-------
-------
capital cost. Fuel cells are an emerging technology that may ultimately
improve the outlook for clean, efficient, and economical energy use of landfill
gas.
Early 1995 marked the completion of a major step toward establishing
fuel cells as a promising new technology for utilizing landfill gas. The first
demonstration using landfill gas in a commercially available phosphoric acid
fuel cell power plant was successfully conducted at the Penrose Landfill in Sun
Valley, California. This was the conclusion of a three-phase program which
began in 1990 when EPA awarded International Fuel Cells Corporation- (IFC) a
contract to demonstrate landfill gas control with energy recovery. The project
addressed two principal issues: (1) a landfill gas cleanup method to remove
contaminants from the gas sufficiently for fuel cell operation, and (2) a
demonstration test of a commercial fuel cell power plant operating on landfill
gas.
This paper will summarize some of the latest test results and the status
of the project. Additional information can be found in EPA reports [2], [3].
LANDFILL GAS FUEL CELL
Overall System
Figure 1 provides a simplified description of the required components of
a fuel cell landfill gas-to-energy system. The system consists of the landfill
gas wells and collection system, a modular gas pretreatment unit (GPU), and a
fuel cell power plant modified for landfill gas operation. The fuel cell power
plant is an adaptation from the mitural-gas-fueled PC25 fuel cell sold by ONSI
Corporation, an IFC subsidiary. Landfill gas collected at the site is
processed to remove contaminants in the GPU. The cleaned medium heat content
landfill gas is then used to fuel the fuel cell power plant to produce
alterating current (AC) for sale to the electric utility. The clean
cogeneration heat produced by the fuel cell may also be utilized if a
requirement for heat exists at the site. If not utilized, it is rejected by
the air cooling module.
-------
UUUty
power
Undiii
FIGURE 1. FUEL CELL LANDFILL GAS ENERGY RECOVERY
Fuel Cell Power Plant
A simplified- functional schematic of the 200 kW fuel cell power unit is
shown in Figure 2. Major sections of the system, include the fuel processing
system, fuel cell stack, and the thermal management system. In the fuel
processing section treated landfill gas is converted to hydrogen (H2) and
carbon dioxide (CO2) for introduction into fuel cell stack. The fuel treatment
process includes a low temperature fuel preprocessor to remove the residual
contaminants from the treated gas, a fuel reformer, and a low temperature shift
converter where the exhaust from the reformer is further processed to provide
additional H2 and C02.
In fhe fuel cell stack, H2 from the process fuel stream is combined
electrochemically with oxygen from the air to produce direct current (DC)
electricity and byproduct water. The product water is recovered and used in
the reformer. The heat generated in the cell stack is removed to an external
heat rejection system. This energy can be either rejected .to the ambient air.
-------
or recovered for use by the customer. The DC power produced in a fuel cell
stack is converted to AC power in a power conditioning package not shown on the
process schematic.
MANAGEMENT
SYSTEM
FIGURE 2. FUNCTIONAL SCHEMATIC FUEL CELL LANDFILL GAS POWER UNIT
-------
A preliminary design of a fuel cell power plant was established to
identify the design requirements which allow optimum operation on landfill gas.
Three issues specific to landfill gas operation were identified which reflect a
departure from a- design optimized for operating on natural gas. A primary •
issue is to protect the fuel cell from sulfur and halide compounds in the
landfill gas. A second issue is to provide mechanical components in the
reactant gas supply systems to accommodate the large flow rates that result
from use of dilute methane fuel. The third issue is an increase in the heat
rate of the power plant by approximately 10% above that anticipated from
operation on natural gas. This is a result of the inefficiency of using the
dilute methane fuel. The efficiency results in an increase in heat recoverable
from the power plant. Because the effective fuel cost is relatively low, this
decrease in power plant efficiency will not have a significant impact on the
overall power plant economics.
The landfill gas power plant design provides a packaged, truck
transportable, self-contained fuel cell power plant with a continuous
electrical rating of 200 kW. It is designed for automatic, unattended
operation, and can be remotely monitored. It can power electrical loads either
in parallel with the utility grid or isolated from the grid.
