United States
Environmental Protection
Agency
Office of Research and
Development
Washington, DC 20460
EPA-600/R-98-0035
April 1998
«>EPA Environmental
Technology Verification
Verification Testing of
Emissions from the
Combustion of
A-55® Clean Fuels
in a Firetube Boiler
-------
EPA Review Notice
This report has been reviewed by the U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily reflect the views and policy of
the Agency, nor does mention of trade names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Information Service,
Springfield, Virginia 22161.
-------
Verification Testing of Emissions from
the Combustion of A-55® Clean Fuels in
a Firetube Boiler
by:
C. Andrew Miller
U.S. Environmental Protection Agency
National Risk Management Research Laboratory
Air Pollution Prevention and Control Division
Research Triangle Park, NC 27711
EPA Cooperative Research and Development Agreement 0138-97
with
A-55 Limited Partnership
5270 Neil Road
Reno, NV 89502
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
-------
Abstract
Two emulsified fuels and one non-emulsified fuel were tested in a small (2.5xl06 Btu/hr [732 kW])
firetube package boiler to determine emissions of carbon monoxide (CO), nitrogen oxide (NO),
particulate matter (PM), and total hydrocarbons (THC), and to calculate the thermal efficiency of the
boiler using each of the fuels. Changes in emissions and thermal efficiency when using the
emulsified fuels were compared to the base fuels from which they were produced, or that they would
replace in normal usage. The fuels tested were a standard #2 fuel oil, the same #2 oil emulsified with
30% water by volume, and a fuel naphtha emulsified with 30% water by volume. The oil/water
emulsions were produced by A-55 Limited Partnership of Reno, Nevada, and were tested at EPA's
National Risk Management Research Laboratory, Air Pollution Prevention and Control Division in
Research Triangle Park, NC, under EPA's Environmental Technology Verification (ETV) Program.
Each of the fuels was tested at three different boiler loads.
NO emission concentrations from combustion of the emulsified #2 oil decreased 15 to 34%
compared to the #2 oil at the same loads. For the emulsified naphtha, NO emissions decreased 33 to
51% compared to the #2 oil. Reductions in NO emission factors (in lb/106 Btu [kg/kJ]) ranged from
22 to 37% for the emulsified #2 oil and from 37 to 54% for the emulsified naphtha, compared to the
#2 oil emission factors. CO and PM emissions from all the fuels were very low, with CO emissions less
than 8 ppm (at 3% Q2) in all cases, and PM emissions less than 5 mg/dscm in all cases (except for the
initial test run, for which higher PM emissions were suspected as being the result of entrainment of
particles previously on the boiler tubes). THC emission concentrations were typically less than 1
ppm for all cases.
Thermal efficiency typically was lower for the emulsified fuels than for the non-emulsified fuel, with
a drop of 2.5 percentage points for the emulsified #2 fuel and 3.4 percentage points for the
emulsified naphtha, compared to the #2 oil.
-------
Preface
The U.S. Environmental Protection Agency (EPA) established the Environmental Technology
Verification (ETV) program as a means to accelerate the commercialization of environmental
technology through objective verification and reporting of technology performance. The ETV
program approach is to evaluate technologies and report their performance characteristics, without
considering regulatory compliance requirements, ranking of performance, labelling as acceptable or
unacceptable, or determining "best available technology." This straightforward reporting of
technology performance is intended only to provide objective and quality-assured data for potential
technology users.
The ETV program is currently in its initial phase of determining the most effective approaches to
technology verification. Two of the verification approaches being evaluated are to arrange for
independent verification entities to conduct testing following standard test protocols developed for
the ETV program according to EPA requirements, or to have EPA conduct the testing directly using
a specified or developed protocol. Although the majority of verification testing is planned to be
conducted by independent verification entities, EPA's Office of Research and Development has both
substantial equipment and expertise to perform a number of verification tests at EPA's facilities. Such
"in-house" tests can be done when the testing is in line with the mission and resources of the
organization within EPA best suited to conduct the tests.
In the Spring of 1997, EPA's Air Pollution Technology Branch (APTB) of the National Risk
Management Research Laboratory's Air Pollution Prevention and Control Division (APPCD) was
asked to conduct a series of verification tests on emulsified fuel oils. APTB has operated combustion
equipment for study of pollution formation and control for over 20 years and was able to provide the
expertise and equipment necessary to conduct the requested tests. A cooperative research and
development agreement was negotiated with A-55 Limited Partnership to conduct the tests and testing
was conducted in the Summer and Fall of 1997.
in
-------
Acknowledgments
Substantial assistance in preparing for testing, reviewing test data, and verifying quality assurance was
given by Richard Shores of the Technical Services Branch of APPCD. Robert Russell of A-55
Limited Partnership provided considerable help in reviewing data. Dahman Touati and Charlie King
of ARCADIS Geraghty & Miller (formerly Acurex Environmental Corporation) provided the vast
majority of technical support in conducting these tests.
IV
-------
Contents
Abstract ii
Preface iii
Acknowledgments iv
List of Figures vii
List of Tables viii
Nomenclature and Symbols ix
Chapter 1 Introduction 1
EPA Environmental Technology Verification Program 1
Emulsified Fuel Oils 1
ETV Testing of A-55® Clean Fuels 2
Limitations of Results 2
Chapter 2 Equipment and Test Approach 4
North American Firetube Boiler 4
Emissions Measurement Instrumentation 4
NO/NOX Analyzer 5
O2 Analyzer 5
CO and CO2 Analyzers 5
THC Analyzer 6
Data Acquisition System 6
Extractive Sampling Methods 7
Total Particulate 7
Scanning Mobility Particle Sizer 7
Cascade Impactor 7
Thermal Efficiency Instrumentation 7
Measurement Method 7
Instrumentation 8
Test Matrix 8
Fuel Composition 10
Chapter 3 Thermal Efficiency Determination 13
Heat Losses 13
Heat Credits 16
Calculation of Thermal Efficiency 16
Chapter 4 Emission Results 17
Test Matrix Modifications 17
-------
Contents (Continued)
Emission Concentrations and Emission Factors 18
Carbon Monoxide 18
Nitrogen Oxide 22
Participate Matter 24
Total Hydrocarbons 27
Chapter 5 Thermal Efficiency Results 28
Energy Inputs 30
Heat Losses 30
Chapter 6 Quality Assurance 33
CEM, Temperature, and Flow Measurements 33
CEM Precision 33
CEM Accuracy 34
CEM Completeness 35
Temperature Data 36
Flow Data 36
Particulate Matter Measurements 36
Discrepancies 37
Audits 37
Chapter 7 Operational Observations 38
Emulsified #2 Oil 38
Emulsified Naphtha 38
References 39
Appendices
A English Engineering - International System Unit Conversions 41
B Fuel Oil Analyses 42
C Discrepancies 46
D Audit Results 47
Top-Loading Pan Balance Evaluation 47
Systems Audit of Particulate Matter Collection Process 47
VI
-------
Figures
2-1. Burner end of the North American Package Boiler 5
2-2. Schematic of North American Package Boiler 6
2-3. Schematic of thermal efficiency instrumentation for the North American Package Boiler ... 9
3-1. Energy flows into and out of the NAPB 14
4-1. Stack O2 concentration in volume percent for each test condition 19
4-2. Stack CO concentrations in ppm for each of the conditions, corrected to 3% O2 21
4-3. Emission factors for CO in lb/106 Btu and lb/1000 gal 21
4-4. Stack NO concentrations in ppm for each of the conditions, corrected to 3% O2 22
4-5. Emission factors for NO in lb/106 Btu and lb/1000 gal 23
4-6. Stack PM concentrations in mg/dscm for each of the conditions,
corrected to 3% O2 24
4-7. Emission factors for PM in lb/106 Btu and lb/1000 gal 25
4-8. SMPS particle size distributions for the three fuels tested at high load 26
5-1. Thermal efficiencies for each of the nine conditions in percent 29
5-2. Heat losses for each of the nine test conditions as a fraction of the total heat input 32
vn
-------
Tables
2-1. Planned test matrix for verification tests 10
2-2. Ultimate analyses of the fuel oils used in the test program 11
2-3. Trace metal content of the fuel oils tested, in (ig/g 12
4-1. Target and actual test conditions achieved during verification testing 17
4-2. Average emission concentrations of O2, CO, NO, PM, and THC
for the #2 oil, emulsified #2 oil, and emulsified naphtha 18
4-3. Average emission factors for CO, NO, and PM for #2 oil,
emulsified #2 oil, and emulsified naphtha 19
4-4. Percentage reduction in emission concentrations of CO, NO, and PM for
emulsified #2 oil and emulsified naphtha compared to the #2 fuel oil 20
4-5. Percentage reduction in emission factors of CO, NO, and PM for emulsified
#2 oil and emulsified naphtha compared to the #2 fuel oil 20
5-1. Parameters for determination of thermal efficiency 29
5-2. Thermal efficiencies and heat inputs and losses for all conditions tested 31
6-1. Data quality indicator goals for the project 33
6-2. Maximum percent RSD values for CEM and temperature measurements for all conditions . 34
6-3. Cross-run RSD values of the average CEM and temperature
measurements for all conditions 35
6-4. Average and maximum deviations of zero and high span CEM
readings from calibration gas values for all runs 35
6-5. Data quality indicator goals for PM measurements 36
6-6. Measurements of DQI goals for PM mass measurements 37
B-l. Reported ultimate analysis results for the fuels tested 43
B-2. Corrected analysis results for the fuels tested 43
B-3. Differences in calculated thermal efficiency values using the different fuel analyses 45
D-l. Performance evaluation results for the Mettler AE240 48
vin
-------
Nomenclature and Symbols
BA energy supplied by the combustion air, Btu/hr
BCM energy supplied by the condensate makeup water, Btu/hr
BF energy supplied by the fuel sensible heat, Btu/hr
Cb pounds of carbon per pound of as-fired fuel
cp A heat capacity of the air in Btu/lb-°F
cp F heat capacity of the fuel in Btu/lb-°F
cp FG heat capacity of the flue gas, Btu/lb-°R
cp; heat capacity of the ith constituent of the flue gas, Btu/lb-°R
cp w specific heat of water, Btu/lb-°F
Dp particle diameter, (am
F fuel heat input, Btu/lb
fH percent hydrogen in the fuel (not including hydrogen associated with the moisture)
fMp percent moisture content of the fuel
href enthalpy of saturated liquid at the reference temperature (68 °F), Btu/lb
hWG enthalpy of the vapor in the flue gases at the stack temperature and vapor partial pressure
(generally assumed to be 1 psia), Btu/lb
KF higher heating value of the as-fired fuel, Btu/lb
KHC heat content of unburned hydrocarbons in the flue gases, Btu/ft3
L total heat loss, Btu/hr
Lc convective heat transfer loss from the boiler surface, Btu/hr
Lco loss of energy due to the failure of all CO to be completely converted to CO2
LFG loss associated with the sensible heat of the dry flue gas, Btu/hr
LL loss of energy through leaks of boiler combustion gas, feedwater, and/or steam, Btu/hr
LMF loss associated with the moisture in the fuel, Btu/hr
LMH loss associated with the conversion of hydrogen to water in the combustion process, Btu/hr
LR radiative heat transfer loss from the surface of the boiler, Btu/hr
LUBC loss associated with unburned carbon in the captured particulate, Btu/hr
LUHC loss of energy associated with emissions of unburned hydrocarbons, Btu/hr
LWG heat loss due to the moisture in the flue gases, Btu/hr
PC concentration of carbon in the ash, %vol
Pco concentration of CO in the flue gas, %vol
Pco2 concentration of CO2 in the flue gas, %vol
PHC concentration of hydrocarbons in the flue gas, %vol
pN2 concentration of N2 in the flue gas, %vol
p02 concentration of O2 in the flue gas, %vol
SFG specific weight of the flue gas at standard conditions, fWlb
TA ambient air temperature, °F
TCA combustion air temperature, °F
TCM condensate makeup water temperature, °F
TF fuel temperature, °F
TFG flue gas temperature, °F
Tre ambient air temperature, °F
V particle volume, cm3
WA mass flow rate of the combustion air, Ib/hr
WCM condensate makeup water flow rate, Ib/hr
WF mass flow rate of the fuel flow, Ib/hr
WFG mass flow rate of flue gas, Ib/hr
Wp mass flow rate of the particulate, Ib/hr
IX
-------
Nomenclature and Symbols (Continued)
%i molar fraction of the ith constituent of the flue gas
r| thermal efficiency, %
-------
Chapter 1
Introduction
EPA Environmental Technology Verification Program
In 1994, the U.S. Environmental Protection Agency (EPA) Office of Research and Development formed
a workgroup to plan the implementation of the Environmental Technology Verification (ETV) Program .
The goal of ETV is "to verify the environmental performance characteristics of commercial-ready
technology through the evaluation of objective and quality assured data, so that potential purchasers and
permitters are provided with an independent and credible assessment of what they are buying and
permitting."! ETV is currently sponsoring 12 verification pilots, covering a range of technology areas
including indoor air products, site characterization and monitoring, drinking water systems, and air
pollution control technologies.
Although these pilots are partially funded by EPA during the initiation of the program, the intent of the
ETV program is to create an on-going program that is primarily funded by program generated funds
mainly from participants. In keeping with this intent, the Air Pollution Prevention and Control Division
(APPCD) of EPA's National Risk Management Research Laboratory (NRMRL) entered into a
cooperative research and development agreement (CRADA) with A-55 Limited Partnership (A-55) of
Reno, Nevada, to conduct verification testing of the emulsified fuels produced by A-55. This testing was
conducted in the Summer and Fall of 1997 at EPA's Environmental Research Center in Research
Triangle Park, North Carolina, by personnel of APPCD's Air Pollution Technology Branch (APTB) and
their on-site contractor, ARCADIS Geraghty & Miller (formerly Acurex Environmental Corporation).
Emulsified Fuel Oils
Emulsions have been proposed for many years as a means of reducing the emissions of criteria pollutants
from the combustion of fuel oils. A number of studies have shown the ability of emulsions of water
suspended in oil to reduce the emissions from combustion sources2-4; however, the impacts of oil/water
emulsions on particular pollutants vary. For heavy fuel oils, oil/water emulsions tend to reduce
particulate, but in general have had a smaller effect on either carbon monoxide (CO) or oxides of
nitrogen (NOX) when operating conditions are kept constant.2 With distillate oils, particulate matter (PM)
and NOX have been shown to be reduced when using an oil/water emulsion compared to using the same
oil without emulsification, but CO emissions were not significantly changed.3 The use of an emulsified
fuel results in improved secondary atomization of the fuels, often allowing operation at a reduced
stoichiometric ratio, and also tends to reduce the peak combustion temperature. Both of these effects
result in lower NOX emissions, and the improved atomization can also result in lower CO and PM
emissions. Emulsified oils appear to have little impact on emissions of hazardous air pollutants (HAPs)
(compounds which are listed as hazardous under Title III of the Clean Air Act Amendments of 19905).
