EPA/600/R-94/128
August 1994
A REVIEW OF COMPUTER PROCESS SIMULATION
IN INDUSTRIAL POLLUTION PREVENTION
by
Jordan M. Spooner
Waste Minimization, Destruction
and Disposal Research Division
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
EPA Grant No. CT901761-01
Project Officer:
Jordan M. Spooner
Waste Minimization, Destruction
and Disposal Research Division
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
Printed on Recycled Paper
-------
DISCLAIMER
The information in this document has been funded wholly or in part by the United
States Environmental Protection Agency. It has been subjected to the Agency's peer and
administrative review, and it has been approved for publication as an EPA document.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use. -^
11
-------
FOREWORD
Today's rapidly developing and changing technologies and industrial products and
practices frequently carry with them the increased generation of materials that, if improperly
dealt with, can threaten both public health and the environment. The U.S. Environmental
Protection Agency is charged by Congress with protecting the Nation's land, air and water
resources. Under a mandate of national environmental laws, the agency strives to formulate
and implement actions leading to a compatible balance between human activities and the
ability of natural systems to support and nurture life. These laws direct the EPA to perform
research to define our environmental, problems, measure the impacts, and search for
solutions.
The Risk „. Reduction Engineering Laboratory is responsible for planning,
implementing, and managing research, development and demonstration programs to provide
an authoritative, defensible engineering basis in support of the policies, programs, and
regulations of the EPA with respect to drinking water, wastewater, pesticides, toxic
substances, solid and hazardous wastes, and Superfund-related activities. This publication
is one of the products of that research and provides a vital communication link between the
researcher and the user community.
Pollution preyejtion (P2) is the use of materials, processes, or practices that reduce
or eliminate the creation of pollutants or wastes at, the source. Once limited to easy
practices such as good housekeeping, P2 is currently evolving to include new methods of
process design and new process technologies. Process simulation, a process design tool once
used only by experts but now developed for a broader user community, has great potential
to contribute to these new P2 efforts. The intent of this document is to foster the use of
process simulation%r P2 by environmental professionals by discussing and demonstrating
the user-friendliness and capabilities of state-of-the-art simulation software.
m
\
-------
ABSTRACT
The objective of this report is to provide environmental professionals with an
understanding of the power and utility of state-of-the-art process simulation software for
industrial pollution prevention (P2) analysis. Process simulators are process design tools
that were once used only by experts in the chemical process industries (CPI), but are now
sufficiently user-friendly to be used by a wider range of people. These tools are important
for P2 efforts because of the potential for application to processes outside the CPI.
To better understand the issues that are discussed, background information is first
provided on process simulation, including historical development, .current applications, and
current research. Also, to better understand the features of state-of-the-art process
simulators, a review is provided on several commercially-available simulation software
packages. A case study is then performed using one of the simulators reviewed, in order
to demonstrate the P2 analysis capabilities of existing process simulators.
State-of-the-art process simulators are shown to have the ability to do rapid and
convenient analysis of process design options leading to P2. Powerful analytical features and
enhanced user environments make this possible. Existing process simulators can also
contribute to U.S. industrial P2 efforts by their ability to model waste water systems, and
to measure P2 progress. However, despite these strengths, existing simulators have some
important weaknesses that must be addressed for P2, such as in modeling fugitive/trace
emissions and dynamic/batch process operation.
This report was submitted in fulfillment of Grant No. CT901761-01 by Jordan M.
Spooner under the sponsorship of the U.S. Environmental Protection Agency. It is the
result of research performed during a one year mentorship at EPA in the National Urban
Fellows, Inc.'s Environmental Science and Management Fellowship Program. This report
covers a period from August 1992 to August 1993, and work was completed as of December
1993.
IV
-------
TABLE OF CONTENTS
Disclaimer . ... . . . . , . ii
Foreword .......; ...;.......... iii
Abstract .-. . iv
List of Tables vi
List of Figures .-.. , . . vii
Acknowledgments^;: viii
SECTION I INTRODUCTION . . 1
SECTION II BACKGROUND ON PROCESS SIMULATION . ... 3
Process Simulator Description . 3
Historical Development 5
Current Applications , 6
Current Research .. 7
SECTION III PROCESS SIMULATION SOFTWARE REVIEW ........... 9
Commercial Software Packages . . . 9
Software Summary 15
SECTION IV CASE STUDY: METHYLENE CHLORIDE
SOLVENT RECOVERY 17
Problem Statement 17
Computed Results ..-. . . . .... . . .; 20
Discussion 20
Summary . . . ....................;. 24
SECTION V PROCESS SIMULATION NEEDS 26
General P2 Needs 26
Specific P2 Needs . 30
SECTION VI CONCLUSIONS . . 31
REFERENCES 32
APPENDIX: BREAKOUT GROUP RESULTS FROM EPA/DOE/AIChE
PROCESS SIMULATION RESEARCH WORKSHOP 34
-------
LIST OF TABLES
1.
2.
3.
4.
5.
Sample of Commercial Process Simulation
Software Packages
Summary of Features Available in Commercial
Process Simulation Software Packages
Example Problems in the Aspen Technology Inc.
Environmental Casebook
Results from Case Study Design Specification
Runs: Process Performance at Various Solvent
Discharge Concentrations
Results from Case Study Sensitivity Runs:
Solvent Discharge Concentration at Various
Flash Tower Steam Flow Combinations . .
10
11
18
21
21
VI
-------
LIST OF FIGURES
1. Building Blocks of a Computer System
to Aid in Engineering Analysis 4
2. Schematic Diagram of a Unit Operation Model . 4
3. Simplified Example of a Mass Exchange Network (MEN)
Problem for Phenol in a Petroleum Refinery ....................... 8
4. Methylene Chloride Solvent Recovery Flowsheet 19
5. Methylene Chloride Solvent Recovery Flowsheet
Modified for Water Recycling 19
6. Surface Plot of Methylene Chloride Discharge
Concentrations Over a Range of Steam Flow
Combinations .... ..................; 22
7. Contour Plot of Methylene Chloride Discharge
Concentrations Over a Range of Steam Flow
Combinations 22
8. Summary of Results from the EPA/DOE/AIChE
Process Simulation Research Workshop Report 27
vn
-------
ACKNOWLEDGMENTS
This report was prepared for the U.S. Environmental Protection Agency, Risk
Reduction Engineering Laboratory, Pollution Prevention Research Branch. Appreciation
is given to many members of the branch who helped in the report preparation, particularly
Harry Freeman, Ivars Licis, Rada Olbina, and Emma Lou George.
Appreciation is also given to many other significant contributors to this report. First,
to Prof. James Noble of Tufts University, who both supplied and assisted in obtaining much
of the information for this report. Second, to Dr. Harry Bostian and Dr. Subhas Sikdar of
Risk Reduction Engineering Laboratory's Water and Hazardous Waste Treatment Research
Division for organizing and managing the process simulation research workshop upon which
much of the information in this report is based. And third, to Prof. David Allen of UCLA
and Dr. Hilaly Ahmed for reviewing this report and offering very valuable input for the final
version.
Last, but not least, appreciation is also extended to Ms. Linda Alexander, Mr. Luis
Alvarez, and the rest of National Urban Fellows, Inc. for making the Environmental Science
and Management Fellowship a great opportunity.
Vlll
-------
SECTION I
INTRODUCTION
Report Objective
The objective of this report is provide environmental professionals with an
understanding of the power and utility of state-of-the-art process simulation software for
industrial pollution prevention analysis. Most environmental professionals have very little
knowledge of these process analysis tools. Until the past few years, process simulation has
been used primarily by a relatively small number of process design experts in the chemical
process industries (CPI). The intent of this document is to foster the use of these tools for
pollution prevention by non-experts by discussing and demonstrating the capabilities and
user-friendliness of the software.
The document does not promote process simulation as a pollution prevention tool
applicable to all industrial processes and problems. Within the CPI, the software has been
available mostly for continuous rather than batch process design^ and therefore this report
explores the use of process simulation for general manufacturing processes that are
continuous or quasi-continuous only. Further, the current software has some other
important needs that must be addressed for pollution prevention purposes. This report
presents some of these needs.
