Acetylene
Acrylonitrile
Aniline
Benzene
Benzonitrile
Butane
Chloroethane
Chloroethylene
Chloroform
Chloromethane
Cyanic acid
Cyclobutane
Cyclohexane
Cyclopentane
Cyclopropane
Diethyl ether
Diethylamine
Dimethylamine
Dioxane - 1 ,4
Ethyl benzene
Ethylene oxide
Heptane
Hexane
Methanol
Octane
Pentane
Toluene
Vinyl acetate
o - Xylene
m - Xylene
p - Xylene
Antoine Equation Constants
A
7.117
7.100
7.039
7.320
6.905
6.746
6.809
6.986
6.891
6.493
7.0933
7.569
6.916
6.841
6.887
6.888
6.920
5.801
7.082
7.432
6.975
7.128
6.897
6.876
7.897
6.919
6.853
6.955
7.210
6.999
7.009
6.991
B
1210.595
711.0
1232.53
1731.515
1211.033
1436.72
935.86
1030.01
905.01
929.44
948.58
1251.86
1054.54
1201.53
1124.16
856.01
1064.07
583.30
960.242
1554.68
1424.255
1054.54
1264.90
1171.17
1474.08
1351.99
1064.84
1344.8
1296.13
1474.679
1462.266
1453.430
C
229.66
253.4
222.47
206.049
220.790
181.0
238.73
238.61
239.48
196.03
249.34
243.79
241.37
222.65
231.36
246.50
228.80
144.1
221.67
240.34
213.21
2371.76
216.54
224.41
229.13
209.15
233.01
219.48
226.66
213.69
215.11
215.31
Valid
Temperature
Range (°F)
Liquid
-116to-98
-4 to 248
216 to 365
46 to 217
Liquid
-107 to 66
-69 to 54
-85 to 9
-31 to 142
-103 to 23
-105 to 21
-76 to 54
68 to 178
-40 to 162
-130 to -26
-78 to 68
88 to 142
-98 to 44
68 to 221
79 to 327
-56 to 54
28 to 255
-13 to 198
7 to 149
66 to 306
-58 to 136
43 to 279
72 to 162
90 to 342
82 to 331
81 to 331
Reprinted with permission from Lange's Handbook of Chemistry (12th edition),
Table 10-8.[15]
8-40
-------
Appendix 8B
Documentation for Gasoline Vapor Recovery
System Cost Data
As mentioned in Section 8.4.3, vendor cost data were obtained that related the equipment cost ($)
of packaged gasoline vapor recovery systems to the process flow capacity (gal/min). These data
needed to be transformed, in order to develop Equation 8.34, which relates equipment cost ($) to
system refrigeration capacity (R, tons), as follows:
ECp = 4,9WR + 212,000
To make this transformation, we needed to develop an expression relating flow capacity to
refrigeration capacity. The first step was to determine the inlet partial pressure (/Voc.m) °f the VOC-
gasoline, in this case. As was done in Section 8.3.1, we assumed that the VOC vapor was saturated
and, thus, in equilibrium with the VOC liquid. This, in turn, meant that we could equate the partial
pressure to the vapor pressure. The "model" gasoline had a Reid vapor pressure of 10 and a
molecular weight of 66 Ib/lb-mole, as shown in Section 4.3 of Compilation of Air Pollutant
Emission Factors (FPA publication AP-42, Fourth Edition, September 1985). For this gasoline, the
following Antoine equation constants were used:
A =12.5733
B = 6386.1
C = 613
These constants were obtained by extrapolating available vapor pressure vs. temperature data for
gasoline found in Section 4.3 of AP-42. Upon substituting these constants and an assumed inlet
temperature of 77 °F (25 °C) into the Antoine equation and solving for the inlet partial pressure
(Pvoc.ir) We obtain:
D
log PI7r,^ . = A -
voc, m T + c
in
8-41
-------
=12.5733 - 6386'1
25 + 613
Pmc_„ = 366 mm Hg.
If the system operates at atmospheric pressure (760 mm Hg), this partial pressure would correspond
to a VOC volume fraction in the inlet stream of:
yvnc , = 366rnm = 0.482
-* voc. in 760
The outlet partial pressure (PVOC|0ur) and volume fraction are calculated in a similar way. The
condensation (outlet) temperature used in these calculations is -80°F (-62°C), the typical operating
Temperature for the gasoline vapor recovery units for which the vendor supplied costs.
logP/or = 12.5733 - 6386'1
a voc, out -62 +
PVOC.O* = 9.62 mm Hg.
This corresponds to a volume fraction in the outlet stream (yVOCjOW) of:
9.62mm
voc, out 760 nun
=
Substitution of PVOCout and yvocin into Equation 8.24 yields the condenser removal efficiency (r\):
'(760x0.482) -9.62 = ^^&
0.482(760 - 9.62)
The next step in determining the inlet and outlet VOC hourly molar flow rates (Mmcin and Mvoc>ma,
respectively). As Equation 8.8 shows, Mmcm is a function of ymc>in and the total inlet volumetric flow
rate, Qm, (scfm).
Now, because the gasoline vapor flow rates are typically expressed in gallons/minute, we have to
convert them to scfm. This is done as follows:
1 ft3
Q. = Q (gal/min) x - = 0.134Q scfm
in 3 vy 7,48 gal
Substituting these variables into Equation 8.8, we obtain:
8-42
-------
0.134QCT
2 (0.482)60 = 0.009892 lb-mole/hr
o Q o y
We obtain Mvoc_OMt from Equation 8.9:
Mvoc,out = 0.00989<2g(l - 0.986) = (1.38 x l(T)Qg Ib-moles/hr
And according to Equation 8.10, the amount of gasoline vapor condensed (Afvoc_COB) is the difference
between Mvocin and Mvocout:
= 0.00975(2. Ib-moles/hr
The final step is to calculate the condenser heat load. This load is a function of the inlet, outlet, and
condensate molar flow rates, the inlet and condensation temperatures, the heat capacities of the VOC
and air, and the VOC heat of condensation. The VOC heat capacity and heat of condensation data used
are based on pentane and butane chemical properties, the largest components of gasoline, and were
obtained from CHRIS Hazardous Chemical Data (U.S. Coast Guard, U.S. Department of
Transportation, June 1985).
Heat capacities (Btu/Ib-mole-°F):
CP.VOC = 26.6
Cp,air = 6.95
Heat of condensation of VOC: 9,240 Btu/Ib-mole
Substitution of these data, the molar flow rates, and the temperatures into Equations 8.12,8.15 and 8.16
yields the following enthalpy changes in Btu/hr:
AHcon = 130.8Qg
0.572Q
***„ = 11.6Qg
The condenser heat load (Hload) is the sum of these three enthalpy changes (Equation 8.1 1):
Hload = 143<2,
The refrigeration capacity (R, tons) is computed from Equation 8.22:
H. ,
R = - i^£_ = 0.01190
12,000 g
This last equation relates the refrigeration capacity (tons) to the inlet gasoline vapor flow rate (gal/min).
Solving for Qg, in terms of R, we obtain:
8-43
-------
Qg = 83.9J?
Finally, we substitute this relationship into the equipment cost ($) vs. vapor flow rate (Qg~) correlation,
which-was developed from the vendor cost data:
ECp = 5&.5Qg + 212,000
= 58.5(83.97?) + 212,000
= 4,9107?+ 212,000
Note that this last expression is identical to Equation 8.34.
8-44
-------
References
[1] Erikson, D.G., Organic Chemical Manufacturing Volume 5: Adsorption, Condensation, and
Absorption Devices, U.S. Environmental Protection Agency. Research Triangle Park, North
Carolina, Publication No. EPA 450/3-80-027, December 1980.
[2] Vatavuk, W.M., and R.B. Neveril, "Estimating Costs of Air Pollution Control Systems: Part XVI.
Costs of Refrigeration Systems", Chemical Engineering, May 16, 1983, pp. 95-98.
[3] McCabe, W.L., and J.C. Smith, Unit Operations of Chemical Engineering (Third Edition),
McGraw-Hill Book Company, New York, 1976.
[4] Perry, R.H. and C.H. Chilton, Eds. Chemical Engineers' Handbook (Sixth Edition), McGraw-Hill
Book Company, New York, 1989.
[5] Kern, D.Q., Process Heat Transfer, McGraw-Hill Book Company, New York, 1950.
[6] Smith, J.M., and M.C. VanNess, Introduction to Chemical Engineering Thermodynamics (Third
Edition), McGraw-Hill Book Company, New York, 1975.
[7] Reid, Robert C, John M. Prausnitz, and Bruce E. Poling, Properties of Gases & Liquids (Fourth
Edition), McGraw-Hill Book Company, New York, 1987.
[8] Letter and attachment from Robert V. Sisk. Jr. of Piedmont Engineering, Pineville, North Carolina,
to Wiley Barbour of Radian Corporation, Research Triangle Park, North Carolina, January 28,
1991.
[9] Letter and attachment from Waldrop, R., and V. Sardo of Edwards Engineering Corp., Pompton
Plains, New Jersey, to Wiley Barbour of Radian Corporation, Research Triangle Park, North
Carolina, October 1, 1990.
[10] Price, Brian C., "Know the Range and Limitations of Screw Compressors," Chemical
Engineering Progress, 87(2):50-56.
[11] Letter and attachment from Bob Hansek of ITT Corporation, Atlanta, Georgia to Wiley Barbour
of Radian Corporation, Research Triangle Park, North Carolina, October 10, 1990.
[ 12] Letter and attachment from Avery Cooke of Liquid Handling Equipment, Inc., Charlotte, North
Carolina to Rich Pelt of Radian Corporation, Research Triangle Park, North Carolina,
September 20, 1990.
[13] Letter from Richard Waldrop of Edwards Engineering Corp., Pompton Plains, New Jersey to
William Vatavuk, P.E., Durham, North Carolina, August 29, 1988.
8-45
-------
[14] Vatavuk, W.M., and R.B. Neveril, "Estimating Costs for Air Pollution Control Systems, Part
II: Factors for Estimating Capital and Operating Costs", Chemical Engineering, November 3,
1980, pp. 157-162.
[15] Dean, John A., Ed., Lange's Handbook of Chemistry (Twelfth Edition), McGraw-Hill Book
Company, New York, 1978.
8-46
-------
Chapter 9
GAS ABSORBERS
Wiley Barbour
Roy Oommen
Gunseli Sagun Shareef
Radian Corporation
Research Triangle Park, NC 27709
William M. Vatavuk
Innovative Strategies and Economics Group, OAQPS
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
December 1995
Contents
9.1 Introduction 9-3
9.1.1 System Efficiencies and Performance 9-3
9.2 Process Description 9-4
9.2.1 Absorber System Configuration 9-4
9.2.2 Types of Absorption Equipment 9-4
9.2.3 Packed Tower Internals 9-7
9.2.4 Packed Tower Operation 9-9
9.3 Design Procedures 9-10
9.3.1 Step 1: Determining Gas and Liquid Stream Conditions 9-14
9.3.2 Step 2: Determining Absorption Factor 9-20
9.3.3 Step 3: Determining Column Diameter 9-21
9-1
-------
9.3.4 Step 4: Determining Tower Height and Surface Area 9-25
9.3.5 Step 5: Calculating Column Pressure Drop 9-27
9.3.6 Alternative Design Procedure 9-28
9.4 Estimating Total Capital Investment 9-32
9.4.1 Equipment Costs for Packed Towers 9-32
9.4.2 Installation Costs 9-34
9.5 Estimating Annual Cost 9-34
9.5.1 Direct Annual Costs 9-34
9.5.2 Indirect Annual Costs 9-38
9.5.3 Total Annual Cost 9-39
9.6 Example Problem #1 9-39
9.6.2 Step 1: Determine Gas and Liquid Stream Properties 9-39
9.6.3 Step 2: Calculate Absorption Factor 9-43
9.6.4 Step 3: Estimate Column Diameter 9-44
9.6.5 Step 4: Calculate Column Surface Area 9-46
9.6.6 Step 5: Calculate Pressure Drop 9-47
9.6.7 Equipment Costs 9-47
9.6.8 Total Annual Costs 9-49
9.7 Example Problem #2 9-53
9.8 Acknowledgments 9-53
Appendix 9A 9-54
Appendix 9B 9-56
Appendix 9C 9-61
9C. 1 Overvie of the Approach 9-61
9C.2 Example Problem Calculation 9-62
References 9-64
9-2
-------
9.1 Introduction
Gas absorbers are used extensively in industry for separation and purification of gas streams, as
product recovery devices, and as pollution control devices. This chapter focuses on the application
of absorption for pollution control on gas streams with typical pollutant concentrations ranging from
250 to 10,000 ppmv. Gas absorbers are most widely used to remove water soluble inorganic
contaminants from air streams.[1, 2]
Absorption is a process where one or more soluble components of a gas mixture are dissolved
in a liquid (i.e., a solvent). The absorption process can be categorized as physical or chemical.
Physical absorption occurs when the absorbed compound dissolves in the solvent; chemical
absorption occurs when the absorbed compound and the solvent react. Liquids commonly used as
solvents include water, mineral oils, nonvolatile hydrocarbon oils, and aqueous solutions.[1]
9.1.1 System Efficiencies and Performance
Removal efficiencies for gas absorbers vary for each pollutant-solvent system and with the type of
absorber used. Most absorbers have removal efficiencies in excess of 90 percent, and packed tower
absorbers may achieve efficiencies as high as 99.9 percent for some pollutant-solvent systems.[1,
3]
The suitability of gas absorption as a pollution control method is generally dependent on the
following factors: 1) availability of suitable solvent; 2) required removal efficiency; 3) pollutant
concentration in the inlet vapor; 4) capacity required for handling waste gas; and, 5) recovery value
of the pollutant(s) or the disposal cost of the spent solvent.[4]
Physical absorption depends on properties of the gas stream and solvent, such as density and
viscosity, as well as specific characteristics of the pollutant(s) in the gas and the liquid stream (e.g.,
diffusivity, equilibrium solubility). These properties are temperature dependent, and lower temper-
atures generally favor absorption of gases by the solvent.[l] Absorption is also enhanced by greater
contacting surface, higher liquid-gas ratios, and higher concentrations in the gas stream.[l]
The solvent chosen to remove the pollutant(s) should have a high solubility for the gas, low
vapor pressure, low viscosity, and should be relatively inexpensive. [4] Water is the most common
solvent used to remove inorganic contaminants; it is also used to absorb organic compounds having
relatively high water solubilities. For organic compounds that have low water solubilities, other
solvents such as hydrocarbon oils are used, though only in industries where large volumes of these
oils are available (i.e., petroleum refineries and petrochemical plants).[5]
Pollutant removal may also be enhanced by manipulating the chemistry of the absorbing solution
so that it reacts with the pollutant(s), e.g., caustic solution for acid-gas absorption vs. pure water as
a solvent. Chemical absorption may be limited by the rate of reaction, although the rate limiting
step is typically the physical absorption rate, not the chemical reaction rate.
9-3
-------
9.2 Process Description
Absorption is a mass transfer operation in which one or more soluble components of a gas mixture
are dissolved in a liquid that has low volatility under the process conditions. The pollutant diffuses
from the gas into the liquid when the liquid contains less than the equilibrium concentration of the
gaseous component. The difference between the actual concentration and the equilibrium
concentration provides the driving force for absorption.
A properly designed gas absorber will provide thorough contact between the gas and the solvent
in order to facilitate diffusion of the pollutant(s). It will perform much better than a poorly designed
absorber.[6] The rate of mass transfer between the two phases is largely dependent on the surface
area exposed and the time of contact. Other factors governing the absorption rate, such as the
solubility of the gas in the particular solvent and the degree of the chemical reaction, are
characteristic of the constituents involved and are relatively independent of the equipment used.
9.2.1 Absorber System Configuration
Gas and liquid flow through an absorber may be countercurrent, crosscurrent, or cocurrent. The
most commonly installed designs are countercurrent, in which the waste gas stream enters at the
bottom of the absorber column and exits at the top. Conversely, the solvent stream enters at the top
and exits at the bottom. Countercurrent designs provide the highest theoretical removal efficiency
because gas with the lowest pollutant concentration contacts liquid with the lowest pollutant
concentration. This serves to maximize the average driving force for absorption throughout the
column.[2] Moreover, countercurrent designs usually require lower liquid to gas ratios than cocur-
rent and are more suitable when the pollutant loading is higher. [3, 5]
In a crosscurrent tower, the waste gas flows horizontally across the column while the solvent
flows vertically down the column. As a rule, crosscurrent designs have lower pressure drops and
require lower liquid-to-gas ratios than both cocurrent and countercurrent designs. They are
applicable when gases are highly soluble, since they offer less contact time for absorption. [2, 5]
In cocurrent towers, both the waste gas and solvent enter the column at the top of the tower and
exit at the bottom. Cocurrent designs have lower pressure drops, are not subject to flooding
limitations and are more efficient for fine (i.e., submicron) mist removal. Cocurrent designs are
only efficient where large absorption driving forces are available. Removal efficiency is limited
since the gas-liquid system approaches equilibrium at the bottom of the tower.[2]
9.2.2 Types of Absorption Equipment
Devices that are based on absorption principles include packed towers, plate (or tray) columns,
venturi scrubbers, and spray chambers. This chapter focuses on packed towers, which are the most
commonly used gas absorbers for pollution control. Packed towers are columns filled with packing
materials that provide a large surface area to facilitate contact between the liquid and gas. Packed
9-4
-------
tower absorbers can achieve higher removal efficiencies, handle higher liquid rates, and have
relatively lower water consumption requirements than other types of gas absorbers.[2] However,
packed towers may also have high system pressure drops, high clogging and fouling potential, and
extensive maintenance costs due to the presence of packing materials. Installation, operation, and
wastewater disposal costs may also be higher for packed bed absorbers than for other absorbers.[2]
In addition to pump and fan power requirements and solvent costs, packed towers have operating
costs associated with replacing damaged packing.[2]
Plate, or tray, towers are vertical cylinders in which the liquid and gas are contacted in step-wise
fashion on trays (plates). Liquid enters at the top of the column and flows across each plate and
through a downspout (downcomer) to the plates below. Gas moves upwards through openings in
the plates, bubbles into the liquid, and passes to the plate above. Plate towers are easier to clean and
tend to handle large temperature fluctuations better than packed towers do.[4] However, at high gas
flow rates, plate towers exhibit larger pressure drops and have larger liquid holdups. Plate towers
are generally made of materials such as stainless steel, that can withstand the force of the liquid on
the plates and also provide corrosion protection. Packed columns are preferred to plate towers when
acids and other corrosive materials are involved because tower construction can then be of
fiberglass, polyvinylchloride, or other less costly, corrosive-resistant materials. Packed towers are
also preferred for columns smaller than two feet in diameter and when pressure drop is an important
consideration.!^, 7]
Venturi scrubbers are generally applied for controlling paniculate matter and sulfur dioxide.
They are designed for applications requiring high removal efficiencies of submicron particles,
between 0.5 and 5.0 micrometers in diameter.[4] A venturi scrubber employs a gradually converging
and then diverging section, called the throat, to clean incoming gaseous streams. Liquid is either
introduced to the venturi upstream of the throat or injected directly into the throat where it is
atomized by the gaseous stream. Once the liquid is atomized, it collects particles from the gas and
discharges from the venturi.[l] The high pressure drop through these systems results in high energy
use, and the relatively short gas-liquid contact time restricts their application to highly soluble gases.
Therefore, they are infrequently used for the control of volatile organic compound emissions in
dilute concentration.[2]
Spray towers operate by delivering liquid droplets through a spray distribution system. The
droplets fall through a countercurrent gas stream under the influence of gravity and contact the
pollutant(s) in the gas.[7] Spray towers are simple to operate and maintain, and have relatively low
energy requirements. However, they have the least effective mass transfer capability of the
absorbers discussed and are usually restricted to paniculate removal and control of highly soluble
9-5
-------
Gaa Out
Mi«t Eliminator
Liquid in —•
Liquid Distributor
^ Spray Nonla
Packing
RMtraintr
Random
Packing
Liquid R»-distnbutor
Packing Support
Gaaln
UqutdOut
Figure 9.1: Packed Tower for Gas Absorption
9-6
-------
gases such as sulfur dioxide and ammonia. They also require higher water recirculation rates and
are inefficient at removing very small particles. [2, 5]
9.2.3 Packed Tower Internals
A basic packed tower unit is comprised of a column shell, mist eliminator, liquid distributors,
packing materials, packing support, and may include a packing restrainer. Corrosion resistant alloys
or plastic materials such as polypropylene are required for column internals when highly corrosive
solvents or gases are used. A schematic drawing of a countercurrent packed tower is shown in
Figure 9.1. In this figure, the packing is separated into two sections. This configuration is more
expensive than designs where the packing is not so divided. [5]
The tower shell may be made of steel or plastic, or a combination of these materials depending
on the corrosiveness of the gas and liquid streams, and the process operating conditions. One alloy
that is chemical and temperature resistant or multiple layers of different, less expensive materials
may be used. The shell is sometimes lined with a protective membrane, often made from a
corrosion resistant polymer. For absorption involving acid gases, an interior layer of acid resistant
brick provides additional chemical and temperature resistance.[8]
At high gas velocities, the gas exiting the top of the column may carry off droplets of liquid as
a mist. To prevent this, a mist eliminator in the form of corrugated sheets or a layer of mesh can
be installed at the top of the column to collect the liquid droplets, which coalesce and fall back into
the column.
A liquid distributor is designed to wet the packing bed evenly and initiate uniform contact
between the liquid and vapor. The liquid distributor must spread the liquid uniformly, resist
plugging and fouling, provide free space for gas flow, and allow operating flexibility.[9] Large
towers frequently have a liquid redistributor to collect liquid off the column wall and direct it toward
the center of the column for redistribution and enhanced contact in the lower section of packing.[4]
Liquid redistributors are generally required for every 8 to 20 feet of random packing depth.[5, 10]
Distributors fall into two categories: gravitational types, such as orifice and weir types, and
pressure-drop types, such as spray nozzles and perforated pipes. Spray nozzles are the most
common distributors, but they may produce a fine mist that is easily entrained in the gas flow. They
also may plug, and usually require high feed rates to compensate for poor distribution. Orifice-type
distributors typically consist of flat trays with a number of risers for vapor flow and perforations in
the tray floor for liquid flow. The trays themselves may present a resistance to gas flow. [9]
However, better contact is generally achieved when orifice distributors are used.[3]
Packing materials provide a large wetted surface for the gas stream maximizing the area
available for mass transfer. Packing materials are available in a variety of forms, each having
specific characteristics with respect to surface area, pressure drop, weight, corrosion resistance, and
9-7
-------
Tellerecce
IncaLox Saddle
Bcrl Saddle
Raichig Ring
Figure 9.2: Random Packing Material
cost. Packing life varies depending on the application. In ideal circumstances, packing will last as
long as the tower itself. In adverse environments packing life may be as short as 1 to 5 years due
to corrosion, fouling, and breakage. [11]
Packing materials are categorized as random or structured. Random packings are usually
dumped into an absorption column and allowed to settle. Modern random packings consist of
engineered shapes intended to maximize surface-to-volume ratio and minimize pressure drop. [2]
Examples of different random packings are presented in Figure 9.2. The first random packings
specifically designed for absorption towers were made of ceramic. The use of ceramic has declined
because of their brittleness, and the current markets are dominated by metal and plastic. Metal
packings cannot be used for highly corrosive pollutants, such as acid gas, and plastic packings are
not suitable for high temperature applications. Both plastic and metal packings are generally limited
to an unsupported depth of 20 to 25. At higher depths the weight may deform the packing.[10]
Structured packing may be random packings connected in an orderly arrangement, interlocking
grids, or knitted or woven wire screen shaped into cylinders or gauze like arrangements. They
usually have smaller pressure drops and are able to handle greater solvent flow rates than random
packings. [4] However, structured packings are more costly to install and may not be practical for
smaller columns. Most structured packings are made from metal or plastic.
9-8
-------
In order to ensure that the waste gas is well distributed, an open space between the bottom of
the tower and the packing is necessary. Support plates hold the packing above the open space. The
support plates must have enough strength to carry the weight of the packing, and enough free area
to allow solvent and gas to flow with minimum restrictions.[4]
High gas velocities can fluidize packing on top of a bed. The packing could then be carried into
the distributor, become unlevel, or be damaged.[9] A packing restrainer may be installed at the
top of the packed bed to contain the packing. The packing restrainer may be secured to the wall so
that column upsets will not dislocate it, or a "floating" unattached weighted plate may be placed on
top of the packing so that it can settle with the bed. The latter is often used for fragile ceramic
packing.
