Performance of an Internal Combustion Engine ... - ACS Publications

ARTICLE pubs.acs.org/IECR

Performance of an Internal Combustion Engine Operating on Landfill Gas and the Effect of Syngas Addition McKenzie P. Kohn, Jechan Lee, Matthew L. Basinger, and Marco J. Castaldi* Department of Earth & Environmental Engineering, Henry Krumb School of Mines, Columbia University, Room 926, Mudd Building, 500 West 120th Street, New York, New York 10027, United States ABSTRACT: The performance of a four-stroke Honda GC160E spark ignition (SI) internal combustion (IC) engine operating on landfill gas (LFG) was investigated, as well as the impact of H2 and CO (syngas) addition on emissions and engine efficiency. Tests were performed for engine loads from 0.2 to 0.8 kW over a range of CO2 to CH4 ratios (0-0.50). In addition, variation across both the syngas content (up to 15%) and the ratio of H2 to CO in the syngas (H2/CO = 0.5, 1, and 2) were tested. Catalytic testing provided reactor data on the amount of syngas and H2/CO ratios that can be obtained by autothermally reforming LFG. The emissions obtained from the test engine fueled with the simulated LFG were found to be comparable to emissions from commercial LFG to energy (LFGTE) systems currently deployed. Syngas addition was found to not only significantly reduce CO, unburned hydrocarbon (UHC), and NOx emissions but also improve brake efficiency of the engine. CO emissions were reduced from 802 to 214 ppm for a 5% syngas addition and to 230 and 247 ppm for 10 and 15% syngas addition, respectively. UHC emissions were reduced from 113 ppm to approximately 12 ppm for all amounts of syngas addition. Syngas addition decreased NOx from 100 to 62 ppm for 5% syngas and 71 and 76 ppm for 10 and 15% syngas, respectively. Finally, the brake efficiency increased by approximately 10% with the addition of 5% syngas.

1. INTRODUCTION Landfills are the second largest source of anthropogenic methane emissions in the United States. Landfill gas (LFG) is generated from anaerobic bacterial decomposition of organic waste materials in municipal solid waste (MSW) landfills, typically consisting of about 50-55% CH4, 40-45% CO2, and other trace gases. Because methane is a potent greenhouse gas, current regulations require methane emissions from MSW landfills to be captured. LFG to energy (LFGTE) projects capture 60-90% of the methane emitted from a MSW landfill and use it primarily for electricity generation in internal combustion engines or turbines, or in direct-use projects.1 An LFGTE project utilizing methane from a landfill reduces atmospheric CH4 emissions and also displaces fossil fuels that would otherwise have been used. There are currently 519 operational LFGTE projects in the United States supplying 13 billion kWh of electricity and 100 billion cubic feet of LFG to direct-use projects annually.2 The displaced CO2 emissions also provide the opportunity to sell greenhouse gas credits, improving the economic feasibility of the LFGTE project. The difficulty in using LFG for energy is that it is plagued by low and fluctuating energy content resulting in lack of flame stability, deteriorating fuel efficiency, and increased CO, unburned hydrocarbon (UHC), and NOx emissions. To mitigate these emissions, there are many conventional postcombustion cleanup methods. However, many of these methods result in reduced power output due to pressure drop increases and could potentially add significant expense. Typically, emission waivers are required before LFGTE projects can be permitted.3 Due to the low energy content of LFG, most engines need to be modified considerably to accept it as a fuel source.4,5 There have been many studies investigating the use of landfill gas or r 2011 American Chemical Society

CH4-CO2 mixtures in spark ignition (SI) engines.6-10 These studies showed that the presence of CO2 in inlet fuel mixtures not only deteriorated engine efficiency but also increased pollutant emissions, compared to methane or natural gas fueling. Furthermore, often fuel conditioning is necessary due to variability in LFG composition. If the methane content drops, the addition of a secondary fuel is necessary to ensure continuous and stable combustion. The resulting economic considerations arising from the cost of upgrading fuel, the additional expense of using specialized power generators, postcombustion emission cleanup, and permitting costs have prevented widespread use of LFG. In view of these factors LFG is commonly flared, converting the CH4 to CO2 without efficiently utilizing the energy in the LFG. One method of increasing the reactivity of LFG and therefore reducing engine emissions is to add hydrogen to the fuel stream. Akansu et al. have done an extensive review of prior work on the injection of H2 into natural gas fuel streams to reduce emissions and improve internal combustion (IC) engine efficiency.11 In general, UHC and CO emissions decrease with increasing H2 due to the high reactivity and flame speed of H2. The effect of H2 injection on NOx emissions varies by study and combustion system. This discrepancy can be attributed to an increase in flame temperature with H2 injection at a given equivalence ratio, resulting in more thermal NOx. However, H2 also enables operation at lean conditions, resulting in lower in-cylinder temperatures and therefore a NOx reduction. Furthermore, Received: September 20, 2010 Accepted: January 6, 2011 Revised: December 29, 2010 Published: February 07, 2011 3570