The Definition of Pollution Prevention
The U.S. Environmental Protection Agency (EPA) defines pollution prevention (P2)
as the use of materials, processes, or practices that reduce or eliminate the creation of
pollutants or wastes at the source. A strict interpretation of this definition has been adopted
in this report by equating P2 to "source reduction" only, and therefore recycling is included
in the realm of P2 only when it occurs before the subject waste is discharged from a process
or facility to the environment. This is not to imply that end-of-pipe techniques such as off-
site recycling are not desirable; it does, however, indicate that while these methods can help,
there are often better approaches.
Report Motivation
A prime motivation for promoting the application of process simulation to industrial
P2 is the technological evolution of P2. Initial P2 efforts, which have focused on the easiest
practices such as good housekeeping, are approaching a limit, and a second phase is
underway in whicliexisting processes will be modified by the use of separation technology.
This second phase w\l also inevitably reach a limit, and a third phase will be necessary in
which new methods of process design and new process technologies will be used specifically
for P2 [1]. Process simulation is a process design tool with great potential to contribute to
these second and third phase P2 efforts.
1
-------
Report Organization
Section n of this report provides background information on process simulation,
including a basic system description, and discussions on historical development, current
applications, and current research. Section HI presents a review of some of the leading,
state-of-the-art commercial process simulation software packages. Section IV provides a
case study demonstrating some of the P2 analysis capabilities of state-of-the-art process
simulators. Process simulation needs for P2 design and analysis are discussed in Section V.
Finally, some conclusions are offered in Section VI.
-------
SECTION II
BACKGROUND ON PROCESS SIMULATION
Process Simulator Description
A process simulator, or flowsheeting system, is a large computer program that aids
in engineering analysis. It can be defined further as a computer system that accepts, as
input, information about a process at the flowsheet level of detail and performs analyses
useful in process development, design, or operation [2]. All flowsheeting systems are
comprised of the basic building blocks of models, algorithms, software, and a user interface,
as shown in Figure 1.
The models are the foundation of a process flowsheeting system. The models
mathematically describe the key process unit operations by relating the inlet and outlet
stream variables, model parameters (i.e., unit operation performance parameters), sizing and
performance requirements (e.g., heat duty for a heat exchanger), and internal variables (e.g.,
internal stage temperatures for a distillation column). A schematic diagram of a unit
operation model is provided in Figure 2. The mathematical relations are all in the form of
algebraic and differential equations that are based on the applicable laws of chemistry and
physics [2].
The algorithms solve the mathematical equations provided by the models. The two
leading types of algorithms used in current commercial process simulators are a sequential-
modular algorithm and an equation-oriented algorithm. In a sequential-modular algorithm,
the unit operation models are implemented sequentially as computer subroutines that
calculate the outlet stream variables as functions of the inlet stream variables and model
parameters. A computation sequence is initially determined, either automatically or by the
user, and the output from one unit serves as input to the next unit. In an equation-oriented
algorithm, all of the flowsheet equations are collected and solved simultaneously as a large
system of nonlinear algebraic equations [2].
The software consists of everything needed to implement the algorithms on a
particular computer and operating system. Included within the category of software are:
the program and system architecture, database structures, file-system interface, programming
language, computer code, and system documentation [2].
The user interface is the window by which the user views and operates the
flowsheeting system. This includes the input language, or other means, whereby the user
describes his or her problem to the system; the reports that contain the results; user
documentation that explains how to use the system; and protocols to interface with other
computer programs or systems [2].
-------
PROBLEM
RESULTS
USER INTERFACE
SOFTWARE
ALGORITHMS
MODELS
COMPUTER HARDWARE
AND
SYSTEM SOFTWARE
—I
I
I
I
—I
Figure 1: Building Blocks of a Computer System to Aid in Engineering Analysis
(Reprinted from Foundations of Computer-Aided Chemical Process Design.
1981, by U.S. EPA with permission of Engineering Foundation.)
MODEL
PARAMETERS
INLET STREAM
VARIABLES
UNIT
OPERATION
MODEL
.1
[ INTERNAL
I (RETENTION) '
1 VARIABLES J
OUTLET STREAM
VARIABLES
RESULTS VARIABLES
(SIZING & PERF.)
Figure 2: Schematic Diagram of a Unit Operation Model
(Reprinted from Foundations of Computer-Aided Chemical Process Design.
1981, by U.S. EPA with permission of Engineering Foundation.)
-------
Historical Development
Current flowsheeting systems have their roots in flowsheet simulators developed in
the 1960's [3]. Computer processing and thermodynamic data limitations in these earlier
systems did not permit the simulation of a wide variety of plant equipment, and as a result
systems were only available for a small number of components, such as distillation columns.
Computer processing limitations also restricted flowsheeting systems initially to large
mainframe computers, and to steady-state and continuous process operation. In steady-state
operation temperatures, pressures, and flowrates are fixed over tune, while in continuous
operation startups and shutdowns are infrequent and intermediate process interruptions such
as handling, sampling or storage are minimal. Steady-state programs use sequential-modular
modular algorithms, and thus have lower computer power requirements.
Over time, significant increases in computer processing and thermodynamic data
availability have led to many improvements in flowsheeting systems. For one, they can now
model a wide variety of plant equipment, allowing one system to simulate many processes.
These systems are often referred to as general process simulators. General process
simulators such as the ones reviewed in the following section of this report usually have a
unit operation library, which provides the user with a large selection of unit operation
models for flowsheet construction, and a large data bank of physical property data for user
selection of pure chemical components.
Increased computer processing power, as seen in the development of the personal
computer (PC), has also enabled the simulation of dynamic and batch processes, although
to a limited degree. Some dynamic simulation can now even be done on PC's, while steady-
state programs are now widely available for both mainframe computers and PC's. In
dynamic operation, the process temperatures, pressures, and flowrates are not fixed over
time, while in batch process operation startups and shutdowns are frequent and intermediate
process interruptions often occur. Dynamic simulators use equation-oriented algorithms,
and since as many as 30,000 differential equations may need to be solved when simulating
an entire process plant, these systems often require massive amounts of computer power and
are more time-consuming [4].
Process optimization is also now possible in some commercial process simulators.
Optimization allows the user to determine plant operating conditions that will maximize or
minimize any objective that is specified by the user. In some simulators, these objectives
can be either technical or economic. Economic evaluation capabilities built into some
simulators allow the performance of equipment sizing calculations, and capital and operating
cost estimation. A design specification feature, used in conjunction with optimization, can
allow the user to set a target value, or design constraint, on any flowsheet variable or
function of a combination of flowsheet variables. In some simulators, a target value can be
set for any unit operation result, stream flow or property, or component flow or purity, and
there is no limit to the number of specifications or constraints that can be established.
-------
Last but not least, a number of user interface developments have resulted in new
features that have significantly increased the utility of some commercial process simulators.
One feature is expert guidance, which can help the user to build a flowsheet model and also
prevent errors. On-line help, prompts, and tutorials are often available to assist the user.
Another feature is interactive simulation, which depending upon the flowsheeting system,
will allow the user to stop a simulation at any point, examine the results, change any of the
specifications, and then repeat or continue the simulation. A third feature is graphic
interfacing, which again depending upon the system, will allow the user to build a flowsheet
graphically, using block symbols or pictorial icons to represent unit operations. Also known
as process flow diagrams (PFD), these graphic flowsheets can often be printed as output
from the computer.
Current Applications
As a result of the development of commercial process simulators as described above,
these flowsheeting systems have rapidly emerged as process design and analysis tools that
are increasingly applied by engineers and scientists in many different fields. In fact, an
extensive review of the published technical literature revealed a wide range of recent
applications of flowsheeting systems for process design and analysis [5]. Many of these
applications had an environmental focus, although the vast majority were for treatment
purposes and did not involve P2. A list highlighting the various fields in which process
simulation has been applied includes:
• power generation/energy distribution
• nuclear fuel production
• chemical processing/production
• mining
• transportation systems
• petroleum/reservoir engineering
• incineration/combustion
• groundwater contamination/remediation
• wastewater treatment
• biotechnology/bioengineering
Of all of these applications, chemical plant processing/production represents the
largest use of process simulation, primarily because it was developed by and for this field.