A study conducted by EPA concluded that emissions of organic HAPs remained relatively unchanged for
an emulsified heavy fuel oil and the same oil that was not emulsified.6 Since metal emissions depend
primarily upon the amount of metal in the fuel, the major impact on emulsified fuel metal emissions (per
unit energy) will depend on the amount of metal (if any) in both the water and the emulsifying agent. In
systems with particulate control equipment, the use of an emulsified fuel may also affect metal emissions
if the particle size distribution changes in such a way that the net particulate removal efficiency is altered.
The effect of these fuels on operating efficiency will vary according to the particular characteristics of the
fuel and the system in which it is used. For combustion systems that rely on the expansion of gases, the
water contained in the emulsified fuel can provide additional expansive energy as it is heated along with
the combustion products. In other systems where heat transfer is the primary mode of energy transfer,
too much water can cause the thermal efficiency to drop because energy is required to heat the water in
-------
the fuel, rather than that energy's being transferred to the process. However, using emulsified fuels can
allow a boiler to be operated with less excess air, which in turn reduces the energy required to heat the
atmospheric nitrogen and excess oxygen. In short, the thermal efficiency of a unit using emulsified fuels
may either increase or decrease compared to the efficiency of the unit using the non-emulsified (base)
fuel, depending upon the combustor type and the characteristics of the fuel.
The key disadvantage to the use of emulsions in the past has been the ability of the water to remain in
suspension during storage. One method of avoiding this problem has been to mix the oil and water
immediately prior to feeding the mixture into the boiler. However, this requires additional fuel and water
handling equipment, as well as a system to mix the two liquids. The additional expenses associated with
this equipment have not usually been considered worth the resulting reductions in pollutant emissions.
As an alternative to separate storage of the oil and water, emulsifying agents that result in a reduced rate
of oil/water separation have been developed, allowing "premixed" emulsified oils to maintain their
properties for extended periods of time when properly stored. This approach eliminates the need for
additional handling and mixing equipment, and utilizes existing fuel handling systems, thereby reducing
the cost of use. Current emulsifying agents are much more effective at inhibiting phase separation,
allowing emulsified fuels to be effectively used in a variety of applications.
ETV Testing of A-55® Clean Fuels
The CRADA between EPA and A-55 was designed to verify the performance of A-55® Clean Fuels in a
small well-instrumented firetube boiler. The A-55® Clean Fuels included "premixed" oil/water emulsions
of a #2 fuel oil (diesel) and a petroleum naphtha. Performance of these fuels was compared to the
performance of the non-emulsified #2 fuel oil.
The intent of the project was to produce objective data on the pollutant emissions and thermal efficiency
of the fuels to provide potential fuel buyers, users, and regulators with information regarding the
environmental characteristics of the A-55® Clean Fuels. This report discusses the testing approach and
calculations used during the tests, the results of the tests, quality assurance (QA) goals and measures, and
operational observations noted during testing.
The objective of these tests is to determine the changes in emissions of CO, nitrogen oxide (NO), and PM,
and in boiler thermal efficiency when an emulsified fuel is used in place of a non-emulsified fuel. The
tests were designed to ensure that the comparison reflected the impact of using the emulsified fuel, rather
than changes in base fuel properties (e.g., nitrogen content) or in the operating characteristics of different
combustion systems. To minimize the number of variables influencing emissions, the (non-emulsified)
fuel was first combusted in the test boiler under typical operating conditions, followed by testing of the
emulsified fuel in the same unit, again under conditions typically maintained when using an emulsified
oil. The results are indicative of the ability of the emulsified fuel to affect pollutant emissions and
thermal efficiency under the particular conditions of the test.
Limitations of Results
Changing combustion conditions or the system in which the fuels are used can have a considerable effect
on emissions and thermal efficiency. It is impossible to develop a limited test protocol that would cover
all the possible permutations of operating conditions and combustion system configurations in which an
emulsified fuel may be appropriate for use. The approach taken for these tests was to limit the test
conditions to a single boiler, and to draw conclusions regarding the performance of the emulsified fuel
using the limited test data and an understanding of the physico-chemical processes associated with the
fuel's use. This approach provided data as well as an indication of the advantages and disadvantages
associated with combustion of emulsified fuels. However, the wide range of possible operating conditions
and combustion systems makes emphasis of the limitations critical. The quantitative results apply directly
-------
only to the system tested under the conditions tested. While it is expected that these fuels will behave
similarly in other systems, use of these fuels in systems with either different hardware or operating
conditions may not result in similar performance. Note that performance may be either better or worse
than the results reported here. In any case, it is anticipated that individual systems will require
optimization to achieve their best results.
-------
Chapter 2
Equipment and Test Approach
The approach taken to verify the performance of the emulsified fuels compared the pollutant emissions
and the thermal efficiencies of the emulsified fuels to the same parameters measured during the
combustion of the same fuels without the water or emulsifying agent. Measuring only the emissions and
performance of the emulsified fuels would not provide any reference with which to compare. Three
different fuels were tested: a #2 (diesel) fuel oil, the same #2 oil emulsified with water (emulsified #2 oil),
and a petroleum-naphtha/water emulsion (emulsified naphtha). The #2 oil/water emulsion and the
naphtha/water emulsion were compared with the #2 oil, since both the #2 oil/water emulsion and the
naphtha/water emulsion are designed to replace the #2 oil in practice.
The fuels were burned in a small firetube package boiler at EPA's Environmental Research Center.
While this unit is typical of many small institutional or commercial boilers used to generate low pressure
steam, it is of a very different design than a large watertube boiler typically used in large industrial or
utility applications. Thus, the performance results cannot be directly compared to these larger systems.
However, significant changes in performance between the base and emulsified oils on this small boiler are
expected to indicate similar changes in terms of direction and magnitude for larger scale systems.
Therefore the results obtained in these tests can be used to evaluate the potential for emission reductions
and thermal performance for other external combustion steam generating systems.
North American Firetube Boiler
The tests were performed on APPCD's North American package boiler (NAPB) which is capable of firing
natural gas or #2 through #6 fuel oils. The boiler is a three-pass firetube "Scotch" marine-type design
built in 1967, model 5-360H-D, shown in Figure 2-1 and schematically in Figure 2-2. The burner is a
North American model 6121-2.5H6-A65 rated at 2.5 x 106 Btu/hr* and has a ring-type natural gas
burner and an air-atomizing center nozzle oil burner capable of firing #2 through #6 oils. The boiler
has 300 ft2 of heating surface and generates up to 2400 Ib/hr of saturated steam at pressures up to 15
psig. Heat is extracted from the steam through a heat exchanger to an industrial cooling water system
that provides the boiler load. Oil temperature can be adjusted using an electric heater to maintain proper
oil viscosity, and both fuel and atomizing air pressures are variable to ensure adequate oil atomization.
The flue gases from the unit pass through a manifold to an air pollution control system (APCS)
consisting of a natural-gas-fired secondary combustion chamber, an acid gas scrubber, and a fabric filter
to ensure proper removal of pollutants generated during tests designed to mimic poor combustion
conditions. During the tests reported here, the APCS was operated to provide a constant draft to the
NAPB to minimize changes in the induced draft. Although this type of boiler normally operates under
forced draft only, the imposition of an induced draft due to the APCS did not significantly affect boiler
emissions.
Emissions Measurement Instrumentation
The NAPB is fully instrumented with continuous emission monitors (CEMs) for NOX, CO, CO2, O2, SO2,
and total hydrocarbons (THC). The CEM panel for the NAPB uses seven gas analyzers, each with
multiple ranges, and flue gas conditioning equipment. Effluents from the stack are carried to the CEM
panel through heated Teflon tubing to a gas dryer and filter.
"See Appendix A for conversion of units to metric system equivalents.
-------
Figure 2-1. Burner end of the North American Package Boiler.
NO/NOX Analyzer
The NO/NOX analyzer is a Rosemount Analytical Model 951A that operates via chemiluminescence. In
the NO measurement mode, NO is directly measured by light emission from the reaction of the NO with
ozone supplied by the analyzer. In the NOX measurement mode, a portion of the sample is diverted to a
converter where the NO2 is dissociated into NO, and the resulting NO then measured using the analyzer.
The monitor has selectable ranges of 0-3; 0-10; 0-30; 0-100; 0-1,000; 0-3,000; 0-10,000; and 0-30,000
ppm of NO or NOX. Testing used either the 0-100 or 0-1000 ppm range. The analyzer is accurate to
0.5% of full scale.
O2 Analyzer
The O2 analyzer is a Rosemount Analytical Model 755R analyzer that operates using the paramagnetic
property of oxygen. Other gases present in significant concentrations in combustion flue gases do not
exhibit paramagnetism. Measurement ranges for the instrument are 0-5, 0-10, and 0-25% of O2. The 0-
5% scale was used during testing. The analyzer accuracy is specified to be 1% of full scale.
CO and CO2 Analyzers
The CO and CO2 analyzers are Rosemount Analytical Model 880A Non-Dispersive Infrared Analyzers.
They operate by directing identical infrared beams through an optical sample cell and a sealed optical
reference cell. A detector located at the opposite end of each cell continuously measures the difference
in the amount of infrared energy absorbed within each cell. The difference is a measure of the
concentration of the component of interest in the sample. The ranges of the CO2 analyzer are 0-5, 0-15,
and 0-25%. The ranges of the low CO monitor are 0-500, 0-1000, and 0-2000 ppm, and the ranges of
the high CO monitor are 0-1, 0-3, and 0-5%. The CO2 analyzer was operated using the 0-25% range.
-------
Computerized
Data Acquisition
System
Fuel
Temperature
& Pressure
Fuel Flow
Steam
Pressure and
Temperature
• OEMs and Flue Gas
Temperature \
Extractive
'Sampling
Combustion Gases
Figure 2-2. Schematic of North American Package Boiler.
The 0-500 ppm scale on the low CO monitor was used during these tests. Analyzer accuracies for the CO
and CO2 monitors are both 1% of full scale.
THC Analyzer
A Rosemount Analytical Model 402 hydrocarbon analyzer was used to measure THC content of the flue
gas. The analyzer uses a flame ionization detector (FID) and a heated temperature-controlled sample line
with associated electronics. The hydrocarbon sensor in the analyzer uses a burner where a regulated flow
of sample gas passes through a flame sustained by regulated flows of fuel and air. Within the flame, the
hydrocarbon components of the sample undergo a complex ionization that produces electrons and
positive ions. Polarized electrodes collect these ions, causing current to flow through the measuring
circuitry of the analyzer. The resulting current is proportional to the concentration of hydrocarbons in
the sample. The heated sample line allows the sample to be maintained at the desired temperature to
prevent condensation of the heavier hydrocarbon fractions. The analyzer ranges are 0-3; 0-10; 0-30; 0-
100; 0-300; 0-1,000; 0- 3,000; 0-10,000; and 0-30,000 ppm. The analyzer was operated using the 0-100
ppm range. Accuracy of the analyzer is 1% of full scale.
Data Acquisition System
A computerized data acquisition system (DAS) was used to record CEM measurements as well as steam
and flue gas temperatures. The DAS uses a Macintosh computer and Strawberry Tree data acquisition
software. An analog-to-digital converter card allows up to eight standard analyzer outputs to be input to
the computer. The software enables preliminary data calculations during data collection, and both raw
data and calculated values can be recorded to a disk file for later retrieval. The system can take readings
at periodic intervals of 0.1 second or more. All recorded data are time and date stamped to ensure that
logged data correlate with data taken manually or otherwise.
-------
Extractive Sampling Methods
Total Particulate
An EPA Method 5 sampling train7 is used to collect participate samples to determine total suspended
particle loading in the flue gases of the boiler. Three samples per test condition were planned to evaluate
the repeatability of the results. Results are reported both as flue gas concentrations (|ig/m3) and as
emission factors (lb/106 Btu or lb/1000 gal fuel).
Scanning Mobility Particle Sizer
Particle size distributions can be determined for a limited range of particle sizes using a scanning
mobility particle sizer (SMPS). The SMPS measures particle size distributions for particles ranging from
0.3 tolO (am in diameter. These distributions were taken to determine the relative changes between
conditions rather than to measure the absolute concentrations of particles of given sizes. Because of the
nature of the SMPS, QA measurements to evaluate the precision and accuracy of the results were not
made for these tests. Results from the SMPS are reported as relative changes between test conditions.
Cascade Impactor
Particle size distributions can also be measured using a cascade impactor. Cascade impactors classify
particles according to their ability to follow gas flow streamlines. Particles of large mass are unable to
follow sudden changes in flow direction as the gas passes from one stage to the next, and are deposited
on the stage filter. Each stage of the impactor is designed to collect different particle sizes, and the final
particle weights on each stage provide data on the size distribution of the sampled particle stream. A
California Air Resources Board (CARB) Method 5018 was used as the basis for the test method, with the
modifications outlined below:
(1) The CARB method recommends several "trash" runs be discarded to allow for unfamiliarity
with equipment, poor initial conditions, or other factors. These trash runs are then to be
followed by seven actual runs to ensure that valid data are collected. Due to the non-critical
nature of the impactor data for these tests, three Method 501 runs were determined to be
adequate to provide the desired information.
(2) Method 501 requires an in situ sampling approach (i.e., the entire impactor is inserted into
the flow). However, the impactor is too large to insert into the NAPB stack without severe
flow disruption. For these tests, the impactor was mounted external to the stack, and a
buttonhook nozzle was used to collect the sample from the flue gas stream. Method 501 also
requires use of a straight sampling nozzle, which is possible only if the in situ sampling
procedure is followed.
(3) Method 501 requires that upper limits of 50 mg total particle mass and 15 mg particle mass
per stage be captured to minimize the possibility of particle carryover from one stage to the
next. The test procedures are such that these limits should be achieved; however, the non-
critical nature of the impactor measurements allows for exceedance of these limits if an
assessment of the test conditions and results indicates that the measurements are valid, even if
the limits are not met.
Thermal Efficiency Instrumentation
Measurement Method
Thermal efficiency can be determined in a number of ways. For boilers, the most widely accepted
method of thermal efficiency measurements for boilers is the American Society of Mechanical
Engineers (ASME) Power Test Code (PTC) 4.1 - Steam Generating Units.9 PTC 4.1 covers units
ranging from large utility steam generating units and combined cycle systems to high temperature
water heaters and is also applicable to firetube package boilers such as the North American unit on
which the current tests were conducted.