9.2.4 Packed Tower Operation
As discussed in Section 9.2.1, the most common packed tower designs are countercurrent. As the
waste gas flows up the packed column it will experience a drop in its pressure as it meets resistance
from the packing materials and the solvent flowing down. Pressure drop in a column is a function
of the gas and liquid flow rates and properties of the packing elements, such as surface area and free
volume in the tower. A high pressure drop results in high fan power to drive the gas through the
packed tower, and consequently high costs. The pressure drop in a packed tower generally ranges
from 0.5 to 1.0 in. H2O/ft of packing.[7]
For each column, there are upper and lower limits to solvent and vapor flow rates that ensure
satisfactory performance. The gas flow rate may become so high that the drag on the solvent is
sufficient to keep the solvent from flowing freely down the column. Solvent begins toaccumulate
and blocks the entire cross section for flow, which increases the pressure drop and present the
packing from mixing the gas and solvent effectively. When all the free volume in the packing is
filled with liquid and the liquid is carried back up the column, the absorber is considered to be
flooded.[4] Most packed towers operate at 60 to 70 percent of the gas flooding velocity, as it is not
practical to operate a tower in a flooded condition.[7] A minimum liquid flow rate is also required
to wet the packing material sufficiently for effective mass transfer to occur between the gas and
liquid.[7]
The waste gas inlet temperature is another important scrubbing parameter. In general, the higher
the gas temperature, the lower the absorption rate, and vice-versa. Excessively high gas
temperatures also can lead to significant solvent loss through evaporation. Consequently, precoolers
(e.g., spray chambers) may be needed to reduce the air temperature to acceptable levels.[6]
For operations that are based on chemical reaction with absorption, an additional concern is the
rate of reaction between the solvent and pollutant(s). Most gas absorption chemical reactions are
relatively fast and the rate limiting step is the physical absorption of the pollutants into the solvent.
However, for solvent-pollutant systems where the chemical reaction is the limiting step, the rates
of reaction would need to be analyzed kinetically.
9-9
-------
Heat may be generated as a result of exothermal chemical reactions. Heat may also be generated
when large amounts of solute are absorbed into the liquid phase, due to the heat of solution. The
resulting change in temperature along the height of the absorber column may damage equipment and
reduce absorption efficiency. This problem can be avoided by adding cooling coils to the
column. [7] However, in those systems where water is the solvent, adiabatic saturation of the gas
occurs during absorption due to solvent evaporation. This causes a substantial cooling of the
absorber that offsets the heat generated by chemical reactions. Thus, cooling coils are rarely
required with those systems. [5] In any event, packed towers may be designed assuming that
isothermal conditions exist throughout the column. [7]
The effluent from the column may be recycled into the system and used again. This is usually
the case if the solvent is costly, i.e., hydrocarbon oils, caustic solution. Initially, the recycle stream
may go to a waste treatment system to remove the pollutants or the reaction product. Make-up
solvent may then be added before the liquid stream reenters the column. Recirculation of the
solvent requires a pump, solvent recovery system, solvent holding and mixing tanks, and any
associated piping and instrumentation.
9.3 Design Procedures
The design of packed tower absorbers for controlling gas streams containing a mixture of pollutants
and air depends on knowledge of the following parameters:
1. Waste gas flow rate;
2. Waste gas composition and concentration of the pollutants in the gas stream;
3. Required removal efficiency;
4. Equilibrium relationship between the pollutants and solvent; and
5. Properties of the pollutant(s), waste gas, and solvent:
• Diffusivity,
• Viscosity,
• Density, and
• Molecular weight.
9-10
-------
The primary objectives of the design procedures are to determine column surface area and pressure
drop through the column. In order to determine these parameters, the following steps must be
performed:
Step 1: Determine the gas and liquid stream conditions entering and exiting the column.
Step 2: Determine the absorption factor (AF).
Step 3: Determine the diameter of the column (D).
Step 4: Determine the tower height (Htower) and surface area (5).
Step 5: Determine the packed column pressure drop (AP).
To simplify the sizing procedures, a number of assumptions have been made. For example, the
waste gas is assumed to comprise a two-component waste gas mixture (pollutant/air), where the
pollutant consists of a single compound present in dilute quantities. The waste gas is assumed to
behave as an ideal gas and the solvent is assumed to behave as an ideal solution. Heat effects
associated with absorption are considered to be minimal for the pollutant concentrations
encountered. The procedures also assume that, in chemical absorption, the process is not reaction
rate limited, i.e., the reaction of the pollutant with the solvent is considered fast compared to the rate
of absorption of the pollutant into the solvent.
The design procedures presented here are complicated, and careful attention to units is required.
Table 9.1 is a list of all design variables referred to in this chapter, along with the appropriate units.
A key is provided to differentiate primary data from calculated data.
9-11
-------
Table 9.1: List of Design Variables
Variable
> Surface to volume ratio of packing
Cross-sectional area of absorber
Abscissa value from plot of
generalized press drop correlation
Absorption factor
Diameter of absorber
*• Diffusivity of pollutant in gas
> Diffusivity of pollutant in liquid
*• Flooding factor
* Packing factor
*• Waste gas flow rate entering
absorber
Waste gas flow rate exiting absorber
Waste gas molar flow rate entering
absorber
Molar flow rate of pollutant free gas
Waste gas superficial flow rate
entering absorber
Height of gas transfer unit
Height of liquid transfer unit
Height of overall transfer unit
Height of packing
Height of absorber
Pressure drop constants
Liquid rate entering absorber
Liquid rate exiting absorber
Liquid molar flow rate entering
Symbol
a
A
ABSCISSA
AF
D
DG
DL
f
Fp
G,
GO
Gmo!
Gs
Gtfr,,
HG
HL
Ht«
"pack
TJ
Blower
kn, k,, k2, k}, k4,
L,
L0
Lmoll
Units
fWft3
ft2
—
—
feet
ft2/hr
ft2/hr
—
—
acfrn
acfrn
Ib-moles/h
Ib-moles/h
lb/sec-ft2
feet
feet
feet
feet
feet
—
gpm
gpm
Ib-moles/h
absorber
Molar flow rate of pollutant free
solvent
Liquid superficial flow rate entering
absorber
Slope of equilibrium line
Molecular weight of gas stream
Molecular weight of the liquid stream
m
MWG
MW,
Ib-moles/h
lb/hr-ft2
Ib/lb-mole
Ib/lb-mole
9-12
-------
Variable
Symbol
Units
Minimum wetting rate
Number of overall transfer units
Ordinate value from plot of
generalized pressure drop
correlation
Surface area of absorber
Temperature of solvent
Mole fraction of pollutant entering
absorber in liquid
Mole fraction of pollutant exiting
absorber in liquid
Pollutant concentration entering
absorber in liquid
Maximum pollutant concentration in
liquid phase in equilibrium with
pollutant entering column in gas
phase
Pollutant concentration exiting
absorber in liquid
Mole fraction of pollutant entering
absorber in waste gas
Mole fraction of pollutant in gas
phase in equilibrium with mole
fraction of pollutant entering in the
liquid phase
Mole fraction of pollutant exiting
scrubber in waste gas
Mole fraction of pollutant in gas
phase in equilibrium with mole
fraction of pollutant exiting in the
liquid phase
Pollutant concentration entering
scrubber in waste gas
Pollutant concentration entering
scrubber in equilibrium with
concentration in liquid phase
Pollutant concentration exiting
scrubber in waste gas
MM?
ORDINATE
S
T
x.
ft2/hr
X*
y,
Jo
*
y o
r,
ft2
K
Ib-mole of pollutant
Ib-mole of total liquid
Ib-mole of pollutant
Ib-mole of total liquid
Ib-moles pollutant
Ib-moles pollutant free solvent
Ib-moles pollutant
Ib-moles pollutant free solvent
Ib-moles pollutant
Ib-moles pollutant free solvent
Ib-moles pollutant
Ib-mole of total gas
Ib-moles pollutant
Ib-mole of total gas
Ib-moles pollutant
Ib-mole of total gas
Ib-moles pollutant
Ib-mole of total gas
Ib-moles pollutant
Ib-moles pollutant free gas
Ib-moles pollutant
Ib-moles pollutant free gas
Ib-moles pollutant
Ib-moles pollutant free gas
9-13
-------
Variable
Symbol
Units
Pollutant removal efficiency
Pollutant concentration exiting
scrubber hi equilibrium with
concentration in liquid phase
Density of waste gas stream
Density of liquid stream
Viscosity of waste gas
Viscosity of solvent
Ratio of solvent density to water
density
Pressure drop
Packing factors
PG
PL
W
AP
a, j, ?, 0, b, /?, c
Ib-moles pollutant
Ib-mole of total gas
lb/ft3
lb/ft3
Ib/ft-hr
Ib/ft-hr
inches H2O/feet of packing
-Denotes required input data.
9.3.1 Step 1: Determining Gas and Liquid Stream Conditions
Gas absorbers are designed based on the ratio of liquid to gas entering the column (L/G,), slope of the
equilibrium curve (m), and the desired removal efficiency (rf). These factors are calculated from the inlet and
outlet gas and liquid stream variables:
• Waste gas flow rate, in actual cubic feet per minute (acfm), entering and exiting column (G, and G0,
respectively);
• Pollutant concentration (Ib-moles pollutant per Ib-mole of pollutant free gas) entering and exiting the
column in the waste gas (F, and Y0, respectively);
• Solvent flow rate, in gallons per minute (gpm), entering and exiting the column (L, and £ ,
respectively); and
• Pollutant concentration (Ib-moles pollutant per Ib-mole of pollutant free solvent) entering and exiting
the column in the solvent (X, and X0, respectively).
This design approach assumes that the inlet gas stream variables are known, and that a specific pollutant
removal efficiency has been chosen as the design basis; i.e., the variables G,, 7,, and T| are known. For dilute
concentrations typically encountered in pollution control applications and negligible changes in moisture
content, G,, is assumed equal to G0. If a once-through process is used, or if the spent solvent is regenerated
by an air stripping process before it is recycled, the value of X, will approach zero. The following procedures
must be followed to calculate the remaining stream variables Yg, L, (and L0), and X0. A schematic diagram of
a packed tower with inlet and outlet flow and concentration variables labeled is presented in Figure 9.3.
The variable Y0, may be calculated from T) using the following equation:
9-14
-------
i -
100;
o.i;
9-15
-------
G°
jfmol,o
yS
°
h
hmol,
I- o
G i
'mol,
mol, o
Figure 9.3: Schematic Diagram of Countercurrent Packed Tower Operation
9-16
-------
The liquid flow rate entering the absorber, L, (gpm), is then calculated using a graphical method. Figure
9.4 presents an example of an equilibrium curve and operating line. The equilibrium curve indicates the
relationship between the concentration of pollutant in the waste gas and the concentration of pollutant in the
solvent at a specified temperature. The operating line indicates the relation between the concentration of the
pollutant in the gas and solvent at any location in the gas absorber column. The vertical distance between the
operating line and equilibrium curve indicates the driving force for diffusion of the pollutant between the gas
and liquid phases. The minimum amount of liquid which can be used to absorb the pollutant in the gas stream
corresponds to an operating line drawn from the outlet concentration in the gas stream (Y0) and the inlet
concentration in the solvent stream (3Q to the point on the equilibrium curve corresponding to the entering
pollutant concentration in the gas stream (7;). At the intersection point on the equilibrium curve, the
diffusional driving forces are zero, the required time of contact for the concentration change is infinite, and
an infinitely tan tower results.
The slope of the operating line intersecting the equilibrium curve is equal to the minimum L/G ratio on
a moles of pollutant-free solvent (L,) per moles of pollutant-free gas basis Gs. in other words, the values Ls and
Gs do not include the moles of pollutant in the liquid and gas streams. The values of L, and G, are constant
through the column if a negligible amount of moisture is transferred from the liquid to the gas phase. The
slope may be calculated from the following equation:
Y. - Y
1 0_
X* - X.
(9.2)
where X*0 would be the maximum concentration of the pollutant in the liquid phase if it were allowed to come
to equilibrium with the pollutant entering the column in the gas phase, Yr The value of X*,, is taken from the
equilibrium curve. Because the minimum L/GS, ratio is an unrealistic value, it must be multiplied by an
adjustment factor, commonly between 1.2 and 1.5, to calculate the actual Z/Gratio:[7]
act
x (adjustment factor]
(9.3)
The variable Gs may be calculated using the equation:
60pGG.
MW (l + y-
(9.4)
where 60 is the conversion factor from minutes to hours, MWG, is the molecular weight of the gas stream
(Ib/lb-mole), and pG is the density of the gas stream (lb/ft3). For pollutant concentrations typically
encountered, the molecular weight and density of the waste gas stream are assumed to be equal to that of
ambient air.
9-17
-------
L =
s
L
—i x
G
s act
s
(9.5)
The total molar flow rates of the gas and liquid entering the absorber (Gmol, and Lmoll are calculated using the
following equations:
G . . = G (i +
mol, i s
L , = L ( 1 + X . ) (Q -i\
mol, i s ^ i1 (?• I )
The volume flow rate of the solvent, L,, may then be calculated by using the following relationship:
7.48 L . MW
60
T
L (9.8)
where 60 is the conversion factor from minutes to hours, MW,, is the molecular weight of the liquid stream
(Ib/lb-mole), p£ is the density of the liquid stream (lb/ft3), and 7.48 is the factor used to convert cubic feet to
gallons. If the volume change in the liquid stream entering and exiting the absorber is assumed to be negligible,
then L, = L0.
Gas absorber vendors have provided a range for the L/G, ratio for acid gas control from 2 to 20 gpm of
solvent per 1000 cfm of waste gas.[12] Even for pollutants that are highly soluble in a solvent (i.e., HC1 in
water), the adjusted L/G, ratio calculated using Equations 9.2 to 9.8 would be much lower than this range,
because these equations do not consider the flow rate of the solvent required to wet the packing.
9-18
-------
O
0
03
o
"o
«
o
Moles of Pollutant/Mole of Solvent
Figure 9.4: Minimum and Actual Liquid-to-Gas Ratios
9-19
-------
Finally, the actual operating line may be represented by a material balance equation over the gas absorber: [4]
X L + Y.G = X L + Y G
IS IS OS OS
(9.9)
Equation 9.9 may then be solved for X0:
X =
o
Y. - Y
+ X.
(9.10)
9.3.2 Step 2: Determining Absorption Factor
The absorption factor (AF) value is frequently used to describe the relationship between the equilibrium line and
the liquid-to-gas ratio. For many pollutant-solvent systems, the most economical value for AF ranges around
1.5 to 2.0.[7] The following equation may be used to calculate AF:[4, 7]
mol, i
m G
(9.11)
mol, i
where m is the slope of the equilibrium line on a mole fraction basis. The value of m may be obtained from
available literature on vaporAiquid equilibrium data for specific systems. Since the equilibrium curve is typically
linear in the concentration ranges usually encountered in air pollution control, the slope, m would be constant
(or nearly so) for all applicable inlet and outlet liquid and gas streams. The slope may be calculated from mole
fraction values using the following equation:[4]
m ~
y.
x - x.
O 1
(9.12)
where y* andy0* are the mole fractions of the pollutant in the vapor phase in equilibrium with the mole fractions
of the pollutant entering and exiting the absorber in the liquid, x, and %, respectively. The slope of the
equilibrium line in Figure 9.4 is expressed in terms of concentration values X,, X0, Y', and Y0'. These values may
be converted to x,, x0, y', andy()* using the equations:
9-20
-------
x. =
(9.13)
X
X =
0 1 + X
(9.14)
y;
i + y;
(9-15)
y0 =
1 +
(9.16)
where the units for each of these variables are listed in Table 9.1.
The absorption factor will be used to calculate the theoretical number of transfer units and the theoretical
height of a transfer unit. First, however, the column diameter needs to be determined.
9.3.3 Step 3: Determining Column Diameter
Once stream conditions have been determined, the diameter of the column may be estimated. The design
presented in this section is based on selecting a fraction of the gas flow rate at flooding conditions. Alternatively,
the column may be designed for a specific pressure drop (see Section 9.3.6.). Eckert's modification to the
generalized correlation for randomly packed towers based on flooding considerations is used to obtain the
superficial gas flow rate entering the absorber, Gsfri (lb/sec-ft2), or the gas flow rate per crossectional area based
on the Lmol/Gmol, ratio calculated in Step 2.[10] The cross-sectional area (A) of the column and the column
diameter (D) can then be determined from Gsfr,. Figure 9.5 presents the relationship between Qr, and the
Lmol/Gmol, ratio at the tower flood point. The abscissa value (X axis) in the graph is expressed as:[10]
ABSCISSA =
L
mol, i
mol,i
MW
Pr
(9.17)
9-21
-------
The ordinate value (Y axis) in the graph is expressed as:[10]
ORDINATE =
2.42
0.2
(9.18)
9-22
-------
Figure 9.5: Eckert's Modification to the Generalized Correlation at Flooding Rate[10]
9-23
-------
where Fp is a packing factor, gc is the gravitational constant (32.2), HL is the viscosity of the solvent (Ib/ft-hr),
2.42 is the factor used to convert Ib/ft-hr to centipoise, and *? is the ratio of the density of the scrubbing liquid
to water. The value of Fp may be obtained from packing vendors (see Appendix 9B, Table 9.8).
After calculating the ABSCISSA value, a corresponding ORDINATE value may determined from the
flooding curve. The ORDINATE may also be calculated using the following equation:[10]
DRDINATE = 10 F '1-668~1-085 (1°9 ABSCISSA) -0.297 dog ABSCISSA)2i
(9.19)
Equation 9.18 may then be rearranged to solve for Gsfr,:
sfr.i
P2PGg COORDINATE)
p \p
p
VL
1 2.42
0.2
(9.20)
The cross-sectional area of the tower (ft2) is calculated as:
A =
. .
rnol, i
3600 G . f
sfr, i
(9.21)
where/is the flooding factor and 3600 is the conversion factor from hours to seconds. To prevent flooding, the
column is operated at a fraction of Gsfri. The value of/typically ranges from 0.60 to 0.75.[7]
The diameter of the column (ft) can be calculated from the cross-sectional area, by:
D =
\
n
(9.22)
If a substantial change occurs between inlet and outlet volumes (i.e., moisture is transferred from the liquid phase
to the gas phase), the diameter of the column will need to be calculated at the top and bottom of the column. The
larger of the two values is then chosen as a conservative number. As a rule of thumb, the diameter of the column
should be at least 15 times the size of the packing used in the column. If this is not the case, the column diameter
should be recalculated using a smaller diameter packing. [10]
The superficial liquid flow rate entering the absorber, Lsfr, (lb/hr-ft2 based on the cross-sectional area
determined in Equation 9.21 is calculated from the equation:
9-24
-------
L
sfr, i
L . . MWT
molt i L
A
(9.23)
For the absorber to operate properly, the liquid flow rate entering the column must be high enough to
effectively wet the packing so mass transfer between the gas and liquid can occur. The minimum value ofLsfri
that is required to wet the packing effectively can be calculated using the equation:[7, 13]
(L , } = MWRpTa
\ s^'Vmin L
(9.24)
where MWR is defined as the minimum wetting rate (ft2/hr), and a is the surface area to volume ratio of the
packing (ft2/ft3). An MWR value of 0.85 ff/hr is recommended for ring packings larger than 3 inches and for
structured grid packings. For other packings, an MWR of 1.3 ft2/hr is recommended.[7,13] Appendix 9B, Table
9.8 contains values of a for common packing materials.
If L^, (the value calculated in Equation 9.23) is smaller than (Lsfr)mm (the value calculated in Equation 9.24),
there is insufficient liquid flow to wet the packing using the current design parameters. The value of Gsfri, and
A then will need to be recalculated. See Appendix 9C for details.
9.3.4 Step 4: Determining Tower Height and Surface Area
Tower height is primarily a function of packing depth. The required depth of packing (H^) is determined from
the theoretical number of overall transfer units (Ntu) needed to achieve a specific removal efficiency, and the
height of the overall transfer unit (Hlu):[4]
H . = N H
pack tu tu
(9.25)
The number of overall transfer units may be estimated graphically by stepping off stages on the equilibrium-
operating line graph from inlet conditions to outlet conditions, or by the following equation:[4]
AT =
tu
In
y - mx .
J i i
y - mx .
r
i
AF
1
AF
l _ 1
AF
(9.26)
where hi is the natural logarithm of the quantity indicated. The equation is based on several assumptions: 1)
Henry's law applies for a dilute gas mixture; 2) the equilibrium curve is linear from x, to x0; and 3) the pollutant
concentration in the solvent is dilute enough such that the operating line can be considered a straight line.[4]
9-25
-------
If *, - 0 (i.e., a negligible amount of pollutant enters the absorber in the liquid stream) and l/AF - 0 (i.e.,
the slope of the equilibrium line is very small and/or the Lmo/Gmo/ ratio is very large), Equation 9.26 simplifies
to:
1
(9.27)
There are several methods that may be used to calculate the height of the overall transfer unit, all based on
empirically determined packing constants. One commonly used method involves determining the overall gas and
liquid mass transfer coefficients (K& K,). A major difficulty in using this approach is that values for KG and KL
are frequently unavailable for the specific pollutant-solvent systems of interest. The reader is referred to the
book Random Packing and Packed Tower Design Applications in the reference section for further details
regarding this method. [14]
For this chapter, the method used to calculate the height of the overall transfer unit is based on estimating
the height of the gas and liquid film transfer units, H, and Ha, respectively: [4]
H = H +
tu G
—
AF
(9.28)
The following correlations may be used to estimate values for H, and //G:[13]
\
(9.29)
L
sfr, i
(9.30)
The quantity ju/pD is the Schmidt number and the variables a, P, ?, $ and b are packing constants specific
to each packing type. Typical values for these constants are listed in Appendix 9B, Tables 9.9 and 9.10. The
advantage to using this estimation method is that the packing constants may be applied to any pollutant-solvent
system. One packing vendor offers the following modifications to Equations 9.29 and 9.30 for their specific
packing:[15]
a
\
(9.31)
9-26
-------
HL =
sfri,
Vr
\
-4.255
286]
(9.32)
where T is the temperature of the solvent in Kelvin.
After solving for Hpack using Equation 9.25, the total height of the column may be calculated from the
following correlation: [ 16]
= 1.40 H + 1.02D + 2.81
tower pack
(9.33)
Equation 9.33 was developed from information reported by gas absorber vendors, and is applicable for column
diameters from 2 to 12 feet and packing depths from 4 to 12 feet. The surface area (S) of the gas absorber can
be calculated using the equation:[16]
S = uD
H
tower
(9.34)
Equation 9.34 assumes the ends of the absorber are flat and circular.
9.3.5 Step 5: Calculating Column Pressure Drop
Pressure drop in a gas absorber is a function of Gs/r, and properties of the packing used. The pressure drop in
packed columns generally ranges from 0.5 to 1 inch of H2O per foot of packing. The absorber may be designed
for a specific pressure drop or pressure drop may be estimated using Leva's correlation:[7, 10]
AP = clO
J L
sfr, i
3,600
(fG
sfr, i'
(9.35)
The packing constants c and/ are found in Appendix 9B, Table 9.11, and 3600 is the conversion factor from
seconds to hours. The equation was originally developed for air-water systems. For other liquids, Lsfri is
multiplied by the ratio of the density of water to the density of the liquid.
9-27
-------
9.3.6 Alternative Design Procedure
The diameter of a column can be designed for a specific pressure drop, rather than being determined based on
a fraction of the flooding rate. Figure 9.6 presents a set of generalized correlations at various pressure drop
design values. The abscissa value of the graph is similar to Equation 9. 1 7:[1 0]
ABSCISSA=
rnol,
mol, i
MW
MW
L ~ P
(9.36)
The ordinate value is expressed as:[10]
ORDINATE =
10.1
2.42,
(9.37)
-------
sfr, i
N
PG) PGgc(ORDINATE)
2.42
0.1
(9.39)
Equation 9.37 can be solved for Gsfr ,:The remaining calculations to estimate the column diameter and Lsfr, are
the same as presented in Section 9.3.3, except the flooding factor (/) is not used in the equations. The flooding
factor is not required because an allowable pressure drop that will not cause flooding is chosen to calculate the
diameter rather than designing the diameter at flooding conditions and then taking a fraction of that value.
9-29
-------
PL - PG'
Figure 9.6: Generalized Pressure Drop Correlations[10]
9-30
-------
200.000
160.000
160.000
140.000
Q
2
120.000
100.000
(3 80.000
5 60.000
O"
111
40.000
20.000
200
400
soo aoo 1.000
Surface Area of Tow* (tt;)
1.200
1.400
Figure 9.7: Packed Tower Equipment Cost[16]
9-31
-------
9.4 Estimating Total Capital Investment
This section presents the procedures and data necessary for estimating capital costs for vertical packed bed gas
absorbers using countercurrent flow to remove gaseous pollutants from waste gas streams. Equipment costs
for packed bed absorbers are presented in Section 9.4.1, with installation costs presented in Section 9.4.2.