dx.doi.org/10.1021/ie101937s | Ind. Eng. Chem. Res. 2011, 50, 3570–3579

Industrial & Engineering Chemistry Research operation in the lean limit of combustion also results in higher thermal efficiencies. Computational studies have shown that the benefits of H2 injection into natural gas streams can be extended to landfill gas or biogas streams with higher concentrations of CO2. Rakapolous et al.12,13 report an increase in second law efficiency as H2 content increases with CO2 acting as a diluent but not changing efficiency trends. While the advantages of H2 spiking of the fuel are numerous, the major hurdle is in demonstrating a feasible and practical means of providing H2 without compromising system efficiency and/or adequately addressing challenges which control aspects of injecting, mixing, and igniting the reactive mixture. Furthermore, onboard H2 storage and transportation pose uniquely difficult problems. In this study, injection of syngas, a mixture of H2 and CO, is investigated as a method of reducing engine emissions and increasing engine efficiency. While there are multiple options for introducing syngas into the system, this work focuses on in situ generation from the incoming LFG stream. This has the benefit of maintaining a simple system that can be retrofitted to existing LFGTE operations. The in situ methods explored are limited to catalytic reforming a portion of the LFG only using dry reforming or a mixture of LFG and air, autothermal reforming. Autothermal reforming (ATR) is the process of introducing small amounts of an oxidant, usually air or O2, into a fuel stream to convert a portion of the fuel in an exothermic reaction (either combustion or partial oxidation), thereby producing heat in situ necessary for the concurrent endothermic reactions for a net ΔH = 0 kJ/mol. This lowers required heat input for the reforming processes, or, in autothermal operation, eliminates the need for constant heat input.14 Dry reforming, on the other hand, requires an external heat supply. Autothermal reforming and dry reforming produce syngas with a range of H2/CO ratios depending on the gas feed composition and temperature of the reactor. The major advantage of this configuration is that it generates the syngas inline or in situ without significantly reducing the power output. This allows for a more reactive mixture to enter the engine, thus allowing stable engine operation on LFG with considerable variations in the CH4 and CO2 ratios.15 Combustion of syngas and syngas-CH4 blends has been explored with respect to pollutant emissions and engine efficiency, but much of the work15-19 has considered syngas as a complete replacement for gasoline or natural gas fuels in SI engines, using oxidation catalysts15 or biomass gasification17,18 to produce the syngas. The emissions and efficiency results vary. Sobyanin et al. used catalytic reactors to produce syngas composed primarily of H2 and CO with small amounts of CO2, H2O, and CH4. They report that for a SI engine under load with syngas injection the CO, UHC, and NOx emissions were very low compared to gasoline, but at the lean limit CO and UHC emissions increase due to low in-cylinder temperatures, and there was little improvement in engine efficiency.15 Mustafi et al. used “powergas,” a synthetic fuel consisting of CO, H2, and CO2 to compare to gasoline and natural gas. The powergas produced lower CO emissions due to more complete combustion, but higher NOx due to high flame temperatures and higher CO2, likely because of the CO2 present in the powergas. The engine efficiency was lower compared to gasoline and natural gas.16 Shah et al. investigated syngas from biomass gasification composed of H2, CO, and CO2 and found decreased NOx and CO emissions, increased CO2 emissions, and higher efficiency

ARTICLE

compared to gasoline.17 In a computational study, Papagiannakis et al. found that syngas from the gasification of wood (composed of approximately 20% H2, 20% CO, 60% incombustibles), produced higher NO and CO emissions and little change in engine efficiency compared to natural gas.18 Smith and Bartley studied mixtures of syngas and natural gas with exhaust gas recirculation (EGR) and found that the syngas addition extended EGR tolerance by 44% compared to natural gas only. UHC and NOx emissions were reduced, and CO emissions were reduced for less than 20% EGR.19 The discrepancy in emissions and efficiency trends for syngas combustion in SI engines is likely due to the primary fuel used, the varying composition of syngas, and the engine operating parameters. To the best of our knowledge there are currently no studies on syngas injection into LFG fuel streams with respect to emissions and efficiency measurements. This paper will present results from a catalytic reactor reforming simulated landfill gas to syngas. These catalytic tests were done to characterize LFG reforming with respect to syngas production and composition. The data served as a guide for engine performance tests operating on LFG and LFG-syngas mixtures. Engine performance tests investigated emissions and efficiency results of a 5-hp, four-stroke Honda GC160E engine operating on simulated LFG and LFG with syngas addition at various H2/CO ratios.