PTC 4.1 provides two primary methods for determining thermal efficiency: the input-output method
-------
and the heat loss method. The input-output method relies on calculations of the energy input from
the fuel and the energy output of the steam and requires accurate measurement of flow rates,
temperatures, pressures, and energy contents of the different process flows. The heat loss method
measures the energy inputs and losses of energy through radiation, flue gases, and other routes. The
difference between the measured energy input and the energy losses is the energy absorbed by the
steam. The heat loss method requires measurements of fuel heat content, composition, flow, and
temperatures; external boiler temperatures; and flow, composition, and temperature of the boiler flue
gases.
Accurate values of the absolute thermal efficiency are meaningful only for the particular boiler being
tested. In the current tests, the change in boiler thermal efficiency is of much more importance than
the absolute thermal efficiency, because each boiler is unique and the particular efficiencies measured
for the NAPB cannot be easily transferred to other units, even those that are of identical design. Thus
the focus for the current tests was to calculate the change in thermal efficiency as fuels are changed.
This will allow the measurements to focus on the major changes of the fuel's energy content and flow
rate and changes in the flue gas composition and temperature. It was expected that other parameters
such as boiler skin temperature would remain relatively constant for the different fuels, and would not
substantially impact the efficiency measurements.
The current tests used the heat loss method for determining the boiler thermal efficiency. Additional
detail on this method is presented in the following chapter.
Instrumentation
Figure 2-3 shows a schematic of the boiler/heat exchanger system with the measurement points and
parameters. The primary quantities required to determine the thermal efficiency of the boiler are the
energy input from the fuel and other sources and the heat losses through the stack (including sensible
heat of the dry flue gas constituents, the energy content of the water vapor, and the energy in the
unburned carbon, CO, and hydrocarbons), leaks, and heat transfer to the surroundings. The CEMs
are used to determine the composition (Cl in Figure 2-3), and extractive sampling methods are used
to determine stack flow rate F2 and fuel heat content HI. Thermocouples are used to measure
temperatures at Tl, T3, and T4, and flow totalizers provide total flow into the unit at Fl and F3.
Secondary measurements are taken at various points on the system to ensure proper operation. These
measurements are shown in the smaller circles in Figure 2-3, and include stack flue gas pressure, fuel
pressure, cooling water inlet and outlet temperatures, and cooling water flow. Also included in the
secondary measurements are steam temperature T2 and pressure PI. These parameters are secondary
because the heat loss method of determining thermal efficiency is used rather than the input-output
method.9
Fuel flow was measured using a Brooks-Oval Mini-Oil Flowmeter, Model LS-21312, which operates
by using a slight pressure drop across the meter to drive a pair of oval gears. The meshed gears seal
inlet from outlet to generate the pressure differential. The meter is designed to remain unaffected by
changes in liquid viscosity, density, and lubricity, allowing the same meter to be used for a wide range
of fuels. Instrument accuracy is specified as ±0.5% of full scale (200 gpm).
Test Matrix
The test matrix was developed to ensure that test results reflected the performance of the emulsified
oil under a range of load conditions and to allow comparison with the performance of the non-
emulsified fuel it was designed to replace. For each test condition, three Method 5 samples were
planned to be taken, and three SMPS sampling runs were planned. CEM and thermal efficiency
measurements were planned for each test run. Each test run was expected to last approximately 2
hours, during which CEM measurements were planned to be taken continuously and logged every 20
seconds. Four test runs were planned for each test condition. The test conditions were chosen based
primarily on load, with target loads of l.SxlO6 Btu/hr (low load), 2.0xl06 Btu/hr (medium load), and
-------
Heat
Exchanger
Condensate
Return
Fuel
Feed
TsYmCondensate
Makeup
Temperature/"^ Primary Measurement
Pr@SSUr© ^^^
Q Secondary Measurement
Energy Content (not continuous)
Composition
Flow
Figure 2-3. Schematic of thermal efficiency instrumentation for the North American Package Boiler. Instruments
used for primary measurement of thermal efficiency are in the large circles, while secondary
instrumentation is shown in the small circles.
2.5xl06 Btu/hr (high load). Table 2-1 shows the test matrix used for each oil tested in the verification
tests.
To the extent possible, the high, medium, and low loads were to be held constant for all the fuels
tested. This would require significant increases (on the order of 50%) in the emulsified fuel flow
relative to the non-emulsified fuel in order to match loads.
In addition to load, it was expected that the stack O2 level would also be varied for the emulsified fuel
test conditions. In general, the secondary atomization created by the water in the emulsified fuels
allows the combustion air, and thus the stack O2, to be reduced relative to the base fuel O2 level. The
target O2 level was 3% for the base fuels. The target O2 levels for the emulsified fuels was determined
based on the minimum O2 level that could be achieved without increasing either CO or PM (as
-------
Table 2-1. Planned test matrix for verification tests. An "X" denotes measurements to be taken during the
specified test run. This matrix is to be repeated for each of the conditions tested.
I Run |
I CEM Measurements (Gaseous emissions) I
I Thermal Efficiency Measurements I
I Method 5 Samples3 (Total particulate) I
I SMPS Samples3 (Particle size distribution) I
1
X
X
X
X
I Cascade Impactor3 (Particle size distribution) I
2
X
X
X
X
X
3
X
X
X
4
X
X
X
X
X
a. Only three Method 5, SMPS, and cascade impactor samples per condition were
planned. The three runs in which these samples were taken were allowed to vary for
each condition.
measured by Bacharach smoke number) over the values measured for the corresponding base fuel.
Fuel Composition
The fuels used in the test matrix varied in their composition and characteristics. Table 2-2 provides
the ultimate analyses of the fuels. Because of the high level of water in the emulsified fuels, there was
some concern regarding the impact of the water on the analyses. The fuel analyses were conducted
by independent laboratories following American Society of Testing and Materials (ASTM) methods.
These methods rely on combustion of the fuel sample and analysis of the combustion gases for CO2
and water (H2O) to determine the carbon and hydrogen contents of the fuel, with oxygen being
determined by difference.10 The moisture content is determined by distillation of the sample prior to
conducting the analysis for hydrogen,n but the possibility arises that the water is not completely
driven off by the distillation process. In such an instance, the subsequent hydrogen analysis (which
relies on the water from the combustion of the sample) may indicate that the fuel contains more
hydrogen than actually present due to the excess water contained in the fuel.
This possibility was considered to be likely due to the unexpectedly high level of oxygen (7.67%)
originally reported in the fuel analysis for the emulsified #2 oil. Even accounting for any oxygen in
the proprietary emulsifying additives, the oxygen level was expected to be no more than 1%. It was
hypothesized that the reported moisture content was lower than actually present in the emulsified fuel
and that this difference was responsible for the higher reported oxygen content. Since the fuel
hydrogen and moisture contents strongly influences the thermal efficiency due to the combustion-
generated water in the flue gas (see Chapter 3), it was important to ensure that the fuel analyses were
consistent with the known base oil analyses and the amounts of water added during the emulsification
process.
Given the fact that the oxygen level was reported to be significantly higher than anticipated,
additional analyses were done to determine whether the reported oxygen values were actually as high
as reported, or whether there was the possibility of error in the analyses. One testing laboratory
reported the hydrogen and oxygen associated with the water in the total fuel hydrogen and oxygen
contents, and also reported water in volume percent. Corrections were made to these reported results
to yield hydrogen and oxygen values separate from that included in the water, and to convert the
water content to weight percent. Details of the analyses and corrections are given in Appendix B.
The #2 oil, emulsified #2 oil, and emulsified naphtha were also analyzed for trace metal content using
standard fuel analysis methods. Concentrations of antimony, arsenic, beryllium, cadmium,
chromium, copper, iron, lead, magnesium, mercury, nickel, selenium, vanadium, and zinc were
measured, and the results of these analyses are given in Table 2-3. For all the fuels, the trace metal
contents were consistently low. The only metals with concentrations over 1 (jg/g in all three
10
-------
Table 2-2. Ultimate analyses of the fuel oils used in the test program.
ION
1
| % Carbona'b
1 % Hydrogen
| % Oxygen0
1 % Nitrogen
| % Sulfur
| % Water
| % Ash
| Higher Heating
j Value, Btu/lb
| Specific Gravity
j (60 °F)
#2 Oil
86.92
13.01
0.42
0.49
0.03
<0.05
0.001
19,450
0.8607
Emulsified #2 Oil
57.40
8.77
2.42
0.48
0.009
30.93
0.003
12,786
0.9050
Emulsified |
Naphtha 1
53.36 |
9.16 |
1.10 |
0.32 |
0.002 |
36.07 |
0.01 |
12,584 |
1
0.8309 1
1
a.Percents are by weight.
b.Methods: Carbon, Hydrogen, Nitrogen-ASTM D 52911°; Sulfur- ASTM D 429412- Water- ASTM
D 9511; Heating Value - ASTM D 24013; Ash - ASTM D 482.14
c.Determined by difference.
fuels were iron and vanadium, although antimony in the emulsified naphtha was measured at 1.5
(ig/g. Since the emulsifying agent was composed of organic hydrocarbons, it would be expected that
the metal concentrations for the emulsified #2 oil would be lower than those for the #2 oil. However,
this was not the case for all metals. For some of the metals, the concentrations were low enough to be
within the measurement error of the analysis method; however, for iron there was a considerable
increase in the emulsified #2 oil compared to the base #2 oil. It is speculated that this increase may
be due to contamination from the oil drum. Although some metals could be introduced by the water
used in the emulsification, the water was deionized prior to mixing, which should have removed
nearly all metals.
11
-------
Table 2-3. Trace metal content of the fuel oils tested, in ug/g.
Antimony
Arsenic
Beryllium
Cadmium
Chromium
Copper
Iron
Lead
Magnesium
Mercury
Nickel
Selenium
Vanadium
Zinc
#2 Oil
0.3
<0.1
<0.5
0.05
0.3
0.10
4
0.2
0.26
0.10
0.82
<0.1
4.68
0.6
Emulsified #2
Oil
0.4
<0.1
<0.5
0.10
0.5
0.19
56
0.5
NAa
0.09
0.19
<0.1
2.77
1.0
Emulsified
Naphtha
1.5
<0.1
<0.5
0.08
0.4
0.17
15
0.3
0.50
0.06
0.17
<0.1
2.60
0.7
a. NA - Not Available
12
-------
Chapter 3
Thermal Efficiency Determination
The heat loss method* relies on measurements of the input energy (the energy flowing into the
system with the fuel and air) and energy losses; i.e., energy that is not absorbed by the steam. Such
losses include energy carried out of the system by the flue gases and unburned fuel, energy
radiated from the boiler skin to the surroundings, and energy escaping the boiler from leaks. The
ASME PTC 4. 1 defines efficiency through the heat loss method as:
,,=100%- ,. losses \
1 ( Heat in fuel + Heat credits J v '
where heat credits involve energy inflow through the boiler feedwater and combustion air. The
heat-in-fuel term is the product of the fuel's higher heating value" and the flow rate of the fuel to
produce energy per unit time (in this case, 106 Btu/hr). A schematic of energy flows for the NAPB
is shown in Figure 3-1.
Heat Losses
Heat losses are illustrated in Figure 3-1 in underlined text. The major heat loss is through the
sensible heat in the flue gases; however, other heat losses may also be significant, depending upon
the operating characteristics of the particular boiler. In addition to flue gas heat loss, energy may
also be lost through leaks of boiler water or combustion gases; the presence of CO, unburned
hydrocarbons, and/or unburned carbon in the flue gases; or the presence of water in the fuel. The
total heat loss is simply the sum of those losses:
L = LFG+ LWG+LL + Lco + LUHC + LUBC + LR + Lc (3-2)
where LFG is the sensible heat loss in the dry flue gases, LWG is the heat loss from the moisture in
the flue gases, LL is the loss due to boiler and combustion gas leaks, LCo is the loss due to the
presence of CO (rather than CO2) in the flue gas, LUHC is the loss associated with the failure of all
the hydrocarbons to completely burn, LUBC is the loss associated with unburned carbon in the
captured particulate, LR is the radiative heat transfer loss from the surface of the boiler, and Lc is
the convective heat transfer loss from the boiler surface. Each of the losses is calculated in Btu/hr.
LFG is calculated by:
LFG = WFGCP|FG(TFG-TA) (3-3)
where WFG is the flow rate of flue gas in Ib/hr, CP;FG is the specific heat of the flue gas in Btu/lb-°F,
and TFG and TA are the temperatures of the flue gas and ambient air, respectively, in °F. The mass
flow rate of the flue gas can be determined by using the following equation based on measurements
*The more accurate term is "energy loss." However, "heat loss" is the term used in ASME PTC 4.1 to include
both actual heat losses and other energy losses such as those due to unburned carbon.
** This is the higher heating value at constant pressure.
13
-------
^ ., , Unburned
Flue Gas (Sensible Carbon, CO,
Heat Loss and and Unburned
Steam Moisture Losses) Hydrocarbons
Fuel
(Chemical
Energy)
F
Combustion
Air Sensible
Heat
BA
1 LFG' LMF' ' LUBC' LCO'
^ | LMH LUHC
Boiler
Conductive and
Radiative Heat Losses
C ' R
Fuel Combustion
(bensioie Qas Leaks LL
Heat)
Bp
Figure 3-1. Energy flows into and out of the NAPB. Following the terminology of ASME PTC-4.1, energy inputs are
shown in bold, energy losses are underlined, and heat credits are in italics.
of the flue gas constituents:
^44.01pc02+32.00p02+28.02pN2+28.01pco
12.01 (pc02+Pco)
12.01fs
°b+ 32.07
WF
(3-4)
where Pco2, Pco Po2, and PN2 are the measured concentrations (in volume percent) of the specified
flue gas constituents, Cb is the pounds of carbon per pound of as-fired fuel, fs is the fraction of sulfur
in the as-fired fuel, and WF is the fuel flow in Ib/hr. The flue gas specific heat can be calculated by
using:
Cp,FG~ -3 ^p,i%i (3-5)
i
where CP;; and %; are the specific heat and molar fraction, respectively, of constituent i of the flue gas.
LWG is the heat loss due to the moisture in the flue gases and is the sum of LMF and LMH, where LMF is
the loss associated with the moisture in the fuel and LMH is the loss associated with the conversion of
hydrogen to water in the combustion process. LMF can be calculated from:
VF (3-6)
where fMF is the percent moisture content of the fuel, hWG is the enthalpy of the water vapor in the
flue gases at the stack temperature and vapor partial pressure (generally assumed to be 1 psia) in
Btu/lb, and href is the enthalpy of saturated liquid water at the reference temperature (68 °F), also in
14
-------
Btu/lb. hWG and href are determined from standard ASME steam tables.