Total capital investment, TCI, includes equipment cost, EC, for the entire gas absorber unit, taxes, freight
charges, instrumentation, and direct and indirect installation costs. All costs are presented in third quarter 1991
dollars*. The costs presented are study estimates with an expected accuracy of ± 30 percent. It must be kept
in mind that even for a given application, design and manufacturing procedures vary from vendor to vendor, so
costs vary. All costs are for new plant installations; no retrofit cost considerations are included.
9.4.1 Equipment Costs for Packed Towers
Gas absorber vendors were asked to supply cost estimates for a range of tower dimensions (/. e., height, diameter)
to account for the varying needs of different applications. The equipment for which they were asked to provide
costs consisted of a packed tower absorber made of fiberglass reinforced plastic (FRP). and to include the
following equipment components:
• absorption column shell;
• gas inlet and outlet ports;
• liquid inlet port and outlet port/drain;
• liquid distributor and redistributor;
• two packing support plates;
• mist eliminator;
• internal piping;
• sump space; and
• platforms and ladders.
The cost data the vendors supplied were first adjusted to put them on a common basis, and then were
regressed against the absorber surface area (S). The equation shown below is a multivariant regression of cost
data provided by six vendors. [16, 12]
Total Tower Cost($) = 115 5 (9.40)
*For information on escalating these prices to more current dollars, refer to the EPA report Escalation Indexes for Air Pollution Control Cosf
and updates thereto, all of which are installed on the OAQPS Technology Transfer Network (CTC Bulletin Board).
9-32
-------
Table 9.2: Random Packing Costs'
Nominal
Diameter
(inches)
1
1
1
2
2
3.5
3.5
Construction
Material
304 stainless steel
ceramic
polypropylene
ceramic
polypropylene
304 stainless steel
polypropylene
Packing Type
Pall rings, Raschig rings, Ballast rings
Rasching rings, Berl saddles
Tri-pack®, Pall rings, Ballast rings,
Flexisaddles
Berl saddles, Raschig rings
Tri-Pack®, Lanpac®, Flexiring,
Flexisaddle, Tellerette®
Ballast rings
Tri-pack®, Lanpac®,Ballast rings
Packing cost (S / ft3
< 100 ft3 > 100 ft3
70-109
33-44
14-37
13-32
3-20
30
6-14
65-99
26-36
12-34
10-30
5-19
27
6-12
"Provided by packing vendors. [17]
®Denotes registered trademark.
where S is the surface area of the absorber, in ft2. Figure 9.7 depicts a plot of Equation 9.40. This equation is
applicable for towers with surface areas from 69 to 1507 ft2 constructed of FRP. Costs for towers made of
materials other than FRP may be estimated using the following equation:
TTCM = CF x TTC (9.41)
where TTCM is the total cost of the tower using other materials, and TTC is the total tower cost as estimated
using Equation 9.40. The variable CF is a cost factor to convert the cost of an FRP gas absorber to an absorber
fabricated from another material. Ranges of cost factors provided by vendors are listed for the following
materials of construction: [12]
304 Stainless steel = 1.10-1.75
Polypropylene = 0.80-1.10
Polyvinyl chloride = 0.50 - 0.90
Auxiliary costs encompass the cost of all necessary equipment not included in the absorption column unit.
Auxiliary equipment includes packing material, instruments and controls, pumps, and fans. Cost ranges for
various types of random packings are presented in Table 9.2. The cost of structured packings varies over a much
wider range. Structured packings made of stainless steel range from $45/ft3 to $405/ft\ and those made of
polypropylene range from $65/ft3 to $350/ft3.[l 7]
9-33
-------
wider range. Structured packings made of stainless steel range from $45/ft3 to $405/fr\ and those made of
polypropylene range from $65/ft3 to $350/ft3.[17]
Similarly, the cost of instruments and controls varies widely depending on the complexity required. Gas
absorber vendors have provided estimates ranging from $1,000 to $10,000 per column. A factor of 10 percent
of the EC will be used to estimate this cost in this chapter, (see eq. 9.42, below.) Design and cost correlations
for fans and pumps will be presented in a chapter on auxiliary equipment elsewhere in this manual. However,
cost data for auxiliaries are available from the literature (see reference [18], for example).
The total equipment cost (EC) is the sum of the component equipment costs, which includes tower cost and
the auxiliary equipment cost.
1C = TTC + Packing Cost + Auxiliary Equipment (9.42)
The purchased equipment cost (PEC) includes the cost of the absorber with packing and its auxiliaries (EC),
instrumentation (0.10 EC), sales tax (0.03 EC), and freight (0.05 EC). The PEC is calculated from the following
factors, presented in Chapter 2 of this manual and confirmed from the gas absorber vendor survey conducted
during this study:[12, 19],
PEC - (1 + 0.10 + 0.03 + 0.05) EC = 1.18 EC (9.43)
9.4.2 Installation Costs
The total capital investment, TCI, is obtained by multiplying the purchased equipment cost, PEC, by the total
installation factor:
TCI - 2.20 PEC (9.44)
The factors which are included in the total installation factor are also listed in Table 9.3.[19] The factors
presented in Table 9.3 were confirmed from the gas absorber vendor survey.
9.5 Estimating Annual Cost
The total annual cost (TAC) is the sum of the direct and indirect annual costs.
9.5.1 Direct Annual Costs
Direct annual costs (DC) are those expenditures related to operating the equipment, such as labor and materials.
The suggested factors for each of these costs are shown in Table 9.4. These factors were taken from Chapter
2 of this manual and were confirmed from the gas absorber vendor survey. The annual cost for each item is
calculated by multiplying the number of units used annually (i.e., hours, pounds, gallons, kWh) by the
associated unit cost.
9-34
-------
Table 9.3: Capital Cost Factors for Gas Absorbers[19]
Cost Item Fact
or
Direct Costs
Purchased equipment costs
Absorber-f packing+auxiliary equipment", EC As estimated, A
Instrumentation* 0.10 A
Sales taxes 0.03 A
Freight 0.05 A
Purchased equipment cost, PEC B = 1.18 A
Direct installation costs
Foundations & supports
Handling & erection
Electrical
Piping
Insulation
Painting
Direct installation costs
Site preparation As required, SP
Buildings As required. Bldg.
Total Direct Costs, DC 1.85 B - SP - Bldg.
Indirect Costs (installation)
Engineering 0.10 B
Construction and field expenses 0.10 B
Contractor fees 0.10 B
Start-up 0.01 B
Performance test 0.01 B
Contingencies 0.03 B
Total Indirect Costs, 1C 0.35 B
Total Capital Investment = DC + 1C 2.20 B -r SP - Bldg.
"Includes the initial quantity of packing, as well as items normallv not in-
cluded with the unit supplied by vendors, such as ductwork, fan, piping, »»tc.
Instrumentation costs cover pH monitor and liquid level indicator in sump.
9-35
-------
Table 9.4: Suggested Annual Cost Factors for Gas Absorber Systems
Cost Item
Factor
Direct Annual Costs. DC
Operating labor3
Operator
Supervisor
Operating materialsb
Solvent
Chemicals
Wastewater disposal
Maintenance8
Labor
Material
Electricity
Fan
Pump
Indirect Annual Costs. 1C
Overhead
Administrative charges
Property' tax
Insurance
Capital recoveryc
Total Annual Cost
1/2 hour per shift
15% of operator
Application specific
(throughput/yr) x (waste fraction)
Based on annual consumption
(throughput/yr) x (waste fraction)
1/2 hour per shift
100% of maintenance labor
All electricity equal to:
(consumption rate) x
(hours/yr) x (unit cost)
60% of total labor and material costs
2% of Total Capital Investment
1% of Total Capital Investment
1% of Total Capital Investment
0.1098 x Total Capital Inventment
DC + IC
"These factors were confirmed by vendor contacts.
blf system does not use chemicals (e.g., caustic), this quantity is equal to annual solvent consumption.
'Assuming a 15-year life at 7%. See Chapter 2.
9-36
-------
Operating labor is estimated at '/i-hour per 8-hour shift. The supervisory labor cost is estimated at 15
percent of the operating labor cost. Maintenance labor is estimated at 1/2-hour per 8-hour shift. Maintenance
materials costs are assumed to equal maintenance labor costs.
Solvent costs are dependent on the total liquid throughput, the type of solvent required, and the fraction of
throughput wasted (often referred to as blow-down). Typically, the fraction of solvent wasted varies from 0.1
percent to 10 percent of tire total solvent throughput.[12] For acid gas systems, the amount of solvent wasted
is determined by the solids content, with bleed off occurring when solids content reaches 10 to 15 percent to
prevent salt carry-over.[12]
The total annual cost of solvent (C,) is given by:
Cs = ii
60min
hr
annual
operating
hours ,
solvent
unit cost,
(9.45)
where WFis the waste (make-up) fraction, and the solvent unit cost is expressed in terms of $/gal.
The cost of chemical replacement (Cc) is based on the annual consumption of the chemical and can be
calculated by:
S~*l
Ibs chemical used
hr
annual
operating
hours
chemical
\unit cost
(9.46)
where the chemical unit cost is in terms of $/lb.
Solvent disposal (€„,,) costs vary depending on geographic location, type of waste disposed of, and
availability of on-site treatment. Solvent disposal costs are calculated by:
C - L WF
WW 1
60 min
hr
annual
operating
hours
solvent
disposal cost/
(9.47)
where the solvent disposal costs are in terms of $/gaI of waste solvent.
where costs are in terms of $/gal of waste solvent.
The electricity costs associated with operating a gas absorber derive from fan requirements to overcome the
pressure drop in the column, ductwork, and other parts of the control system, and pump requirements to
recirculate the solvent. The energy required for the fan can be calculated using Equation 9.48:
9-37
-------
Energy
1.17
~l G. AP
fan
(9.48)
where Energy (in kilowatts) refers to the energy needed to move a given volumetric flow rate of air (acfrn), G,
is the waste gas flow rate entering the absorber, AP is the total pressure drop through the system (inches of H2O)
and e is the combined fan-motor efficiency. Values for e typically range from 0.4 to 0.7. Likewise, the electricity
required by a recycle pump can be calculated using Equation 9.49:
;0.746) (2.52
nergy.
I/i (pressure
pump
(9.49)
where 0.746 is the factor used to convert horsepower to kW, pressure is expressed in feet of water, and e is the
combined pump-motor efficiency.
The cost of electricity (Ce) is then given by:
C = Energy
fan + pump
annual
operating
hours
cost of \
electricity/
(9.50)
where cost of electricity is expressed in units of $/KW-hr.
9.5.2 Indirect Annual Costs
Indirect annual costs (1C) include overhead, taxes, insurance, general and administrative (G&A), and capital
recovery costs. The suggested factors for each of these items also appear in Table 9.4. Overhead is assumed to
be equal to 60 percent of the sum of operating, supervisory, and maintenance labor, and maintenance materials.
Overhead cost is discussed in Chapter 2 of this manual.
The system capital recovery cost, CRC, is based on an estimated 15-year equipment life. (See Chapter 2 of
this manual for a discussion of the capital recovery cost.) For a 15-year life and an interest rate of 7 percent, the
capital recovery factor is 0.1098 The system capital recovery cost is then estimated by:
CRC = 0.1098 TCI
(9.51)
G&A costs, property tax, and insurance are factored from total capital investment, typically at 2 percent, 1
percent, and 1 percent, respectively.
9-38
-------
9.5.3 Total Annual Cost
Total annual cost (TAG) is calculated by adding the direct annual costs and the indirect annual costs.
TAG = DC + 1C (9.52)
9.6 Example Problem #1
The example problem presented in this section shows how to apply the gas absorber sizing and costing
procedures presented in this chapter to control a waste gas stream consisting of HC1 and air. This example
problem will use the same outlet stream parameters presented in the thermal incinerator example problem found
in Chapter 3 of this manual. The waste gas stream entering the gas absorber is assumed to be saturated with
moisture due to being cooled in the quench chamber. The concentration of HC1 has also been adjusted to
account for the change in volume.
9.6.1 Required Information for Design
The first step in the design procedure is to specify the conditions of the gas stream to be controlled and the
desired pollutant removal efficiency. Gas and liquid stream parameters for this example problem are listed in
Table 9.5. The quantity of HC1 can be written in terms of Ib-moles of HC1 per Ib-moles of pollutant-free-gas (7,)
using the following calculation:
= 0.001871
~
1 - 0.001871
Ib-moles HC1
= 0.00187
Ib-mole pollutant free gas
The solvent, a dilute aqueous solution of caustic, is assumed to have the same physical properties as water.
9.6.2 Step 1 : Determine Gas and Liquid Stream Properties
Once the properties of the waste gas stream entering the absorber are known, the properties of the waste gas
stream exiting the absorber and the liquid streams entering and exiting the absorber need to be determined. The
pollutant concentration in the entering liquid (X,) is assumed to be zero. The pollutant concentration in the
exiting gas stream (Y0) is calculated using Equation 9.1 and a removal efficiency of 99 percent.
9-39
-------
0002
00002 -
I,I,[-
UJ2 004 006 008 01 0.12
to-motes HCl/lb-moles Solvent
014
I
016
0 18
Figure 9.8: Equilibrium Curve-Operating Line for HC1-Water System[7]
9-40
-------
Table 9.5: Example Problem Data
Parameters Values
Stream Properties
Waste Gas Flow Rate Entering Absorber 21,377 scfm (22,288 acfm)
Temperature of Waste Gas Stream 100°F
Pollutant in Waste Gas HC1
Concentration of HC1 Entering Absorber in Waste Gas 1871 ppmv
Pollutant Removal Efficiency 99% (molar basis)
Solvent Water with caustic in solution
Density of Waste Gas" 0.0709 lb/ft3
Density of Liquid[7] 62.4 lb/ft3
Molecular Weight of Waste Gas" 29 lb/lb-mole
Molecular Weight of Liquid[7j 18 lb/lb-mole
Viscosity of Waste Gas" 0.044 Ib/ft-ht
Viscosity ofLiquid(7j 2.16 Ib/ft-hr
Minimum Wetting Rate[7] 1.3 ft2/hr
Pollutant Properties*
Diffusivity of HC1 in Air 0 725 ft'/hr
Diffusivity of HC1 in Water 1.02 x 10'4 R'2/hi
Packing Propertiesr
Packing fype 2-inch ceramic Raschig rings
Packing factor: Fr 65
Packing constant: a 3.82
Packing constant: 0 0.41
Packing constant: •» 0.45
Packing constant: 0 0125
Packing constant: b 0.22
Surface Area to Volume Ratio 28
"Reference [7], at 100°F.
6 Appendix 9A.
rAppendix 9B.
9-41
-------
Y = 0.00187 11 - -
1 100
= 0.0000187
The liquid flow rate entering the column is calculated from the L/GS ratio using Equation 9.2. Since ¥„ Y0,
and X, are defined, the remaining unknown, X0', is determined by consulting the equilibrium curve. A plot of
the equilibrium curve-operating line graph for an HCl-water system is presented in Figure 9.8. The value of X0'
is taken at the point on the equilibrium curve where 7, intersects the curve. The value of, Y intersects the
equilibrium curve at an Xvalue of 0.16.
The operating line is constructed by connecting two points: (X,, Y0) and (X0\ Y,). The slope of the operating
line intersecting the equilibrium curve, (L/GJmin, is:
L
0.00187 - 0.0000187
0.16 - 0
= 0.0116
The actual L/GS ratio is calculated using Equation 9.3. For this example, an adjustment factor of 1.5 will
be used.
L
= (0.0116) (1.5) = 0.0174
act
The value of G, may be calculated using Equation 9.4.
G
(60min/hr) (0.0709 lb/ft3) (22,288 acfm)
= 3,263
(29 Ib/lb-mole) (1 + 0.00187;
Ib-moles
^—
The flow rate of the solvent entering the absorber may then be calculated using Equation 9.5.
L = 0.0174 3,263 —
s \ ^r
, 56.8
hr
9-42
-------
The values of Gmol, and Lmoll are calculated using Equations 9.6 and 9.7, respectively:
mol, i
3,263
Ib-moles
hr
(1 + 0.00187) = 3,269
Ib-moles
hr
= I56.8
hr
(1.0) = 56. 8
hr
The pollutant concentration exiting the absorber in the liquid is calculated using Equation 9.10.
X =
o
0.00187 - 0.0000187 0.106 Ib-moles HC1
0.0174
Ib-mole solvent
9.6.3 Step 2: Calculate Absorption Factor
The absorption factor is calculated from the slope of the equilibrium line and the Lmol/Gmo/l ratio. The slope of
the equilibrium curve is based on the mole fractions of x,, x0,y', and_y0*, which are calculated from^f,, X0, Y'
and Y* from Figure 9.8. From Figure 9.8, the value of Y* in equilibrium with the Rvalue of 0.106 is 0.0001.
The values of 7,* and X are 0. The mole fraction values are calculated from the concentration values using
Equations 9.13 through 9.16.
x .
0 1 + 0.106
=0.096
y*= °-0001 =0.0001
0 1 + 0.0001
The slope of the equilibrium fine from x, to x0 is calculated from Equation 9.12:
0.096 - 0
=0.00104
9-43
-------
Since HC1 is very soluble in water, the slope of the equilibrium curve is very small. The absorption factor is
calculated from Equation 9.11.
AF - °-°174 = 17
0.00104
9.6.4 Step 3: Estimate Column Diameter
Once the inlet and outlet stream conditions are determined, the diameter of the gas absorber may be calculated
using the modified generalized pressure drop correlation presented in Figure 9.5. The abscissa value from the
graph is calculated from Equation 9.17:
ABCISSA - 0.0174
18
29
\
0.0709
62.4
= 0.000364
Since this value is outside the range of Figure 9.5, the smallest value (0.01) will be used as a default value. The
ordinate is calculated from Equation 9.19.
ORDINATE = -J_Q[-I.668-1.085 (Io9 °-01> -0.297 (log o.oi)2]
= 0.207
The superficial gas flow rate, Gsfri, is calculated using Equation 9.20. For this example calculation, 2-inch
ceramic Raschig rings are selected as the packing. The packing factors for Raschig rings are listed in Appendix
9B.
sfr, i
;0.207) (62.4) (0.0709 lb/ft3) (32.2 ft/sec2)
^
= 0.681 lb/sec-ft
'65) (1) (0.893
0.2
2
Once GSfr, is determined, the cross-sectional area of the column is calculated using Equation 9.21.
A = (3,263 Ib-mol/hr) (29 Ib/lb-mol) = 55 x f t 2
(3600 sec/hr) (0.681 lb/sec-ft2)' (0.7)
The superficial liquid flow rate is determined using Equation 9.23.
9-44
-------
(56.8 Ib-mol/hr) (18 Ib/lb-mol) _ 1Q g 1K/K 2
1, . = - its. b ±D/nr~rt
sfr'J 55.1 ft2
At this point, it is necessary to determine if the liquid flow rate is sufficient to wet the packed bed. The
minimum value ofLsfrl is calculated using Equation 9.24. The packing constant (a) is found in Appendix 9B.
(Zs/, ,)„,,„ = (1.3 ft2/hr)(62.4 Ib/ft3)(28 fWft3) = 2,271 lb/hr-ft2
The Lsfr, value calculated using the L/G ratio is far below the minimum value needed to wet the packed bed.
Therefore, the new value, (Lsfr ,)„,,„ will be used to determine the diameter of the absorber. The calculations for
this revised diameter are shown in Appendix 9C. Appendix 9C shows that the cross-sectional area of the column
is calculated to be 60 ft2, Lmol, is 7572, and (fri is 0.627 Ib/sec^ft . (The diameter of the column is then
calculated using Equation 9.22:A)
=8.74ft
\ n
9-45
-------
The value of X0 is then:
X .0.00187-0.0000187
7,572
3,263
Expressed in terms of mole fraction:
, = Q-OOQ8 .0.0008
0 1-0.0008
The value ofy0 in equilibrium with x0 cannot be estimated accurately. However, the value will
approach zero, and the value of AF will be extremely large:
A Z7 7»572
AF=-
(3,263)(«0)
9.6.5 Step 4: Calculate Column Surface Area
Since xt = 0 and AF is large, Equation 9.26 will be used to calculate the number of transfer units:
, 1 0.00187 \ , ,,
\f =\n\ =4.61
'" \ 0.0000187]
The height of a transfer unit is calculated from, AF, HL, and HG. The values of HG and HL are
calculated from Equations 9.29 and 9.30:
3.82[(3,600)(0.7)(Q.627)]°
41
2,27 1
0-45
0.044
(0.725)(0.0709)
=2.24/?
2.16
022
2.16
(0.000102)(62.4)
-1.06/r
The height of the transfer unit is calculated using Equation 9.28:
tu
=(2.24//)+(1.06//)=2.24//
9-46
-------
The depth of packing is calculated from Equation 9.25.
The total height of the column is calculated from Equation 9.33:
The surface area of the column is calculated using Equation 9.34:
S=(3.14)(8.74)(26.1+8.74/2)=836//2
9.6.6 Step 5: Calculate Pressure Drop
The pressure drop through the column is calculated using Equation 9.35.
(0 17X2 271)
-
0.0709
= 0.83 inches water/foot packing
The total pressure drop (through 10.3 feet of packing) equals 8.55 inches of water.
9.6.7 Equipment Costs
Once the system sizing parameters have been determined, the equipment costs can be calculated.
For the purpose of this example, a gas absorber constructed of FRP will be costed using Equation
9.40.
TTC($) = 115(836) = $96,140
The cost of 2-inch ceramic Raschig rings can be estimated from packing cost ranges presented in
Section 9.5. The volume of packing required is calculated as:
Volume of packing = (60 ft2)(10.3 ft) = 618 ft3
Using the average of the cost range for 2-inch ceramic packings, the total cost of packing is:
Packing cost = ($20/ft3)(618 ft3) = $12,360
9-47
-------
For this example problem, the cost of a pump will be estimated using vendor quotes. First, the
flow rate of solvent must be converted into units of gallons per minute:
)= 2,271
' [ h-fiiy 8.34/6 60min
= 272 gpm
The average price for a FRP pump of this size is $16/gpm at a pressure of 60 ft water, based on
the vendor survey. [12] Therefore, the cost of the recycle pump is estimated as:
cum =(272gpw)($ 1 61 gpm) =$4,350
pump
For this example, the cost for a fan (FRP, backwardly-inclined centrifugal) can be calculated using
the following equation: [1 8]
where d is the impeller (wheel) diameter of the fan expressed in inches. For this gas flow rate and
pressure drop, an impeller diameter of 33 inches is needed. At this diameter, the cost of the fan is:
Cfan =57.9(33)' 38 =$7,2 10
The cost of a fan motor (three-phase, carbon steel) with V-belt drive, belt guard, and motor
starter can be computed as follows: [18]
As will be shown in Section 9.6.8, the electricity consumption of the fan is 32.0kW. Converting
to horsepower, we obtain a motor size of 42.6 hp. The cost of the fan motor is:
Cmoror=104(42.6)0821 =$2,260
The total auxiliary equipment cost is:
$4,350 + $7,210 + $2,260 = $13,820
The total equipment cost is the sum of the absorber cost, the packing cost, and the auxiliary
equipment cost:
EC - 96,140 + 12,360 + 13,820 = $122,320
9-48
-------
The purchased equipment cost including instrumentation, controls, taxes, and freight is estimated
using Equation 9.43:
PEC = 1.18(122,320) = $144,340
The total capital investment is calculated using Equation 9.44:
TCI = 2.20(144,340) = $317,550 * $318,000
9.6.8 Total Annual Cost
Table 9.6 summarizes the estimated annual costs using the suggested factors and unit costs for the
example problem.
Direct annual costs for gas absorber systems include labor, materials, utilities, and wastewater
disposal. Labor costs are based on 8,000 hr/year of operation. Supervisory labor is computed at
15 percent of operating labor, and operating and maintenance labor are each based on 1/2 hr per
8-hr shift.