In summary, a landfill gas fuel power plant can be designed to provide
200 kW of electric output without need for technology developments.. The design
would require selected components to increase reactant flow rates with a
minimum pressure drop. To implement the design would require non-recurring
expenses for system and component design, verification testing of the new
components, and system testing to verify the power plant performance and
overall system integration. A thermodynamic analysis of the fuel cell power
plant optimized for operating on landfill gas was completed. The resulting
performance of the landfill, gas power plant is-compared to a power plant
operating on natural gas in Table 1.
-------
TABLE 1. PERFORMANCE COMPARISON FOR NOMINAL 200 KW OUTPUT
Fuel
Electrical Efficiency (LHV) - % . .
Heat Rate (HHV) -kcal/kWhr
Available Heat - kcal/hr
Ambient Temperature for Fuel Water Recovery - °C
Startup Fuel
Natural Gas
Power Plant
Natural Gas
40
2,390
192,000
35
Natural Gas
Landfill Gas
Power Plant
Landfill Gas
36.4
2,620
208,000
35
Natural Gas
Landfill Gas Cleanup
Landfill gases consist primarily of CO2', methane (CH4) , and nitrogen (N2) ,
plus trace amounts of hydrogen sulfide (H2S), organic sulfur, organic halides,
and non-methane hydrocarbons. N2 is not a true constituent of landfill gas,
but is drawn into the mix when vacuuming gas into the collection system. The
concentration varies widely, with highest concentrations occurring from the
landfill's perimeter wells. The specific contaminants in the landfill gas of
concern to the fuel cell are sulfur and halides. Both of these components can
"poison" and, therefore, reduce the life of the power plant's fuel processor.
The fuel processor is the unit which converts CH4 in the landfill gas into H2
and CO2 in an endothermic reaction over a catalyst bed. The catalyst in this
bed can react with the halides and sulfides and lose its activity. This
reaction, when it occurs, is irreversible.
The GPU, designed to remove fuel cell contaminants, is shown in Figure 3.
H2S is first removed by adsorption on a packed .bed. The material which
performs this function is a specially treated carbon, activated to catalyze the
conversion of H2S into elemental sulfur which is deposited on the bed. The
reactions for the conversion of sulfur, the Claus reactions, are:
H2S + 3/2 Oa --> H2O + SO2
, SO2 + 2H2S --> 2H2O + 3S
The bed is not regenerable on site, but must be removed to another site if
regeneration is desired.
-------
UMM
1
L
F1U.QA8
HI*
MCMOUU.
^
eaoun
COHMNSI8
RAM
^i
I^H»
BHT1H
•CD
•"JBW?0"
CYCLE)
ORVCR
nHt«Sa
-------
regenerable dryer and hydrocarbon removal beds operate on a nominal 16-hour
cycle, with each set of beds operating in the adsorption mode for 8 hours and
in the regeneration mode for 8 hours.
The gas then passes through a particulate filter and is warmed indirectly
by an ambient air, finned-tube heat exchanger to attain a temperature above O°C
before being fed to the fuel cell unit.
GPU TEST RESULTS
The GPU depicted in Figure 3 has been successfully tested. Detailed
results of the tests, along with the test plan, are contained in an EPA report
[3] . The results of those tests will be summarized here.
After completing 216 hours of continuous operation and a total of 616
hours since the first GPU startup, performance testing was conducted over a 3
day period at the beginning, middle, and end of the regenerative cycle to
evaluate performance over the normal 8-hour cycles on each of the two
regenerative beds (dryer bed and carbon bed). At specific times in the
regeneration cycles, Tedlar bag samples were collected from sampling manifolds
located at the GPU inlet and outlet as well as at the flare inlet and outlet.
These bag samples were analyzed off-site using gas chromatography/mass
spectrometry (GC/MS) analysis for the volatile organic compounds (VOCs) and
gas chromatography/flame photometric detection (GC/FPD) analysis for sulfur
compounds. Additionally, sulfur compounds were measured at the GPU's inlet and
outlet using on-line GC/FPD. No on-line or on-site measurements were utilized
for VOCs because it was found that the landfill gas matrix could bias the
results. This was determined via^-an audit using certified cylinder gases.:.
The results from one of the cycles are summarized in Table 2. All the
measurements demonstrated that the GPU was very effective in removing the
target sulfur compounds and VOCs. For sulfur compounds, the GPU outlet
concentrations were either below detection limits (0.01 ppm for the on-line
method and 0.004 ppm for the off-site analyses) or in the parts per billion
concentration range. Likewise halogenated and other VOCs were below 0.002 ppm
with the exception of methylene chloride, which was detected in trace levels of
less than 0.02 ppm.