LMH is calculated using:
LMH = 8.936fH(hWG-href)WF (3-7)
where fH is the fraction of hydrogen in the fuel (not including hydrogen associated with the
moisture) and hWG, href, and WF are as defined above. The 8.936 is the number of pounds of water
produced from the complete combustion of a pound of hydrogen.9 From Eqs. (3-6) and (3-7), LWG
is then given by:
LwG=(fMF + 8.936 fH)(hWG-href)WF (3-8)
LL represents the loss of energy through leaks of boiler combustion gas, feedwater, and/or steam.
This value is calculated from an estimated leak rate times the energy content of the leaking material.
It is assumed that this value is very small relative to the other losses, particularly for a small unit where
such leaks can be easily spotted, and LL is neglected for these calculations.
LCO is the loss of energy due to the failure of all CO in the flue gas to be completely converted to
CC>2, and is given by:
Lco = 10,160 — CbWF (3-9)
con+n
Pco2 Pco
where pco and pco2 are the percent by volume concentrations of CO and CO2, respectively, in the flue
gas, and Cb is the pounds of carbon burned per pound of as-fired fuel. The 10,160 value is the heat
released in Btu when burning 1 Ib of CO to CO2.9
LUHC is the loss of energy associated with emissions of unburned hydrocarbons, and is given by:
= PHcWFGKHC
UHC 100sFG ( '
where pHc is the concentration of hydrocarbons in the flue gas in percent, WFo is the mass flow rate
of the flue gas in Ib/hr, KHc is the heating value of the hydrocarbons in Btu/ft3, and SFG is the specific
weight of the flue gas in fWlb. KHc is usually considered to be that for methane, roughly 1010
Btu/ft3. SFG (at 68 °F, 14.7 psia) can be calculated from:9
5^=0.0401
'FG
PcO2 . Po2 . PcO . PN2 . PsO2
1545
35.11 48.28 55.16 55.14 24.12 16.00
(3-11)
is the energy loss associated with the emission of unburned carbon in the fly ash, and is
calculated by:
LUBC=14,500Wppc (3-12)
where Wp is the mass flow rate of the particulate in Ib of ash/lb of fuel, and pc is the mass fraction of
carbon in the ash. The 14,500 value is the heating value in Btu of 1 Ib of carbon as it occurs in
refuse.9
LR and Lc are the losses due to radiative and convective heat transfer, respectively, from the boiler
15
-------
surface to the surroundings. These values are highly unit specific and can also change substantially
as conditions surrounding the unit changes (e.g., increase or decrease in ambient air temperature). In
instances such as the current tests, where the surroundings are relatively constant* and no significant
changes are expected to occur due to changes in the fuels, LR and Lc are expected to be negligible
compared to the other loss components.
Heat Credits
Heat credits are measures of energy flows into the boiler other than through the fuel (shown in italics
in Figure 3-1) and are summarized as:
B= BA + BF (3-13)
where B is the total heat credit, BA is the energy supplied by the combustion air, and BF is the energy
supplied by the fuel sensible heat, all in Btu/hr. BA is calculated by:
BA = WACpiA(TCA-Tref) (3-14)
where WA is the flow rate of the combustion air in Ib/hr, CP;A is the specific heat of the air in Btu/lb-°F,
TCA is the temperature of the combustion air in °F, and Tref is the ambient air temperature, also in °F.
BF is given by:
BF = WFcpF(TF-Tref) (3-15)
where WF is the flow rate of the fuel in Ib/hr, cp F is the specific heat of the fuel in Btu/lb-°F, and TF is
the temperature of the fuel in °F.
Calculation of Thermal Efficiency
To calculate the efficiency, the above losses are calculated and the thermal efficiency is then
determined using Eq. (3-1) written as:
T|=100%-|— ^=— 1 100% (3-16)
where F is the heat input through the fuel, given by:
F = WFKF (3-17)
where KF is the higher heating value of the as-fired fuel in Btu/lb, measured at constant pressure.
*The unit is located inside a completely enclosed building and is not exposed to weather.
16
-------
Chapter 4
Emission Results
Tests were conducted in blocks of four runs per condition. Due to the limited availability of the test
oils, each of the four replicate test runs per condition was conducted sequentially. CEM data reported
below are corrected to account for CEM drift using:
XC = (A-X0)-
(4-1)
where Xc is the corrected gas concentration, A is the average value from CEM measurements, X0 is
the average of the pre- and post-test zero calibration readings, Xm is the average of the pre- and post-
test high span calibration readings, and Xma is the actual high span calibration gas concentration.
Reported CEM concentration data are also corrected to 3% O2 (all gas concentrations are given in
either volume % or ppmv, at dry conditions).
Test Matrix Modifications
During the course of testing, several changes were made to the test matrix shown in Table 2-1. The
reasons for these changes are given below and are also discussed in Appendix B. Table 4-1 shows the
actual test conditions achieved during the tests.
Target O2 concentration for the baseline #2 oil was originally expected to be 3% for all loads.
However, very little measurable CO or particulate was noted at this O2 level, and so it was decided to
reduce the O2 level of the medium load condition to improve the baseline thermal efficiency of the
unit. The use of the lower O2 level for the #2 oil at medium load therefore allowed a direct
comparison of the effects of emulsifying the oil, since the O2 levels at medium load for the emulsified
Table 4-1. Target and actual test conditions achieved during verification testing.
Condition
Fuel
Target Load
(1Q6Btu)
Actual Load
(1Q6Btu)
Target O2 (%)c
Actual O2 (%)c
Test Runs
1
#2 Oil
2.0
2.20
3.0
2.95
4
2
#2 Oil
1.5
1.63
1.5
1.48
4
3
#2 Oil
1.0
1.46
3.0
2.99
4
4
Ems#2a
2.0
2.23
1.5
2.10
4
5
Ems #2
1.5
1.59
1.5
1.50
4
6
Ems #2
1.0
1.45
1.5
2.42
4
7
Ems Napb
2.0
2.09
1.5
1.97
5
8
Ems Nap
1.5
1.58
1.5
1.49
4
9
Ems Nap
1.0
1.42
1.5
2.43
4
a. Emulsified #2 oil
b. Emulsified naphtha
c. Target O2 levels were estimates. Actual test O2 levels were determined by matching particulate emissions (as
measured by smoke number) of the base fuel and the emulsified fuel.
17
-------
#2 and the emulsified naphtha were set at the same nominal O2 level as the baseline #2 medium load
tests.
Emission Concentrations and Emission Factors
Emission results for the nine conditions (the #2 oil, the emulsified #2 oil, and the emulsified naphtha,
each at high, medium, and low loads) are presented in Table 4-2 in terms of pollutant concentrations.
Figure 4-1 presents the average O2 concentrations for each of the test conditions. Since one of the
advantages normally associated with the use of emulsified fuels in external combustion applications is
the ability to operate at reduced O2 levels, the emulsified fuels are operated at O2 levels below those of
the base fuel, with the exception of the #2 oil medium load condition noted above.
Table 4-3 presents the calculated emission factors in lb/106 Btu and lb/1000 gal of fuel for the
different conditions. For the emulsified fuels, the 106 Btu and the 1000 gal represent the heat content
and the volume, respectively, of the oil/water/emulsifying agent mix, not of just the oil in the
emulsified fuel. Note that although the volumetric fuel flow increased, the net Btu input remained
constant for constant load, since negligible energy was provided by the water and emulsifying agent.
The effect of using the emulsified fuel is seen in Tables 4-4 and 4-5, which present the percent
change in emissions of the emulsified fuels compared to the corresponding base (non-emulsified)
fuels. Table 4-4 presents the percent change in emission concentrations for the emulsified #2 oil, and
the emulsified naphtha compared to the #2 oil, and Table 4-5 shows the percent change in emission
factors for the emulsified fuels compared to the #2 oil.
Carbon Monoxide
In general, the CO emissions were very low for all conditions tested, with average values for all
conditions falling below 10 ppmv (see Tables 4-2 and 4-3). Considerable fluctuation was noticed
between the individual test runs, leading to relatively large relative standard errors (See Chapter 6) for
each condition. In addition, there were large percent changes in CO emissions when comparing the
emulsified fuels with their corresponding base fuel; however, these large percentage changes are not
highly significant, due to the very low absolute CO concentration levels. CO levels of less than 10
ppmv are typically considered quite low for practical combustion systems. In addition, the CO CEM
is accurate only to 1% of full scale (500 ppmv), meaning that measurements of less than 5 ppmv are
essentially the same. Nevertheless, values less than 5 ppmv are reported here for completeness.
Table 4-2. Average emission concentrations of O2, CO, NO, PM, and THC for the #2 oil, emulsified #2 oil, and
emulsified naphtha for the different conditions tested.
| Fuel
ILoad
02 (%)
1 CO(ppmv@3%O2)
1 NO(ppmv@3%O2)
1 PM (mg/dscm @ 3% O2)
|THC(ppmv@3%O2)
#2 Oil
High3
2.95
3.63
127.
11.67
0.52
Medb
1.48
2.71
96.1
0.77
0.24
Low0
2.99
1.64
105.
0.95
NDd
Emulsified #2 Oil
High3
2.10
2.30
84.3
4.58
0.70
Medb
1.50
1.71
79.5
3.11
ND
Low0
2.42
2.87
88.7
3.26
ND
Emulsified Naphtha
High3
1.97
7.76
61.9
2.87
ND
Medb
1.49
2.72
62.2
4.67
ND
Low0
2.43
2.06
70.3
4.88
0.33
a. 2.1x106 Btu/hr
b. 1.6x106 Btu/hr
c. 1.4x106 Btu/hr
d. Not Detected
18
-------
I I High Load
• MedLoad
0.0
12 Oil
Em ulsif i e d #2 Oil Bin ilsi fi ed Nap hit ha
Figure 4-1. Stack O2 concentration in volume percent for each test condition. Error bars denote relative
standard deviation for the four test runs at each condition.
Table 4-3. Average emission factors for CO, NO, and PM for #2 oil, emulsified #2 oil, and emulsified naphtha for
the different conditions tested.
Fuel
Load
CO(lb/106Btu)
CO(lb/1000gal)
NO(lb/1#Btu)
NO(lb/1000gal)
PM(lb/1#Btu)
PM(lb/1000gal)
#2 Oil
High3
0.0025
0.35
0.102
14.3
0.0077
1.08
Medb
0.0022
0.31
0.078
10.9
0.0006
0.09
Low0
0.0013
0.19
0.084
11.8
0.0007
0.10
Emulsified #2 Oil
High3
0.0018
0.18
0.067
6.41
0.0031
0.30
Medb
0.0014
0.13
0.063
6.05
0.0019
0.18
Low0
0.0023
0.22
0.070
6.74
0.0028
0.27
Emulsified Naphtha 1
High3
0.0068
0.58
0.048
4.10
0.0019
0.16
Medb
0.0021
0.18
0.049
4.15
0.0038
0.32
Low0 I
0.0016 I
0.14 I
0.055 I
4.68 I
0.0035 I
0.30 I
a. 2.1x106 etu/hr
b. 1.6x106 Btu/hr
c. 1.4x106 Btu/hr
19
-------
No substantial changes were noted in CO emissions for the fuels tested (#2 oil, emulsified #2 oil, and
emulsified naphtha). In all cases but one, the emissions were below 5 ppmv. As seen in Table 4-2
and Figure 4-2, the CO emissions from the emulsified naphtha at high load were somewhat higher
than for either the #2 oil or the emulsified #2 oil. Even so, the CO emissions for these conditions
were still less than 10 ppm, and could likely be reduced further if necessary by optimizing the
combustion O2 level. This potential was illustrated during setup testing for the emulsified naphtha,
when the boiler was operated at the nominal 2xl06 Btu/hr load and 3.2% O2, resulting in CO
emissions less than 2 ppm. CO emission factors are shown in Table 4-3 and Figure 4-3. On a per
unit energy basis, CO emissions were typically 0.002 lb/106 Btu or less, except for the emulsified
naphtha at high load, which was roughly 0.007 lb/106 Btu. The emissions per volume of fuel had a
higher variation, due primarily to the differences in fuel energy content per unit volume between the
base #2 oil and the emulsified fuels. Per unit volume, the #2 oil CO emission factors ranged between
Table 4-4. Percent reduction in stack gas concentrations of CO, NO, and PM for emulsified #2 oil and
emulsified naphtha, compared to the #2 fuel oil. Comparisons are made between conditions at
similar boiler load.
| Fuel
| Load
| CO
| NO
1 PM
Emulsified #2 Oil
High3
36.6
33.6
60.8
Medb
36.9
17.2
-301.
Low0
-74.8
15.2
-243.
Emulsified Naphtha
High3
-114
51.2
75.4
Medb
-0.36
35.3
-503.
Low0
-25.2
32.7
-413.
a. 2.1x106 Btu/hr
b. 1.6x106 Btu/hr
c. 1.4x106 Btu/hr
Table 4-5. Percent reduction in emission factors of CO, NO, and PM for emulsified #2 oil, and emulsified naphtha,
compared to the #2 fuel oil. Emulsified naphtha results are in comparison to the #2 fuel oil.
Comparisons are made between conditions at similar boiler load.
| Fuel
I Load
Emulsified #2 Oil
High3
Medb
Low0
Emulsified Naphtha
High3
Medb
Low0
| lb/106Btu
| CO
| NO
| PM
26.3
34.8
59.3
38.0
18.9
-208.
-71.7
16.8
-305.
-176.
53.0
74.9
2.99
37.4
-519.
-21.3
34.9
-416.
| lb/1000gal
| CO
| NO
I PM
49.4
55.3
72.0
57.4
44.3
-111
-17.9
42.9
-178
-68.2
71.4
84.7
40.8
61.8
-275.
26.0
60.3
-215.
a. 2.1x106 Btu/hr
b. 1.6x106 Btu/hr
c. 1.4x106 Btu/hr
20
-------
High Load
fifed Load
Em u Isifi ed #2 Oil Em ul sifi ed N aphth a
Figure 4-2. Stack CO concentrations in ppm for each of the conditions, corrected to 3% O2. Error bars denote
relative standard deviation for the four test runs at each condition.
u.uu/.
0.006^-
c
o
u
I
c.
O
D
HighLoad
MedLwd
Lav.-1 Load
Brnl3ifiEd#2Oi! EmubifiedNaphtha
Figure 4-3. Emission factors for CO in lb/106 Btu and lb/1000 gal.