The electricity required to run the fan is calculated using Equation 9.48 and assuming a combined
fan-motor efficiency of 70 percent:
r (1.17*10 4)(22,288)(8.55) „ n, „,
Energy, =- —— — -=32.0kW
fa" 0.70
The energy required for the liquid pump is calculated using Equation 9.49. The capital cost of the
pump was calculated using data supplied by vendors for a pump operating at a pressure of 60 feet
of water. Assuming a pressure of 60 ft of water a combined pump-motor efficiency of 70 percent:
9-49
-------
Table 9.6: Annual Costs for Packed Tower Absorber Example Problem
?ost Item
Calculations
Cost
direct Annual Costs. DC
Operating Labor
Operator
Supervisor
Operating materials
Solvent (water)
Caustic Replacement
Wastewater disposal
Maintenance
Labor
Material
Electricity
Total DC
Indirect Annual Costs. 1C
Overhead
Administrative charges
Property tax
Insurance
Capital recovery3
Total 1C
Total Annual Cost (rounded)
0.5hr x shift x S.OOOhr x SI5.64
shift 8hr yr hr
15% of operator = 0.15 x 7,820
7.16 gpm x 60 min x S.OOOhr x S0.20
hr yr 1 OOOgal
3.06lb-mole x 621b x S.OOOhr x ton x 1 x S300
hr Ib-mole yr 20001b 0.76 ton
7.16gpm x 60 min x 8.000 hr x S3.80
hr yr 1 OOgal
0.5 x shift x S.OOOhr x $17.21
shift 8hr yr hr
100% of maintenance labor
36.4kw x S.OOQhr $0.0461
yr kWh
60% of total labor and maintenance material:
= 0.6(7,820+ 1,170 + 8,610 + 8,610)
2% of Total Capital Investment = 0.02(S317,550)
1% of Total Capital Investment = 0.01(S317,550)
1% of Total Capital Investment = 0.01(S317,550)
0.1315 * $317,550
$7,820
1,170
690
299,560
13,060
8,610
8,610
13.420
$352,940
15,730
6.350
3,180
3,180
41.760
$70,200
$423,000
The capital recovery cost factor, CRF, is a function of the absorber equipment life and the opportunity cost of the
capital (i.e., interest rate). For this example, assume a 15-year equipment life and a 10% interest rate.
9-50
-------
Energy -(0-746)(2.52xlQ-<)(272)(60)(l)
"" 0.70
The total energy required to operate the auxiliary equipment is approximately 36.4 kW. The cost of electricity, Ce, is
calculated using Equation 9.50 and with the cost per kWh shown in Table 9.6.
Ce = (36.4kW)(8,000 n/yr)($0.0461/kWh) = $13,420/yr
The costs of solvent (water), wastewater disposal, and caustic are all dependent on the total
system throughput and the fraction of solvent discharged as waste. A certain amount of solvent
will be wasted and replaced by a fresh solution of water and caustic in order to maintain the
system's pH and solids content at acceptable levels. Based on the vendor survey, a maximum
solids content of 10 percent by weight will be the design basis for this example problem. [ 12] The
following calculations illustrate the procedure used to calculate how much water and caustic are
needed, and how much solvent must be bled off to maintain system operability.
From previous calculations, Lmoll = 7,572 Ib-moles/hr. The mass flow rate is calculated as:
L = 7,572 " 18 lb = 136,300-^-
mass { hr )( Ib-mole) hr
With Gmol, at 3,263 Ib-moles/hr, the mass flow rate of the gas stream is calculated as:
G =( 3,263 lb-mole\ ( 29 lb } =94,800^
mass ( hr )( Ib-mole) hr
The amount of HC1 in the gas stream is calculated on a molar basis as follows:
On a mass basis:
lb lb HC1
=223.4-
-^ _
hr Ib-mole hr
For this example problem, the caustic is assumed to be Na^, with one mole of caustic required
for neutralizing 2 moles of HCL. Therefore, 3.06 Ib-moles/hr of caustic are required.
The unit cost of a 76 percent solution of NajO is given in Table 9.6. The annual cost is calculated
from:
9-51
-------
r ( * nfilb-moles}( „ Ib }( S.OOOArV ton }( 1 }( $300
C =| 3.UO o2
I hr )( lb-mole ){ yr ]( 2,000/6)( 0.76){ ton)
= $299,560 yr
Mass of the salt formed in this chemical reaction, NaCl, is calculated as:
ji^cc I VYI Alb~HC1\( lb-mole \( 1 Ib-mole NaCl\( 58.5 Ib NaCl \
Mass., r,=\ 223 .4 - 1| - )| - ) -
c
s., r,= . - - - -
NaCI ( hr )( 36.5lb HC1 )( lb-mole HC1 )( lb-mole NaCl }
Ib NaCl
=358.1-
hr
If the maximum concentration of NaCl in the wastewater (ww) is assumed to be 10 weight
percent, the wastewater volume flow rate is calculated as:
IT/ * , fi AtoC/V I Ib ww }( gal ww }( 1 hr }
- - — £ - -
hr )( 0.1 Ib NaCl)( 8.34 Ib ww)( 60 minj
= 7.16 gpm
00M| S3.80 \ =$
yr}{ 1,000 gal)
The cost of solvent (water) is:
/- /-TI/: \f 60 minV 0 nftnArW $0.20 "\ c/;nn/
C =(7.16 gpw) - 8,000 — - =$690/yr
^ '( ^r J( jrj( 1,000 gal)
'Because the wastewater stream contains only NaCl, it probably will not require pretreatment before discharge to a
municipal wastewater treatment facility. Therefore, the wastewater disposal unit cost shown here is just a sewer usage
rate. This unit cost ($3.80/1,000 gal) is the average of the rates charged by the seven largest municipalities in North
Carolina.[20] These rates range from approximately $2 to $6/1,000 gal. This wide range is indicative of the major
differences among sewer rates throughout the country.
Indirect annual costs include overhead, administrative charges, property tax, insurance, and capital recovery. Total
annual cost is estimated using Equation 9.52. For this example case, the total annual cost is estimated to be $423,000
per year (Table 9.6).
9-52
-------
9.7 Example Problem #2
In this example problem the diameter of a gas absorber will be estimated by defining a pressure
drop. A pressure drop of 1 inch of water per foot of packing will be used in this example
calculation. Equation 9.38 will be used to calculate the ordinate value relating to an abscissa
value. If the Lmo, /Gmoll ratio is known, the abscissa can be calculated directly. The ordinate
value is then:
ORDINATE = exp [-4.0950-1.00121n(0.0496)-0.1587(ln 0.0496)2 +
0.0080(ln 0.0496)3 + 0.0032(ln 0.0496)4]
- 0.084
The value of Gsfr is calculated using Equation 9.39.
(62.4-0.0709)(0.0709)(32.2)(0.084)
65(0.893)°'
= 0.43 Ib/ft2-sec
The remaining calculations are the same as in Section 9.3.4, except the flooding factor is not used
in the equations.
9.8 Acknowledgments
The authors gratefully acknowledge the following companies for contributing data to this chapter:
• Air Plastics, Inc. (Cincinnati, OH)
• April, Inc. (Teterboro, NJ)
• Anderson 2000, Inc. (Peachtree City, GA)
• Calvert Environmental (San Diego, CA)
Ceilcote Air Pollution Control (Berea, OH)
• Croll-Reynolds Company, Inc. (Westfield, NJ)
• Ecolotreat Process Equipment (Toledo, OH)
• Glitsch, Inc. (Dallas, TX)
• Interel Corporation (Englewood, CO)
• Jaeger Products, Inc. (Spring, TX)
• Koch Engineering Co., Inc. (Wichita, KS)
• Lantec Products, Inc. (Agoura Hills, CA)
• Midwest Air Products Co., Inc. (Owosso, Ml)
• Monroe Environmental Corp., (Monroe, MI)
• Norton Chemical Process Products (Akron, OH)
9-53
-------
Appendix 9A
Properties of Pollutants
9-54
-------
Table 9.7: Physical Properties of Common Pollutants3
Pollutant
Ammonia
Methanol
Ethyl Alcohol
Propyl Alcohol
Butyl Alcohol
Acetic Acid
Hydrogen Chloride
Hydrogen Bromide
Hydrogen Fluoride
Molecular
Weight
f lh }
{ Ib-molej
17
32
46
60
74
60
36
36
20
Diffusivity in
Air
at25°C
(cmVsec)
0.236
0.159
0.119
0.100
0.09
0.133
0.187
0.129
0.753
Diflusivity in
Water
at20°C
(cm2/sec)x!05
1.76
1.28
1.00
0.87
0.77
0.88
2.64
1.93
3.33
"Diffusivity data taken from Reference [7, 21].
9-55
-------
Appendix 9B
Packing Characteristics
9-56
-------
Table 9.8: Packing Factors for Various Packings [3, 7, 10, 13]
Packing
Type
Raschig rings
Raschig rings
Pall rings
Pall rings
Berl saddles
Intalox saddles
Tri-Packs®
Construction
Level
ceramic
metal
metal
polypropylene
ceramic
ceramic
plastic
Nominal
Diameter
(inches)
1/2
5/8
3/4
1
1 1/2
2
3
1/2
5/8
3/4
1
1 1/2
2
3
5/8
1
1 1/2
2
3 1/2
5/8
1
1 1/2
2
1/2
3/4
1
1 1/2
2
1/2
3/4
1
1 1/2
2
3
2
3 1/2
FP
640
380
255
160
95
65
37
410
290
230
137
83
57
32
70
48
28
20
16
97
52
32
25
240
170
110
65
45
200
145
98
52
40
22
16
12
a
111
100
80
58
38
28
118
72
57
41
31
21
131
66
48
36
110
63
39
31
142
82
76
44
32
190
102
78
60
36
48
38
9-57
-------
Table 9.9: Packing Constants Used to Estimate HG [1, 3, 7, 13]
Packing
Type
Raschig Rings
Berl Saddles
Partition Rings
LanPac®
Tri-Packs®
Size
(inches)
3/8
1
1
1 1/2
1 1/2
2
1/2
1/2
1
1 1/2
3
2.3
2
3 1/2
Packing Constants
a (3 Y
2.32 0.45 0.47
7.00 0.39 0.58
6.41 0.32 0.51
1.73 0.38 0.66
2.58 0.38 0.40
3.82 0.41 0.45
32.4 0.30 0.74
0.81 0.30 0.24
1.97 0.36 0.40
5.05 0.32 0.45
640. 0.58 1.06
7.6 0.33 -0.48
1.4 0.33 0.40
1.7 0.33 0.45
Applicable Range3
Gsfr Lsfr
200-500 500-1,500
200-800 400-500
200-600 500-4,500
200-700 500-1,500
200-700 1,500-4,500
200-800 500-4,500
200-700 500-1,500
200-700 1,500-4,500
200-800 400-4,500
200-1,000 400-4,500
150-900 3,000-10,000
400-3,000 500-8,000
100-900 500-10,000
100-2,000 500-10,000
aUnitsoflb/hr-ft2
9-58
-------
Table 9.10: Packing Constants Used to Estimate HL [1, 3, 13]
Packing
Type
Raschig Rings
Berl Saddles
Partition Rings
LanPac®
Tri-packs®
Size
(inches)
3/8
1
1 1/2
2 1/2
2
1/2
1
1 1/2
3
2.3
3.5
2
3 1/2
Packing Constants
b
0.00182 0.46
0.00357 0.35
0.0100 0.22
0.0111 0.22
0.0125 0.22
0.00666 0.28
0.00588 0.28
0.00625 0.28
0.0625 0.09
0.0039 0.33
0.0042 0.33
0.0031 0.33
0.0040 0.33
Applicable Range
Ta
L sft
400-15,000
400-15,000
400-15,000
400-15,000
400-15,000
400-15,000
400-15,000
400-15,000
3,000-14,000
500-8,000
500-8,000
500-10,000
500-10,000
"Units of lb/hr-ft2
9-59
-------
Table 9.11: Packing Constants Used to Estimate Pressure Drop [1,7, 13]
Packing
Type
Raschig rings
Raschig rings
Pall rings
Berl saddles
Intalox saddles
Construction
Material
ceramic
metal
metal
ceramic
ceramic
Nominal
Diameter
(inches)
1/2
3/4
1
1 1/4
1 1/2
2
5/8
1
1 1/2
2
5/8
1
1 1/2
2
1/2
3/4
1
1 1/2
1/2
3/4
1
1 1/2
c
3.1
1.34
0.97
0.57
0.39
0.24
1.2
0.42
0.29
0.23
0.43
0.15
0.08
0.06
1.2
0.62
0.39
0.21
0.82
0.28
0.31
0.14
j
0.41
0.26
0.25
0.23
0.23
0.17
0.28
0.21
0.20
0.135
0.17
0.16
0.15
0.12
0.21
0.17
0.17
0.13
0.20
0.16
0.16
0.14
9-60
-------
Appendix 9C
Minimum Wetting Rate Analysis
As explained in the design procedures, the liquid flow rate entering the column must be high
enough to effectively wet the packing. If the liquid flow rate, as determined theoretically in
Equation 9.23, is lower than the flow rate dictated by the minimum wetting rate, calculated in
Equation 9.24, then the packing will not be wetted sufficiently to ensure mass transfer between
the gas and liquid phases. The minimum liquid flow rate should then be used as a default value.
The superficial gas flow rate, Gsfr , and cross-sectional area of the column must then be
recalculated to account for the increased liquid flow rate. The approach necessary to recalculate
these variables is explained in Section 9C.1 of this Appendix. The calculation of these variables
using the results from Example Problem #1 are presented in Section 9C.2 of this Appendk.
9C.1 Overview of the Approach
1. The value of L mo!l must be recalculated from the value of (Lsfr,) mm using the equation:
£
mol.i
(MW)L
The value of A (the cross-sectional area of the absorber column) is the only unknown in the
equation.
1. The ABSCISSA value is calculated in terms of A by substituting the new Lmo,, into Equation
9.17.
9-61
-------
1. The value of Gsfr, is recalculated by rearranging Equation 9.21, with A as the only unknown.
1. The ORDINATE value is calculated in terms of A from the new Gsfr<, using the Equation 9.18.
1. An iterative process is used to determine A, ABSCISSA, and ORDINATE. Values of A are
chosen and the ABSCISSA and ORDINATE values are calculated. The ORDINATE value
corresponding to the ABSCISSA value is determined from Figure 9.5 (or Equation 9.19), and
this value is compared to the ORDINATE value calculated using Equation 9.18. This process
is continued until both ORDINATE values are equal.
9C.2 Example Problem Calculation
Step 1 : The first step is to recalculate the liquid flow rate. The liquid molar flow rate may be
calculated using Equation 9.23.
L , =(2,271 Iblhr-ft
mol,, ( 18 /ft
-(126.2 lb-mole/hr-ft2)A
Step 2: The abscissa value from Figure 9.5, and presented in Equation 9. 1 7, is calculated as:
A /JCY^TVC/l — ^L ^•^••*-ltJ iriunsfftr Jl /-^ I lo 1 U.U/U:7
3,263lb~mole/hr \ 29 j\ 62.4
=8.09x10 -4A
Step 3: The value of Gsfr , is then recalculated in terms of the cross- sectional area of the
column.
lb-mole/hr)(29 Ibllb-mole) _37.6
sfr (3,600 sec/Jrr)(0.7>4 A
Step 4: The ordinate value from Figure 9.5, and presented in Equation 9. 1 8, is calculated as:
9-62
-------
Step 5: At this point the simplest solution is an iterative approach. Choose a value for A,
calculate the ABSCISSA value using Equation 9.53, and find the corresponding
ORDINATE value off the flooding curve in Figure 9.5 (or use Equation 9.19 to
calculate the ORDINATE value). Compare the calculated ORDINATE value from
Equation 9.54 to the value obtained from the graph or from Equation 9.19. By
continuing this process until the ORDINATE values converge the value of A is
determined to be 60 ft2. The following table illustrates the intermediate steps in the
calculational process.
Assumed
Value
ofA
65
62
60
ABSCISSA
Calculated
From Eqn. 9.53
0.0526
0.0503
0.0485
ORDINATE
Calculated
From Eqn. 9.19
0.1714
0.1740
0.1757
ORDINATE
Calculated
From Eqn. 9.54
0.1493
0.1642
0.1752
The value of Gcfr is then:
Gf=±±2-= 0.627 /6/sec-//2
sfr 60
The liquid molar flow rate is:
IWO/I=(126.2) (60)=7,572 Ib-mole/hr
The diameter and height of the column using the results of this calculation are presented in
Example Problem #1.
9-63
-------
References
[ 1 ] Control Technologies for Hazardous A ir Pollutants, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Par,, North
Carolina, Publication No. EPA 625/6-91-014.
[2] Mclnnes, R., K. Jameson, and D. Austin, "Scrubbing Toxic Inorganics", Chemical
Engineering, September 1990, pp. 116-121.
[3] Letter from Jose L. Bravo of Jaeger Products, Inc., to William M. Vatavuk, U. S.
Environmental Protection Agency, June 8, 1992.
[4] Treybal, Robert E., Mass Transfer Operations (Third edition), McGraw-Hill Book
Company, New York, 1980.
[5] Letter from Jack D. Brady of Anderson 2000, Inc., to William M. Vatavuk, U.S.
Environmental Protection Agency, June 9, 1992.
[6] Letter from S. Raymond Woll of Air Products, Inc., to William M. Vatavuk, U.S.
Environmental Protection Agency, June 25, 1992.
[7] Perry, R.H. and C.H. Chilton, Eds., Chemical Engineers 'Handbook (Sixth edition),
McGraw-Hill Book Company, New York, 1984.
[8] Crowe, Charles R., and D. Cooper, "Brick/Membrane Linings Pass the Acid Test",
Chemical Engineering, July 1988, pp. 83-86.
[9] Harrison, Mark E., and John J. France, "Distillation Column Troubleshooting, Part 2:
Packed Columns", Chemical Engineering, April 1989, pp. 121-128.
[10] Coker, A.K., "Understanding the Basics of Packed-Column Design", Chemical
Engineering Progress, November 1991, pp. 93-99.
[11] Telephone conversation between Roy Oommen, Radian Corporation and Gerald Nealon,
Norton Process Equipment, April 4, 1992.
[12] Gas absorber questionnaire responses from nine gas absorber vendors to Radian
Corporation August-December, 1991.
9-64
-------
[13] Buonicore, A.J., and L. Theodore, Industrial Control Equipment for Gaseous Pollutants,
Volume I, CRC Press, Inc., Cleveland, Ohio, 1975.
[14] Strigle, Ralph F., Random Packings and Packed Towers, Design Applications, Gulf
Publishing Company, Houston, Texas, 1987.
[15] Questionnaire response from K.C. Lang of Lantec Products, Inc. To R.V. Oommen,
Radian Corporation, August 23, 1991.
[16] Memorandum from Vatavuk, W.M. of U.S. Environmental Protection Agency to Martha
Smith, U.S. EPA, March 27, 1992.
[17] Packing vendor questionnaire responses from seven packing vendors to Radian
Corporation, August 1991 - January 1992.
[18] Vatavuk, W.M., "Pricing Equipment for Air-Pollution Control", Chemical Engineering,
May 1990, pp. 126-130.
[19] Vatavuk, W.M., and R.B. Neveril, "Estimating Costs of Pollution Control Systems, Part
II: Factors for Estimating Capital and Operating Costs", Chemical Engineering,
November 3, 1980, pp. 157-162.
[20] Telephone conversation between William M. Vatavuk, U.S. Environmental Protection
Agency, and Cindy Kling, City of Raleigh, N.C., July 16, 1992.
[21] "Air Pollution Engineering Manual" (AP-40), (Second Edition), Danielson, John A., Los
Angeles County Air Pollution Control District, CA, May 1973.
9-65
-------
Chapter 10
HOODS, DUCTWORK, and STACKS
William M. Vatavuk
Innovative Strategies and Economics Group, OAQPS
U.S. Environmental Protection Agency
Research Triangle Park, NC
December 1995
Contents
10.1 Introduction 10-4
10.2 Equipment Description 10-4
10.2.1 Hoods 10-4
10.2.1.1 Types of Hoods 10-5
10.2.2 Ductwork 10-8
10.2.2.1 Ductwork Components 10-9
10.2.3 Stacks 10-13
10.3 Design Procedures 10-13
10-1
-------
10.3.1 Design Fundamentals 10-14
10.3.1.1 The Bernoulli Equation 10-14
10.3.1.2 Pressure: Static, Velocity, and Total 10-17
10.3.1.3 Temperature and Pressure Adjustments 10-19
10.3.2 Hood Design Procedure 10-20
10.3.2.1 Hood Design Factors 10-20
10.3.2.2 Hood Sizing Procedure 10-25
10.3.3 Ductwork Design Procedure 10-27
10.3.3.1 Two Ductwork Design Approaches 10-27
10.3.3.2 Ductwork Design Parameters 10-27
10.3.3.3 Ductwork Pressure Drop 10-31
10.3.4 Stack Design Procedures 10-34
10.3.4.1 Calculating Stack Diameter 10-34
10.3.4.2 Calculating Stack Height 10-35
10.3.4.3 Calculating Stack Draft 10-36
10.4 Estimating Total Capital Investment 10-37
10.4.1 Equipment Costs 10-38
10.4.1.1 Hood Costs 10-38
10.4.1.2 Ductwork Costs 10-40
10.4.1.3 Stack Costs 10-46
10.4.2 Taxes, Freight, and Instrumentation Costs 10-49
10.4.3 Purchased Equipment Cost 10-49
10.4.4 Installation Costs 10-49
10-2
-------
10.5 Estimating Total Annual Cost 10-50
10.5.1 Direct Annual Costs 10-50
10.5.2 Indirect Annual Costs 10-51
10.5.3 Total Annual Cost 10-52
10.6 Acknowledgements 10-52
References 10-53
10-3
-------
10.1 Introduction
Most control devices are located some distance from the emission sources they control. This
separation may be needed for several reasons. For one thing, there may not be enough room to
install the control device close to the source. Or, the device may collect emissions from several
sources located throughout the facility and, hence, must be sited at some convenient, equidistant
location. Or, it may be that required utility connections for the control device are only available at
some remote site. Regardless of the reason, the waste gas stream must be conveyed from the source
to the control device and from there to a stack before it can be released to the atmosphere.
The kinds of equipment needed to convey the waste gas are the same for most kinds of control
devices. These are: (1) hoods, (2) ductwork, (3) stacks, and (4) fans. Together, these items
comprise a ventilation system. A hood is used to capture the emissions at the source; ductwork, to
convey them to the control device; a stack, to disperse them after they leave the device; and a fan,
to provide the energy for moving them through the control system. This chapter covers the first
three kinds of equipment. However, because they constitute such a broad and complex subject, fans
will be dealt with in a future Manual chapter. Also, the kinds of stacks covered are short stacks
(100-120 feet high or less). Typically, these are included with packaged control systems or added
to them. So-called "tall stacks" ("chimneys"), used at power plants or other sources where the
exhaust gases must be dispersed over great distances, will not be discussed in this chapter.
This chapter presents all the information one would need to develop study (± 30%-accurate) cost
estimates for hoods, ductwork, and stacks. Accordingly, the following sections include: (1)
descriptions of the types of equipment used in air pollution control ventilation systems, (2)
procedures for sizing (designing) this equipment, and (3) methodologies and data for estimating
their capital and annual costs. Also, sprinkled throughout the chapter are several illustrations
(example problems) that show the reader how to apply the various sizing and costing methodologies.
10.2 Equipment Description
In this section, the kinds of hoods, ductwork, and stacks used in air pollution control systems are
described, each in a separate subsection. These descriptions have been based on information
obtained from standard ventilation and air pollution control references, journal articles, and
equipment vendors.
10.2.1 Hoods
Of the several components of an air pollution control system, the capture device is the most
important. This should be self-evident, for if emissions are not efficiently captured at the source
they cannot be conveyed to and removed by a control device. There are two general categories of
capture devices: (1) direct exhaust connections (DEC) and (2) hoods. As the name implies, a DEC
is a section of duct (typically an elbow) into which the emissions directly flow. These connections
10-4
-------
often are used when the emission source is itself a duct or vent, such as a process vent in a chemical
manufacturing plant or petroleum refinery. (See discussion below on "Ductwork".)
Hoods comprise a much broader category than DECs. They are used to capture participates,
gases, and/or mists emitted from a variety of sources, such as basic oxygen steelmaking furnaces,
welding operations, and electroplating tanks. The hooded processes are generally categorized as
either "hot" or "cold", a delineation that, in turn, influences hood selection, placement, and design.
The source conditions also influence the materials from which a hood is fabricated. Mild
(carbon) steel is the material of choice for those applications where the emission stream is
noncorrosive and of moderate temperature. However, where corrosive substances (e.g., acid gases)
are present in high enough concentrations, stainless steels or plastics (e.g., fiberglass-reinforced
plastic, or FRP) are needed. As most hoods are custom-designed and built, the vendor involved
would determine which material would be optimal for a given application.