In the CPI, process simulation is typically used for:
• process design and economic evaluation of a new plant
• evaluation of different design configurations for a new plant
• optimization of operating conditions for a new plant
• simulation of the operation of an existing plant
• optimization of the operation of an existing plant
• retrofit studies for an existing plant [6],
-------
Current Research
Despite the varied and growing application of flowsheeting tools in process design
and analysis, these tools have yet to be sufficiently employed for industrial P2. This
conclusion is based on the small number of published documents in the literature on this
topic, as mentioned above, and on the results of a December 1992 process simulation
research workshop jointly sponsored by EPA, the U.S. Department of Energy (DOE), and
the American Institute of Chemical Engineers (AIChE). The workshop report states that
"currently ... process simulators do not effectively integrate the technical and economic
considerations of environmental needs." [7].
Attendance at the EPA/DOE/AIChE workshop consisted of 50 leading practitioners
in the fields of process design and simulation, including university professors, simulation
software developers, process designers, and federal research and development (R&D)
managers. The overall objective of the workshop was to identify process simulation R&D
areas addressing environmental needs in the CPI. The workshop results - the R&D needs
with some perspective on priorities and development time periods - are included in the
workshop report. A number of critical areas for R&D emerged from the results of the
workshop, and these are presented and discussed in Section V. It was also noted in the
workshop report that many of the areas identified for R&D have already been the subject
of some level of research to date.
One R&D area identified by the workshop, and in fact the leading area of R&D
activity in process simulation over the past several years, is known as process synthesis.
Process synthesis has also been the leading area of simulation R&D activity for P2 purposes,
as evident by the literature. Process synthesis can be defined as "determining the optimal
interconnection of processing units as well as the optimal type and design of the units within
a process" [8]. One of the most significant P2 activities in this area in recent years is a
concept developed at the University of California at Los Angeles known as mass exchange
network (MEN) synthesis.
MEN synthesis involves "the systematic generation of a cost-effective network of mass
exchangers (i.e., separation units) with the purpose of preferentially transferring certain
species from a set of rich streams to a set of lean streams" [9]. Limiting the amount
transferred are mass balance and equilibrium constraints. A simplified example of an MEN
problem is shown in Figure 3, in which phenol in petroleum refinery waste water is
transferred from rich streams (Rj) to lean streams (L,), with the goal of identifying the
process configuration that minimizes the amount of phenol that appears as a pollutant.
MEN synthesis is somewhat analogous to the well-studied process synthesis topics of heat
exchange networks, and optimal distillation column trains.
-------
1
\
P
2
a
3
.^
o
s
PQ
1 1 i,
t
o
Sa
SM • ' • ^ 3! .
S^o §g
ifiE 2
w "-
a |
1
),
§ -
3 3
g
o
s
I
J
a.
I
&
.
u
ll
s
A #
sl
- s
« w
1 1
1
05 M
1 3*1
Ed t«
3
J I • J ,
s I
i
O
SI
Ed M
l^!
1
s
1 .
g
i-
y S
a sp
o «
If*.
fill
i
^
tx
o
1
aj
fa
ll
^ •<-•
ee^J
S.a
•E|
s ••§
' i*
TS O
tl
Jo
C * * ^^
O fO
rzi 53 «_
Oo s
!g.i? -ai'a *
•§ 1^1
I-H (S crt
C4&&,
^^^^^
Sp^ |
ll -si
.ill *
J<0
II II II
233
^™v^^
aB.
s w »3 OT
8
-------
SECTION HI
PROCESS SIMULATION SOFTWARE REVIEW
The objectives of this review are to: (1) present a sample of the state-of-the-art
process simulation software that is commercially-available; and (2) highlight the specific
features of these process simulators that can be used for P2. The review focuses on process
simulators that can be used to model a wide variety of chemical and petrochemical
processes (i.e., general process simulators), since this software is the most applicable to
achieving widespread P2. It includes some dynamic and batch process simulators, although
the majority are for steady-state and continuous operation since these simulators'are still
much more prevalent. Exclusion of similar commercial process simulators is not meant to
imply that they have limited P2 capabilities; rather, those included are ones for which
sufficient information was available.
The process simulators that are reviewed are listed in Table I, along with each
supplier's name, address, and phone number. The important features of each process
simulator are summarized in Table II, and discussed in the following section. All of the
information provided for these software packages was obtained from suppliers' marketing
literature, phone calls to suppliers, and journal literature [10, 11, 12]. However, it is
recommended that a supplier be contacted directly for the most up-to-date information.
Commercial Software Packages
1. ASPEN PLUS
ASPEN PLUS is a steady-state process simulator that is used for numerous industrial
processes involving complex chemistry. This is due to having comprehensive libraries of unit
operation and physical property models than can handle all media, including solids and
electrolytes. There is also a physical property data bank of over 4000 chemical components,
as well as a data regression system to determine property parameters using laboratory data,
and a property constant estimation system to estimate missing property parameters. The
user can also incorporate their own unit operation models, physical property models and
data, and in-line FORTRAN into the program without flowsheet size limitations.
Design specifications can be used to specify a target value for a unit operation
output, stream flow or property, or component flow or purity. An optimization capability
can determine process operating conditions for any type of objective function, such as
technical or economic, with no limit on the number of constraints. ASPEN PLUS costing
methodologies can then perform a full process economic evaluation, including capital and
operating costs, using equipment size and operating data as input.
-------
TABLE 1
Sample of Commercial Process Simulation
Software Packages
Name
ASPEN
PLUS
BATCHES
ChemCAD
m
DESIGN
n
ESP
HYSIM
MAX
PD-PLUS
PRO/H
PROSIM
SPAN
SPEEDUP
Supplier
Aspen Technology Inc.
Batch Process
Technology
Chemstations Inc.
ChemShare Corp.
OLI Systems Inc.
Hyprotech Ltd.
Aspen Technology Inc.
Deerhaven Software
Simulation
Sciences Inc.
Bryan Research &
Engineering
Kesler
Engineering Inc.
Aspen Technology Inc.
Address
10 Canal Park
Cambridge, MA 02141
1291E Cumberland Ave.
W. Lafayette, IN 47906
10375 Richmond Ave.
Houston, TX 77042
P.O. Box 1885
Houston, TX 77251-1885
108 American Rd.
Morris Plains, NJ 07950
10333 Richmond Ave.
Houston, TX 77042
10 Canal Park
Cambridge, MA 02141
7 Shady Lane Dr.
Burlington, MA 01803
1051 W. Bastanchury Rd.
FuUerton, CA 92633
P.O. Box 3403
Bryan, TX 77805
1200 Tices Ln.
E. Brunswick, NJ 08816
10 Canal Park
Cambridge, MA 02141
Phone #
617-577-
0100
317-463-
6473
713-954-
4100
713-267-
5600
201-539-
4996
713-780-
7087
617-577-
0100
617-229-
2541
800-854-
3198
409-846-
8771
908-249-
4100
617-577-
0100
10
-------
=sr:.
3
(M
a S
C gjo
C CQ
O ^^
O °
£
•S
2J
— < cs
03 ^j
41 _^j
5 «
5 «« °
^^ ^j «p^
s-s
** ss
fe CO
O <»
09
& g
es >M
gftH
S
a
|
c
Q i
&
s 1
111
^ Q E
*S *E
Q ^
"1 s •!
s N
& •§ «
W !§ <»
O
01
£
.-3^
.^ T3
"o c
. w S
f si
3?^
^ 2
| .A c
£ ^ s
1 ta
£ £°
§
|||
0 -
' 0
H
c <
o
" O
o) •'
1 1
03 £
-------
A user interface system, ModelManager, provides interactive building of flowsheets
with the help of an expert guidance system. Graphics are available to build a process flow
diagram (PFD) with either block symbols or pictorial icons. Interactive simulations can then
be performed, with the user free to stop at any point to examine the results or to change
a specification, and then either continue or repeat the process. Sensitivity analyses can also
be performed. ASPEN PLUS can be run on mainframes, workstations, and personal
computers (PCs).