21
-------
0.19 and 0.35 lb/1000 gal, while for the emulsified fuels emission factors ranged between 0.13 and
0.58 lb/1000 gal, again with the emulsified naphtha at high load exhibiting the highest emissions.
These values compare to an emission factor of 5 lb/1000 gal for distillate fuel combustion in a
commercial/institutional boiler listed in EPA's AP-42.16
As discussed above, the percent change in CO emissions should not be regarded as highly significant
with respect to the performance of the emulsified fuels, due to the substantial fluctuations in CO
emissions across the four test runs at each condition and the low absolute levels for CO emissions for
all conditions. As noted above, optimization of the operating conditions beyond what was done for
these verification tests is likely to result in lower CO emissions, if such low values are desired.
Nitrogen Oxide
Emission measurements were much more stable for NO than for CO, as measured by the relative
standard error across the individual test runs for each condition. The average values for each
condition therefore have a substantially higher degree of confidence with respect to the average
values than the CO emission values.
Emissions of NO are presented in Table 4-2 and Figure 4-4. NO emissions from the emulsified fuels
showed significant reductions compared to the base #2 oil. The baseline emissions of NO averaged
127 ppm at high load, with a low of 96 ppm at medium load (unless otherwise noted, all
concentrations are corrected to 3% O2). At low load, NO emissions from the #2 oil were 105 ppm,
slightly higher than the medium load emissions, due to the lower O2 level at the medium load (1.48%
vs. 2.99%). These are compared to the average emulsified #2 oil emissions of 84, 80, and 89 ppm at
high, medium, and low loads, respectively. These values represent reductions of approximately 34,
(VI
O
£
Tf
o
u-
1
cf
D
X3
I
ft.
E
8
o
z:
ii
i
*2 Oil
Emulsified $0. Gil Emulsified Naphtha
Figure 4-4. Stack NO concentrations in ppm for each of the conditions, corrected to 3% O2. Error bars denote
relative standard deviation for the four test runs at each condition.
22
-------
17, and 15% compared to the base #2 oil emissions at high, medium, and low loads, respectively. For
the emulsified naphtha, emissions were even lower, with 62 ppm and a 51% reduction at high load, 62
ppm and a 35% reduction at medium load, and 70 ppm and a 33% reduction at low load. It is
significant to note that at medium load, where the Q^ level remained nearly constant for the three
fuels, NO emissions fell 17% for the emulsified #2 oil and 35% for the emulsified naphtha compared
to the base #2 oil. These differences can be attributed to the use of the emulsified fuels rather than to
any difference in operating conditions.
Emission factors for NO are presented in Table 4-3 and Figure 4-5. On a per unit energy basis,
emission factors for NO ranged from 0.078 lb/106 Btu at medium load to 0.102 lb/106 Btu at high
load for the #2 oil, from 0.063 lb/106 Btu at medium load to 0.070 lb/106 Btu at low load for the
emulsified #2 oil, and from 0.048 Ib/lO6 Btu at high load to 0.055 lb/106 Btu at low load for the
emulsified naphtha. Although the NO concentrations are significantly lower for the emulsified
naphtha than for either of the other two fuels, the emissions per unit energy input are roughly the
same.
On a per unit volume basis, NO emission factors for the #2 oil ranged from 10.9 lb/1000 gal at
medium load to 14.3 lb/1000 gal at high load. As was the case for CO, the emission factors
calculated per unit volume dropped significantly for the emulsified fuels due to the water in the fuel.
For the emulsified #2 oil, the emission factors varied from 6.1 lb/1000 gal at medium load to 6.7
lb/1000 gal at low load, and for the emulsified naphtha, the emission factors ranged from 4.1 lb/1000
gal at high load to 4.7 lb/1000 gal at low load. These values compare to the AP-42 emission factor of
20 lb/1000 gal for distillate fuel oil in both utility and industrial boilers.16
I I High Load |g| Low Load
HI Med Load
o 0.00
£
UJ
o
Oil Emulsified #2 ai Emulsified Naphtha
Figure 4-5. Emission factors for NO in lb/106 Btu and lb/1000 gal.
23
-------
Particulate Matter
PM emissions, presented in Table 4-2 and Figure 4-6, were very low, ranging from a high of 11.7
mg/dscm at high load for the #2 oil to a low of 0.77 mg/dscm at medium load for the #2 oil. There is
some uncertainty regarding the value of the PM concentration at high load, because the boiler tubes
were cleaned immediately prior to the start of the test program. During the initial test runs (#2 oil at
high load), the particulate captured on the Method 5 filter dropped considerably from the first to the
last run, even though the test conditions remained relatively constant. The PM emissions measured
during runs 1, 2, and 3 of condition 1 were 17.26, 12.14, and 5.60 mg/dscm, respectively. The
following six runs resulted in PM concentrations of between 0.65 and 1.31 mg/dscm during the
medium and low load tests. For this reason, it is believed that the high PM emissions seen during
condition 1 were due to entrainment of particles already present on the boiler tubes that were
loosened during the cleaning process. The emulsified #2 oil PM emissions were between 3.11
mg/dscm at medium load and 4.58 mg/dscm at high load, while the PM emissions from the emulsified
naphtha ranged from 2.87 mg/dscm at high load to 4.88 mg/dscm at low load. The increases in PM
emissions for the two emulsified oils may have been due in part to the lower O2 levels used during
these test conditions. However, for the medium load cases, the O2 levels were very consistent for all
three fuels, with average O2 levels ranging between 1.48 and 1.50% for the three test conditions. The
PM emissions at medium load for both emulsified fuels increased over those measured for the base
#2 oil at medium load, leading to the conclusion that factors other than the O2 level were responsible
for the observed differences in PM emissions for this load. PM emissions can be affected by the
atomization at the nozzle, which is a function of fuel flow rate, pressure, and viscosity; the flow rate
and pressure of the atomizing fluid (in this case, air); and the design of the nozzle. Optimization of
these parameters may reduce PM emissions, but such optimization was not conducted during these
tests.
Emission factors for PM are presented in Table 4-3 and Figure 4-7. On a per unit energy basis, the
PM emission factors for the #2 oil ranged from 6. IxlO-4 lb/106 Btu at medium load to 8.0xlO-3
18-
O
V
O
O
to
50
E
UJ
High Load
Med Load
Low Load
#2 Oil
Em u Is if ied #2 O il Em u Is F ted Nap htha
Figure 4-6. Stack PM concentrations in mg/dscm for each of the conditions, corrected to 3% O2.
24
-------
lb/106 Btu at high load. For the emulsified #2 oil, the emission factors ranged from 1.9xlO-3 lb/106
Btu at medium load to S.lxlO-3 lb/106 Btu at high load, while the emulsified naphtha PM emission
factors ranged from 1.9xlO-3 at high load to 3.8xlO3 at medium load. PM emission factors in terms
of mass per volume of fuel feed for the #2 oil ranged from 0.09 lb/1000 gal at medium load to 1.12
lb/1000 gal at high load. For the emulsified #2 oil, the PM emission factors ranged from 0.19
lb/1000 gal at medium load to 0.32 lb/1000 gal at high load. The PM emission factors for the
emulsified naphtha ranged from 0.16 lb/1000 gal at high load to 0.32 lb/1000 gal at medium load.
The AP-42 emission factor for filterable PM is listed at 2.0 lb/1000 gal for distillate oil for both
utility and industrial boilers.16
Although current regulations limit emission rates of PM, in terms of either mass or opacity, the
distribution of particle sizes may become an important factor with recent concern over health effects
associated with ambient concentrations of particles smaller than 2.5 pn. For that reason, these tests
also measured particle size distributions where possible.
The emissions of total particulate were so low that determining size distributions was very difficult,
and consistent size distributions with the cascade impactors were not possible. In most cases, the
cascade impactor substrates exhibited weight loss from the initial to the final weighings. This was due
to the very low levels of particulate passing through the impactor.
Repeatable data obtained from the SMPS indicated that the particle sizes in the flue gas from the fuels
were in the size range from 0.01-0.1 (am (10-100 nm). Figure 4-8 shows representative particle size
distributions from the SMPS for all fuels at high load. This figure shows there are clear differences
0,
0,
0,
= 0.
ufO,
o
50-
o,
0,
c
o
s
5
CL
008
007-:
006-i
00 5-j
004 J
QQ3-;
002J
001
0
1.2
High Load %# Low Load
Med Load
1.0:-
1* °-8^-
§ 0.6
S 0.44—
#2 Oil
Em iJsi i ed # 2 Oil Em y teif i e d Na pht ha
Figure 4-7. Emission factors for PM in lb/106 Btu and lb/1000 gal.
25
-------
in the distributions between the emulsified and non-emulsified fuels. Figure 4-8 plots dV/d[log(Dp)],
where dV is the differential particle volume and Dp is the particle electrostatic diameter. By assuming
a constant particle density, one can use this plot as an indication of the distribution of particle mass as
a function of particle size.
The greatest difference in the size distributions is between the #2 oil and the emulsified #2 oil. The
size distribution for the #2 oil shows a distinct fine mode (peak) near 0.02 (jm (20 nm), consistent
with a nucleation mode of particle formation. From the peak near 0.02 (am, the particle mass
decreases with increasing particle size until a minimum is reached near 0.06 (jm. At this point, the
mass begins to increase with increasing particle size, which suggests the presence of combustion
chars. These particles cover a significant size range, up to 100 (jm. The emulsified #2 oil, in contrast,
does not show a significant decrease in particle mass with increasing size, but shows a shallow
irregular increase in mass as the particle size increases from about 0.02 (am to approximately 0.2 (am,
at which point it begins to decrease slowly. This distribution is consistent with the secondary
atomization process characteristic of emulsified fuels, in which the mean fuel droplet size decreases
due to the microexplosions of the water inside the fuel. Unlike mechanical atomization, which forms
relatively large droplets that usually have a characteristic peak in the droplet size distribution,
secondary atomization tends to create a continuum of droplet sizes. This results in a particle size
distribution with fewer and less distinct maxima or minima. In terms of particle mass, the emulsified
1x10
12.
1X101
s
Q
V
1x10*
-=-*
V Emulsified #2 QI
A Emulsified Naphtha
0.01
0.1
Diameter,
Figure 4-8. SMPS particle size distributions for the three fuels tested, at high load.
26
-------
fuels tend to shift the average size of particles toward the smaller sizes. This behavior is seen in the
difference between the size distributions of the #2 oil and the emulsified #2 oil. The lack of a distinct
minimum near 0.02 (am in the emulsified #2 oil distribution indicates that there are significantly
more particles in the 0.02-0.2 (am range than for the #2 oil.
The size distribution for the emulsified naphtha does not behave in a manner similar to that for the
emulsified #2 oil, but retains the bimodal shape of the #2 oil, although at a higher total mass and at a
larger average size. It is not clear how this distribution compares to non-emulsified naphtha, but
there are clearly more particles in the 0.02-0.1 (am range for the emulsified naphtha than for the #2
oil.
While the secondary atomization of the fuels by the water in the emulsification most likely is the
major influence on particle size, it must also be noted that there were substantial changes in the fuel
flow rates between the non-emulsified and emulsified fuels. For instance, the #2 oil flow rate
averaged 15.8 gal/hr at high load (2.204xl06 Btu/hr), while the measured flow rate of the emulsified
#2 oil at nearly identical load (2.228x106 Btu/hr) averaged 23.2 gal/hr. Such changes in flow rates
can affect the particle size distribution, as can changes in atomizing air pressure (which were held at
relatively constant levels for the corresponding fuels). Additional testing would be required to
determine the impacts of each of these parameters; however, the emulsification is believed to play the
dominant role in changing the particle size distributions.
Total Hydrocarbons
Emissions of THCs were very difficult to measure, particularly for the emulsified fuels. The
difficulties arose due to the low concentrations of THCs and high levels of water in the flue gases. It
has been shown that the presence of water in flue gases can impact the performance of THC monitors
using flame ionization detection, by introducing a small negative bias to the reading. Estimates of the
negative bias were on the order of 1-5 ppm.17 This bias would have the effect of reducing the
instrument reading below the actual value, but would not mask changes or high levels (>15 ppm) in
THC concentrations. The behavior was noted when burning the emulsified #2 oil and the emulsified
naphtha, resulting in average THC readings consistently below zero, even though the CEM was
calibrated before and after each test day. The negative values therefore indicate that the THC levels
were likely to be below 5 ppm.
For the #2 oil, THC emissions were less than 1.5 ppm for all cases. The only measured value greater
than zero for the emulsified #2 oil was at high load, at a concentration of 0.7 ppm. THC emissions
were measured at quantities greater than zero in the emulsified naphtha only at the low load condition
and then only at a level of 0.3 ppm.
Given the low measured values and the impact of the high water content on the performance of the
THC analyzers, the measurements given above have a low level of confidence as quantitative values.
Qualitatively, however, it can be stated that the THC emissions from the fuels at the tested conditions
were found to be very low.
27
-------
Chapter 5
Thermal Efficiency Results
Changes in thermal efficiency associated with the use of alternative fuels often play a major role in
the acceptability of those fuels. In the case of fuels emulsified with water, there may be a significant
efficiency penalty associated with the use of those fuels, due to the physical requirement of heating
the water in the fuel to steam. As the steam exits the boiler stack, any energy that was expended in
the phase change from liquid to steam is not used to heat the process fluid, thereby becoming an
energy loss and reducing the thermal efficiency of the process. Conversely, the use of emulsified
fuels can allow an operator to reduce the amount of combustion air. The water included in the
emulsion creates "microexplosions" as it evaporates in the flame, resulting in a secondary
atomization of the fuel and producing smaller fuel droplets that can burn more efficiently. This
allows less excess O2 to be used and, because each mole of O2 carries along with it 3.76 moles of
nitrogen (N2) when using air, it also reduces the amount of the relatively inert N2 that passes through
the combustion system. Since the N2 does not react, it acts only as a heat sink that is heated and
carries that energy out the boiler stack. As with the steam from the water in the fuel, this heat is lost
to the process and thus a reduction in thermal efficiency results. By reducing the amount of
combustion air required to burn the same amount of fuel, an increase in thermal efficiency may be
realized. Finally, the introduction of water into the combustion process may also reduce the flue gas
exit temperature. Since the amount of energy leaving the boiler stack depends upon the volume and
temperature of the flue gas, a drop in flue gas temperature will result in an efficiency gain. The
degree to which these competing effects cancel each other can depend upon the unit design, the
amount of water being used in the emulsified fuel, operating practices, and other factors.