10.2.1.1 Types of Hoods
Although the names of certain hoods vary, depending on which ventilation source one consults,
there is general agreement as to how they are classified. There are four types of hoods: (1)
enclosures, (2) booths, (3) captor (capture) hoods, and (4) receptor (receiving) hoods.1'2
Enclosures are of two types: (1) those that are completely closed to the outside environment and
(2) those that have openings for material input/output. The first type is only used when handling
radioactive materials, which must be handled by remote manipulators. They are also dust- and gas-
tight. These kinds of enclosures are rarely used in air pollution control.
Total enclosures, the second type, have applications in several areas, such as the control of
emissions from electric arc furnaces and from screening and bin filling operations. They are
equipped with small wall openings (natural draft openings—"NDOs") that allow for material to be
moved in or out and for ventilation. However, the area of these openings must be small compared
with the total area of the enclosure walls (typically, 5% or less).
Another application of total enclosures is in the measurement of the capture efficiency of VOC
(volatile organic compound) control devices. Capture efficiency is that fraction of all VOCs
generated at, and released by, an affected facility that is directed to the control device. In this
application, a total enclosure is a temporary structure that completely surrounds an emitting process
so that all VOC emissions are captured for discharge through ducts or stacks. The air flow through
the total enclosure must be high enough to keep the concentration of the VOC mixture inside the
enclosure within both the Occupational Safety and Health Administration (OSHA) health
requirement limits and the vapor explosive limits. (The latter are typically set at 25% of the lower
explosive limit (LEL) for the VOC mixture in question.) In addition, the overall face velocity of air
flowing through the enclosure must be at least 200 ft/min3.
The surfaces of temporary total enclosures are usually constructed either of plastic film or of
such rigid materials as insulation panels or plywood. Plastic film offers the advantages of being
lightweight, transparent, inexpensive, and easy to work with. However, it is flimsy, flammable, and
10-5
-------
has a relatively low melting point. In addition, the plastic must be hung on a framework of wood,
plastic piping, or scaffolding. ±
Although rigid materials are more expensive and less workable than plastic, they are more ^^
durable and can withstand larger pressure differentials between the enclosure interior and exterior.
Total enclosure design specifications (which have been incorporated into several EPA emission
standards) are contained in the EPA report, The Measurement Solution: Using a Temporary Total
Enclosure for Capture Testing.4
Booths are like enclosures, in that they surround the emission source, except for a wall (or
portion thereof) that is omitted to allow access by operators and equipment. Like enclosures, booths
must be large enough to prevent particulates from impinging on the inner walls. They are used with
such operations (and emission sources) as spray painting and portable grinding, polishing, and
buffing operations.
Captor Hoods: Unlike enclosures and booths, captor hoods (also termed active or external hoods)
do not enclose the source at all. Consisting of one to three sides, they are located at a distance from
the source and draw the emissions into them via fans. Captor hoods are further classified as side-
draft/backdraft, slot, downdraft, and high-velocity, low-volume (HVLV) hoods. A side-draft/back-
draft hood is typically located to the side/behind of an emission source, but as close to it as possible,
as air velocities decrease inversely (and sharply) with distance. Examples of these include snorkel-
type welding hoods and side shake-out hoods.
A slot hood operates in a manner similar to a side-draft/back-draft. However, the inlet opening
(face) is much smaller, being long and narrow. Moreover, a slot hood is situated at the periphery
of an emission source, such as a narrow, open tank. This type of hood is also employed with bench
welding operations.
While slot and side-draft/back-draft hoods are located beside/behind a source, a downdraft hood
is situated immediately beneath it. It draws pollutant-laden air down through the source and. thence,
to a control device. Applications of down-draft hoods include foundry shake-out and bench
soldering and torch cutting operations.
HVLV hoods are characterized by the use of extremely high velocities (capture velocities) to
collect contaminants at the source, and by the optimal distribution of those velocities across the hood
face. To maintain a low volumetric flow rate, these hoods are located as close to the source as
possible, so as to minimize air entrainment.
Receptor hoods: The last category is receptor hoods (a.k.a. passive or canopy hoods). A receptor
hood typically is located above or beside a source, to collect the emissions, which are given
momentum by the source. For example, a canopy hood might be situated directly above an open
tank containing a hot liquid (a buoyant source). With entrained air, vapors emitted from the liquid
would rise into the hood. Here, the canopy hood would function as a passive collector, as the rising
gases would be drawn into the hood via natural draft. (See Figure 10.1.)
10-6
-------
Source: tank or process
Figure 10.1: Typical Canopy Hood Installation
10-7
-------
Receptor hoods are also used with nonbuoyant sources, sources from which emissions do not
rise. However, the emissions can be "thrown off" from a process, such as a swing grinder. The
initial velocity of the emissions typically is high enough to convey them into a receiving hood.5
10.2.2 Ductwork
Once the emission stream is captured by either a hood or a direct exhaust connection, it is conveyed
to the control device via ductwork The term "ductwork" denotes all of the equipment between the
capture device and the control device. This includes: (1) straight duct; (2) fittings, such as elbows
and tees; (3) flow control devices (e.g., dampers); and (4) duct supports. These components are
described in Section 10.2.2.1.)
In air pollution control systems, the fan is usually located immediately before or after the control
device. Consequently, most of the ductwork typically is under a negative static pressure, varying
from a few inches to approximately 20 inches of water column. These pressure conditions dictate
the type of duct used, as well as such design parameters as the wall thickness (gauge). For instance,
welded duct is preferable to spiral-wound duct in vacuum applications.6
Ductwork is fabricated from either metal or plastic, the choice of material being dictated by the
characteristics of the waste gas stream, structural considerations, purchase and installation costs,
aesthetics, and other factors. Metals used include carbon steel (bare or galvanized), stainless steel,
and aluminum. The most commonly used plastics are PVC (polyvinyl chloride) and FRP
(fiberglass-reinforced plastic), although polypropylene (PP) and linear polyethylene (LPE) also have
been applied. However, one serious drawback to PP and LPE is that both are combustible?
PVC and other plastic ductwork are resistant to a variety of corrosive substances, from aqua
regia to 95% sulfuric acid. But plastic ductwork cannot tolerate environmental temperatures above
150°F.8 Metal ductwork can handle temperatures up to approximately 1000°F, but only certain
alloys can tolerate corrosive streams.
In terms of construction, ductwork can be either rigid or flexible. As the name implies, rigid
ductwork, whether metal or plastic, has a fixed shape. Conversely, flexible ductwork can be bent
to accommodate situations where space is limited or where the layout is so convoluted that rigid
fittings cannot meet construction requirements. Usually circular in cross-sectional shape, flexible
duct can be fabricated from metals or plastic and can be either insulated or uninsulated.
10-8
-------
Rigid ductwork is fabricated into circular, flat oval, or square/rectangular cross-sectional shapes.
Of these, circular duct is most commonly used in air pollution control systems. Although
square/rectangular duct is advantageous to use when space is limited, round duct offers several
advantages. It resists collapsing, provides better transport conditions, and uses less metal than
square/rectangular or flat oval shapes of equivalent cross-sectional area.9 Unless otherwise noted,
the following discussion will pertain to rigid, circular duct, as this is the type most commonly used
in air pollution control.
Rigid metal circular duct is further classified according to method of fabrication. Longitudinal
seam duct is made by bending sheet metal into a circular shape over a mandrel, and buttwelding the
two ends together. Spiral seam duct is constructed from a long strip of sheet metal, the edges of
which are joined by an interlocking helical seam that runs the length of the duct. This seam is either
raised or flush to the duct wall surface.
Fabrication method and cross-sectional shape are not the only considerations in designing
ductwork, however. One must also specify the diameter; wall thickness; type, number, and location
of fittings, controllers, and supports; and other parameters. Consequently, most ductwork
components are custom designed and fabricated, so as to optimally serve the control device. Some
vendors offer prefabricated components, but these are usually common fittings (e.g., 90° elbows)
that are available only in standard sizes (e.g., 3- to 12-inch diameter).10'11
If either the gas stream temperature or moisture content is excessive, the ductwork may need to
be insulated. Insulation inhibits heat loss/gain, saving energy (and money), on the one hand, and
prevents condensation, on the other. Insulation also protects personnel who might touch the
ductwork from sustaining bums. There are two ways to insulate ductwork. The first is to install
insulation on the outer surface of the ductwork and cover it with a vapor barrier of plastic or metal
foil. The type and thickness of insulation used will depend on several heat transfer-related
parameters. For instance, one vendor states that 4 inches of mineral wool insulation is adequate for
maintaining a surface ("skin") temperature of 140°F (the OSHA workplace limit) or lower, provided
that the exhaust gas temperature does not exceed 600°F. n
The second way to insulate ductwork is by using double-wall, insulated duct and fittinqs.
Double-wall ductwork serves to reduce both heat loss and noise. One vendor constructs it from a
solid sheet metal outer pressure shell and a sheet metal inner liner with a layer of fiberglass
insulation sandwiched between. The insulation layer is typically 1-inch, although 2- and 3-inch
thicknesses are available for more extreme applications. The thermal conductivities of these
thicknesses are 0.27, 0.13, and 0.09 Btu/hr-ft2-°F, respectively.13
10.2.2.1 Ductwork Components
As discussed above, a ductwork system consists of straight duct, fittings, flow control devices, and
supports. Straight duct is self-explanatory and easy to visualize. The "fittings" category, however,
encompasses a range of components that perform one or more of the following functions: change
the direction of the ducted gas stream, modify the stream velocity, tie it to another duct(s), facilitate
the connection of two or more components, or provide for expansion/contraction when thermal
stresses arise.
10-9
-------
The most commonly used fittings are elbows ("ells"). These serve to change the gas stream
direction, typically by 30°, 45°, 60°, or 90°, though they may be designed for other angles as well.
The elbow centerline radius determines the rate at which this directional change occurs. (See Figure
10.2.) The standard centerline radius (R^) is 1.5 x the elbow cross-sectional diameter (D ) .
However, in "long-radius" elbows, in which the directional change is more gradual than in standard
elbows, R,, = > 2DC.14
Tees are used when two or more gas streams must be connected. In straight tees, the streams
converge at a 90° angle, while in angle tees ("laterals", "wyes") the connection is made at 30°, 45 °,
60°, or some other angle. (See Figure 10.2.) Tees may have one "tap" (connection) or two, and may
have either a straight or a "conical" cross-section at either or both ends. Crosses are also used to
connect duct branches. Here, the two branches intersect each other at a right angle.
Reducers (commonly called "expansions" or "contractions") are required whenever ducts of
different diameter must be joined. Reducers are either concentric or eccentric in design. In
concentric reducers, the diameter tapers gradually from the larger to smaller cross section.
However, in eccentric reducers, the diameter decreases wholly on one side of the fitting.
To control the volumetric flowrate through ventilation systems, dampers are used. Dampers are
usually delineated according to the flow control mechanism (single blade or multiblade), pressure
rating (low/light or high/heavy), and means of control (manual or automatic). In single blade
dampers, a circular plate is fastened to a rod, one end of which protrudes outside the duct. In the
most commonly used type of single blade damper (butterfly type), this rod is used to control the gas
flow by rotating the plate in the damper. Fully closed, the damper face sits perpendicular to the gas
flow direction; fully open, the face is parallel to the gas flow lines. Several single blade "control"
dampers are depicted in Figure 10.2.
With blast gate dampers, a second type, the flow is controlled by sliding the damper blade in
and out of the duct. Blast gates are often used to control the flow of air streams containing
suspended solids, such as in pneumatic conveyors. In these respects, butterfly dampers and blast
gates are analogous, respectively, to the globe valves and quick-opening gate valves that are used
to regulate liquid flow in pipes.
Multiblade (louvered) dampers operate by means of the same principal. However, instead using
a single blade or plate to control the gas flow, multiblade dampers employ slats that open and close
like Venetian blinds.15 Louvered dampers typically are used in very large ducts where a one-piece
damper blade would be too difficult to move.
Manually-controlled dampers simply have a handle attached to the control rod which is used to
adjust the gas flow by hand. If automatic control is needed, a pneumatic or electronic actuator is
used. The actuator receives a pneumatic (pressurized air) or electrical signal from a controller and
converts it to mechanical energy which is used, in turn, to open/close the damper via the
10-10
-------
LONGITUDINAL
SEAM DUCT
I Fully w*ia«d
longitudinal seam)
DIMENSIONS:
90* m«i*rium
DOMED ELBOW
DIMENSIONS:
B . 1 iA
STRAIGHT TEE
STRAIGHT W CROSS
DIMENSIONS:
V . C • 2
Mumum C • A
-
r*-
R-
,E
• ]
—
*. —
_j.
T
1
T
DIMENSIONS:
V . C . 2
MAjtfttum C Or D » A
HEAVY-DUTY
CONTROL DAMPER
CONCENTRIC
REDUCER
ECCENTRIC REDUCER
f Reference: "Single-Wall Round and Flat Oval Duct and
Fittings." In: Sheet Metal Division Catalog. Groveport, OH:
United McGill Corporation. 1990.
Figure 10.2: Selected Circular Ductwork Components1
10-11
-------
damper rod. In this respect, an actuated damper is analogous to an automatic control valve.16 For
example, an automatic damper may be used to control the dilution air flow rate to an incinerator
combustion chamber. This flow rate, in turn, would depend on the combustibles concentration (i.e.,
percentage of lower explosive limit—%LEL) in the inlet waste gas stream. If this concentration
deviates from a predetermined amount ("set point"), a signal is sent from the measuring device via
the controller to the automatic damper to increase/decrease the dilution air flow rate so as to
maintain the desired %LEL.
Expansion joints are installed, especially in longer metal duct runs, to allow the ductwork to
expand or contract in response to thermal stresses. These fittings are of several designs. One type,
the bellows expansion joint, consists of a piece of flexible metal (e.g., 304 stainless steel) that is
welded to each of two duct ends, connecting them. As the temperature of the duct increases, the
bellows compresses; as the duct temperature decreases, the bellows expands.
Another commonly used expansion joint consists of two flanges between which is installed a
section of fabric. Like the bellows expansion joint, it compresses as the duct temperature increases,
and vice-versa. The temperature dictates the type of fabric used. For instance, silicone fiberglass
and aramid fiber cloth can be used for duct temperatures of up to 500°F., while coated fiberglass
cloth is needed to accommodate temperatures of 1,000°F.17
The last component to consider is the ductwork support system. However, it is far from being
the least important. As the SMACNA (Sheet Metal and Air Conditioning Contractors' National
Association) HVAC Duct Construction Standards manual states, "The selection of a hanging system
should not be taken lightly, since it involves not only a significant portion of the erection labor, but
also because [the erection of] an inadequate hanging system can be disastrous." As a rule, a support
should be provided for every 8 to 10 feet of duct run.18 Ductwork can be suspended from a ceiling
or other overhead structure via hangers or supported from below by girders, pillars, or other
supports.
A suspension arrangement typically consists of an upper attachment, a hanger, and a lower
attachment. The upper attachment ties the hanger to the ceiling, etc. This can be a concrete insert.
an eye bolt, or a fastener such as a rivet or nailed pin. The hanger is generally a strap of galvanized
steel, round steel rod, or wire that is anchored to the ceiling by the upper attachment. The type of
hanger used will be dictated by the duct diameter, which is proportional to its weight per lineal foot.
For instance, wire hangers are only recommended for duct diameters up to 10 inches. For larger
diameters (up to 36 inches), straps or rods should be used. Typically, a strap hanger is run from the
upper attachment, wrapped around the duct, and secured by a fastener (the lower attachment). A
rod hanger also extends down from the ceiling. Unlike strap hangers, they are fastened to the duct
via a band or bands that are wrapped around the circumference. Duct of diameters greater than 3
feet should be supported with two hangers, one on either side of the duct, and be fastened to two
circumferential bands, one atop and one below the duct.19 Moreover, supports for larger ductwork
should also allow for both axial and longitudinal expansion and contraction, to accommodate
thermal stresses.20
10-12
-------
10.2.3 Stacks
Short stacks are installed after control devices to disperse the exhaust gases above ground level and
surrounding buildings. As opposed to "tall" stacks, which can be up to 1000 feet high, short stacks
typically are no taller than 120 feet.
Certain packaged control devices come equipped with short ("stub") stacks, with heights ranging
from 30 to 50 feet. But if such a stack is neither provided nor adequate, the facility must erect a
separate stack to serve one or more devices. Essentially, this stack is a vertical duct erected on a
foundation and supported in some manner. For structural stability, the diameter of the stack bottom
is slightly larger than the top diameter, which typically ranges from 1 to 7 feet.21
A short stack may be fabricated of steel, brick, or plastic (e.g., fiberglass-reinforced plastic, or
FRP). A stack may be lined or unlined. The material selection depends on the physical and
chemical properties of the gas stream, such as corrosiveness and acidity, as well as the temperature
differential between the gas stream and the ambient air. Liners of stainless steel, brick, or FRP
usually are used to protect the stack against damage from the gas stream. They are much easier and
less expensive to replace than the entire stack. Alternatively, the interior of an unlined stack may
be coated with zinc (galvanized), aluminum, or another corrosion-resistant material, but a coating
does not provide the same protection as a liner and does not last as long.
Short stacks are either self-supporting (free-standing), supported by guy wires, or fastened to
adjacent structures. The type of support used depends on the stack diameter, height and weight, the
wind load, local seismic zone characteristics, and other factors.
Auxiliary equipment for a typical stack includes an access door, a sampling platform, ladders,
lightning protection system, and aircraft warning lights. The access door allows for removal of any
accumulated materials at the bottom of the stack and provides access to the liner for repair or
replacement. Local and state air pollution control regulations also may require the permanent
installation of sampling platforms for use during periodic compliance tests, while ladders are used
both during stack sampling and maintenance procedures. The lightning protection system is needed
to prevent damage to the stack and immediate surroundings during electrical storms. Lastly, aircraft
warning lights are required by local aviation authorities.23 Altogether, these auxiliaries can add a
large amount to the base stack cost.
10.3 Design Procedures
As stated above, a hood, ductwork, and a stack are key elements in any air pollution control system.
Because each of these elements is different, both in appearance and function, each must be designed
separately. But at the same time, these elements comprise a system, which is governed by certain
physical laws that serve to unite these elements in "common cause". Thus, before the individual
design procedures for hoods, ductwork, and stacks are described, ventilation fundamentals will be
presented. These fundamentals will cover basic fluid flow concepts and how they may be applied
to air pollution control ventilation systems. Nonetheless, these concepts will be given as
10-13
-------
straightforwardly as possible, with the aim of making the design parameters easy to understand and
compute.
10.3.1 Design Fundamentals
10.3.1.1 The Bernoulli Equation
The flow of fluids in any hood, duct, pipe, stack, or other enclosure is governed by a single
relationship, the familiar Bernoulli equation. Put simply and ideally, the Bernoulli equation states
that the total mechanical energy of an element of flowing fluid is constant throughout the system.
This includes its potential energy, kinetic energy, and pressure energy. However, as no system is
ideal, the Bernoulli equation must be adjusted to take into account losses to the surroundings due
to friction. Gains due to the energy added by fans, pumps, etc., also must be accounted for. For a
pound mass (lbro) of fluid flowing in a steady-state system the adjusted Bernoulli equation is:24
IP + AWCI + Al/tyltf = t - f (10.1)
where
v = specific volume of fluid (ft3/lbm)
p = static pressure—gauge (lb/ft2)
z = height of fluid above some reference point (ft)
u = fluid velocity through duct, hood, etc. (ft/sec)
g = gravitational acceleration (ft/sec2)
gc = gravitational constant (32.174 ([Ibm-ft/sec2]/lbf)
W = work added by fan, etc. (ft-lb/lbj
F = energy lost due to friction (ft-lb/lbm)
Each of the terms on the left hand side of equation 10.1 represents an energy change to a pound
mass of fluid between two locations in the system—points "1" and "2". The work (W) and friction
(F) terms denote the amounts of energy added/lost between points 1 and 2.
Note that the units of each term in equation 10.1 are "ft-lb/lbm," energy per unit mass. In the
English system of units, "lbf" and "lbm" are, for all intents, numerically equivalent, since the ratio
of the gravitational acceleration term (g) to the gravitational constant (gc) is very close to 1. In
effect, therefore, the equation unit are "feet of fluid" or "fluid head in feet". In air pollution control
situations, the fluid often has the properties of air. That is because the contaminants in the waste
gas stream are present in such small amounts that the stream physical properties approximate those
of pure air.
Because air is a "compressible" fluid, its specific volume is much more sensitive to charges in
pressure and temperature than the specific volume of such "incompressible" fluids as water. Hence,
the "vdp" term in the equation has to be integrated between points 1 and 2. However, in most air
10-14
-------
pollution control ventilation systems neither the pressure nor the temperature changes appreciably
from the point where the emissions are captured to the inlet of the control device. Consequently,
the specific volume is, for all practical purposes, constant throughout the ventilation system, and one
does not have to integrate the vdp term. With this assumption, the first term in equation 10.1
becomes simply:
J2vdp = vJ2dp = vAp (10.2)
Illustration: VOC emitted by an open tank is captured by a hood and conveyed, via a blower,
through 150 feet of 12-inch diameter ductwork to a refrigerated condenser outdoors. The blower,
which moves the gas through the hood, ductwork, and condenser, is located immediately before the
inlet to the condenser. Thus, the entire ventilation system is under vacuum. The stream temperature
and absolute pressure are 100°F and approximately 1 atmosphere (14.696 lb/in2), respectively. The
elevation of the refrigerated condenser inlet is 30 feet below that of the tank. The air velocity at the
source is essentially zero, while the duct transport velocity is 2,000 ft/min. The static gauge
pressure increases from -0.50 in. w.c. (water column) at the source to 4.5 in. w.c. at the blower
outlet. Finally, the calculated friction loss through the ductwork and hood totals 1.25 in. w.c.
Calculate the amount of mechanical energy that the blower adds to the gas stream. Assume that the
gas temperature remains constant throughout.
Solution:
•^ First, develop a factor to convert "inches of water" to "feet of air":
Feet of air = (Inches of water) (1 ft/12 in)(va]00/vwloo) (10.3)
where
vwloo = specific volume of water @ 100°F = 0.01613 ftVlbm
vaioo = specific volume of air @ 100°F, 1 atmosphere
Because the system absolute pressure is close to atmospheric, the waste gas behaves as an ideal
gas. Thus, the specific volume can be calculated from the ideal gas law:
va = RT/pM (10.4)
where
R = ideal gas constant = 1,545 ft-lb/(lbm-mole)(°R)
T = absolute temperature of gas = 100 +460 = 560°R
M = molecular weight of gas (air) = 28.85 lbm/lbm-mole
p = absolute pressure = 2,116 lb/ft2
Substituting, we obtain:
10-15
-------
va=14.17ftVlbffi
Finally, substitution of these values for va and vw into equation 10.3 yields:
Feet of air (@ 100°F, 1 atm.) = 73.207 x Inches of water
*& Compute the changes in the mechanical energy terms and the friction losses between the hood
inlet (point 1) and the blower outlet/condenser inlet (point 2):
Pressure: vAp = (4.5 - [-0.50] in. w.c.)(73.207 ft air/in, w.c.) = 366.0 ft air
Potential: Az = -30 ft air (point 2 is below point 1)
Kinetic: Au2/2gc = ([2,000 ft/min]/[60 ft/min/1 ft/sec]2 x (l/2)(32.174[lbm-ft/sec2]/lbf)'1
= 17.3 ft air
Friction losses: F =1.25 in. w.c. x 73.207 = 91.5 ft air
"^ Substitute above results into equation 10.1 and solve for W, the fan energy added:
366.0 + (-30) + 17.1 = W - 91.5, or
W = 444.6 ft-lb/lbm air = 6.07 in. w.c.
To convert the fan energy input, W, to horsepower (hpf), we would have to multiply it by the
air mass flow rate (lbm/sec). and divide the result by the horsepower conversion factor, 550 ft-
lbf/sec-hp. However, the mass flow rate is just the volume flow rate (Q, ft3/sec) divided by the
specific volume:
hpf = W(Q/v.) (1/550) = 0.001818WQ/va (10.5)
(The reader may wish to compare this equation to the fan horsepower equation in Chapter 3 of this
manual.)
In turn, Q is a function of the duct velocity (x^, ft/sec) and duct diameter (Dd, ft):
Q = u.frD/M) (10.6)
Equation 10.6 applies, of course, only to circular ducts.
If we combine equations 10.5 and 10.6 and substitute the inputs for this illustration, we obtain:
hpf = (444.6) (2,000/60) (7t/4)(l)2(l/14.17) (1/550) = 1.49 hp
* * *
10-16
-------
Some observations about this illustration:
"^ Recall that the precise units for W and the other terms in equation 10.1 are "ft-lb/lbm air," which,
for convenience, have been shortened to "ft air". Thus, they measure energy, not length.