2. BATCHES
BATCHES is a process simulator for managing multiple product, recipe-driven batch
and semicontmuous flow processes in the biochemical, food/pharmaceutical, and specialty
chemical industries. It enables optimization of process configurations and operating
procedures, sizing of process equipment, and process scheduling. A graphical user interface
facilitates data entry and analysis through case comparison and animation. It is available
for workstations and mainframes.
3. ChemCAD III
ChemCAD m is a steady-state process simulator with large unit operation and
physical property model libraries, including models for solids and electrolytes processing.
A chemical components data bank contains physical property data for approximately 1450
components, while a data regression system permits the use of laboratory data. There is
also a facility for the user to supply their own unit operation models, physical property data,
and in-line FORTRAN.
TALK, an interactive program within ChemCAD HI, handles all input, calculations,
and output for a simulation. TALK allows the user to stop the simulation at any point,
review the results, edit the data, and rerun the program. The user may perform calculations
for the entire flowsheet, or for individual unit operations within the flowsheet, permitting
sensitivity analyses. Graphics enables the drawing of PFDs, while an interface to
spreadsheet software enables supplemental analysis, such as for an economic evaluation.
ChemCAD HI is for use only on PCs.
4. DESIGN II
DESIGN II is a steady-state process simulator that is used for oil/gas production,
petroleum refining, petrochemicals, and other chemical processing systems. The software
includes a unit operation library that contains over 20 process equipment models, a chemical
component data bank with over 850 pure components, and a petroleum crude library with
over 150 published crudes. The user may also add proprietary unit operation models,
physical property data, and in-line FORTRAN subroutines with no limits on flowsheet size,
components, feeds, and products.
12
-------
An optional Windows user interface allows process flowsheets to be created
graphically using pictorial icons, and the user can then perform the flowsheet simulations
interactively by stopping the calculations in progress and reviewing interim results. Also
available are features for sensitivity analyses and optimization. Spreadsheet interfacing is
permitted. DESIGN H is for use on PCs, workstations, and mainframes.
5. Environmental Simulation Program (ESP)
ESP is both a steady-state and dynamic flowsheet simulator that can be applied for
a variety of processes involving complex chemistry, although it was developed primarily for
environmental treatment and remediation processes. It has a unit operation model library,
and a chemical component data bank of approximately 2,000 components. ESP has the
ability to model aqueous and non-aqueous reactive systems. A data estimation and
regression system allows users to add new components to the data bank.
Dynamic simulations, which are run through ESP by a program component known
as DynaChem, may be conducted interactively. And ESP simulation output files may be
exported to other software to do supplemental analyses. ESP runs on PCs and workstations.
6. HYSIM
HYSIM is a steady-state process simulator that is used for design and evaluation in
the gas processing, petroleum refining, petrochemical, and chemical industries. There are
comprehensive unit operation and physical property model libraries, as well as a component
data bank that contains approximately 1,500 components. There are models for solid
processing, but not for electrolytes. A data regression and estimation system is available for
using experimental data, and proprietary unit operation models, data, and in-line
programming ("C" language) may be used.
Simulations can be performed interactively with the help of built-in intelligence in
order to perform process sensitivity and case study analyses. Graphic interfacing allows for
PFD display and output, and software interfacing provides access to spreadsheet software.
HYSIM runs on PCs, workstations, and mainframes.
7. MAX
.MAX is a steady-state flowsheet simulator that is built on core ASPEN PLUS
modeling technology, and is designed so that engineers new to simulation can get fast and
meaningful results. It has full upward compatibility with ASPEN PLUS. MAX has large
unit operations and physical property model libraries, and a pure component data bank of
over 1,300 components. There are also data regression and property constant estimation
systems. Proprietary models and data, and in-line FORTRAN, can be used with no limits
on the number of components, blocks, or size of the flowsheet.
13
-------
Aspen's ModelManager user interface system provides icons for problem specification
and PFD-style diagram generation, as well as a patented expert guidance system. Interactive
control of process calculations permits sensitivity analyses for any process variable.
Simulation results may be exported to a spreadsheet for additional analysis. MAX is run
only on PCs.
8.PD-PLUS
PD-PLUS is a steady-state process simulator for chemical processes including refinery
systems and non-ideal chemicals. It has a large unit operation library, in which all
operations allow multiple feed streams and in general can produce multiple product streams,
with some exceptions. There are two pure component data banks, one small and one large.
The small data bank contains 59 components (mostly hydrocarbons and water) and is
standard, while the large data bank contains 1,284 components and is optional.
PD-PLUS has an interactive ability that allows the user to stop the program as the
next calculation step is about to begin. At this point the user can display unit operation and
streamflow conditions, change unit operation and streamflow specifications, and rerun the
simulation at any point in the flowsheet. It also interfaces to other software, such as
spreadsheets, enabling the ability to run economic analyses, if desired. The software may
be run on PCs.
9.PRO/II
PRO/II is a steady-state flowsheet simulator that is used for gas processing,
petroleum refining, and many other chemical and petrochemical processes. It has
comprehensive unit operation and physical property model libraries, and a component data
bank that has over 1,450 pure components, including physical property data for solids and
electrolytes. In addition, the program has a data estimation and regression system, as well
as the ability to handle mixture data. Proprietary models, data, and in-line FORTRAN can
also be used, and any number of components, unit operations, and streams can be simulated.
PRO/IE has the ability to set process design specifications and operating constraints,
and to then optimize the flowsheet. Simulations can be performed interactively, with the
ability to view any stream or unit operation, change any unit operation or design
specification, and then automatically do case comparisons. Graphics interfacing, an expert
system, and links to third party software such as spreadsheets are part of the user
environment. PRO/n runs on PCs, workstations, and mainframes.
10.PROSIM
PROSIM is a steady-state flowsheet simulator for designing and optimizing plants in
the gas, oil, and petroleum industries. It has multiple unit operation models, two physical
property models, and a pure component data bank with approximately 100 components.
14
-------
' Graphic interfacing enables the user to build a flowsheet PFD on the screen and
enter operating data on forms with the help of an expert system. Interactive simulations
may then be performed, allowing for interruption and recalculation. Simulation results may
be exported to third party software. PROSIM software runs on PCs, workstations, and
mainframes.
11. SPAN
SPAN is a steady-state process simulator for use in the gas processing, petroleum,
and petrochemical industries. It has a large unit operation library, and can accept user-
supplied models. There is a comprehensive physical property data bank, as well as several
physical property models, with facilities for blending of petroleum streams. Simulations can
handle up to 100 streams, 50 components, and 50 unit operations in one flowsheet.
Process flowsheeting is interactive, with diagnostics and expert guidance. The user
can selectively execute any portion of the flowsheet, and there are automated facilities for
economic analysis and parametric studies of the effect of changing process variables on
profitability. The user can also create files that can be interfaced with CAD packages in;
order to generate process flow diagrams including the simulation results. SPAN runs on
PCs.
12. SPEEDUP
SPEEDUP is an equation-based process modeling system that offers facilities for
both steady-state and dynamic simulation and optimization of chemical processes. The
program allows the user a great deal of flexibility in the way he or she defines a model for
their process, since it will deal with an arbitrary mixture of equations and procedures
(subroutines) relating the variables in the process.
There is a unit operations library and a physical property data bank, and the user can
import unit operation models and physical property data into the program. It is available
for use on workstations and mainframes.
Software Summary
Several state-of-the-art process simulators are now available that enable the
simulation of a wide range of industrial processes. These simulators have extensive libraries
of unit operation models and physical property data, and also allow for the importation of
user-supplied models and data. Most existing state-of-the-art process simulators also have
good user interfacing capacity through interactive simulation and graphic display/output
features, and can do sensitivity analyses and set design specifications using any process
variable. All of these features make process simulation very useful for P2 analysis of many
industrial processes, as will be demonstrated in Section IV.