Thermal efficiencies were calculated for each run using the heat loss method described in Chapter 3.
Energy losses through the flue gases are subtracted from the total heat input via the fuel and other
sources (such as the energy in the air) to determine the amount of energy that is transferred to the
steam. This figure is then divided by the total heat input to determine the boiler thermal efficiency
(see Eq. 3-17). The energy losses are calculated for each of the primary routes of energy loss from
the system. Because the efficiency relative to the baseline condition is more critical in this study than
the absolute efficiency, losses due to convection and radiation from the boiler or from boiler leaks
were not calculated, since it was felt that these losses would remain relatively constant for all cases, and
in any case would not be substantially affected by the choice of fuel.
Thermal efficiency is significantly affected by the composition of the fuel, particularly the moisture
and hydrogen contents, as seen in Equations 3-6 and 3-7, respectively. A number of fuel analyses
were conducted for this study, particularly for the emulsified fuels. The results of these analyses
varied considerably in their hydrogen, oxygen, and moisture contents, differences that potentially
could change the calculated thermal efficiency results. A discussion of the different results and the
final analyses used are presented in Appendix B. The fuel analyses used in the efficiency
calculations are given in Table 2-3.
The average thermal efficiency for each test condition is shown in Figure 5-1. As seen in Figure 5-1,
the variation between test runs for each condition was small, as measured by the relative standard
deviation for each condition. Table 5-1 shows the parameters that are used (in combination with fuel
and flue gas compositions) to calculate the average thermal efficiency. The heat inputs and losses
due to each of the major parameters influencing the efficiency are shown for each test condition in
Table 5-2.
28
-------
100
98
96J
High Load
Med Load
Low Load
#20!
Emulsified $2 QI Emulsified Naphtha
Figure 5-1. Thermal efficiencies for each of the nine test conditions. Error bars indicate the relative standard
deviation. High load averaged 2.08x106 Btu/hr, medium load averaged 1.55X106 Btu/hr, and low
load averaged 1.38x106 Btu/hr.
Table 5-1. Parameters used for determination of thermal efficiency.
I I
1 Condition 1
| #2 Oil, High Load |
| #2 Oil, Medium Load |
| #2 Oil, Low Load |
| Emuls#2Oil, High Load |
1 Emuls #2 Oil, Medium Load 1
1 Emuls #2 Oil, Low Load 1
1 Emuls Naphtha, High Load 1
1 Emuls Naphtha, Medium Load 1
1 Emuls Naphtha, Low Load 1
Fuel Flow,
Ib/hr
113.
83.9
75.2
174.
124.
113.
168.
128.
115.
Air
Temperature, °F
85a
91.0
95.7
93.9
93.0
83.6
81.5
85.5
89.7
Stack
Temperature, °F
429.
347.
340.
397.
344.
341.
397.
344.
324.
Flue Gas Flow,
scfmb
392.
266.
261.
377.
261.
249.
350.
261.
245.
a. Measurements for air temperature were not available for condition 1 (#2 oil at high load). An estimate
of 85°F was made for each of the test runs at condition 1.
b. Standard conditions are 77 °F, 1 atm.
29
-------
It should be emphasized that the thermal efficiency results presented here are highly specific to
the test conditions, fuels, and equipment used, although it is expected that similar trends would be
found for these fuels used in other systems. The thermal efficiency values presented here are
measurements of the thermal efficiencies determined only for the package boiler used in the tests:
efficiency measurements using other types of units, or even for similar units in different condition,
operating under different parameters, are likely to be different. However, the addition of water
and the emulsification agent to a fuel oil is expected to result in similar relative changes when used
in other units.
For the #2 oil, emulsified #2 oil, and emulsified naphtha, the impacts on thermal efficiency of
using the emulsified fuels can be clearly seen in comparison to the efficiency for the base #2 oil.
The thermal efficiency was lowest for the high load cases, primarily due to the higher flue gas
temperature which resulted in significantly higher energy flows out the stack. For the base #2 oil,
the thermal efficiency was 85.1% at high load compared to 87.6% at medium load and 87.3% at
low load. For the emulsified #2 oil, the efficiencies were 83.3% at high load, 84.7% at medium
load, and 84.5% at low load. For the emulsified naphtha, thermal efficiencies ranged from a low
of 82.1% at high load to 83.9% at low load.
Energy Inputs
Energy into the boiler was primarily from the fuel, although a small amount of energy entered the
system from the combustion air. In general, the energy in the combustion air was between 3,900
and 14,200 Btu/hr for all tests, while the energy in the fuel accounted for between 1.41 and
2.25xl06 Btu/hr for the three fuels. Relatively little difference was noted in total energy input
between the three fuels at similar loads.
Heat Losses
The major heat losses for all cases were the heat loss in the dry flue gases and the heat loss due to
moisture from hydrogen in the fuel. In all cases, the heat loss due to moisture in the fuel was no
more than the third largest heat loss. However, the major difference between the emulsified fuels
and the base #2 oil was the heat loss due to moisture in the fuel. The average heat loss due to
moisture in the fuel for the three base #2 oil conditions is roughly 50 Btu/hr, compared to between
41,100 and 64,800 Btu/hr for the emulsified #2 oil and between 47,500 and 71,400 Btu/hr for the
emulsified naphtha.
Heat losses in the dry flue gases decreased for the emulsified fuels compared to the base #2 oil.
This is due to two causes: (1) lower excess air was used during emulsified fuel operation, leading
to lower total mass flow of flue gases; and (2) the flue gas temperatures decreased, resulting in
lower energy flow out the stack in the dry flue gases. Other losses, from incomplete combustion
of hydrocarbons, CO, and carbon in the ash, accounted for no more than 6,900 Btu/hr in all cases.
Figure 5-2 presents the heat losses as a percent of the total energy input for each of the nine
conditions.
30
-------
Table 5-2. Thermal efficiencies and heat inputs and
of four test runs. Heat inputs and losses
losses for all conditions tested. Values represent the average
are in Btu/hr.
1
1
1 Condition
1
|#2Oil, High Load
|#2Oil, Medium Load
| #2 Oil, Low Load
|Emuls#2Oil, High Load
1 Emuls #2 Oil, Medium Load
1 Emuls #2 Oil, Low Load
1 Emuls Naphtha, High Load
| Emuls Naphtha,
j Medium Load
1 Emuls Naphtha, Low Load
Efficien-
cy, %
85.1
87.6
87.3
83.3
84.7
84.5
82.1
83.8
83.9
Heat input
through fuel
2,204,000
1,632,000
1,462,000
2,228,000
1,590,000
1,445,000
2,111,000
1,616,000
1,446,000
Other heat
inputs
10,800
9,400
11,300
16,000
11,100
7,800
10,100
9,100
9,100
Losses in
dry flue
gases
165,000
86,000
82,500
142,000
82,100
78,500
135,000
83,700
73,600
Losses
from
moisture
in fuel
69
49
44
64,800
45,300
41,100
72,800
54,600
48,400
Losses
from
hydrogen
moisture
161,000
115,000
103,000
164,000
115,000
104,000
165,000
124,000
110,000
I
Other |
losses j
1
3,500 |
2,100 |
1,400 |
3,800 |
2,600 |
1,500 |
7,400 |
1,900 |
1
1,800 I
31
-------
&
Y,
ii
ra
i
^
FG
HF
WF
UBC
UHC
CO
Figure 5-2. Heat losses for each of the nine test conditions as a fraction of the total heat input. FG is the heat
loss of the dry flue gases out the stack, HF is the heat loss of the water from the hydrogen in the
fuel, WF is the heat loss due to the water in the fuel, UBC is the heat loss due to unburned carbon in
the particulate, UHC is the heat loss due to the unburned hydrocarbons in the flue gases, and CO is
the heat loss due to CO in the flue gases. Losses due to unburned carbon were negligible and are
not shown.
32
-------
Chapter 6
Quality Assurance
This project was conducted under an approved APPCD Level II Quality Assurance (QA) Project Plan.
The plan set forth the operating, sampling, and analysis procedures to be used during the testing, as
well as the data quality indicator (DQI) goals for the project. The DQI goals for the project are
shown in Table 6-1.
CEM, Temperature, and Flow Measurements
CEM Precision
Table 6-2 presents the maximum relative standard deviation (RSD) values calculated for each of the
test runs for each test condition. RSD is determined by calculating the standard deviation of the CEM
data for a particular run divided by the average measurement value of that run. Data for CO2 and NO
were all within the DQI goals for precision.
The data for O2 met the DQI precision goal of <7% RSD for only 6 of 13 test conditions, indicating
much higher variation in O2 levels during testing than for either CO2 or NO. The failure of the O2
measurements to achieve the DQI precision goal for these runs does not impact the conclusions of
this report concerning NO or PM emissions. However, the O2 measurements are less precise than
desired. This impacts the NO and PM emissions that are reported as concentrations corrected to 3%
02.
THC and CO data did not meet the precision DQI precision goal for any of the test conditions. The
goals for these two compounds were met for none of the test runs for CO, and for only one test run
for THC. For both THC and CO, this large variation was most likely due to the fact that both
compounds were detected at levels near zero. The measured values for THC and CO do not meet the
DQI goals set for this project, making the quantitative values questionable; however, the qualitative
results that both THC and CO emissions were near zero in all cases remain valid.
The variation of the four replicate test runs was also measured by calculating the RSD of the four test
runs' average values for each condition. This value, the cross-run RSD, was calculated by dividing
Table 6-1. Data quality indicator goals for CEM, temperature, and fuel flow measurements.
Measurement
02
CO2
CO
NOX
THC
Temperature
Fuel Flow
Method
CEM
CEM
CEM
CEM
CEM
Thermocouple
Volume Totalizer
Precision (RSDa), %
<7
<5
<7
<5
<5
10
10
Bias, %
<±10
<±15
<±15
<±10
<±10
±10
±10
Completeness, %
>90
>90
>90
>90
>90
>90
>90
aRelative standard deviation
33
-------
the standard deviation of the four runs' average value by the average of the four runs' average
values. The cross-run RSD between test runs was not a critical DQI. These values are shown in Table
6-3 for each of the test conditions. The CO and THC measurements again showed the greatest
variation across the tests.
For these three measurements, the highest cross-run RSD was 14.9% for O2 measured from the
emulsified naphtha at high load. The other conditions showed maximum values typically less than
2% for CO2, and less than 6% for O2 and NO.
For CO and THC, the cross-run RSD values were quite high, as high as 418%. This was due to the
very low values of both CO and THC. These measurements were essentially at the noise level of the
instruments, leading to relatively high variability in the measurements and to the large cross-run RSD
values.
CEM Accuracy
The accuracy of the CEMs was determined by daily pre- and post-test calibration of the instruments.
The deviations from the zero and high span calibration gas concentrations are shown in Table 6-4 for
the five flue gas constituents measured. NO met the DQI accuracy goal of less than ±10% deviation
in accuracy as measured by the percent deviation from the high span calibration gas concentration,
both for the average and maximum deviations for any one run. The O2 values met the DQI accuracy
goal of less than ±10% deviation for all test days. The average deviation for the O2 analyzer high
span reading was -0.01% and 0.23% for the NO analyzer high span reading. The maximum
deviation from the zero point was a CEM reading of -0.20% for O2 and 17 ppm for the NO. For O2
the high span gas concentration was 8.01%, and for NO the high span gas concentration was 974
ppm for the first few tests and 997 for the remaining tests.
The remaining CEMs showed slightly greater deviations. For CO2 the average deviation of the high
span reading was -0.35%, with two runs having a deviation of 3.31%, but all measurements were
within the ±15% DQI goal. The maximum deviation represented a CEM reading of 15.9% compared
Table 6-2. Maximum RSD values, in percent, for CEM and temperature measurements for each condition. The
values shown are the maximum values of the individual run RSDs measured during the four replicate test runs for
each condition.
Condition
#2 Oil, High Load
#2 Oil, Medium Load
#2 Oil, Low Load
Emuls#2Oil, High Load
Emuls #2 Oil, Medium Load
Emuls#2Oil, Low Load
Emuls Naphtha, High Load
Emuls Naphtha, Medium Load
Emuls Naphtha, Low Load
CO
85.5
164.
NAa
348.
192.
29.4
125.
553.
70.6
CO2
1.41
1.01
0.73
1.19
1.47
1.11
1.18
0.98
1.03
NO
2.10
2.54
2.71
1.91
2.51
0.73
2.27
2.71
3.44
02
6.74
9.55
3.35
8.68
14.9
6.73
8.13
10.5
6.97
THC
33.8
22.6
240.
106.
6.95
NA
43.9
82.5
153.
Stack
Temp
0.64
0.86
0.45
0.43
0.48
0.59
0.63
0.69
0.66
a. Not available
34
-------
to the high span calibration gas concentration of 15.0%. The zero points for CO2 were very close to
zero for all runs, with an average zero reading of 0.002% and a maximum of 0.09%. The THC
deviations from the high span calibration gas concentration averaged -0.54% from the high span
value of 91 ppm, with a maximum of 5.49% deviation. The zero readings for the THC analyzer
averaged 0.53 ppm, with a maximum of 5 ppm. Again, all THC bias values were within the ±10%
range specified in the DQI goals.
In all cases, the reported gas concentrations have been corrected for CEM bias using Eq. 4-1 to
account for the effect of CEM drift during the test runs.
CEM Completeness
The completeness DQI goal of greater than 90% was met for all CEM measurements (except THC)
for all tests. The THC analyzer was not operating during the first two tests of the #2 oil. For all other
tests, however, the CEMs were fully operational. During one test, the data acquisition system (DAS)
stopped logging data due to a "Disk Full" error. The CEM results reported for this test cover only
the period of time during which the DAS was logging data (see Appendix D).
Table 6-3. Cross-run RSD values, in percent, of the average CEM and temperature measurements for all
conditions.
Condition
1
#2 Oil, High Load
#2 Oil, Med Load
#2 Oil, Low Load
Emuls#2Oil, High Load
Emuls#2Oil, Med Load
Emuls#2Oil, Low Load
Emuls Naphtha, High Load
Emuls Naphtha, Med Load
Emuls Naphtha, Low Load
CO
41.9
29.2
67.1
66.2
126.
16.3
36.1
43.2
65.6
CO2
0.35
1.11
0.55
0.78
0.55
2.02
1.17
0.26
0.76
NOX
0.37
6.25
3.69
5.38
4.27
2.41
4.68
4.07
5.30
°2
0.48
2.74
2.44
2.92
3.31
4.34
14.9
2.13
3.48
THC
53.8
223.
240.
134.
228.
42.0
418.
NAa
242.