The fuel test data for the GPU flare have been summarized elsewhere [3] .
It suffices to state that the flare destruction of VOCs and sulfur compounds
exceeded 99%. Nitrogen oxides and carbon monoxide concentrations at the flare
-------
outlet average 10.4 and 3.0 ppm, respectively, over three test periods.
Particulate matter averaged 0.03 mg/m3.
TABLE 2. GPU INLET/OUTLET EMISSION TEST DATA
Sampling Location
Reduced Sulfur Compounds (ppm)
Sample Type
hydrogen sulfide
carbonyl sulfide
methyl mercaptan
ethyl mercaptan
dimethyl sulfide
carbon disulfide
dimethyl disulfide
Total Reduced Sulfur - see note 1
Volatile Organic Halogens -
GC/US Analysis (ppm)
Sample Type
Compound
dichlorodifluoromethane
vinyl chloride
methylene chloride
cis- 1.2-dichloroethene
1, 1-dichloroethane
trichloroetnene
tetrachloroethene
chlorobenzene
Total Halogens (as halide) - see note 2
Volatile Organic Compounds -
GC/MS Analysis (ppm)
benzene
toluene
xylenes
ethyl benzene
styrene
acetone
2-butanone
ethyl acetate
ethyl butyrate
alpha-pinene
d-limonene
tetrahydrofuran
1800-1900
Hour 1
GPU
Inlet
bag
92.7
0.197
2.91
0.48
6.51
<0.07
<0.07
104
bag
0.83
1.1
6.6
4.3
1.9
1.3
2.7
0.91
46.7
1.1
28
14.9
6.1
0.6
<1.2
5.2
8.1
6.3
10.8
12.6
1.3
GPU
Outlet
bag
<0.004
0.017
<0.004
<0.004
<0.004
<0.002
<0.002
0.017
bag
<0.002
<0.002
0.016
<0.002
<0.002
<0.002
<0.002
<0.002
0.032
<0.002
0.005
<0.002
<0.002
<0.002
0.01
<0.004
<0.002
<0.002
<0.002
-------
FUEL CELL TESTING
Fuel Cell Performance
The power plant used in the demonstration is a commercial ONSI PC25 200-
kW phosphoric acid fuel cell. The power plant, designed to operate on pipeline
natural gas, was shipped to and installed at the Penrose Landfill during 1994.
The field test began on December 7, 1994, and ended on February 19, 1995, with
eight operational test runs. During this period, the fuel cell operated for
707 hours on landfill gas. Power produced by the unit was fed into- the
electrical grid for sale to the local electrical utility, the Los
Angeles Department of Water and Power (LADWP). It was the first fuel cell ever
connected to the LADWP grid. The revenue produced by the sale of electricity
was used to help offset program costs.
Because landfill gas has approximately half the heat content of natural
gas, this fuel cell power plant cannot produce 200 kW of net power when
operated on landfill gas. To increase the net power, higher flows of landfill
gas would be required to obtain an equivalent natural gas fuel content and -
heating value. However, the required modifications to achieve fuel rated power
were not accomplished for this project. The only modifications to the fuel
cell were those that could be field installed under the direction of an IFC
engineer, from modification kits designed and fabricated by IFC. The
modifications included a larger fuel control valve and fuel flow venturi, a new
process fuel recycle orifice, a new cathode exit orifice, and a new redundant
start fuel shut-off valve. Additionally, there were modifications to the
control software.
The fuel cell was operated at up to 137 kW, which is 3 kW below the goal
for operation on the Penrose Landfill gas. An-endurance operating condition of
120 kW was selected for the bulk of the field test operations to provide a
margin for steady fuel cell operation during periods of sub-standard gas
quality, which occur periodically due to upsets in gas quality from the active
landfill (Bradley) which supplies gas to the Penrose sife. The fuel cell
efficiency was calculated over two periods during the field test. The first
period covered 6 days from January 24 through 30, 1995. Efficiency during this
6-day period of continuous operation was 37.1%. The second period covered 8
days from February 9 through 17, 1995. "Average efficiency for this 8-day
period, which included a brief shutdown, was 36.5%.
11
-------
Table 3 presents results of emission testing of the PC25 power plant at
the Penrose Landfill conducted during February 1995. The emission levels
indicate that fuel cells can operate on landfill gas while maintaining the low
emission levels characteristic of natural-gas-based fuel cells.