15
-------
Despite possessing the abilities stated above, however, many state-of-the-art process
simulators lack certain features that place a limit on the application of process simulation
to P2. For one, most existing process simulators can only model steady-state and continuous
process operation, as opposed to dynamic and batch process operation. For another, most
of these simulators can not model either electrolytes (i.e., ionic species) or solids. And two
other critical limitations of most current process simulators is the inability to do optimization
and economical analysis. These and several other critical process simulation needs will be
discussed further in Section V.
16
-------
SECTION IV
CASE STUDY: METHYLENE CHLORIDE SOLVENT RECOVERY
In order to demonstrate the power and utility of state-of-the-art process simulation
for industrial P2, a case study was performed using the process simulator MAX. The ease
study illustrates how some of the important features of existing process simulators can be
used to do rapid and convenient process analysis leading to P2. MAX was chosen because
it possesses most of the features reviewed in Section m, and because it is marketed as user-
friendly. No prior experience or training with process simulation software was held by the
author before this case study analysis was attempted.
The case study is based on one of ten examples provided in an Environmental
Casebook written by Aspen Technology Inc. (AspenTech) to illustrate the use of process
simulation to solve environmental problems [13]. The examples, listed in Table IE, use the
process simulator ASPEN PLUS to solve these problems. Three important points need to
be made about the AspenTech examples in order to understand their value for P2. First,
most of the examples illustrate waste treatment, although the first two examples can be
considered cases of in-process recycling and therefore P2. Second, hah7 of the examples
involve waste water, which is beneficial since over 90% of U.S. hazardous waste generated
is in this form [14]. And third, the Zero Discharge Waste Water Treatment System example
involves the simulation of zero water discharge; not zero pollution discharge.
Problem Statement
The problem in the case study was to examine one of the AspenTech example
processes for possible P2 opportunities or improvements. This was done using the process
simulator MAX to analyze process design alternatives.
The AspenTech simulation example that was examined in this case study is
Methylene Chloride Solvent Recovery. A flowsheet is shown in Figure 5. Two steam-
injection flash towers (TOWER1 and TOWER2) remove methylene chloride solvent (1.4%)
from a combined waste water stream (FEED) before discharge to the sewer (BOT2). In
the AspenTech example, the minimum total steam flow (STEAM1 + STEAM2) required
to meet a solvent concentration limit in BOT2 of 150 parts per million (ppm) was
determined. Some P2 is achieved in the example, since approximately 99% of the solvent
is recovered in stream MECL with minimum steam flow, but there is opportunity for more
P2 because stream WATER becomes a waste stream with a higher solvent concentration
than waste stream FEED (.019 vs. 014).
The process simulator MAX was used to study two possible P2 opportunities for the
recovery process: (1) recycling of stream WATER into stream FEED; and (2) increasing
steam flow. A simulation model of a modified methylene chloride solvent recovery process
17
-------
TABLES
Example Problems in the Aspen Technology Inc.
Environmental Casebook [13]
Case Study
Methylene Chloride Solvent Recovery
Acetone Solvent Recovery
Water Absorber for Hydrogen Chloride
Sour Water Stripping System
Flue Gas Desulfurization
Sweetening Natural Gas by Diglycolamine Absorption
Nitric Acid Absorption
Waste Water Treatment
Mobile Incineration of Heavy Oil-Laden Soil
Zero Discharge Waste Water Treatment System
Subject
Media
water
water
gas
water
gas
gas
gas
water
soil
water
18
-------
Figure 4: Methylene Chloride Solvent Recovery Flowsheet
(Reprinted from ASPEN PLUS APPLICATIONS; Environmental Casebook
by U.S. EPA with permission of Aspen Technology, Inc.)
Figure 5: Methylene Chloride Solvent Recovery Flowsheet
Modified for Water Recycling (created with MAX)
19
-------
with water recycling was built with MAX, as seen in Figure 6. Several design
specificationruns were then performed with this modified model to determine: (1) the effect
of recycling WATER on process performance (e.g., steam required, solvent recovery, and
solvent purity) at the 150 ppm solvent discharge limit; and (2) process performance at 1 ppm
and 1 ppb (part per billion) discharge levels. Sensitivity runs were also performed to
characterize the impact of the steam flows on the solvent discharge concentration.
Computed Results
A summary of the results from four design specification runs completed with the
methylene chloride solvent recovery flowsheet models is provided in Table IV. In the first
row of the table are results of a specification run with the original process flowsheet (i.e.,
no water recycling) to determine process performance at the 150 ppm solvent discharge
limit. In the second row are the process performance results at the 150 ppm limit, with the
modified process flowsheet (i.e., water recycling). The third and fourth rows are the
computed results, with the modified flowsheet, of the process performance at 1 ppm and 1
ppb solvent discharge limits.
The performance results in Table IV include both technical and economic process
parameters, and a benchmarking parameter for the purpose of comparison. The technical
parameters include: the minimum total steam flow required; the percentage of the solvent
recovered in methylene chloride-rich stream MECL; and the purity of the recovered solvent.
The economic parameter is the steam generation cost per gallon of waste water treated in
stream FEED, assuming fuel oil at $ I/gallon or $6/MMBTU. The benchmarking
parameter, dilution water, is the amount of water that would need to be added to stream
FEED to reach the design specification discharge concentration if dilution was used as the
treatment instead of flash stripping.
A summary of the results of six sensitivity runs completed with the modified flowsheet
model is provided in Table V. The table is a six-by-six array of solvent discharge
concentrations at various combinations of flash tower steam flows. In the first column are
the steam flows to the first flash tower, STEAM1. In the remaining columns are the solvent
discharge concentrations at the corresponding steam flows to the second flash tower,
STEAM2. For instance, when there is no flow to either flash tower the solvent discharge
concentration is 14,000 ppm. These data are also presented in the form of surface and
contour maps in Figures 7 and 8, respectively. These results help to illustrate the impact
of the combined flash tower steam flows on the solvent discharge concentration.
Discussion
The design specification and sensitivity runs performed in the case study
demonstrated the capabilities of the process simulator MAX to do rapid and convenient
analysis of process design options. These capabilities enabled the examination of the
impacts of waste water recycling and flash tower steam flow on process performance.
20
-------
TABLE 4
Results from Case Study Design Specification Runs:
Process Performance at Various
Solvent Discharge Concentrations
Design Spec.
Solvent
Discharge
Concentration
150 ppm1
,150 ppm2
1 ppm2
Ippb2
Minimum
Steam
Required
(Ib/hr)
12,892
13,080
33,170
456,000
Solvent
Recovery
(%)
97.3
98.8
99.99
-
Solvent
Purity
(%)
99.8
99.8
99.8
-
Steam
Cost
(0/gal
waste)
0.98
0.99
2.52
34.67
Dilution
Water
(Ib/hr)
9.3 E+6
9.3 E+ 6
1.4 E+9
1.4E+12
1 No water recycling.
2 Water recycling.
TABLES
Results from Case Study Sensitivity Runs:
Solvent Discharge Concentration (ppm) at
Various Flash Tower Steam Flow Combinations
STEAM1
(Ib/hr)
0
10,000
20,000
30,000
40,000
50,000
STEAM2 (Ib/hr)
0
14,000
1,500
45.2
16.3
10.5
8.07
10,000
1,530
23.4
1.43
.668
.484
.404
20,000
129
14.4
1.08
.44
.304
.248
30,000
65.6
12.2
1.11
.382
.251
.199
40,000
46.8
11.5
1.29
.372
.23
.178
50,000
37.9
11.2
1.59
.387
.225
.168
21
-------
STEAM2
Gb/hr)
i-ur-
1*10
5'104
STEAM1 (Ib/hr)
Figure 6: Surface Plot of Methylene Chloride Discharge
Concentrations (base 10) Over a Range of Steam Flow Combinations
4.146
LPPM
S1 S2
^0.775
Figure 7: Contour Plot of Methylene Chloride Discharge
Concentrations (base 10) Over a Range of Steam Flow Combinations
22
-------
Comparision of the first two specification runs in Table IV indicates that
approximately 1.4% more steam is needed to meet the 150 ppm methylene chloride
discharge limit with the change to water recycling. This is due to the higher solvent
concentration in the recycled water stream than in the influent waste water stream. It can
also be seen that 1.5% more solvent is recovered, while solvent purity is maintained. The
comparison of the two design specification runs indicates that P2 can be achieved with
recycling of stream WATER, because of the higher level of solvent recovery and the
elimination of one waste stream.