Stack
Temp
0.37
0.39
0.43
2.26
0.50
1.42
3.35
0.94
0.33
a. Not available
Table 6-4. Average and maximum deviations of zero and high span CEM readings from calibration gas values for
all runs. Zero span values are given in ppm for CO, NO, and THC, and in percent for CO2 and O2. High
span values are given in percent difference from calibration gas values.
I
1 Zero
I
1
1 High Span
1
Average
Maximum
Average
Maximum
CO
-1.55 ppm
-8 ppm
4.01%
-27.5%
CO2
0.00%
0.09%
-0.35%
6.00%
NO
2.91 ppm
17 ppm
0.23%
1.61%
°2
0.00%
-0.20%
-0.01%
4.61%
THC
0.53 ppm
5 ppm
-0.54%
5.49%
35
-------
Temperature Data
No bias checks were made of the thermocouples during the test program. However, accuracy was
measured prior to beginning the test program. Measurements from the 14 thermocouples used in the
test program were compared to measurements from a thermometer standard using an ice bath, boiling
water at ambient pressure, and ambient temperature as measurement points. Of the 42 readings (14
thermocouples times 3 readings per thermocouple), 6 readings indicated a difference between the
thermocouple and the thermometer standard, each difference being 1 °F. The maximum percent
difference was 1.3%. No thermocouple failures occurred during the test program, so the
completeness goal of greater than 90% was met.
Flow Data
As discussed in Chapter 2, the fuel flow rate was determined using a fuel totalizer and a stop watch.
The totalizer was calibrated to determine the accuracy of the instrument for the #2 oil, and the
emulsified #2 oil. The calibration was done using a container of known volume, and comparing the
known volume to the totalizer reading. Calibrations with #2 oil and emulsified #2 oil were done after
all the tests had been completed. For the #2 oil, the average deviation was slightly higher in
magnitude at -3.3%, with a maximum deviation of-5.4%. The emulsified #2 oil's average deviation
was -4.7%, with a maximum deviation of-5.4%. These values are well within the DQI accuracy goal
of ±10%. Since the totalizer was operational during the entire test series, the completeness goal of
greater than 90% was also met. The values reported for fuel flow rates have been corrected to
account for the totalizer deviation, and the reported efficiencies and emission factors, both of which
depend upon fuel flow rates, were also corrected to account for the difference between totalizer
readings and measured volumes during the calibration procedures. The emulsified naphtha flow rate
was corrected using the -4.7% factor for the emulsified #2 oil flow rate.
Particulate Matter Measurements
PM values were dependent upon two primary measurements - the mass of the particulate captured
and the volume of the gas sampled. These two measurements were combined to determine the
concentration of PM in the flue gases. DQI goals for PM measurements are presented in Table 6-5.
For the parti culate mass, off-center error and precision of the scale used are determined by
calculating the standard deviation of the five measurements of the difference between a certified mass
standard and the measured value. Accuracy is determined by the calculation of a linear regression
based on the measurement of 10 certified mass standards. The measured values are used to calculate
Table 6-5. Data quality indicator goals for PM measurements.
Mass Measurements
Capacity
Range, g
40
200
Display Drift,
mg/min
<0.02
<0.2
Off-Center Error,
mg
Std. Dev. <0.05
Std. Dev. < 0.5
Precision, mg
Std. Dev. < 0.05
Std. Dev. < 0.5
Accuracy
Linear Regression:
y=mX+b
m: 0. 99998 0.99998
Linear Regression:
y=mX+b
m: 0.99980.9998
Complete-
ness, %
>70
>70
36
-------
Table 6-6. Measurements of DQI goals for PM mass measurements.
Capacity
Range, g
40 g
200 g
Display Drift,
mg/min
0.01 mg/min
0.01 mg/min
Off-Center Error,
mg
Std. Dev. = 0.04 mg
Std. Dev. = 0.3 mg
Precision, mg
Std. Dev. = 0.02 mg
Std. Dev. = 0.1 mg
Accuracy
Linear Regression:
y=mX+b
m = 1.00000
b = 0.00000
r= 1.00000
Linear Regression:
y=mX+b
m = 1.00000
b = 0.00000
r= 1.00000
Complete-
ness, %
100%
100%
the linear response in the form:
y = mX + b
(6-1)
where y is the measured value, m is the slope, X is the certified mass, and b is the intercept. The
regression coefficient, r, is also determined and evaluated as one of the DQI goals.
Measurements were made by APPCD's QA group to determine the DQI values for the scale, as part
of a systems audit performed during the test program. Those measurements are presented in Table
6-6, and show that the scale met all the DQI goals for PM mass measurements.
Discrepancies
A number of relatively minor discrepancies between the test plan and the actual testing occurred
and are listed in Appendix C along with the action taken to resolve the discrepancy and the impact
on data quality. More significant discrepancies between the test plan and the actual testing were
the change in planned O2 level for the medium load cases discussed in section 4.1. These
discrepancies are discussed from the perspective of their impacts on data quality in more detail
below.
As discussed in Chapter 4, the target O2 level for the medium load cases was changed from the
originally planned 3% to about 1.5%, because little or no particulate matter was observed during
the high load tests at 3% O2. Although the use of a different O2 level makes it impossible to
directly compare the medium load results with the high and low load results, it does allow
evaluation of the impact of changing only the fuel without any changes in the excess air (the
primary operating variable that impacts NO, CO, and PM levels). In addition, the difference in
load between the medium and low load cases was relatively small, 1.6xl06 Btu/hr for medium load
and 1.4xl06 Btu/hr for low load. Since the tests were conducted following the same procedures as
for the high and low loads, there was no change in data quality arising from the use of a different
O2 setpoint. In addition, since the primary objective of the tests was to evaulate the impact of the
use of the emulsified fuels in comparison to the base fuels, the change in target O2 levels did not
deviate from evaluating that original objective. Thus, there was no significant impact on data
quality or the ability to derive conclusions due to the use of a different O2 setpoint for the tests in
question.
Audits
A number of performance audits were conducted by APPCD's Quality Assurance staff. The
results of those audits are discussed in the QA report prepared for the project, and are provided in
Appendix D.
37
-------
Chapter 7
Operational Observations
This chapter discusses observations regarding the handling and general combustion behavior of the
emulsified fuels noted during operation of the boiler. While these observations do not impact the
quantitative verification results, they are included here to provide information concerning the
performance of the fuels from an operability perspective, at least to the extent that these issues were
applicable to the test unit.
Emulsified #2 Oil
The emulsified #2 oil was milky in appearance, with a pink tint from the dye used to distinguish
between on-road and off-road use. When the emulsified #2 oil was introduced into the fuel feed
system, some difficulty was encountered in maintaining steady fuel flow due to the solvent properties
of the fuel. Following the introduction of the emulsified fuel into the system, small particles were
picked up by the fuel from the fuel feed piping, turning its appearance to a milky gray. As a remedy
to this problem, the fuel supply system was flushed for approximately 30 minutes until the pinkish
color was seen in the return line, and normal operations were continued.
The flame was shorter and less luminous at all loads with the emulsified #2 oil than for the base #2
oil, but after the fuel system had been flushed, no problems were noted with the operation of the
boiler using the emulsified #2 oil. The O2 and CO2 levels did fluctuate more often and more rapidly
when using the emulsified #2 oil than they did when using the base #2 oil. As O2 increased, CO2
simultaneously decreased by a similar degree and vice versa. It was felt that this behavior was due to
changes in the fuel composition. Since the O2 and CO2 measurements are based on dry flue gas, the
fluctuations were believed to be the result of reduced carbon content (and higher water content) of
the fuel at that time. However, no substantial changes were noted in emissions of CO, NO, or THC
during these fluctuations.
Emulsified Naphtha
The emulsified naphtha was milky white in appearance, similar to the emulsified #2 oil but without
the pink tint. No difficulties were noted with the feed system such as occurred when using the
emulsified #2 oil for the first time. The combustion behavior of the emulsified naphtha was very
similar to that of the emulsified #2 oil. As with the emulsified #2 oil, the flame was less luminous at
all loads than for the #2 oil. The emulsified naphtha also resulted in a region within the flame that
exhibited a light blue color, similar to what one would expect from a natural gas flame. This change
in appearance did not seem to affect the performance or emissions.
38
-------
References
1. U.S. Environmental Protection Agency, "Environmental Technology Verification Program:
Verification Strategy," EPA-600/K-96-003 (NTIS PB97-160006), Office of Research and
Development, Washington, DC, February 1997.
2. Hall, R.E., "The Effect of Water/Residual Oil Emulsions on Air Pollutant Emissions and Efficiency
of Commercial Boilers," Journal of Engineering for Power, pp. 425-434, October 1976.
3. Hall, R.E., "The Effect of Water/Distillate Oil Emulsions on Pollutants and Efficiency of Residential
and Commercial Heating Systems," presented at the 68th Annual Meeting of the Air Pollution
Control Association, Boston, MA, June 15-20, 1975, paper 75-09.4.
4. Adiga, K.C., "On the Vaporization Behavior of Water-in-Oil Microemulsions," Comb. Flame, Vol.
80, p. 214, 1990.
5. Public Law 101-549, Clean Air Act Amendments of 1990, November 15, 1990.
6. Miller, C.A., "Hazardous Air Pollutants from the Combustion of an Emulsified Heavy Fuel Oil in a
Firetube Boiler," EPA-600/R-96-019 (NTIS PB96-168281), U.S. Environmental Protection
Agency, National Risk Management Research Laboratory, Research Triangle Park, NC, February
1996.
7. EPA Test Method 5 - Determination of Particulate Emissions from Stationary Sources, in 40 CFR
Part 60 Appendix A, Government Institutes, Inc., Rockville, MD, July 1994.
8. CARB Method 501 - Determination of Size Distribution of Parti culate Matter Emissions from
Stationary Sources. State of California Air Resources Board Stationary Source Test Methods:
Volume 1 - Methods for Determining Compliance with District Nonvehicular (Stationary Source)
Emission Standards, Sacramento, CA, adopted March 23, 1998; amended September 12, 1990.
9. American Society of Mechanical Engineers, "Power Test Code PTC 4.1 - Steam Generating
Units," ASME, New York, NY, 1991.
10. American Society of Testing and Materials, "Standard Test Methods for Instrumental
Determination of Carbon, Hydrogen, and Nitrogen in Petroleum Products and Lubricants," ASTM
D 5291-92, Philadelphia, PA, June 1992.
11. American Society of Testing and Materials, "Standard Test Method for Water in Petroleum
Products and Bituminous Materials by Distillation," ASTM D 95, Philadelphia, PA, 1990.
12. American Society of Testing and Materials , "Standard Test Method for Sulfur in Petroleum
Products by Energy Dispersive X-Ray Fluorescence Spectroscopy," ASTM D 4294-90,
Philadelphia, PA, July 1990.
13. American Society of Testing and Materials, "Standard Test Method for Heat of Combustion of
Liquid Hydrocarbon Fuels by Bomb Calorimeter," ASTM D 240-92, Philadelphia, PA, June 1992.
14. American Society of Testing and Materials, "Standard Test Method for Ash from Petroleum
Products,"ASTM D 482-95, Philadelphia, PA, June 1995.
15. Miller, C.A., Linak, W.P., King, C., and Wendt, J.O.L., "Fine Particle Emissions from Heavy Fuel Oil
Combustion in a Firetube Package Boiler," Combustion Science and Technology, in press, 1998.
16. U.S. Environmental Protection Agency, Compilation of Air Pollutant Emission Factors, Edition 5,
Volume I, AP-42 (GPO 055-000-005-001), Office of Air Quality Planning and Standards, Research
Triangle Park, NC, January 1995.
39
-------
17. Ryan, J.V., Lemieux, P.M., and Groff, P.W., "Evaluation of the Behavior of Flame lonization
Detection Total Hydrocarbon Continuous Emission Monitors at Low Concentrations," presented at
the 1997 International Conference on Incineration and Thermal Treatment Technologies, May 12-
16, 1997, Oakland, CA.
18. American Society of Testing and Materials, "Standard Method for Testing Top-Loading, Direct-
Reading Laboratory Scales and Balances," ASTM E 898-82, Philadelphia, PA, 1982 (latest version
ASTM E 898-88, published November 1988).
40
-------
APPENDIX A
English Engineering to International System Unit Conversions
°C = (°F-32) x 5/9
kg = Ib x 0.454
kg/hr = lb/hrx 0.454
kJ/kg = Btu/lb x 2.326
kg/kJ = lb/106 Btu x 4.299x10-7
kg/kl = lb/1000galx0.1198
kPa = psix6.895
kW = Btu/hrx 2.93x10-4
1/min = gpm x 0.2642
m2 = ft2x 0.0929
m3 = ft3x 0.028317
m3/min = cfmx 0.028317
41
-------
APPENDIX B
Fuel Oil Analyses
The fuel oils were analyzed to determine their chemical composition for use in calculating the thermal
efficiencies of the different fuels at the conditions tested. Fuel composition plays a significant role in
determination of thermal efficiency through equations (3-6) and (3-7) which account for the moisture in
the fuel and the moisture from the hydrogen in the fuel, respectively. The ultimate analyses were
conducted initially by one laboratory (Lab A). However, there was concern that the weight percent water
reported in the Lab A results for the emulsified fuels did not match the expected values based on the
composition of the fuels as mixed by A-55. A second set of analyses was then conducted by a second
laboratory (Lab B) on the same fuels. While these results were closer to the expected compositions,
concern still remained that the percent of water reported did not agree well with the percent water mixed
with the fuel during preparation. At the request of EPA, Lab B re-evaluated their original analyses, and
reported revised ultimate analyses for the emulsified #2 and emulsified naphtha samples (see Table B-l).
Even with the additional analyses, concerns about the accuracy of the analyses remained.
A further difference between the results reported by Lab A and Lab B was that the hydrogen and oxygen
contents reported by Lab B included the hydrogen and oxygen associated with the moisture. This
resulted in higher values for both hydrogen and oxygen than were present in the fuel fraction of the
emulsification. This is important since the thermal efficiency equations [equations (3-6) and (3-7)]
account for the losses due to moisture and those due to moisture generated from the fuel hydrogen
separately, and care must be taken not to count these losses more than once.
The method used to determine hydrogen (ASTM D 5291) is based on a measurement of the amount of
water generated during combustion of the sample. If the water in the fuel is not completely driven off
prior to the combustion step, the amount of hydrogen in the hydrocarbon portion of the emulsified fuel
may be overestimated. Second, the oxygen in the fuel is determined by difference, after accounting for
the remaining constituents of the hydrocarbon portion of the fuel.