TABLE 3. EMISSION TEST* RESULTS FROM FUEL CELL
OPERATING ON LANDFILL GAS
Pollutant
Fuel Cell Emission
Sulfur Dioxide (S02)
Nitrogen Oxides (NOX)
Carbon Monoxide (CO)
Non-detect**
0.12 ppm, avg.
0.77 ppm, avg.
*Tests used EPA Methods 6c, 7e, and 10.
**Detection limit for SO2 was 0.23 ppm; all data are dry measurements corrected
to 15% O,. .
GPU Operation
The GPU consistently removed total halides (as HC1) from inlet levels of
45 to 60 ppm in the raw landfill gas to very low or undetectable levels at the
outlet. During seven tests conducted between January 19 and February 17, 1995,
there were no detectable halides (individual species detection limits range
from 0.001 to 0.020 ppm). The results are consistent with the data contained
in Table 2.
Total sulfur (as H2S) was reduced from about 110 ppm (about 10 ppm £rom
organic sulfur, plus 100 ppm H2S) to between non-detectable and 0.385 ppm. The
only sulfur species detected was carbonyl sulfide. The elevated levels of
0.173 to 0.385 ppm of cabonyl sulfide measured on February 9 and 10, 1995, are
believed due to a slight increase in H2S exiting the non-regenerable H2S
removal bed, since H2S at the H2S bed exit was measured' at 1.0 to 2.7 ppm on
February 14. Earlier laboratory work at IFC showed that H2S is converted to
carbonyl sulfide over the activated alumina in the downstream drier bed by the
reaction H2S -t- C02 = COS + H2O, due to the removal of the product water by the
alumina. The resulting carbonyl sulfide is not readily removed by the low
temperature carbon bed. The non-regenerable H2S removal bed was switched over
to a fresh bed on February 15, 1995, and the exit H2S level returned to non-
detectable. The carbonyl sulfide level measured shortly after, on February 17,
also fell to just 0.061 ppm.
12
-------
DISCUSSION AND CONCLUSIONS
Emissions from landfills are potential contributors to global warming and
hazardous air pollutants. Conventional methods to mitigate these emissions,
such as flaring, produce other greenhouse gases, such as CO2. Operating a fuel
cell at a landfill site can eliminate CH4, hazardous air pollutants,and other
universal secondary emissions (CO, SO2, NOX) , lowers total CO2 emissions, and
can permit efficient generation of electric power. In order to operate a fuel
cell on landfill gas, the gas roust be sufficiently purified. A landfill gas
cleanup pretreatment module designed, constructed, and tested in this
demonstration was successful. The combined gas cleanup system and power plant
produced electrical power with low levels of air pollution.
Currently, the fuel cell/GPU is located on a landfill in Groton,
Connecticut, as a follow-up to the earlier demonstration of the technology at
the Penrose Landfill. This demonstration is scheduled to run for approximately
18 months during which time a significant amount of operating data will be
acquired. Additionally, research will be conducted to refine the design of the
GPU in order to simplify the required equipment. This demonstration is a
partnership between EPA, the Town of Groton, and Connecticut Light and Power,
with technical support from IFC.
It is expected that at least 1.6 million kWh of electricity will be
generated over the test period. The fuel cell has produced a record 165 kW
from the Groton Landfill gas, but the system is being operated at a
conservative 140 kW to minimize inadvertent shutdown due to gas heat
fluctuations.
13
-------
REFERENCES
[1] Air Emissions from Municipal Solid Waste Landfills - Background
Information for Proposed Standards and Guidelines (EPA-450/3-90-011a;-
NTIS PB91-197061), March 1991.
[2] Sandelli, G. J., Demonstration of Fuel Cells to Recover Energy from
Landfill Gas, Phase I Final Report: Conceptual Study (EPA-600/R-92-007;
NTIS PB92-137520), January 1992.
[3] Trocciola, J. C. and Preston, J. L., Demonstration of Fuel Cells to
Recover Energy from Landfill Gas, Phase II, Pretreatment System
Performance Measurement (EPA-600/R-95-155; NTIS PB96-103601), October
1995.
[4] Graham. J. R. and Ramaratnam, M., Recovery of VOCs Using Activated
Carbon, Chemical Engineering, pp. 6-12, February 1993.
14
-------
-------
------- |