Other ways of assessing the P2 impact of the process change include examining the
treatment cost and dilution water values in Table IV. For the case of the first two runs, the
dilution water parameter does not change because the solvent discharge limit is the same.
However, the value indicates that more than 9 million Ib/hr of water would be required if
dilution were used, rather than steam stripping, and also serves as a benchmark for the
other runs. The dilution water parameter indicates that steam stripping is the better choice
from a water usage standpoint. The treatment cost increases a very small amount because
of the small increase in the steam flow required.
The third specification run in Table IV gives the results for a 1 ppm methylene
chloride discharge. For this level of discharge, the steam required is much higher,
approximately 33,000 Ib/hr, but the solvent recovery is 99.99%. The purity of the methylene
chloride-rich stream recovered, MECL, does not increase above 98.8% because the water
in this stream is below it's solubility limit in methylene chloride and therefore can not be
separated out. This is essentially the purity limit for the recovered solvent.
The dilution water required in the third specification run is 1.4 billion Ib/hr, which
again shows why steam stripping is the preferred treatment method. The treatment cost
increases by a factor of 2.5, however, because of the similar increase in the steam flow.
These results all help to show the cost-benefit trade-off involved in decreasing the solvent
discharge to 1 ppm.
The surface plot, Figure 7, contains curves of constant solvent discharge
. concentrations (in the logarithm of ppm), for a range of steam flow combinations from: no
flow to either flash tower (i.e., the origin point in the plot); to a flow of 50,000 Ib/hr of
steam to each tower (i.e., the point [5xl04,5xl04]). From this plot, the optimum steam flow
combination (i.e., the one for which the total steam flow is the minimum) for each
concentration level can be estimated by constructing a 45 degree line tangent to the curve
of interest, as seen in Figure 7. The tangency point is the optimum. This is true because
a 45 degree line is a line of constant total steam flow, and the tangency point will give the
smallest steam total. For instance, the optimum steam flow combination for a 1 ppm
solvent discharge is at approximately STEAM1 = 23,000 Ib/hr and STEAM2 = 10,000
Ib/hr, as seen in Figure 7.
23
-------
The contour plot, Figure 8, shows the solvent discharge level contour (again in the
logarithm of ppm) over the same steam flow range as the surface plot. The contour plot
shows how the concentration level changes (increases or decreases) along the vertical axis
with a varying steam flow combination. The darker the shade of the contour plot, the
smaller the discharge concentration. Both the contour plot and the surface plot indicate
that STEAM1 has a greater impact on solvent discharge concentration than STEAM2, which
is seen by a greater decrease in concentration level along the STEAM1 axis. Both plots also
show that as both STEAM1 and STEAM2 are increased, the solvent discharge concentration
is minimized.
The fourth design specification run in Table IV provides the results for a 1 ppb, or
"nominal zero," solvent discharge level. Much discussion has transpired in recent years
concerning the technical and economic feasibility of achieving near-zero or zero waste
discharge for many industrial processes. These results indicate that to reduce the solvent
concentration to this level would require 456,000 Ib/hr of combined steam flow. The steam
cost and dilution water values for this level are also very high, at approximately 35 cents per
gallon and one trillion gallons, respectively. Solvent recovery and purity values are not given
in the table because at the steam flow required, the solvent is below it's solubility limit in
water for the stream entering the decanter, and therefore can not be separated out.
The results indicate that for the process analyzed "nominal zero" solvent discharge
is neither technically nor economically feasible. Both the level of steam flow required and
the cost to generate this steam would be enormous. One must also consider the
atmospheric emissions and other life cycle impacts associated with the high steam
generation. These impacts may present an even greater health risk than simply discharging
the waste water to the sewer.
The solvent discharge concentrations provided in the results of both the design
specification and the sensitivity runs are waste generation data that can be used to measure
the P2 potential of the hypothetical process changes. These are relatively accurate and
reliable data that would otherwise need to be either collected from pilot or actual facility
testing, or estimated using engineering judgement. The data are generated using process
material balances, although some emissions, such as fugitive and trace emissions, are not
included in the material balances because they are several orders of magnitude smaller than
the main process flow streams.
Summary
The case study demonstrated the ability of MAX to do rapid and convenient analysis
of process design alternatives leading to P2. With no prior training or experience by the
author in using process simulation software, a process simulation model of the solvent
recovery example was easily constructed, and process design alternatives for P2 were
analyzed.
24
-------
The design specification and sensitivity analysis features, along with user interface
features such as graphic display/output and an expert guidance system, all facilitated easy
and reliable process analysis. Design specifications were used to determine the process
operation required to meet a desired waste discharge level. Sensitivity analyses provided
a means to determine the impact of key process variables on system waste generation. And
the user interface features allowed the powerful process analysis capabilities to be used
rapidly with no training or experience.
The case study analysis also demonstrated the ability of MAX to model aqueous
systems, and to generate waste generation data that can be used for P2 measurement. Both
of these capabilities are critical to existing industrial P2 efforts. Modeling aqueous systems
is critical because industrial aqueous waste represents the largest volume of hazardous waste
in the U.S. [14]. And P2 measurement is important in order to meet existing regulatory and
practical requirements that justify continued P2 efforts.
25
-------
SECTION V
PROCESS SIMULATION NEEDS
Despite possessing many features that make them powerful and convenient tools for
process design and analysis, as discussed in the previous two sections, current process
simulators still lack many critical aspects required to be widely effective for P2. A number
of these aspects were pointed out already during the process simulation software review of
Section HI, such as shortfalls in dynamic and batch process simulation, process optimization,
and economic analysis. Also, several critical needs were identified in the December 1992
EPA/DOE/AIChE process simulation research workshop discussed in Section II. The
purpose of this section is to discuss these and other needs important for P2 design and
analysis in the CPI.
General P2 Needs
Possibly the best source of information on process simulation needs for P2 is the
report from the EPA/DOE/AIChE workshop discussed in Section II. The objective of this
workshop was to identify the critical areas for R&D needed in order to incorporate
environmental factors into process simulation and design tools for the CPI. With this in
mind, several areas were identified and they are summarized here in a table reproduced
from the report as Figure 8. These areas represent common themes that were agreed upon
by most participants in the workshop, and are neither specific in nature nor inclusive of all
ideas presented by the attendees. Specific ideas from workshop participants are included
in the breakout group results found in the Appendix.
It is important to note that this list essentially reflects the R&D interests of the
process simulation community, and therefore includes mostly complex and long-term goals,
such as process synthesis, rather than more immediate and practical needs such as fugitive
emissions estimation. For instance, process synthesis is at the top of the list in Figure 8
because it received the highest number of "top" priority votes from the workshop attendees.
This high vote count for process synthesis is not surprising since it has been the leading area
of interest in process simulation research for several years. Also, there may have been a
disproportionate number of workshop attendees who had this topic as their primary research
interest.
Despite the probable bias towards process synthesis at the workshop, there is no
disputing it's potential value for P2. Process synthesis could possibly help to determine
alternative chemical reaction pathways and catalysts, determine alternative chemical
separation sequences, and efficiently incorporate waste treatment units into a process design.
The objective in developing process synthesis according to the workshop report would be
to incorporate P2 concepts into chemical process design, but there would undoubtedly be
many applications beyond those in the environmental field.
26
-------
Topic
Findings
Identified Needs
Process
Synthesis
Developing new and improved methods
for synthesizing chemical processes
that meet environmental objectives is
one of the most important needs for
incorporating pollution prevention
concepts into process design. This will
enable alternative process flow sheets
and has applications beyond process
simulators.