The first area of concern was the oxygen level reported for the emulsified fuels. For example, the Lab A
and initial Lab B analyses of the emulsified #2 oil found 7.67 and 6.27% oxygen (corrected to account
for the oxygen in the water), respectively, in the samples. Although the emulsifying agent used in the
preparation of the emulsified fuels contained a small amount of oxygen, a mass balance on the known
inputs of oxygen via the fuel, water, and emulsifying agent predicted oxygen values much less than those
reported. The expected oxygen value for the emulsified #2 oil was less than 1%.
A second concern was the accuracy of the hydrogen value. The method used to determine hydrogen
(ASTM D 5291) is based on a measurement of the amount of water generated during combustion of the
sample. If the water in the fuel is not completely driven off prior to the combustion step, the amount of
hydrogen in the sample may be overestimated. Finally, the Lab B analyses reported water in percent
volume (vol%), while the Lab A analyses reported water in percent weight (wt%), making it more difficult
to compare the results.
The Lab B results were revised to account for the hydrogen and oxygen from the water, and to change
the vol% water content to wt%. These results are shown in Table B-2. The approach used in developing
the revised table was to first convert the vol% water to an equivalent wt%. The vol% value is given by:
voL
vol% = — ^ x100% (B-l)
where volw is the volume of water and volf is the volume of fuel. For 1 gallon of fuel, volw is equal to
42
-------
Table B-1. Reported ultimate analysis results for the fuels tested.
II
Fuel
Laboratory
Carbon (wt%)
Hydrogen (wt%)
Nitrogen (wt%)
Sulfur (wt%)
Ash (wt%)
Water (wt%)
Water (vol%)
Oxygen (wt%)
(by difference)
Btu/lb
Btu/gal
Specific Gravity
Density, Ib/gal
#2 Oil
A
86.76
10.68
0.22
0.07
0.01
<0.10
NA
<2.16
19,462
139,873
0.8630
7.186
B
86.92
13.01
0.49
0.0289
0.001
NA
<0.05
0.42
19,450
139,592
0.8607
7.167
Emulsified #2 Oil
A
54.86
8.46
0.18
0.03
<0.01
28.80
NA
<7.67
13,032
97,883
0.9019
7.510
B
55.16
12.59
0.21
0.009
0.003
NA
26.7
32.47C
12,368
93,205d
0.9050
7.536
Ba
57.40
12.21
0.48
NAb
NA
NA
28.0
29.92C
12,786
96,355d
0.90506
7.536
Emulsified Naphtha
A
54.25
9.26
0.20
0.01
0.01
25.60
NA
10.67
13,046
88,491
0.8145
6.782
B
48.90
13.47
0.35
0.002
0.002
NA
31.1
37.98C
11,657
80,655d
0.8309
6.919
Ba
53.36
13.17
0.32
NA
NA
NA
30.0
33.16C
12,584
87,069d
0.8309'
6.919
a. Revised analyses
b. Not available
c. Includes the oxygen in the water
d. Calculated from Btu/lb and density values
e. Same as Lab B unrevised analysis
Table B-2. Corrected analysis results for the fuels tested.
Fuel
Laboratory
Carbon (wt%)
Hydrogen (wt%)
Water (wt%)
Oxygen (wt%)
(by difference)
#2 Oil
A
86.76
10.68
<0.10
<2.16
B
86.92
13.01
< 0.058C
0.42
Emulsified #2 Oil
A
54.86
8.46
28.80
<7.67
B
55.16
9.32b
29.47C
6.27d
Ba
57.40
8.77b
30.93C
2.42d
Emulsified Naphtha
A
54.25
9.26
25.60
10.67
B
48.90
9.32b
37.39C
4.74d
Ba
53.36
9.16b
36.07C
1.10d
II
a. Revised analysis
b. Corrected to account for hydrogen in water
c. Calculated from vol%
d. Corrected to account for oxygen in water
43
-------
vol%, and the mass of water, mw, in 1 gallon of fuel is given by
mw = pwvolw (B-2)
where pw is the density of water.
Since the pw is 8.3385 Ib/gal at 15.6 °C (60 °F), and the density of the fuel, pf, is 8.3385yf Ib/gal,
where Yf is the specific weight of the fuel, then the wt% of water in 1 gallon of fuel is given by:
mw vol%Apw vol%
Wt% = = = (B-3)
mf Pr Yr
One can then calculate the wt% water for the Lab B analyses.
The second step is to account for the hydrogen and oxygen from the water in order to directly
compare the different analyses. Since 1 Ib-mole of water (H2O) weighs 18 Ib, and since 16 Ib of that
is from the oxygen and the remaining 2 Ib is from hydrogen, then 8/9 is the fraction of water's mass
attributable to oxygen and the remaining 1/9 is attributable to hydrogen. The oxygen content of the
fuel, corrected to remove the oxygen from the water, is then given by:
Ocorr = Orep-(wt%H20H-|) (B-4)
V 3 /
where Ocorr is the corrected oxygen content of the fuel in wt%, Orep is the reported oxygen content in
wt%, and wt%H2o is the weight percent of water. Likewise, the corrected hydrogen content is given
by:
Hcorr = Hrep-Lt%H20H|) (B-5)
V 3 /
where Hcorr is the corrected hydrogen content of the fuel in wt% and Hrep is the reported hydrogen
content in wt%. The corrected values for oxygen and hydrogen and the wt% water values are given
in Table B-2.
Based on the results shown in Table B-2, those most consistent with the known amounts of water
added to the fuel and with the calculated oxygen contents of the fuels are the revised Lab B analyses,
corrected to account for the oxygen and hydrogen from the water. These values are reported in
Table 2-2 and used in the thermal efficiency calculations reported in Chapter 5.
Effects of Changing Fuel Composition
Due to the large impact that the hydrogen and moisture contents of a particular fuel have on thermal
efficiency, it is important to understand the relationship between fuel composition and thermal
efficiency.
Changes in the fuel composition can affect the calculated value of losses due to energy in the dry flue
gas (LFG), hydrogen in the fuel (LMH), and moisture in the fuel (LMF). Eqs. 3-3 and 3-4 show that the
primary fuel related variable influencing LFG is the fraction of carbon in the fuel, 0,. Therefore,
significant changes in Q, can result in significant changes in thermal efficiency. Likewise, LMH and
LMF directly affect the losses due to hydrogen in the fuel, LMH, and moisture in the fuel, LMF. It is
important, then, to understand the changes in thermal efficiency associated with the changes in the
fuel analysis. In general, as carbon increases, LFG will increase due to the higher dry flue gas flow,
LMH will increase as the fuel hydrogen content increases, and LMF will increase as the fuel moisture
content increases.
The thermal efficiency was calculated separately using the three fuel analyses discussed above (shown
in Table B-2) and the measured values for fuel input, flue gas composition and temperature, and oil
44
-------
and air temperatures. The results of these calculations are shown in Table B-3. Calculations using
Lab A analyses resulted in thermal efficiencies that were higher than those calculated using Lab B
analyses, by 0.80-1.32 percentage points. The difference between Lab B and revised Lab B results
was much smaller, with calculations using the revised Lab B analyses resulting in thermal efficiencies
that were lower by 0.07-0.11 percentage point than those using the unrevised Lab B analyses, for the
emulsified #2 oil. For the emulsified naphtha, calculations using the revised Lab B analyses resulted
in thermal efficiencies that were slightly higher (0.17-0.30 percentage point) than those calculated
using the unrevised Lab B analyses.
These results emphasize the importance of the fuel analyses when evaluating thermal efficiency,
particularly when using emulsified fuels that are high in water content. It is suggested that
comparisons of oxygen, carbon, or energy contents of the emulsified and non-emulsified fuels be
conducted to determine whether the reported analyses are consistent with the addition of known
amounts of water. Such comparisons can provide an indication of the reliability of the analyses.
Table B-3. Differences in calculated thermal efficiency values using the different fuel analyses. All other
parameters were held constant. Differences are in comparison to the revised Laboratory B analyses.
1
Fuel
Laboratory
High Load
Medium Load
Low Load
#2 Oil
A
86.43
88.88
88.59
B
85.11
87.61
87.33
Emulsified #2 Oil
A
84.34
85.68
85.47
B
83.23
84.60
84.39
Ba
83.30
84.71
84.49
II
Emulsified Naphtha
A
82.89
84.56
84.74
B
82.39
83.93
84.11
Ba
82.09
83.75
83.94
Difference from reported value (Revised Laboratory B data)
High Load
Medium Load
Low Load
1.32
1.27
1.26
_b
-
-
1.04
0.97
0.98
-0.07
-0.11
-0.10
-
-
-
0.80
0.81
0.80
0.30
0.18
0.17
-
-
-
a. Revised analysis
b. Not applicable
45
-------
APPENDIX C
Discrepancies
The discrepancies below are listed in chronological order. The discrepancy is described, followed by
the action taken to resolve the discrepancy, and the impact of the discrepancy on data quality.
1. Date: 8/5/97
Discrepancy: The THC analyzer was not operating during Condition 1, Test 1.
Action: The analyzer was repaired and was in service for the remaining tests.
Impact: Given the extremely low THC values measured during the remaining tests at Condition 1,
no substantial impact resulted from the loss of the instrument. The average THC value reported is
the average of three rather than four test runs.
2. Date: 8/5/97
Discrepancy: No SO2 analyzer was available for use in the #2 oil, emulsified #2 oil, or emulsified
naphtha tests.
Action: No action taken.
Impact: The levels of sulfur in the #2 oil, emulsified #2 oil, and emulsified naphtha are extremely
low. The use of an oil/water emulsion does not impact the emissions of SC>2 (in terms of mass per
unit energy input); thus the total emissions per unit time at a given load will not change between the
base oil and the emulsified oil. The lack of an SC>2 analyzer had no impact on the test data quality.
3. Date: 8/7/97
Discrepancy: The fuel oil drum being used as the primary feed for the boiler ran dry during Test
2.1, causing boiler shutdown.
Action: The test was discontinued.
Impact: The data from the test were not used. The problem did not impact data quality as the
results from this test run were not used.
4. Date: 9/3/97
Discrepancy: The data acquisition system stopped operating due to a "Disk Full" error during Test
6.2, resulting in a temporary loss of data while the data acquistion software was being restarted.
The data from this test were lost between the time the program halted and it could be restarted, a
period of approximately 6 minutes.
Action: Only the data from the period from the beginning of the test to the program "crash" were
used.
Impact: Because the boiler operation and the extractive sampling were not impacted, no significant
loss of data was suffered. Although the data from the time of the program restart to the end of the
test were gathered (approximately 50 minutes), it was decided that, to avoid any uncertainty
regarding the impact of the program restart, only the initial data would be used. Because the boiler
operations were very stable with relatively low fluctuations in conditions, no adverse impact is
believed to have resulted, from the use of data from only the initial 70 minutes of testing.
5. Date: 9/8/97
Discrepancy: The filter in the Method 5 sampling train tore during particulate sampling.
Action: The test was repeated on 9/17/97'.
Impact: Because the test was repeated, there was no impact on data quality.
46
-------
Appendix D
Audit Results
TOP-LOADING PAN BALANCE EVALUATION,
The Mettler AE240 top-loading pan balance in room H-202 at the ERC Building of the EPA was
audited on August 14, 1997 according to an audit procedure developed by Research Triangle
Institute from ASTM Standard E-898-82.18 Standard class S weights were used to evaluate such
characteristics of the balance as the display-drift, off-center error, precision, and accuracy. The
working weights used to evaluate the Mettler AE240 were certified on March 8, 1997, by the North
Carolina Department of Agriculture, Standards Division.
The performance of the Mettler AE240 is considered satisfactory, and the balance will pass the audit
if all of the following are true:
1. None of the following malfunctions occur during the audit.
a. Balance cannot be zeroed.
b. Air movement and/or vibration interfered with scale clarity or resolution.
2. Display-drift did not exceed 1.5 times the resolution (R) of the balance per minute.
3. The standard deviation of the differences obtained for the off-center error test did not exceed
5 times the resolution (R) of the balance.
4. The standard deviation of the differences obtained for the precision test did not exceed 5
times the resolution (R) of the balance.
5. The results of the least square linear regression equation (y = mX + b) for the accuracy test
were within audit limits where:
y = the display value
m = the slope of the line [audit limit: 1.0 + 1.5(R)]
X = the total certified mass on the pan load
b = the intercept [audit limit: 0.0 + 1.5(R)]
and the correlation coefficient (r) of the regression is greater than 0.9998.
Instrument performance for a particular characteristic is considered satisfactory if the measured value
for that characteristic is equal to or less than the audit limit value. If the value for the characteristic
exceeds the audit limit, then the performance of the Mettler AE240 is considered unsatisfactory for
that characteristic and fails the audit. If a balance equipped with two weighing ranges passes the audit
on one range and fails the audit on the other range, it will be evaluated for use limited to the passing
range.
The overall results of the Environmental Evaluation, the Pre-Audit Test, the Display-Drift Test, the
Off-Center Error Test, the Precision Test, and the Accuracy Test were all satisfactory. The air
movement and vibration along the counter top were well within the requirements. The scale image
was clear, and the zero adjustment and leveling leg screws worked properly. Performance evaluation
results are summarized in Table D-l.
SYSTEMS AUDIT OF PARTICULATE MATTER COLLECTION PROCESS
On December 3, 1997, a systems audit of the procedures used to determine the particulate catch using
the EPA Method 5 train and the California Air Resources Board Method 501 cascade impactor was
47
-------
conducted in room H-202 at the ERC Building of the EPA/RTP. Charly King of ARCADIS Geraghty
& Miller first explained the use of the Method 5 train and demonstrated the use of the drying
chamber and how glass beakers containing acetone probe rinses were processed including how the
residue was weighed and the results were recorded. He also explained and demonstrated the
conditioning of Method 5 quartz fiber filters, the assembly and disassembly of the CARB Method
501 cascade impactor (Andersen Mark III). A detailed explanation and demonstration of the
conditioning and weighing of the filters and impactor surfaces was also given.
It was concluded that proper procedures were being followed in the collection and gravimetric
determination of particulate matter using EPA Method 5 and CARB Method 501. Control of
humidity in the balance room is less than desirable when the ambient air is very moist. King stated
that no weighings are made in the summer when the weather is rainy.
Table D-1. Performance evaluation results for the Mettler AE240
Capacity Range = 40 grams, Readability = 0.00001 grams
Test Type
Display Drift
Off-Center Error
Precision
Accuracy
Audit Limits
< 0.00002 g/min
Standard deviation < 0.00005 g
Standard deviation < 0.00005 g
Linear regression:
m: 0.99998
------- |