"S* Compared to the pressure energy and friction terms, the potential and kinetic energy terms are
small. Had they been ignored, the results would not have changed appreciably.
i®1 The large magnitude of the pressure and friction terms clearly illustrates the importance of
keeping one's units straight. As shown in step (1), one inch of water is equivalent to over 73 feet
of air. However, as equation 10.3 indicates, the pressure corresponding to equivalent heights of air
and water columns would be the same.
"^ The fan power input depends not just on the total "head" (ft air) required, but also on the gas
flow rate. Also, note that the horsepower computed via equation 10.5 is a theoretical value. It
would have to be adjusted to account for the efficiencies of the fan and fan motor. As mentioned
in Chapter 3, the fan efficiency ranges from 40 to 70 percent, while the motor efficiency is typically
90 percent. These efficiencies are usually combined into a single efficiency (e, fraction), by which
the theoretical horsepower is divided to obtain the actual horsepower requirement.
10.3.1.2 Pressure: Static, Velocity, and Total
Although it is more rigorous and consistent to express the Bernoulli equation terms in terms of feet
of air (or, precisely, ft-lb/lbm of air), industrial ventilation engineers prefer to use the units "inches
of water column (in. w.c.)." These units were chosen because, as the above illustration shows,
results expressed in "feet of air" are often large numbers that are cumbersome to use. In addition,
the total pressure changes in ventilation systems are relatively small, compared to those in liquid
flow systems. Total pressure changes expressed in inches of mercury would be small numbers
which are just as awkward to work with as large numbers. Hence, "inches of water" is a
compromise, as values expressed in this measurement unit typically range from only 1 to 10.
Moreover, practical measurement of pressure changes is done with water-filled manometers.
In the previous paragraph, a new quantity was mentioned, total pressure (TP). Also known as
the "impact pressure", the total pressure is the sum of the static gauge (SP) and velocity pressures
(VP) at any point within a duct, hood, etc., all expressed in in. w.c.25 That is:
TP = SP + VP (10.7)
where
SP = (cf)vp
VP = (cf)u2/2gc
10-17
-------
The "cf ' in the expressions for SP and TP is the factor for converting the energy terms from "ft
air" to "in. w.c.", both at standard temperature and absolute pressure (70°F, 1 atmosphere). (Again,
keep in mind that, regardless of what units SP or VP are expressed in, the actual units are "energy
per unit mass".) This conversion factor would be obtained via rearranging equation 10.3:
cf = in. w.c./ft. air = 12(vw70/va70) (10.8)
where
vw70 = specific volume of water at 70°F = 0.01605 (ftVlbJ
va70 = specific volume of air at 70°F = 13.41 (ftVlbJ
Thus: cf = 0.01436 in. w.c./ft air
Clearly, "cf" varies as a function of temperature and pressure. For instance, at 100°F and 1
atmosphere, cf = 1/73.207 = 0.01366. Nevertheless, unless noted otherwise, all quantities
henceforth in this chapter will reflect conditions at 70 °F and 1 atmosphere.
Conspicuously absent from equation 10.7 is the potential energy term, "z(g/gc)". This omission
was not inadvertent. In ventilation systems, the potential energy (P.E.) is usually small compared
to the other terms. (For example, see illustration above.) The P.E. is, of course, a function of the
vertical distance of the measurement point in question from some datum level, usually the ground.
At most, that distance would amount to no more than 20 or 30 feet, corresponding to a P.E. of
approximately 0.3 to 0.4 in. w.c. Consequently, we can usually ignore the P.E. contribution in
ventilation systems without introducing significant error.
The static gauge pressure in a duct is equal in all directions, while the velocity pressure, a
function of the velocity, varies across the duct face. The duct velocity is highest at the center and
lowest at the duct walls. However, for air flowing in a long, straight duct, the average velocity (ut)
approximates the center line velocity (uc]).26 This is an important point, for the average velocity is
often measured by a pitot tube situated at the center of the duct.
By substituting for "cf" in equation 10.7, we can obtain a simple equation that relates velocity
to velocity pressure at standard conditions:
= 0.01436u,2/2gc (10.9)
Solving:
u, (ft/sec) = 66.94(VPf (10. 10)
Or:
ut(ft/min) = 4,016(VPf (10.11)
Incidentally, these equations apply to any duct, regardless of its shape.
10-18
-------
As Burton describes it, static gauge pressure can be thought of as the "stored" energy in a
ventilation system. This stored energy is converted to the kinetic energy of velocity and the losses
of friction (which are mainly heat, vibration, and noise). Friction losses fall into several
categories:27
• Losses through straight duct
• Losses through duct fittings — elbow tees, reducers, etc.
• Losses in branch and control device entries
• Losses in hoods due to turbulence, shock, vena contracta
• Losses in fans
• Losses in stacks
These losses will be discussed in later sections of this chapter. Generally speaking, much more
of the static gauge pressure energy is lost to friction than is converted to velocity pressure energy.
It is customary to express these friction losses (ASPf) in terms of the velocity pressure:
F = ASPf = kVP (10.12)
where: k = experimentally-determined loss factor (unitless)
Alternatively, equations 10. 1 1 and 10.12 may be combined to express F (in. w.c.) in terms of
the average duct velocity, u, (ft/min):
F = (6.200 xlO-s)ku,2 (10.13)
10.3.1.3 Temperature and Pressure Adjustments
Equations 10.8 to 10.13 were developed assuming that the waste gas stream was at standard
temperature and pressure. These conditions were defined as 70°F and 1 atmosphere (14.696 lb/in2),
respectively. While 1 atmosphere is almost always taken as the standard pressure, several different
standard temperatures are used in scientific and engineering calculations: 32°F, 68°F, and 77°F,
as well as 70°F. The standard temperature selected varies according to the industry or engineering
discipline in question. For instance, industrial hygienists and air conditioning engineers prefer 70 °F
as a standard temperature, while combustion engineers prefer 77 °F, the standard temperature used
in Chapter 3 ("Thermal and Catalytic Incinerators").
Before these equations can be used with waste gas streams not at 70°F and 1 atmosphere, their
variables must be adjusted. As noted above, waste gas streams in air pollution control applications
obey the ideal gas law. From this law the following adjustment equation can be derived:
Q2 = Q.OyT.XP./Pj) (10.14)
10-19
-------
where:
Q2,Qj = gas flow rates at conditions 2 and 1, respectively (actual ftVmin)
T2,T, = absolute temperatures at conditions 2 and 1, respectively (°R)
P2,Pj = absolute pressures at conditions 2 and 1, respectively (atm)
However, according to equation 10.6:
If equations 10.6 and 10.14 were combined, we would obtain:
,) (P,/P2) (Dd22/Ddl2) (10.15)
This last expression can be used to adjust u, in any equation, as long as the gas flow is in circular
ducts.
10.3.2 Hood Design Procedure
10.3.2.1 Hood Design Factors
When designing a hood, several factors must be considered:28
v^ Hood shape
^ Volumetric flow rate
"^ Capture velocity
^ Friction
Each of these factors and their interrelationships will be explained in this section.
As discussed in section 10.2.1, the hood shape is determined by the nature of the source being
controlled. This includes such factors as the temperature and composition of the emissions, as well
as the dimensions and configuration of the emission stream. Also important are such environmental
factors as the velocity and temperature of air currents in the vicinity.
The hood shape partly determines the volumetric flow rate needed to capture the emissions.
Because a hood is under negative pressure, air is drawn to it from all directions. Consider the
simplest type of hood, a plain open-ended duct. Now, envision an imaginary sphere surrounding
the duct opening. The center of this sphere would be at the center of the duct opening, while the
10-20
-------
sphere radius would be the distance from the end of the duct to the point where emissions are
captured. The air would be drawn through this imaginary sphere and into the duct hood. Now, the
volume of air drawn through the sphere would be the product of the sphere surface area and the
hood capture velocity, uc:29
Q = uc(47ix2) (10.16)
where: x = radius of imaginary sphere (ft)
Equation 10.16 applies to a duct whose diameter is small relative to the sphere radius. However,
if the duct diameter is larger, the capture area will have to be reduced by the crosssectional area of
the duct (Dd), or:
Q = uc(4nx2 - 7iDd2/4) (10.17)
Similarly, if a flange were installed around the outside of the duct end, the surface area through
which the air was drawn—and the volume flow rate—would be cut in half. That occurs because
the flange would, in effect, block the flow of air from points behind it. Hence:
Q = uc(2Ttx2) (10.18)
From these examples, it should be clear that the hood shape has a direct bearing on the gas flow
rate drawn into it. But equations 10.16 to 10.18 apply only to hoods with spherical flow patterns.
To other hoods, other flow patterns apply—cylindrical, planal, etc. We can generalize this
relationship between volumetric flow rate and hood design parameters as follows:
Q = f(u,,x. Sh) (10.19)
where
"f(...)" denotes "function of..."
"Sh" indicates hood shape factors
u, = design velocity—capture, face, slot
Table 10.1 lists design equations for several commonly used hood shapes. As this table shows,
Q is a function of x, the hood shape, and, in general, the capture velocity (uc). But in one case
(booth hood), the design velocity is the hood face velocity (uf). And in the case of slotted side-draft
and back-draft hoods, the slot velocity (us) is the design velocity. In reality, both the hood face and
slot velocities are the same, as each measures the speed at which the gas passes
10-21
-------
Table 10.1: Design Equations, Loss Factors, and Coefficients of
Entry for Selected Hood Types*
Hood Type
Duct end (round)
Ranged duct end
(round)
Free-standing slot
hood
Slot hood w/sides,
back
Tapered hood
Booth hood with tapered
take-off duct (round)
Canopy hood
Canopy hood
w/insert
Dip tank hood
(slotted)
Paint booth hood
Design Equation
Q = 4irx2ur
Q = 2nx\
Q = 2irxLuc
Q = 0.5irxLuc
0 = 2UXU,
Q = ufAh
0 = 1.4Pxu,
Q=1.4Pxuc
Q=125A,
Q = 100Ah
Loss Factor
OO
0.93
0.50
1.78
1.78
0.06 c£
0.25
0.25
1.0
1.78
0.25
Coefficient of Entry
(CJ
0.72
0.82
0.55
n.a.§§
0.97
0.89
0.89
0.71
n.a.
n.a.
*
* Reference: Burton. D. Jeff. Industrial Ventilation Work Book. Salt Lake City: DJB A. Inc. 1989.
• In the equations: Q = flow rate drawn into hood (ft'/min)
x = distance from hood to source (ft)
uc = hood capture velocity (ft/min)
u, = hood face velocity (ft/min)
u, = hood slot velocity (ft/min)
Ah = hood face area (ft2)
P = perimeter of source (ft)
L = width of hood slot (ft)
A, = tank + drainboard surface area (ft2)
Ab = booth cross-sectional area (ft2)
" Not applicable
eE Both kh and Cc pertain to round ducts and to hoods with a 45° taper. At other angles, kh and Cc will differ.
10-22
-------
through the hood inlet opening(s).
When gas enters a hood, there is mechanical energy loss due to friction. This friction loss is
calculated using equations 10.1 and 10.2, assuming that the potential energy contribution from
gravity, Az (g/gc), and the work added to the system, W, are both zero. Thus:
vp2 - vp, + u22/2gc - Ul2/2gc = -F (10.20)
Replacing these terms with the corresponding ones from equations 10.7 and 10.12, we obtain:
SP2 - SP, + VP2 - VP, = - Hc = - \VP2 (10.21)
where
SP, = static gauge pressure at point i (in. w.c.)
VP, = velocity pressure at point i (in. w.c.)
Hc = hood entry loss (in. w.c.)
kh = hood loss factor (unitless)
In this equation, subscript 1 refers to a point just outside the hood face. Subscript 2 denotes the
point in the duct, just downstream of the hood, where the duct static pressure, SP2 or SPh and the
duct transport velocity, u2 or \\ are measured. At point 1, the hood velocity pressure, yP , is
essentially zero, as the air velocity there is negligible. Moreover, the static gauge pressure, SP1?
will be zero, as the absolute pressure at point 1 is assumed to be at one atmosphere, the reference
pressure. After these simplifications are made, equation 10.21 can be rearranged to solve for the
hood loss factor (kh):
kh = (-SPh/VP2) - 1 (10.22)
At first glance, it appears that kh could be negative, since VP is always positive. However, as
the air entering the hood is under a vacuum created by a fan downstream, SPh must be negative.
Thus, the term "-SP/VTV must be positive. Finally, because the absolute value of SPh is larger than
VP2, kh > O.
The hood loss factor varies according to the hood shape. It can range from 0.04 for bell mouth
hoods to 1.78 far various slotted hoods. A parameter related to the hood loss factor is the coefficient
of entry (ce).30 This is defined as:
ce={l/(l+kh)}y' (10.23)
10-23
-------
Table 10.2:Hood Design Velocities"
Operation/Hood Type
Tanks, degreasing
Drying oven
Spray booth
Canopy hood
Grinding, abrasive
blasting
Slot hood
Velocity Type
Capture
Face
Capture
Capture
Capture
Slot
Velocity Range (ft/min)
50 - 100
75 - 125
100 - 200
200 - 500
500 - 2,000
2,000
** Reference: Burton, D. Jeff. Industrial Ventilation Work Book. Salt Lake City. DJBA, Inc. 1989.
ce depends solely on the shape of the hood, and may be used to compute kh and related parameters.
Values of kh and ce are listed in Table 10. 1 .
Illustration: The static gauge pressure, SPh, is -1.75 in. w.c. The duct transport velocity (u_) is
3,500 ft/min. Calculate the loss factor and coefficient of entry for the hood. Assume standard
temperature and pressure.
Solution: First, calculate the duct velocity pressure. By rearranging equation 10.11 and substituting
for ut, we obtain:
VP = (u/4,016)2 = (3,500/4,01 6)2 = 0.76 in. w.c.
Next, substitute for VP in equation 10.22 and solve:
kh = (-SPh/VP) - 1 = (-[-1.75]/0.76) - 1 = 1.30.
Finally, use this value and equation 10.23 to calculate the coefficient of entry:
c . =
. 30) }¥t = 0.66
Hood design velocities are listed in Table 10.2. Three kinds of velocities are shown: (1) capture
(defined in Section 10.2.1), (2) face, and (3) slot. As stated in Section 10.2.1, the capture velocity-
is, the air velocity induced by the hood to capture contaminants emitted at some distance from the
10-24
-------
hood inlet. The face velocity is the average velocity of the air passing through the hood inlet (face).
A similar parameter is the slot velocity, which is the average air velocity through the hood slot
openings, whose area is only a fraction of the entire hood face area. Consequently, the slot velocity
is usually much higher than the face velocity.31
Note that these velocities range from 50 to 100 ft/min (tank and degreasing hoods) to 2,000
ft/min, the recommended slot velocity for slotted side-draft/back-draft hoods. As a reference point,
the velocity of air in industrial operations due to thermal mixing alone is 50 ft/min. Thus, hood
design velocities must exceed this value if effective capture is to occur.32
Two other velocities are also discussed in the industrial hygiene literature, although they do not
have as much bearing on hood design as the capture, face, or slot velocities. These are the plenum
velocity and the transport velocity. The first is the velocity of the gas stream as it passes through
the tapered portion of a hood (plenum) between the hood opening and the duct connection. This
plenum is a transition area between the hood opening and duct. Consequently, the plenum velocity
is higher than the hood face velocity, but lower than the duct (transport) velocity. The transport
velocity—the gas velocity through the duct—varies according to the waste gas composition. It is
a crucial parameter in determining the duct diameter, the static pressure loss, and the sizes of the
system fan and fan motor. (For more on transport velocity, see Section 10.3.3.)
10.3.2.2 Hood Sizing Procedure
As with many control devices and auxiliaries, there are several approaches to sizing hoods. Some
of these approaches are quite complex, entailing a series of complex calculations that yield
correspondingly accurate results. For instance, one hood sizing method in the literature involves
first determining the hood dimensions (length and width for rectangular hoods; diameter, for
circular). The next step is to estimate the amount of metal plate area (ft2) required to fabricate a
hood of these dimensions, via parametric curves. (No curves are provided for nonmetal hoods.)
This plate area is input to an equation that includes a "pricing factor" and the per-pound price of
metal. The cost of labor needed to fabricate this hood is estimated from equations similar to the
plate-area relationships. Finally, the metal and labor costs are summed to obtain the total fabricated
hood cost.31
This method does yield reasonably accurate hood cost—or rather, it did. Unfortunately, the
labor cost data are outdated—1977 vintage—which makes them unescalatable. (The rule-of-thumb
time limit for escalating costs is five years.) Even if the costs were up-to-date, the procedure is
cumbersome to use, especially if calculations are made by hand.
A simpler sizing method—yet one sufficiently accurate for study estimating purposes—involves
determining a single dimension, the hood face area (Af). This area, identical to the hood inlet area,
can be correlated against the fabricated hood cost to yield a relatively simple cost equation with a
single independent variable. To calculate Af, the following information is needed:
• Hood type
• Distance of the hood face from source (x)
10-25
-------
• Capture (uc), face (uf), or slot velocity (us)
• Source dimensions (for some hood types).
As the equations in Table 10.1 indicate, these same parameters are the ones that are used to
determine the volumetric flow rate (Q) through the hood and ductwork. With most control devices
and auxiliaries being sized, Q is given. For hoods, however, Q usually must be calculated.
Illustration: A circular canopy hood is being used to capture emissions from a chromium
electroplating tank. The hood face is 6 feet above the tank, an 8-foot diameter circular vessel. The
capture velocity for this example is 200 ft/min. Assuming that the tank surroundings are at standard
conditions, calculate the required volumetric flow rate drawn into the hood, the hood face area, and
the hood face velocity.
Solution: Obtain the canopy hood equation from Table 10.1:
Q = 1.4Pxuc (10.24)
where
P = perimeter of tank (ft)
x = distance of hood above tank (ft)
uc = capture velocity (ft/min)
Because the tank is circular, P = 71(8) = 25.1 ft.
Therefore:
Q = (1.4) (25.1) (6) (200) = 42,200 ft3/min.
For this type of canopy hood, the hood diameter is 40% greater than the tank diameter (hence.
the "1.4" factor in equation 10.24). Thus:
Af=(7i/4)([1.4][8])2 = 98.5ft2
Finally, the hood face velocity (uf) would be:
Uf = Q/Af = 42,200/98.5 = 428 ft/min.
In this example, note that the hood face velocity is higher than the capture velocity. This is
logical, given the fact that the hood inlet area is smaller than the area through which the tank fumes
10-26
-------
are being drawn. The face velocity for some hoods is even higher. For example, for slotted hoods
it is at least 1,000 ft/min.34 In fact, one vendor sizes the openings in his slotted
hoods so as to achieve a slot velocity equal to the duct transport velocity.35
10.3.3 Ductwork Design Procedure
The design of ductwork can be an extremely complex undertaking. Determining the number,
placement, and dimensions of ductwork components—straight duct, elbows, tees, dampers,
etc.—can be tedious and time-consuming. However, for purposes of making study-level control
system cost estimates, such involved design procedures are not necessary. Instead, a much simpler
ductwork sizing method can be devised.
10.3.3.1 Two Ductwork Design Approaches
There are two commonly used methods for sizing and pricing ductwork. In the first, the total weight
of duct is computed from the number and dimensions of the several components. Next, this weight
is multiplied by a single price (in $/lb) to obtain the ductwork equipment cost. To determine the
ductwork weight, one needs to know the diameter, length, and wall thickness of every component
in the system. As stated above, obtaining these data can be a significant effort.
The second method is a variation of the first. In this technique, the ductwork components are
sized and priced individually. The straight duct is typically priced as a function of length, diameter,
and wall thickness, as well as, of course, the material of construction. The elbows, tees, and other
fittings are priced according to all of these factors, except for length. Other variables, such as the
amount and type of insulation, also affect the price. Because it provides more detail and precision,
the second method will be used in this chapter.
10.3.3.2 Ductwork Design Parameters
Again, the primary ductwork sizing variable are length, diameter, and wall thickness. Another
parameter is the amount of insulation required, if any.
Length: The length of ductwork needed with an air pollution control system depends on such
factors as the distance of the source from the control device and the number of directional changes
required. Without having specific knowledge of the source layout, it is impossible to determine this
length accurately. It could range from 20 to 2,000 feet or more. It is best to give the straight duct
cost on a $/ft basis and let the reader provide the length. This length must be part of the
specifications of the emission source at which the ductwork is installed.
Diameter: As discussed in Section 10.2.2., circular duct is preferred over rectangular, oval, or other
duct shapes.
Therefore:
Ad = :rDd2/4 (10.25)
where
10-27
-------
Ad = cross-sectional area of duct (ft2)
Dd = duct diameter (ft)
The duct cross-sectional area is the quotient of the volumetric flow rate (Q) and the duct transport
velocity (u,):
(10.26)
Combining equations 10.25 and 10.26 and solving for Dd:
Dd = 1.128(Q/u,)* (10.27)
As Q is usually known, the key variable in equation 10.27 is the duct transport velocity. This
variable must be chosen carefully. If the u, selected is too low, the duct will be oversized and, more
importantly, the velocity will not be high enough to convey the paniculate matter in the waste gas
stream to the control device. However, if ^ is too high, the static pressure drop (which is
proportional to the square of u,) will be excessive, as will be the corresponding fan power
consumption.
Cost is also a consideration when determining the optimum duct diameter. The equipment cost
increases with increasing duct diameter. However, the fan power cost changes inversely with
diameter. Nonetheless, for study-estimating purposes, the optimum duct diameter does not have to
be determined. It is sufficient to calculate the duct diameter merely by using the transport velocity
values contained in this section.
The transport velocity typically varies from 2,000 to 6,000 ft/min, depending on the waste gas
composition. The lower duct velocity would be adequate for a waste gas containing gaseous
pollutants or very fine, light dusts, while the higher velocity would be needed to convey a stream
with a large quantity of metals or other heavy or moist materials. The following velocities may be
used as general guidance :36
Material(s) Conveyed
Gases; very fine, light dusts
Fine, dry dusts and powders
Average industrial dusts
Coarse dusts
Heavy or moist dust loading
Minimum Transport Velocity
(u,, ft/min)
2,000
3,000
3,500
4,000 - 4,500
> 4,500
Table 10.3 supplements these values with recommended duct velocities for a variety of conveyed
materials.
10-28
4
-------
Table 10.3: Minimum Duct Velocities for Selected Materials5
Material
Aluminum dust (coarse)
Brass turnings
Cast iron boring dust
Clay dust
Coal dust (powdered)
Cocoa dust
Cotton dust
Flour dust
Foundry dust
Grain dust
Lead dust
Limestone dust
Magnesium dust (coarse)
Metal turnings
Plastics dust (buffing)
Rubber dust
Silica dust
Soap dust
Soapstone dust
Spray paint
Starch dust
Stone dust
Tobacco dust
Minimum Transport Velocity
(ft/min)
4,000
4,000
4,000
3,500
4,000
3,000
3,000
2,500
3,000 - 5,000f
2.500 - 3,000
4.000
3,500
4.000
4,000 - 5,000
3,000
2,500 (fine) - 4,000 (coarse)
3,500 - 4.500
3,000
3,000
2,000
3,000
3,500
3,500
! Reference: Burton, D. Jeff. Industrial Ventilation Work Book. Salt Lake City: DJBA, Inc. 1989.
t Transport velocity varies with foundry operation.
10-29
-------
Table 10.4: Wall Thicknesses of Steel and Aluminum Duct1
Gauge
Number
28
26
24
22
20
18
16
14
12
10
Nominal Thickness (inches)
Carbon Steel
Galv*
0.0187
0.0217
0.0276
0.0336
0.0396
0.0516
0.0635
0.0785
0.1084
0.1382
Nongalv
0.0149
0.0179
0.0239
0.0299
0.0359
0.0478
0.0598
0.0747
0.1046
0.1345
Stainless Steel
(304 or 3 16)
0.0156
0.0188
0.0250
0.0313
0.0375
0.0500
0.0625
0.0781
0.1094
0.1406
Aluminum
3003-H14T
0.025
0.032
0.040
0.050
0.063
0.080
0.090
! Reference: Engineering Design Reference Manual for Supply Air Handling Systems. Groveport, OH: United
McGill Corporation. 1992.
f To provide equivalent strength and stiffness, the nominal thickness of aluminum is approximately 150% of the
nominal thickness of galvaized carbon steel of the same gauge.
* Galvanized and paintable galvanized carbon steel.