Develop expert systems for synthesis.
Develop tools to identify alternative
reaction pathways and catalysts.
Pursue non-conventional technology
alternatives.
Develop methods for defining "ultimate"
limiting process efficiencies.
Determine if barriers lie in models or
heuristics.
Couple synthesis and simulation.
Use mathematical programming to
synthesize processes.
Dilute
Streams
Improving characterization and the
ability to separate dilute components of
streams through acquisition of data and
enhanced modeling is critical for
developing cost-effective process
designs.
• Improve simulation models and tools to
better handle dilute components of
process streams.
• Determine reaction equilibrium
partitioning constants.
• Determine reaction rates and byproducts.
• Improve measurement capability to meet
process control, regulatory, and other
needs.
• Determine data needs for modeling in the
dilute region.
• Use computational chemistry to estimate
properties and behaviors of mixtures in
the dilute region.
Optimization
Methodologies
Development of new optimization
strategies will allow users to identify
process designs that best satisfy a
range of environmental, cost, and
operating requirements.
Develop large-scale optimization
methodologies.
Develop non-linear optimization strategies.
Develop methods for dynamic
optimization of processes.
Enhance stochastic modeling and
optimization.
Optimize agregate process models.
Develop on-line optimization methods.
Modeling
Techniques
Better modeling is needed to
accommodate process synthesis and
optimization methodologies. The
probabilistic nature of much
environmentally-based data makes
stochastic modeling essential in
developing effective design tools and
simulators. Greater flexibility in
modeling is needed to accommodate
different levels of detail, data, and rigor.
• Develop large-scale modeling
methodologies (larger than unit
operations.)
• Improve probabilistic and stochastic
modeling techniques.
1 Develop better hierarchical models.
Develop dynamic simulation models with
process control.
Develop heuristic modeling capabilities to
accommodate uncertainties and provide
flexibility.
Take advantage of parallel computing
techniques.
Figure 8: Summary of Results from the EPA/DOE/AIChE
Process Simulation Research Workshop Report [7],
27
-------
Topic
Findings
Identified Needs
Rate-Based
(Non-Equilibrium)
Processes
Data and research are needed
to characterize rate-based
processes that are not
adequately modeled in current
simulators.
Characterize non-equilibrium
phenomena.
Improve interfacing and sequencing of
rate-based processes.
Environmental
Costs
The lack of environmentally-
related cost information in
current simulators is a key
barrier in identifying cost-
effective process designs. Until
process simulators accurately
account for waste treatment
costs and intangible costs on a
process-by-process basis,
pollution prevention approaches
may not appear cost
competitive with designs based
on end-of-pipe treatment.
Define and quantify intangible
Costs.
Develop methods to allocate costs to
specific processes and products.
Develop a flexible cost estimating
system.
Define cost of various end-of-pipe
treatments and associate residuals.
Develop a library of cost models.
Develop environmental cost
factors/integrate environmental
considerations with cost.
Environmental
Impact
Assessment
Better methods are needed to
determine the environmental
impact of alternative process
designs. Tools to quickly
determine whether processes
will meet environmental
standards are needed.
Develop an environmental impact
index.
Develop quick risk assessment
techniques.
Include environmental regulations in
process simulators.
Quantify or weigh competing
environmental, cost, and energy
concerns.
Link ecological and process models.
Process
Characterization
Better characterization and
modeling of unit operations and
process streams is needed to
understand the environmental
implications of alternative
process configurations.
• Improve characterization and simulation
of trace components (of environmental
concern) in process streams.
• Integrate property data into models and
simulators.
• Characterize and define benefits of
hybrid units.
• Predict environmentally-troublesome
byproducts.
• Characterize and simulate alternative
waste treatment and recycling
technologies.
Figure 8 (cont.): Summary of Results from the EPA/DOE/AIChE
Process Simulation Research Workshop Report [7].
28
-------
The high priority given to the areas of dilute streams and process characterization
reflects the fact that in most instances, the hazardous components in chemical process
streams are present hi very small (i.e, trace) concentrations that are difficult to predict,
quantify, and separate. Process simulation tools could potentially be very helpful in this
area, especially in evaluating alternative reaction pathways to prevent these troublesome
byproducts. According to the workshop report, however, current process simulators can not
adequately handle dilute components of process streams primarily because of existing data
gaps for many of these species, and also because of the lack of byproduct tracking models.
In order to strengthen the ability of process simulators to handle dilute components, there
is a need for good measured data in the dilute region, and for reliable data estimation
techniques using computational chemistry.
The areas of optimization methodologies, modeling techniques, and rate-based
processes all relate to the need for improved mathematical methodologies. Optimization
methodologies such as sequential quadratic programming (SQP) are already available in
some current process simulators, as seen in Section HI, but according to the workshop report
there is a need to develop new strategies, especially for large-scale or dynamic optimization.
Under modeling techniques, improvements are needed for: dynamic simulation of process
transients such as start-ups and shut-downs; stochastic modeling to deal with non-routine
events such as accidents, upsets, and spills; and large-scale modeling to understand the
environmental conditions that result from interactions among unit operations. And for rate-
based processes, process simulators need to improve upon their ability to handle the various
non-equilibrium phenomena (e.g., reaction kinetics, sorption, and transport) impacting waste
generation.
The emphasis on the areas of environmental costing and environmental impact
assessment is due to the inability of current process simulators to determine the true (i.e.,
total) impact, both environmental and economic, of a chemical process design. Both of
these are often key barriers to incorporating P2 approaches because if these true impacts
are not known, then a design incorporating P2 may not seem as attractive (e.g., as cost
competitive) as one incorporating end-of-pipe treatment. For environmental costing, high
priority was given by the workshop to developing and incorporating total cost accounting
models and factors that help to quantify and allocate intangible costs such as liability and
public relations. For environmental impact assessment, priority was given to developing the
ability of process simulators to quickly determine the risk or impact of alternative process
designs.
As previously stated, the areas discussed above and presented in Figure 8 are
common themes that were ageed upon by most participants in the workshop. Because of
the size of the workshop and the breadth of interests, the topics agreed upon are not very
specific, and in fact most of them, such as process synthesis and optimization, have
widespread application beyond P2. Many specific ideas were offered by individual workshop
participants, however, and these can be found in the breakout group results in the Appendix.
Some of these specific ideas are discussed in the following subsection.
29
-------
Specific P2 Needs
The following list contains some more specific capabilities that would be desirable
in process simulators for P2 purposes:
1. Fugitive emissions estimation. These emissions have become increasingly important for
industrial P2 in recent years because of regulations requiring their reporting and reductions.
Current fugitive emissions estimation methods are frequently criticized as inadequate and
costly. Current process simulators do not have the ability to estimate these emissions, but
possible simulation methodologies do exist, such as incorporating emissions factors into
simulation architecture; application of deterministic emissions correlations; and application
of equipment failure analysis [15].
2. P2 technology databases. A large number of P2 case studies have revealed a series of
effective equipment and process modifications. These technologies can be organized by
chemical, process, or unit operation, and can be made available in the form of an expert
system to the industrial designer using a process simulator.
3. Access to public domain data. The Toxic Release Inventory, the Resource Conservation
and Recovery Act (RCRA) biennial survey, the Chemical Manufacturers Association waste
data bank, and a number of other sources of data could be useful to an industrial designer
in benchmarking process configurations. Process simulators should have the ability to query
these data banks.
4. Life cycle and ancillary operation analysis. Simulation tools could be useful in evaluating
the upstream and downstream impacts of alternative process designs and modifications, as
well as the impacts of process ancillary operations such as maintenance, cleaning, and
storage.
5. Combustion byproduct estimation. Stack air emissions from hazardous waste incinerators
and combustors typically contain trace quantities of products of incomplete combustion
(PICs), such as chlorinated dioxins and furans, and unburned principal organic hazardous
constituents (POHCs). These emissions are difficult to both predict and measure. Process
simulators do not currently offer sufficient data support to model these trace species, but
they have the potential to do so.
6. Biological process modeling. Biological processes are increasingly being applied for the
treatment, remediation, and separation of hazardous wastes in air emissions, waste waters,
sludges, soils, and sediments. Very few process simulators currently contain unit operation
models for these processes.
30
-------
SECTION VI
CONCLUSIONS
1. Most existing state-of-the-art process simulators provide many features that make
them powerful tools for the analysis of P2 design alternatives in a wide range of industrial
processes. These features include: extensive libraries of unit operation models and physical
property data; the ability to incorporate user-supplied models and data; and the ability to
perform sensitivity analyses and set design specifications using any process variable. Other
important features that are available in only some of the existing simulators include:
process optimization; electrolytes modeling; and solids modeling.
2. Most existing state-of-the-art process simulators are now sufficiently user-friendly
that they can be used with little or no training or experience to do rapid process P2 analysis.
Features such as an expert guidance system and graphic display/output have greatly
enhanced the user environment of current process simulators compared to the earlier
versions that were used only by people with specialized training or experience.
3. Existing process simulators can significantly contribute to U.S. industrial P2 efforts
because of the capability to easily model and analyze waste water systems. This is important
because industrial waste water is the largest volume of hazardous waste in the U.S. Waste
water treatment is probably the largest application of process simulation currently.
4. Existing process simulators can significantly contribute to U.S. efforts to measure
progress in P2. Current measurement obstacles of data collection and data quality are
overcome by the accurate and reliable waste generation data provided by simulation models.
The obstacle of material balance closure is also overcome with the material balances done
by these simulators.
5. Despite possessing many features that make them powerful and convenient tools
for process design and analysis, current process simulators still lack many critical aspects
required to be widely effective for P2. Some of these shortcomings are general in that they
have potentially widespread applications other than the environment, while some of these
are specific to P2. Specific needs include, but are not limited to: .
• Fugitive emissions estimation
• P2 technology databases
• Access to public domain data
• Life cycle and ancillary operation analysis
• Combustion byproduct estimation
• Biological process modeling
31
-------
REFERENCES
1) Y. Cohen and D.T. Allen, "An Integrated Approach to Process Waste
Minimization Research." Journal of Hazardous Materials. 29(1992): 237-253.
2) L.B. Evans, "Advances in Process Flowsheeting Systems." In: Foundations of
Computer-Aided Chemical Process Design. R.S.H. Mah and W.D. Seider, eds.,
Engineering Foundation, New York, NY, 1981, pp. 425-469.
3) R.S. Butner, U.S. DOE Office of Industrial Technologies' Advanced Process
Design Initiative. Presentation to U.S. EPA Risk Reduction Engineering Lab,
Cincinnati, OH, December 8, 1992.
4) K. Fouhy, "Process Simulation Gains a New Dimension." Chemical
Engineering J.. Vol. 98, No. 10, Oct. 1991, pp. 47-52.
5) R. Olbina and J. Spooner, Bibliographic Database on Process Simulation
Applications. U.S. EPA Risk Reduction Engineering Lab, Pollution
Prevention Research Branch, 140 documents, 1992.
6) J.D. Seader, Computer-Aided Process Simulation Fundamentals and Applications.
Presentation to U.S. EPA Risk Reduction Engineering Lab, Cincinnati, OH, March
16, 1992.
7) U.S. Environmental Protection Agency, U.S. Department of Energy, and American
Institute of Chemical Engineers. Workshop report: Environmental Considerations
in Process Design and Simulation. Energetics, Inc., Columbia, MD, March 1993.
8) N. Nishida, G. Stephanopoulos, and A.W. Westerberg, "A Review of Process
Synthesis." AIChE Journal. Vol. 27, No. 3, May 1981, pp. 321-351.
9) M.M. El-Halwagi and V. Manousiouthakis, "Synthesis of Mass Exchange
Networks." AIChE Journal. Vol. 35, No. 8, Aug. 1989, pp. 1233-1244.
10) K. Simpson, ed., 1992-1993 CEP Software Directory. Supplement to Chemical
Engineering Progress Journal, Vol. 88, No. 10, Oct. 1992.
11) K. Simpson, ed., 1994 CEP Software Directory. Supplement to Chemical Engineering
Progress Journal, Vol. 89, No. 12, Dec. 1993.
12) J.D. Perkins and R.W.H. Sargent, "SPEED-UP: A Computer Program for Steady-
State and Dynamic Simulation and Design of Chemical Processes." In: Computer-
Aided Process Design and Analysis. AIChE Symposium Series, Vol. 78, No.
214, 1982, pp. 1-11.
32
-------
13) Aspen Technology, Inc., ASPEN PLUS APPLICATIONS: Environmental Casebook.
November 1992. .
14) R.D. Baker and J.L. Warren, "Generation of Hazardous Waste in the United
States." Hazardous Waste & Hazardous Materials J.. Vol. 9, No. 1, 1992, pp. 19-35.
15) J. Spooner, "The Role of Computer Process Simulation in Industrial Pollution
Prevention." Tufts University Masters Report, February, 1994.
33
-------
APPENDIX:
BREAKOUT GROUP RESULTS FROM EPA/DOE/AIChE
PROCESS SIMULATION RESEARCH WORKSHOP
34
-------
Work Group f: Envfrornnental Considerations In Process Design
o
O
"8
I
i
i
s
35
-------
Work Group 1: Environmental Considerations hi Prdcftw Design (cent)'
" ', . . f . •« . v **• '-xA. \ ~ *. *" / t. .: *~ "• < '"" jt •; ff $. ft'
36
-------
Work Group 2: Modal Needs
37
-------
"Work Group 2: Model Nwd* (
-------
Work Group 3: Design Toofe and Simulators
39
-------
WorkGroup 3: Design Tool* and Simulators (cont)
^ ji.^v X *• > f J . <• ? . f-, * t * f* '
40
-------
Work Group 4: Data Needs
. v/ '•>,
^ 'Sfftr f "- f
^rl^H
CO
o>
ffi
i§ «•
i
''','/„'''/',,';, ,'"'''>"' '--s, ;,™,%lr*,4 ;.
,-X ,/,',',', V-. •• ~ •• •• •.-.'.•.•••. '•-•:
*'",, *"f, ,,,',^,/ '•. <••*;/•:,••,''*•**'• .
; -,?..£"', x /-',; »:'*%'jr-r^rt-
"• . S -Jr. . *•. ' * * j. -f, f,-f ••j.'.
•'•y-s v
^;-
a
Si*
' V '" _, ^ "•<••* ' ; ^ ^ fr <,f f fff ' f * ffJ*Sflf '"'~'-' ' -'77/sH*"^v' ^r-'^vS^*^-
-, - , ,«/!!?:''"-~-.'^'^;' W» . v'\','4>ai«^fi-"''~>wi^'i^/^4fe
vjtstrfe r f%e;,: '^-^t,-, ^^'4f'' ^^^v-^,:--- -* ;^»^^
^u '^V^^Y-J-:: '<-^fe*>^ t^4^;v;^^
'* f-t' ," ^ .. ^ y,^f¥ /* , ' «• '# s X 5 ?' X' ^iStj:*"'
^ - ;t'ffS~'%y-'V*V^^;;'^'^5^^|^
'^'l'^/1'S^^f^,^^^
'',''-/ ^:-i v"?'^:;7^^'''^^^'
' -'^ *r?«5^-i^XV''-/fi'S
i—1.1 1- -\ ^'*^;; - v ,,^v"' 4«;"^,
'-,>», ' ^/'ff«"fff'f^,L',/^:-'",'•'-}''::y-',^'/
- <••;/?,"„':, ...^.:•'•" •"'••! *••"•••••• "i™11 K •',^-' ',;',.•• ^ f <,ff S "•'•" *' ^ *" '* ^
•si ^ } £•« '<:>^'^^^Hi^C
I II 1 11 ,:-;-;':-:vfc'i;;'£^?^
• •
m
a
LU .S
&
41
-------
Work Group 4: Data Needs (cont)
te
CA
03
IS
42
* U.S. OOVEraWENT PRINTING OFFICE: 1994-550-001/00176
------- |