' Nongalvamzed carbon steel.
Wall thickness: The wall thickness of a duct depends on several factors—internal pressure,
diameter, material of fabrication, and other structural parameters. Nonetheless, duct of a given
diameter can be fabricated of a range of wall thicknesses, and vice-versa. For instance, 24-in.
diameter 304 stainless steel "fully-welded longitudinal seam duct" is fabricated in thicknesses
ranging from 22 to 14 gauge (0.0313 to 0.0781 in.). This same range of gauges is used with duct
diameters ranging from 3 to 36 in.37
Note that the gauge number decreases with increasing wall thickness. This measure, which is
traditionally used in the metal fabricating industries, is more convenient to deal with than the
thickness expressed in inches, as the latter are usually small numbers less than 0.25. Moreover, the
gauge number varies according to the metal used—carbon steel (galvanized or nongalvanized),
10-30
-------
stainless steel, or aluminum. Gauges for these metals are given in Table 10.4 for a wide range of
nominal thicknesses.
The gauge measure is not used with plastic duct, as the wall thickness is typically expressed in
inches. In any event, the wall thickness usually does not need to be known to estimate duct cost,
as this parameter is already accounted for in the cost equations. (See Section 10.4.)
•5" Insulation: As discussed in Section 10.2.2., insulation can be either installed on the outer
surface of ductwork or the ductwork itself can be fabricated with built-in insulation. In the first
case, the amount of insulation required will depend on several heat transfer variables, such as: the
temperature, velocity, composition, and other properties of the waste gas; the ambient temperature;
the duct diameter, wall thickness, and thermal conductivity; and the desired surface ("skin")
temperature. Determining these variables involves making a series of complex calculations that,
while well-established, are beyond the scope of this chapter. Such standard references as Perry's
Chemical Engineers' Handbook and Plant Design and Economics for Chemical Engineers present
these calculations, as do heat transfer bibliographies.38 39
The second approach is to select pre-insulated ductwork. As mentioned previously, it can be
equipped with any type and thickness of insulation. However, 1, 2, or 3 inches is typical. (Prices
for these are presented in Section 10.4.)
10.3.3.3 Ductwork Pressure Drop
As mentioned in Section 10.3.1, ventilation system energy losses due to friction are traditionally
computed as fractions of the velocity pressure, VP. In most cases, equation 10.12 can be used to
estimate these losses. Technically, though, these equations apply only to those regions in the
ventilation system where there are no changes in the velocity pressure (i.e., where the duct diameter
is constant). These regions would include straight duct, hoods, and such fittings as couplings and
simple elbows. But, with tees, wyes, and other divided flow fittings, the velocity — and velocity
pressure — are not constant between the fitting inlet and outlet. The corresponding friction loss (Fb)
is a function of both the upstream (inlet) and branch VPs, as the following equation indicates:40
Fb = VPu(kb-l) + VPb (10.28)
where
VPU, VPb = upstream and branch velocity pressures, respectively (in. w.c.)
kb = branch loss coefficient
However, divided flow fittings generally are not used with simple pollution control ventilation
systems, except in those cases where a tee might be needed, say, for purposes of adding dilution air.§
Divided flow fittings are needed with more complex control systems that collect waste gases from several emission points. The design of sue
k jtion systems is beyond the scope of this chapter, however.
10-31
-------
As any fluid mechanics textbook would attest, the friction loss for ductwork is a complex
function of several variables: duct diameter and length, transport velocity, and gas viscosity and
density. Specifically, the Darcy-Weisbach and Colebrook Equations are typically used to make this
calculation, the latter being used to compute the Reynolds number.41 Traditionally, the friction loss
has been obtained from a nomograph or, more recently computer programs. A typical nomograph
is found in Burton.42 Also, to simplify the calculation, empirical equations have been derived for
certain kinds of commercially-available ductwork. For instance, to estimate the friction loss per 100
ft (F^lOO ft) at standard conditions for round, spiral, galvanized ductwork having 10 joints per 100
ft, use the following equation:43
•vl.8
ft = 0.136(1/IV 18(u/l,000)
where: Dd = duct diameter (ft), and: 0.25 < Dd < 5
(10.29)
Clearly, this equation provides the total friction loss, not the loss factor (k). However, the reader
may compute k for a given diameter (Dd) and flow rate (Q) by simply dividing the equation 10.29
results by VP and multiplying by 100.
To estimate the friction loss for other duct materials, multiply the value from equation 10.29 by
a roughness correction factor, approximate values of which are:44
Material
Roughness Correction Factor
Non-spiral-wound galvanized
Fiberglass (smooth finish)
ABS and PVC plastic
Concrete
Corrugated flex duct
0.9
0.8
0.8
1.4
2.3
Loss factors for fittings have also been compiled, based on experimental data. Mainly of interest
are those for 90° elbows, arguably the most commonly used fitting in air pollution control systems.
The "k90" values for elbows vary according to the diameter and radius of curvature, which is
expressed as a multiple of the elbow diameter. Typical ranges of these values are as follows:45
Radius of Curvature
Friction Loss Factor
0.50
1.00
1.25
1.50
2.00
2.50
0.80
0.35
0.30-0.55
0.27 - 0.39
0.24-0.27
0.22-0.24
10-32
-------
As these values indicate, the higher the radius of curvature, the lower the friction loss. This
stands to reason, as the higher the radius of curvature, the more gradually the gas stream changes
direction. For an elbow having of angle less than 90°, multiply the above k^ value by an
adjustment factor (6/90), so that:
ke = (6/90)k90 (10.30)
where
ke = loss factor for 0 < 90°
Illustration: A control device at a cosmetic factory is connected to a source by 250 feet of round
spiral duct. The duct run includes three 90° elbows and two 45° elbows, each with a 1.50 radius
of curvature. The volumetric flow rate (Q) of the waste gas (which contains entrained face powder)
is 15.000 ft3/min at standard conditions. Calculate the friction loss for the ductwork.
Solution: Because the material being conveyed in the ductwork (face powder) is light, an
appropriate transport velocity (i\) in this case is 2,000 ft/min. (See text table above.) Upon
substituting this value and the volumetric flow rate into equation 10.27 we obtain the duct diameter
Dd = 1.128(15,000/2,000)05 = 3.09 ft
Next, substitute the diameter and velocity into equation 10.29 to compute the straight duct friction
(static pressure) loss, Fd:
Fd = 0.136(l/3.09)118(2,000/l. ,000)' "(250/100)
= 0.313 in. w.c.
The 250/100 factor in this expression adjusts the friction loss from 100 feet (the basis of equation
10.29) to 250 feet (the length of the duct system in this illustration). The rest of the friction loss
occurs through the five elbows (three 90°, two 45°), each with a 1.50 radius of curvature. These
losses (Fc) are computed via equation 10.12:
F, = k«VP
c
where
VP = (2,000/4,016)2 (equation 10.11, rearranged)
= 0.248 in. w.c.
For the 90° elbows, ke = kc,0 = 0.33 (average of table range), and:
Fc = 3 x 0.33 (0.248) = 0.246 in. w.c.
For the 45° elbows, k0 = (45/90)k90 = 0.165 (equation 10.30), and:
10-33
-------
Fc = 2 x 0.165(0.248) = 0.0818 in. w.c.
The total friction loss is, therefore:
F = 0.313 + 0.246 + 0.0818 = 0.641 in. w.c.
* * *
From this illustration, two observations may be made: (1) the static pressure loss through the
straight duct is not large, even at this length (250 ft.) and (2) the losses through the elbows—which
total 0.328 in. w.c.—are larger than the straight duct loss. Though it may be tempting to neglect
fittings losses for the sake of expediency, doing so can cause a significant underestimation of the
ventilation system static pressure loss.
10.3.4 Stack Design Procedures
As with ductwork, the design of stacks involves a number of stream, structural, and site-specific
parameters.46'47 These include:
Waste gas variables: inlet volumetric flow rate, temperature, and composition;
Site-specific data: elevation above sea level, ambient temperature fluctuations, topographic and
seismic data, meteorological records, and building elevations and layout;
Structural parameters: thickness of stack wall and liner, location of breaching opening, type of
supports, load capacity of foundation, modulus of resistance, and natural vibration frequency.
Fortunately, for study cost-estimating purposes, the only two stack design parameters that need
to be determined are: (1) the stack diameter and (2) the stack height. The other variables (e.g.. wall
thickness) are incorporated into the equipment cost correlations. The stack diameter is relatively
easy to determine, as it depends primarily on waste stream conditions. The stack height is more
difficult to arrive at, as it is influenced by several site-specific variables. Nonetheless, ample
guidance has been developed to allow the estimator to determine an acceptably accurate stack
height.
10.3.4.1 Calculating Stack Diameter
Because most stacks have circular cross-sections, the stack diameter (Ds, ft) can be calculated via
the duct diameter formula (equation 10.27):
Ds = 1.128(Qc/uc)m (10.31)
10-34
-------
where
uc = stack exit velocity (ft/min)
Qc = exit volumetric flow rate (actual ftVmin)
It should be noted that the stack diameter in this formula is measured at the stack exit, not at the
entrance. That is because, for structural reasons, the diameter at the bottom of the stack typically
is larger than the top diameter. Also note that the stack exit velocity does not necessarily equal the
duct transport velocity. Finally, Qc may be different from the volumetric flow rate used to size the
ductwork. Because the stack always follows the control device, the flow rate entering the device
may not equal the flow rate entering the stack, either in standard or actual ft3 /min terms. For
instance, in a thermal incinerator, the outlet standard waste gas flow rate is almost always higher
than the inlet flow rate due to the addition of supplemental fuel.
The stack exit velocity, uc, affects the plume height, the distance that the plume rises above the
top of the stack once it exits. In a well-designed stack, uc should be 1.5 times the wind speed.
Typically, design exit velocities of 3,000 to 4,000 ft/min are adequate.48 This range corresponds to
wind speeds of 34 to 45 mi/hr.
10.3.4.2 Calculating Stack Height
Estimating the stack height is more difficult than calculating the stack exit diameter. The stack
height depends on several variables: the height of the source; the stack exit velocity; the stack and
ambient temperatures; the height, shape, and arrangement of the nearby structures and terrain; and
the composition of the stack outlet gas. Some of these variables are straightforward to determine,
while others (such as the dimensions and layout of nearby structures) are difficult to determine
without performing on-site modeling and monitoring studies.
This height has two components: the height of the stack itself (Hs) and the plume rise height
(Hpr). Together these components comprise the effective stack height (He). That is:
He = Hs + Hpr (10.32)
However, the cost of the stack is a function of Hs alone. (See Section 10.4.) As discussed
above, the plume rise is a function of the stack exit velocity. It also depends on the temperature
differential between the stack gas and the ambient air. Specifically, a 1 °F temperature difference
corresponds to approximately a 2.5-ft. increase in Hp,..49
For those sources subject to State Implementation Plans (SIPs), the stack height (Hs) should be
determined according to "good engineering practice" (GEP). GEP is defined as "the height
necessary to insure that emissions from the stack do not result in excessive concentrations of any
air pollutant in the immediate vicinity of the source as a result of atmospheric downwash, eddies,
or wakes which may be created by the source itself, nearby structures, or nearby terrain obstacles."50
In this respect, GEP establishes the maximum allowable stack height credit for purposes of
calculating the ambient air quality impact of the emitting source. A source may build a stack to any
10-35
-------
height, but only a certain amount of stack height will be allowed in determining environmental
impacts.51
For stacks constructed after January 12, 1979, the GEP stack height shall be the greater of: (1)
65 meters (213 ft); (2) the height demonstrated by an approved fluid model or field study that
ensures that stack emissions do not cause excessive pollutant concentrations from atmospheric
downwash, wakes, eddy effects., etc; or (3) the height determined by the following equation:52
Hs = Hb+1.5L (10.33)
where
Hs = GEP stack height, measured from the ground level elevation at the stack base (ft)
Hb = height of nearby structure(s) measured from this ground level elevation (ft)
L = lesser dimension (height or projected width of nearby structure(s))
10.3.4.3 Calculating Stack Draft
As discussed previously, waste gas flowing through hoods and ductwork loses static pressure due
to friction. In the case of stacks, however, the gas stream can actually gain static pressure, as a
result of stack draft, which is the draft created by the stack gas-ambient air temperature differential.
Stack draft (SPS, in. w.c.) can be calculated as follows:53
SPS = 0.034(HS - Hbr)7t(VTamb - 1/TJ (10.34)
where
Hbr = height of stack breaching (inlet duct connection) above stack base (ft)
TC= barometric pressure(in. w.c.)
T^b = ambient temperature (°R)
Tsa = average stack gas temperature (°R)
Illustration: The waste gas from a thermal incinerator has an outlet flow rate and temperature of
21,700 actual ftVmin. and 550°F, respectively. The maximum wind speed in the vicinity is 42
mi/hr, while the stack exit and ambient temperatures are 450°F and 70°F, in turn. The barometric
pressure is 1 atm. (29.92 in. Hg). The incinerator is near a 35-ft tall brick building, while the
"projected width" of an adjacent building is 40 ft. For a stack to disperse the incinerator offgas,
calculate the required: (1) exit velocity, (2) diameter, (3) height, and (4) draft.
Solution:
Exit velocity: According to the above guideline, the velocity should be 1.5 times the wind speed,
or:
uc = 1.5 x 42 mph x 88 fpm/mph = 5,540 ft/min.
10-36
-------
Stack diameter: The exit volumetric flow rate is measured at the stack exit temperature, namely
450°F. However, the above flow rate was measured at 550°F, the incinerator outlet temperature.
Correcting to the stack exit temperature, we obtain:
Qc = 21,700 x (450 + 460) / (550 + 460) = 19,600 actual ftVmin.
Substituting this value into equation 10.31:
Ds = 1.128(19,600/5,540)* = 2.12 ft.
Stack height: As a first approximation, estimate the GEP stack height from equation 10.33, where
the variables Hb and L are 35 ft and 40 ft, respectively:
Hs = 35+ 1.5(40) = 95 ft.
Clearly, this Hs is less than the GEP maximum height (213 ft), so it will be used in this example.
Stack draft: All of the inputs needed to compute the stack draft via equation 10.34 are known
except the stack breaching height, Hbr. However, a minimum of 5 ft is recommended for this
parameter.54 This value will be used in this calculation. Also, the average stack temperature is:
T^ = (450 + 550)72 + 460 = 960 °R.
Finally, the barometric pressure expressed in inches of water is:
7i = 29.92 in. Hg x 13.6 in. water/in. Hg = 407 in. w.c.
Upon substitution, we obtain:
SPS = 0. 034 (118 - 5)(407) (1/[70 + 460] - 1/960) = 1.32 w. c.
10.4 Estimating Total Capital Investment
This section presents the information needed for estimating the total capital investment (TCI) for
hoods, ductwork, and stacks. The TCI includes the equipment cost (EC) for the hood, ductwork,
or stack; taxes; freight charges; instrumentation (if applicable); and direct and installation costs. All
costs are presented in second quarter 1993 dollars, and are of "study" estimate accuracy (± 30
percent).* Moreover, the costs are for new facility installations; no retrofit costs are included.
For information on escalating these prices to more current dollars, refer to the EPA report Escalation Indexes for Air Pollution Control Costs am
is thereto, all of which are installed on the OAQPS Technology Transfer Network (CTC Bulletin Board).
10-37
-------
The equipment costs are presented in Section 10.4.1, while the installation costs are shown in
Section 10.4.2. In each of these sections, the three categories of equipment are covered in separate
subsections.
10.4.1 Equipment Costs
Several vendors provided costs (prices) for each of the three equipment categories. Their responses
reflected a range of sizes, designs, and materials of construction. These prices have been correlated
against some easy-to-determine design (sizing) parameter via least-squares regression analysis.
Each of these correlations pertains to a certain type of equipment (e.g., circular canopy hoods)
within a specified size range of the parameter in question (e.g., 2 to 200 ft2 inlet area). For that
reason, a cost correlation should not be extrapolated outside the parameter range specified.
Some of the prices the vendors provided pertain to stock ("off-the-shelf") items, while other
costs are for custom-fabricated equipment. Vendors tend to specialize in either stock or custom
items. Most hoods and stacks are custom-made, either fabricated in the vendor's factory or erected
on-site. Conversely, ductwork components usually are stock items, though larger pieces have to be
custom-made. (Of course, there are exceptions to this.) Finally, all prices given in the following
section are "F.O.B. (free-on-board) vendor," meaning that they include neither freight nor taxes.
10.4.1.1 Hood Costs
In all, four vendors provided prices for hoods.55 These prices covered the following types of hoods:
• Canopy—circular
• Canopy—rectangular
• Push-pull
10-38
-------
Table 10.5: Parameters for Hood Cost Equation*
Type of
Hood
Canopy-
circular
Canopy-
rectangular
Push-pull
Side-draft
Backdraft
(slotted)
Backdraft
(slotted)
Backdraft
(slotted)
Backdraft
(slotted)
Backdraft
(slotted)
Fabrication
Material
FRPf
FRP
FRP
FRP
PVC*'*
PVCn
pptt
FRP
Galvanized
Steel
Equation Parameter
a
123
294
595
476
303
789
645
928
688
b
0.575
0.505
0.318
0.332
1.43
0.503
0.714
0.516
0.687
Equation
Range
(A, ft2)
2-200
2-200
2-200
2-200
0.6-2.0§§
1.1-2.1
1.1-2.1
1.1-2.1
0.5-1.3
5 Based on data received from hood vendors. (See Reference 55.)
+ Fiberglass-reinforced plastic.
* Polyvinyl chloride.
• Each hood is equipped with two rows of slots, but no dampers.
51 For each slotted hood, "equation range" denotes the range in the area of the slot openings, which is much less than
the total hood face area.
n Each hood is equipped with manual slot dampers and four rows of slots.
** Polypropylene.
10-39
-------
• Slide-draft
• Back-draft (slotted)
Descriptions and design procedures for these hoods are given in Sections 10.2.1 and 10.3.2,
respectively. As explained in Section 10.3.2, hood costs have been found to correlate well with the
hood inlet or face area (Af, ft2). Furthermore, the functional form that best fits the cost-face area
correlation (equation) is the "power function," or:
Q = aA," (10.35)
where
Ch = hood cost ($)
a,b = equation regression parameters
The values of the equation parameters vary according to hood type and material of construction.
These parameters are shown in Table 10.5.
Illustration: What would be the cost of the electroplating tank canopy hood sized for the
illustration in Section 10.2.2.? Assume that the hood is fabricated of FRP.
Solution: Recall that the face area (Af) calculated for that hood was 98.5 ft2. Because this is a
circular canopy hood, the equation parameters from Table 10.5 are: a = 123 and b = 0.575. (Note
that this hood area falls within the equation range of 2 to 200 ft2.) Substituting these parameters into
equation 10.35, we obtain:
Ch= 123(98.5)a575 = $1,720.
10.4.1.2 Ductwork Costs
Several vendors provided ductwork prices, also for a range of sizes, materials, and designs.56 These
prices covered the following equipment items:
• Straight ductwork:
— Circular
- Steel sheet (galvanized carbon, w/ & w/o insulation; 304 stainless;)
- Steel plate (coated carbon; 304 stainless)
- Plastic (FRP; PVC)
— Square
- Steel (aluminized carbon; w/ & w/o insulation)
• Elbows (90°):
— Steel (galvanized carbon, w/ & w/o insulation; 304 stainless)
— Plastic (FRP; PVC)
10-40
-------
• Dampers:
— Butterfly
- Steel (galvanized carbon, w/ & w/o insulation)
- Plastic (FRP; PVC, w/ & w/o actuators)
— Louvered
- Steel (aluminized carbon w/ & w/o actuators)
— Blast gate
- Steel (carbon)
- PVC
These prices were regressed against the diameter of the equipment item (straight duct, elbow,
or damper). The regression correlations were of three forms: power function (primarily),
exponential, and linear. Equation 10.35 depicts the power function, while the other forms are:
Exponential: C, = aebD (10.36)
Linear: C, = a + bD (10.37)
where
C, = cost of equipment item in question
a,b = regression parameters
The regression parameters are listed in Tables 10.6 to 10.8, along with the size applicability
ranges for the respective correlations. (Note: The correlations should not be extrapolated outside
these ranges.) The following paragraphs contain additional information about the price data and
the correlations:
Straight duct: As indicated above, vendors provided prices for steel plate, steel sheet (spiral-
wound and longitudinal seam), and plastic straight duct. The major difference between the two steel
duct types lies in the wall thickness. Steel plate duct typically has wall thicknesses of 3/16 in. to 1/2
in., while steel sheet duct wall thicknesses usually range from 28 gauge to 10 gauge. As Table 10.4
shows, this range corresponds to wall thicknesses of 0.0149 in. to 0.1406 in., respectively, although
the exact thicknesses will vary with the type of steel used (e.g., carbon vs. stainless). Also, as
discussed in Section 10.3.3.2, each duct diameter can be fabricated with a range of wall thicknesses.
Most of the steel duct vendors supplied prices for a minimum and a maximum wall thickness
for a given diameter. However, to simplify matters for cost estimators, these "low" and "high"
prices first were averaged, and then the average prices were regressed against the diameters. This
averaging was necessary, because those making study cost estimates usually do not have enough
information available to predict duct wall thicknesses.
Prices for both circular and square insulated steel sheet duct were among the data received. The
insulated circular steel duct is "double-wall, spiral-wound" in construction, wherein the insulation
10-41
-------
Table 10.6: Parameters for Straight Steel Ductwork Cost Equations§
Duct
Type
Circujar-
spiral
Circular-
spiral
Circular-
spiral
Circular-
spiral
§§
Circular-
longitudinal
Circular-
longitudinal
Circular-
longitudinal
Circular-
longitudinal
Square
Square
Material
Sheet-
galv CS*
Sheet-
304 SS'
Sheet-
galvCS
Sheet-
galvCS
Sheet-
galvCS
Sheet-
304 SS
Plate-
coat CS**
Plate-
304 SS**
Sheet-
alum CS"
Sheet-
alum CS
Insulation
Thickness
(in.)
None
None
1
3
None
None
None
None
None
4
Equation
Type
Power
function
Power
function
Power
function
Power
function
Power
function
Power
function
Power
function
Power
function
Linear
Linear
Equation
Parameter
a
0.322
1.56
1.55
2.56
2.03
2.98
2.49
6.29
0.254
21.1
b
1.21
1.00
0.936
0.937
0.784
0.930
1.15
1.23
2.21
5.81
Equation
Range
(D,in.)
3-84
3-84
3-82
3-82
6-84
6-84
6-84
6-84
18-48
18-48
1 Based on data from ductwork vendors. (Reference 56.)
T Spiral-wound and welded circular duct.
* Galvanized carbon steel sheet.
• 304 stainless steel sheet.
H Circular duct welded along the longitudinal seam.
n Carbon steel plate with one coat of "shop paint."
** 304 stainless steel plate.
• • Aluminized carbon steel sheet.
10-42
-------
Table 10.7: Parameters for Steel Elbows and Dampers Cost Equations§
Ductwork
Item
Elbows*
Elbows
Elbows-
insulated^
Dampers-
butterfly1"1'
Dampers- . .
bunerflv /insulated
Dampers-
louvered"
Dampers-
louvered w/actuators
Dampers-
blast gates
Material
Galv CS*
304 SS
GalvCS
GalvCS
GalvCS
Alum CS§§§
AlumCS
Carbon steel
Equation
Type
Exponential
Exponential
Exponential
Exponential
Exponential
Power function
Power function
Power function
Equation
Parameter
a
30.4
74.2
53.4
23.0
45.5
78.4
208.
17.2
b
0.0594
0.0668
0.0633
0.0567
0.0597
0.860
0.791
0.825
Equation
Range
(D, in.)
6-84
6-60
3-78
4-40
4-40
18-48
18-48
3- 18
s Based on data received from ductwork vendors. (See Reference 56.)
f Single-wall "gored" 90° elbows, uninsulated.
* Galvanized carbon steel sheet.
88 Double-wall "gored" 90c elbows with 1-inch fiberglass insulation.
Single-wall "opposed blade" type manual butterfly dampers.
** Double-wall "opposed blade" butterfly dampers with 1-inch fiberglass insulation.
• • Louvered dampers with 95-98%- sealing.
888 "Aluminized" carbon steel sheet.
tn Louvered dampers with electric actuators (automatic controls).
10-43
-------
Table 10.8: Parameters for Plastic Ductwork Cost Equations5
Ductwork
Item
Straight duct
Straight duct
Elbows-90°
Elbows-90°
Dampers-
butterfly
Dampers-
butterfly
Dampers-
butterfly w/actuators
Dampers-
blast gate
Material
PVC*
FRP*
PVC
FRP
PVC
FRP
PVC
PVC
Equation
Type
Power function
Exponential
Power function
Exponential
Power
function
Power
function
Exponential
Power
function
Equation
Parameter
a
0.547
11.8
3.02
34.9
10.6
35.9
299.
8.14
b
1.37
0.0542
1.49
0.0841
1.25
0.708
0.0439
1.10
Equation
Range
(D, in.)
6-48
4-60
6-48
4-36
4-48
4-36
4-48
4-48
8 Based on data received from ductwork vendors. (See Reference 56.)
+ Polyvinyl chloride.
* Fiberglass-reinforced plastic.
•Butterfly dampers with pneumatic actuators (automatic controls). All other dampers listed in this table are manually
controlled.
10-44
-------
is installed between the inner and outer walls. Costs were provided for both 1-in. and 3-in.
fiberglass insulation thicknesses. For the square duct, prices were given for a 4-in. thickness of
mineral wool insulation appb'ed to the outer surface of the duct. The correlation parameters in Table
10,6 reflect these specifications.
Prices for both carbon steel (galvanized, painted, or aluminized) and 304 stainless steel duct
were received. The carbon steel duct is used in situations where "mild" steel is suitable, while the
stainless steel duct is required whenever the gas stream contains high concentrations of corrosive
substances.
Vendors gave prices for plastic (FRP and PVC) duct also (Table 10.8). However, for a given
diameter this duct is fabricated in a single wall thickness, which varies from approximately 1/8 in.
to 1/4 in. Consequently, the estimator is not required to select a wall thickness when costing plastic
duct.
Elbows: Prices for steel sheet and plastic 90° elbows were also submitted. The steel sheet elbows
were "gored" (sectioned) elbows fabricated from five pieces of sheet metal welded together. Like
the straight duct, the steel elbows were priced at two wall thicknesses: "minimum" and "maximum".
These prices were averaged before being regressed against the elbow diameter. Prices were also
given for both galvanized carbon steel elbows (with and without 1-in. fiberglass insulation) and 304
stainless steel elbows. Correlation parameters for steel elbows are listed in Table 10.7.
Costs for both PVC and FRP 90° elbows were also given. The PVC ells were fabricated from
three sections ("three-piece miter"), while the FRP elbows were one-piece molded fittings. As with
the plastic straight duct, each elbow of a given diameter was fabricated in a single wall thickness.
Table. 10.8 contains correlation parameters for plastic elbows.
Dampers: Prices were obtained for three types of dampers: butterfly, louvered, and blast gates.
The galvanized carbon steel butterfly dampers were priced with and without 1-in. fiberglass
insulation, while prices for the aluminized carbon steel louvered dampers were based on either
manual or automatic control (via electric actuators). Similarly, the PVC butterfly dampers were
manual or equipped with pneumatic actuators. Both the carbon steel and the PVC blast gates were
manual. Correlation parameters for the steel and plastic dampers are shown in Tables 10.7 and 10.8,
in turn.
Illustration: A fabric filter handling 16,500 ftVmin of 200°F waste gas laden with noncorrosive
cocoa dust is located 95 ft across from and 20 ft above, the emission source (a drying oven).
Straight duct with four 90° elbows (all fabricated from spiral-wound, galvanized carbon steel sheet)
and a butterfly damper (also galvanized CS) will be required to convey the gas from the source to
the control device. Assume that the ductwork contains 1-in. thick insulation to prevent
condensation. Estimate the cost of these items.
Solution: First, determine the diameter of the straight duct, elbows, and damper. From Table 10.3,
the minimum transport velocity (uj for cocoa dust is 3,000 ft/min. Substituting this value and the
gas volumetric flow rate into equation 10.27, we obtain:
10-45
-------
Dd = 1.128(16,500/3,000)% = 2.65 ft = 31.7 in.
Next, obtain the costs of the ductwork items as follows:
Straight ductwork: From Table 10.6, select the equation parameters for galvanized circular spiral-
wound duct (1-in. insulation) and substitute them and the diameter into the appropriate equation type
(power function, equation 10.35).
Straight duct cost ($/ft) = 1.55(31.7)°'936 = $39.4/ft.
However, a total of 115 ft (95 + 20) of duct is required, so:
Straight duct cost = $39.4/ft x 115 ft = $4,531.
Elbows: The Table 10.7 correlation parameters for galvanized
carbon steel, insulated elbows are 53.4 (a) and 0.0633 (b). However, the regression correlation form
is exponential (equation 10.36). Thus:
Elbow cost ($) = 53.4e°'0633(317) = $397 ea.
For four elbows, the cost is: $397 x 4 = $1,588.
Damper: Also from Table 10.7, select the correlation parameters for galvanized carbon steel
"dampers-butterfly/insulated" and substitute into equation 10.36:
Damper cost ($) = 45.5e°0597(3! 7) = $302.
After summing the above three costs, we obtain:
Total ductwork cost = $6,421 ~ $6,420.
10.4.1.3 Stack Costs
Prices for steel and PVC short stacks were obtained from four vendors.57 The steel stack costs were
for those fabricated from carbon and 304 stainless steels, both plate and sheet metal. As with
ductwork, the difference between steel sheet and plate lies in the thickness. For these stacks, the
sheet steel thickness ranged from 18 to 16 gauge (0.05 to 0.06 in., approximately). Steel plate
thicknesses were considerably higher: 0.25 to 0.75 in, a fact that makes them more resistant to wind
and other loadings than stacks fabricated of steel sheet. This is especially true for taller stacks. The
major drawback is that plate steel stacks are more costly than those fabricated from steel sheet.
Another feature that increases costs is insulation. As the correlation parameters show (Table
10.9), insulated stacks cost as much as three times more per foot than uninsulated. With or without
insulation, a typical short (15-ft) steel stack consists of the following components:58
• Longitudinal seam duct (12-ft section)
10-46
-------
Table 10.9: Parameters for Stack Cost Equations5
Material
pvc§§
Plate-coated CSn
Plate-304 SS**
Sheet-galv CS "
Sheet-304 SS w
Sheet-insul CS/DW w
Sheet-uninsul CS/DW ***
Sheet-insul CS/DW*"
Equation Parameter*
a
0.393
3.74
12.0
2.41
4.90
143.
10.0
142.
b
1.61
1.16
1.20
1.15
1.18
0.402
1.03
0.794
Equation Ranee
D. (in)*
12-36
6-84
6-84
8-36
8-36
18-48
18-48
24-48
ri (ft)'
<10
20-100
20-100
<75
<75
<15
<15
30-75
1 Based on data received from ductwork vendors. (See Reference 54.)
f All cost equations are power functions. (See equation 10.35.) Except where noted, costs are expressed in terms of
$/ft of stack height.
* Stack diameter range to which each equation applies.
• Stack height range to which each equation applies.
11 Polyvinyl chloride.
n Carbon steel plate with one coat of "shop paint."
** 304 stainless steel plate.
• • Galvanized carbon steel sheet.
111 304 stainless steel sheet.
m Aluminized carbon steel sheet covered with 4 inches of fiberglass insulation (double-wall construction).
*** Uninsulated aluminized carbon steel sheet (double-wall construction).
1 Costs for these stacks are expressed in $, and are correlated with the stack surface area. (S,f ft2).
10-47
-------
• Reducer fitting (3-ft)
• Drip pan
• Support plate (1/4-in, welded to stack)
• Rectangular tap (for connecting to fan discharge)
• Ring (for attaching guy wires)
Taller stacks may require additional components, such as ladders and platforms, guy wires or other
supports, and aircraft warning lights. (See Section 10.2.3.) TablelO.9 lists the parameters and
applicable ranges of the stack cost correlations. The correlations cover short PVC stacks, and taller
stacks fabricated from plate steel (carbon and 304 stainless types) and sheet steel (insulated and
uninsulated). Except for three double-wall sheet steel designs, these stacks are of single-wall
construction. Note that all of the correlations are power functions. Also note that the equations
apply to various ranges of stack height. In all but one of these equations the cost is expressed in $/ft
of stack height. The exception is the cost equation for insulated carbon steel sheet stacks of heights
ranging from 30 to 75 feet. In this equation the cost is expressed in $.
This last cost equation is different in another respect. The other six equations in Table 10.9
correlate stack cost ($/ft) with stack diameter (Ds, in.). However, this seventh equation correlates
stack cost with stack surface area (Ss, ft2), a variable that incorporates both the stack diameter and
the stack height (Hsft.). The surface area is calculated via the following equation:
Ss = (7i/12)DsHs (10.38)
where
1/12 = stack diameter (Ds) conversion factor
Illustration: Estimate the cost of the stack sized in the Section 10.3.4.3 illustration.
Solution: Recall that the stack dimensions were: Hs = 95 ft and D = 2.12 ft = 25.4 in. Both
dimensions fall within the ranges of the cost correlations for steel plate stacks. Because the previous
illustration did not indicate whether the waste gas was corrosive, we will estimate the prices for both
carbon steel and 304 stainless steel plate stacks.
Upon substituting the equation parameters and stack dimensions into equation 10.35, we obtain:
Price (carbon steel) = 3.74 (25.4)116 ($/ft) x 95 ft
= $15,100.
Price (304 stainless) = 12.0 (25.4)120 ($/ft) x 95 ft
= $55,300.
Notice that the price of the stainless steel stack is nearly four times that of the carbon steel stack.
In view of this difference, the estimator needs to obtain more information on the waste gas stream
10-48
-------
properties, so that he/she can select the most suitable stack fabrication material. Clearly, it would be
a poor use of funds to install a stainless steel stack where one is not needed.
10.4.2 Taxes, Freight, and Instrumentation Costs
Taxes (sales, etc.) and freight charges apply to hoods, ductwork, and stacks, as they do to the control
devices that these auxiliaries support. As discussed in Chapter 2, these costs vary, respectively,
according to the location of the ventilation system and the site's distance from the vendor. Typical
values are 3% (taxes) and 5% (freight) of the total equipment cost.
Unlike the control devices, ventilation systems generally are not instrumented. The exception
would be an electric or pneumatic actuator for a butterfly or louvered damper. In such a case,
however, the cost of the instrument (actuator and auxiliaries) would be included in the damper price.
Thus, no supplementary instrumentation cost is included.
10.4.3 Purchased Equipment Cost
With ventilation systems, the purchased equipment cost (PEC,) is the sum of the equipment, taxes,
and freight costs. Incorporating the typical values listed in Section 10.4.2, we obtain:
PECt = ECt + 0.03EQ + 0.05EC, = 1.08 (EQ (10.39)
where
EC, = total cost of hood(s), ductwork, and stack(s)
10.4.4 Installation Costs
When making a cost estimate for an air pollution control system according to the procedure in this
manual, the estimator first determines the cost of the control device, then estimates the costs of such
auxiliaries as the hood, ductwork, stack, fen and motor, and other items. To these items he/she adds
the costs of instrumentation, taxes, and freight, to obtain the PEC. Finally, the estimator multiplies
the PEC by the installation factor appropriate to the control device (e.g., 2.20 for gas absorbers) to
obtain the total capital investment. In these cases, the installation factor incorporates all direct and
indirect costs needed to install and start up the control system equipment, including, of course, the
hood, ductwork, and stack. (See Chapters 3 to 9 for more information about these factors.)
For this reason, it usually is unnecessary to estimate the installation cost of the ventilation system
separately. However, there may be occasions where a hood, a stack, or ductwork has to be installed
10-49
-------
alone, either as replacement equipment or to augment the existing ventilation system. In those
instances, the estimator may want to estimate the cost of installing this item.
As might be imagined, these installation costs vary considerably, according to geographic
location, size and layout of the facility, equipment design, and sundry other variables. Nonetheless,
some of the vendors (and a peer reviewer59) provided factors for hoods and ductwork, which, when
multiplied by their respective purchased equipment costs, will yield approximate installation costs.
These are:
• Hoods: 50 to 100%
• Ductwork: 25 to 50%
If one or both of the latter factors is used, the total capital investment (TCI) of the hood and/or
ductwork would be:
= (l+rFh/d)xPECh/u (10.40)
where
IFh/d = installation factor for hood(h)/ductwork(d)
PECh/d = purchased equipment cost of hood (h)/ductwork (d)
10.5 Estimating Total Annual Cost
10.5.1 Direct Annual Costs
Ventilation systems incur few, if any, direct annual costs, as they function to support control devices.
There are no costs for operating or supervisory labor, operating materials, or waste
treatment/disposal allocated to ventilation systems. Maintenance costs would also be minimal, except
for such minor expenses as painting, insulation repair, or calibration of automatic damper controls.
The only utilities cost would be the incremental electricity needed for the waste gas stream to
overcome the static pressure loss in the hood, ducting, and stack.13 The incremental electricity cost
(Cc, $/yr) can be calculated as follows:
Cc = (1.175 x 1O-VQFje/e (10.41)
where
Technically, this direct annual cost should be allocated to the ventilation system fan, not to the hood, ductwork, and stack. The fen power cost equation
will be included in the Manual chapter on fens. However, as the fans chapter has yet to be written, this equation has been provided as a temporary
kLence to Manual users.
c^^er
10-50
-------
pc = electricity price ($/kwh)
Q = waste gas flow rate (actual fWmin)
F = static pressure drop through ventilation system (in. w.c.)
0 = operating factor (hr/yr)
e = combined fan-motor efficiency
Illustration: In the cosmetic factory ventilation system illustration above (Section 10 .3), what
would be the cost of the electricity consumed by the fan needed to convey the e nrough the
ductwork? Assume an electricity price of $0.075/kwh, a combined fan-motor effic y of 0.6, and
an 8,000-hr/yr operating factor.
Solution: Recall that the pressure drop and gas flow rate for this illustration were 0.313 in. w.c. and
15,000 actual ftVmin, respectively. Upon substituting these values and the other parameters into
equation 10.40, we obtain:
Cc = (1.175 x 1Q-4) (0.075) (15,000) (0.313) (8,000)70.6 = $552/yr.
10.5.2 Indirect Annual Costs
The indirect annual costs for ventilation systems include property taxes, insurance, general and
administrative (G&A), and capital recovery costs. (Overhead—a fifth indirect annual cost—is not
considered, because it is factored from the sum of the operating, supervisory, maintenance labor and
maintenance materials costs, which is negligible.) When a ventilation system is part of a control
system, these costs are included in the control system indirect annual cost. However, if the ventilation
equipment has been sized and costed separately, these costs can be computed from the total capital
investment (TCI) via standard factors, as follows:
Indirect Annual Cost
Property taxes
Insurance
General and Administrative
Capital Recovery
Computation Equation
0.0 Ix TCI
0.01 x TCI
0.02 x TCI
CRF x TCI
The "CRF" term in the capital recovery equation is the capital recovery factor, which is a function
of the economic life of the ventilation system and the interest rate charged to the total capital
investment. (See Section 2.3 of this manual for more discussion of the CRF and the formula used for
computing it.)
For a ventilation system, the economic life varies from at least 5 to 10 years to 15 to 20 years or
more.60'61 In general, the ventilation equipment should last as long as the control system it supports.
As discussed in Section 2.3, the interest rate to use in the CRF computation should be a "pre-tax,
marginal (real) rate of return" that is appropriate for the investor. However, for those cost analyses
10-51
-------
related to governmental regulations, an appropriate "social" interest (discount) rate should be used.
For these kinds of analyses, the Office of Management and Budget (OMB) directs that a real annual
interest rate of 7% be used.62 (This replaces the 10% rate OMB previously had mandated.)
10.5.3 Total Annual Cost
The total annual cost (TAG) is calculated by adding the direct (DC) and indirect (1C) annual costs:
TAG = DC + 1C (10.42)
10.6 Acknowledgements
Several firms and individuals provided very useful technical and cost information to this chapter.
Foremost among these was Todd N. Stine of United McGill Corporation (Raleigh, NC), who
submitted current prices for a variety of ductwork items, as well as a comprehensive product catalog
and engineering design manual. In addition, Mr. Stine patiently replied to the author's many
questions, providing supplemental data when requested. Samir Karandikar of EPCON Industrial
Systems (The Woodlands, TX) and Gregory P. Michaels of Piping Technology & Products (Houston,
TX) also were very helpful in submitting data and responding to inquiries.
The author also would like to thank the following firms for their valuable contributions:
• Air Plastics, Inc. (Mason, OH)
• General Resource Corporation (Hopkins, MN)
• Harrington Industrial Plastics, Inc. (Chino, CA)
• Intellect Systems & Marketing, Inc. (Bohemia, NY)
• Wer-Coy Metal Fabrication Co. (Warren, MI)
In addition, several individuals reviewed the draft chapter and provided valuable suggestions,
supplemental information, or both. The EPA peer reviewers, all located at Research Triangle Park,
NC, were:
• James C. Berry (OAQPS/ESD)
• Peter A. Eckhoff (OAQPS/TSD)
• Norman Kaplan (ORD/AEERL)
10-52
-------
• James H. Maysilles (OAQPS/ESD)
• Larry Sorrels (OAQPS/ESD)
Finally, Howard Goodfellow of Goodfellow Consultants, Inc. (Mississauga, Ontario, Canada) also
reviewed the chapter and supplied helpful comments.
References
[1] Goodfellow, H.D. "Ancillary Equipment for Local Exhaust Ventilation Systems". In: Air
Pollution Engineering Manual. New York: Van Nostrand Reinhold/Air and Waste
Management Association. 1992, pp. 155-205.
[2] Burton, D. Jeff. Industrial Ventilation Work Book. Salt Lake City: DJBA, Inc. 1989.
[3] The Measurement Solution: Using a Temporary Total Enclosure for Capture Efficiency
Testing. Research Triangle Park, NC: U.S. Environmental Protection Agency. August 1991
(EPA-450/4-91-020), pp. 3,11.
[4] The Measurement Solution, pp. 11-29.
[5] Heinsohn, Robert Jennings. Industrial Ventilation: Engineering Principles. New York:
John Wiley & Sons, Inc. 1991.
[6] Telephone conversation between William M. Vatavuk, U.S. Environmental Protection Agency
(Research Triangle Park, NC) and Todd N. Stine, United McGill Company (Raleigh, NC),
May 24, 1993.
[7] Thermoplastic Duct (PVC) Construction Manual. Vienna, VA: Sheet Metal and Air
Conditioning Contractors' National Association, Inc. (SMACNA). May 1987, pp. 61-85.
[8] Thermoplastic Duct Construction Manual, p. 64.
[9] Burton, p. 6-7.
[10] Dust Control System Accessories Price List. Huntington Park, CA: Murphy-Rodgers, Inc.
July 1992.
[11] Price and Data Catalog: Standard Ductwork Components. Warren, MI: Wer-Coy Metal
Fabrication Co. 1992-93.
10-53
-------
[12] Letters from Samir Karandikar, EPCON Industrial Systems (Woodlands, TX) to William M.
Vatavuk, U.S. Environmental Protection Agency (Research Triangle Park, NC). May 21 and
June 9, 1993.
[13] "Double Wall Insulated Duct and Fittings." In: Sheet Metal Division Catalog. Groveport,
OH: United McGill Corporation. 1990.
[14] "Single-Wall Round and Flat Oval Duct and Fittings." In: Sheet Metal Division Catalog.
Groveport, OH: United McGill Corporation. 1990.
[15] HVAC Duct Construction Standards: Metal and Flexible. Vienna, VA: Sheet Metal and Air
Conditioning Contractors' National Association, Inc. (SMACNA). 1985, pp. 2-15 to 2-17.
[16] Wherry, T.C. and Peebles, Jerry R., "Process Control". In: Perry's Chemical Engineers'
Handbook, Sixth Edition. New York: McGraw-Hill, Inc. 1984.
[17] Product catalog. Rio, WI: Gaskets, Inc. 1994.
[ 18] HVAC Duct Construction Standards, pp. 4-2 to 4-3.
[ 19] HVAC Duct Construction Standards, pp. 4-2 to 4-7.
[20] Letter from Howard D. Goodfellow, Goodfellow Consultants (Mississauga, Ontario, Canada)
to William M. Vatavuk, U.S. Environmental Protection Agency (Research Triangle Park,
NC). February 23, 1994.
[21] Guide for Steel Stack Design and Construction. Vienna, VA: Sheet Metal and Air
Conditioning Contractors' National Association, Inc. (SMACNA). 1983.
[22] Goodfellow, pp. 192-193.
[23] Goodfellow, p. 193.
[24] Peters, Max S. and Timmerhaus, Klaus D. Plant Design and Economics for Chemical
Engineers, Third Edition. New York: McGraw-Hill, Inc., 1980, pp. 508-510.
[25] Burton, pp. 2-10 to 2-11.
[26] Burton, p. 2-11.
[27] Burton, pp. 4-5 to 4-8.
[28] Burton, p. 5-12.
[29] Burton, pp. 5-15 to 5-16.
10-54
-------
[30] Burton, p. 5-5.
[31] Burton, pp. G-2, G-5.
[32] Burton, p. 5-18.
[33] Vatavuk, William M. and Neveril, Robert B., "Estimating Costs of Air-Pollution Cor1
Systems, Part III: Estimating the Size and Cost of Pollutant Capture Hoods," Chem^ ./
Engineering, December 1, 1980, pp. Ill to 115.
[34] Telephone conversation between William M. Vatavuk, U.S. Environmental Protection
Agency (Research Triangle Park, NC) and Dennis Woll, Air Plastics, Inc. (Mason, OH),
August 10, 1993.
[35] Telephone conversation between William M. Vatavuk, U.S. Environmental Protection
Agency (Research Triangle Park, NC) and Pat Caputo, Intellect Systems & Marketing, Inc.
(Bohemia, NY), October 22, 1993.
[36] Burton, "Chart 9".
[37] Letter from Todd N. Stine, United McGill Corporation (Raleigh, NC) to William M. Vatavuk,
U.S. Environmental Protection Agency (Research Triangle Park, NC). June 10, 1993.
[38] Green, Don W. and Maloney, James O. Perry's Chemical Engineers' Handbook, Sixth
Edition. New York: McGraw-Hill, Inc. 1984.
[39] Peters, Max S. and Timmerhaus, Klaus D. Plant Design and Economics for Chemical
Engineers, Fourth Edition. New York: McGraw-Hill, Inc. 1991.
[40] Engineering Design Reference Manual for Supply Air Handling Systems. Groveport, OH:
United McGill Corporation. 1992, pp. 34.
[41] Engineering Design Reference Manual, p. 8.
[42] Burton, "Chart 5".
[43] Engineering Design Reference Manual., p.7.
[44] Burton, p. 6-6.
[45] Burton, "Chart 13".
[46] Goodfellow, p. 193.
[47] Guide for Steel Stack Design and Construction, pp. 39 to 50.
10-55
-------
[48] Goodfellow, p. 193.
[49] Carlton-Jones, Dennis and Schneider, H.B., "Tall Chimneys," Chemical Engineering, October
14, 1968, p. 167.
[50] Guideline for Determination of Good Engineering Practice: Stack Height (Technical
Support Document for Stack Height Regulations) (Revised). Research Triangle Park, NC:
U.S. Environmental Protection Agency. June 1985 (NTIS PB-85-225241), p.l.
[51] Guideline for Determination of Good Engineering Practice, pp. 50-51.
[52] Guideline for Determination of Good Engineering Practice, pp. 1-2.
[53] Goodfellow, p. 194.
[54] Guide for Steel Stack Design and Construction, p. 4.
[55] Hood cost data request responses from four hood vendors to William M. Vatavuk, U.S.
Environmental Protection Agency (Research Triangle Park, NC). June-July 1993.
[56] Ductwork cost data request responses from six vendors to William M. Vatavuk, U.S.
Environmental Protection Agency (Research Triangle Park, NC). May-July 1993.
[57] Stack cost data request responses from four vendors to William M. Vatavuk, U.S.
Environmental Protection Agency (Research Triangle Park, NC). May-July 1993.
[58] Op. cit., Stine-Vatavuk letter, June 10, 1993.
[59] Goodfellow-Vatavuk letter.
[60] Goodfellow-Vatavuk letter.
[61] Telephone conversation between William M. Vatavuk, U.S. Environmental Protection
Agency (Research Triangle Park, NC) and Todd N. Stine, United McGill Company (Raleigh,
NC), December 10, 1993.
[62] Darman, Richard. Guidelines and Discount Rates for Benefit-Cost Analysis of Federal
Programs (OMB Circular No. A-94 Revised). Washington, DC: Office of Management and
Budget. October 29,1992.
10-56
-------
--••••'1::.' r>-'--^ion Agency
:..--.•,vu>:cJ, 12th Floor
Ci.icago, IL 605C4-3u90
------- |