2. EXPERIMENTAL METHODS The following experimental work was conducted to evaluate the performance of a dry and autothermal catalytic reactor reforming LFG to produce syngas, and to examine the direct use of landfill gas and the addition of syngas (H2 and CO) in a small internal combustion engine. The engine performance and exhaust emissions were evaluated as a function of electrical load. Reforming experiments were performed at atmospheric pressure using a quartz flow-through reactor, shown in Figure 1, and a 4% Rh/γ-Al2O3 wash coated cordierite monolith (400 cpsi). The monolith, obtained from BASF, had a bulk density of 0.44 g cm-3 and wash-coat loading of 0.07 g cm-3. The quartz-reactor assembly was placed in a two-stage furnace (Applied Test Systems, Inc., 3210 Series) controlled with Omega temperature controllers (CN9000A Series). Temperature readings were acquired with thermocouples in the monolith (Omega K-type, KMQIN-020U) and a continuous data acquisition system (Omega, OMB-DAQ-55). The mass flow rates of reactant gases into the reactor were controlled with Aalborg mass flow controllers (GFC17) that were supplied from gas cylinders of ultrahigh purity CH4, CO2, N2, and zero grade O2 (TechAir). The reactor was coupled to an online Agilent Micro GC (QUAD) instrument to measure the gas product composition. To simulate landfill gas, a gas mixture with a 1:1 ratio of CH4: CO2 in a balance of N2 was used. For autothermal reforming experiments, an 8% CH4, 8% CO2, 5% O2, and 79% N2 mixture was used to study outlet gas composition as a function of temperature. For comparison, dry reforming experiments were performed with a gas mixture of 8% CH4, 8% CO2, and 84% N2. Engine experiments were performed using a Honda GC160EQHA engine. The mixture of fuel and air was fed directly into the engine cylinder. Four gases were used to create various simulated gaseous fuels: CH4, CO2, H2, and CO. Air was used as an oxidant. CH4, CO2, H2, and CO were fed from individual gas cylinders (TechAir, 99.97, 99.995, 99.999, and 99.9%, respectively), and 3571

dx.doi.org/10.1021/ie101937s |Ind. Eng. Chem. Res. 2011, 50, 3570–3579

Industrial & Engineering Chemistry Research

ARTICLE

Figure 1. Schematic of catalytic reforming apparatus.

Table 1. Experiment Parameters for LFG Mixture with Syngas Addition load (kW)

rpm

Tad (K)

phi

LHV (MJ/kg)

brake efficiency (%)

50% CH4, 50% CO2 5% syngas (H2/CO = 2)

0.8 0.8

3384 3330

2240.2 2119.2

0.827 0.733

50.016 47.557

11.58 12.64

10% syngas (H2/CO = 2)

0.8

3324

2128.2

0.732

45.232

12.57

15% syngas (H2/CO = 2)

0.8

3318

2144.8

0.735

43.208

12.42

laboratory air was used. The supplied gases were monitored using rotameters (Fisher & Porter Co.). Pressure gauges were connected with each in order to accurately calculate mass flow rates of the gases. The range of conditions tested without syngas included 100, 85, 75, and 50% CH4 with a balance of CO2. To determine the effect of added syngas, mixtures of H2 and CO were added to the 50% CH4, 50% CO2 mixture in amounts of 5, 10, and 15%. For the mixtures of H2 and CO, the H2/CO ratios were chosen on the basis of results from the catalytic testing and consisted of 0.5, 1.0, and 2.0. Each condition was repeated three times. Error bars are included on the figures showing one standard deviation from the mean of the three runs. As just discussed, the engine operated on simulated LFG mixtures from pure gases metered via mass flow controllers and rotameters. The tests that included syngas were done by adding pure H2 and CO from separate cylinders, each individually controlled via rotameters. The primary objective for the syngas addition tests was to maintain a constant lower heating value (LHV) of the total fuel mixture, equal to that of the 50% CH4, 50% CO2 mixture in order to obtain a consistent adiabatic temperature (Tad) for a given equivalence ratio (Φ). The equivalence ratio is defined as the stoichiometric air to fuel ratio divided by the actual air to fuel ratio used in the engine. A constant LHV of fuel mixture was chosen to mimic energy content variations as if the engine were operating on a pure fuel. However, the complexity of controlling five different gas inputs (CH4, CO2, H2, CO, air) led to a variation of LHV (or equivalence ratio) by about 10% for conditions targeted to be constant. The achieved experimental parameters are shown in Table 1. Specifically, the difference in LHV input was approximately 5% between 50% CH4 -50% CO2 and 50% CH4 -50% CO2 with 5% syngas. This resulted in a difference in equivalence ratio of approximately 11% when the air was added to complete the mixture. The engine was directly connected to a PRAMAC EG2800 electric generator. In this work, the term “load” is used to mean the

measured electric power produced by the electric generator. A bank of several light bulbs was used to vary the electric load produced by the generator. To increase or decrease the engine loading, the number of bulbs was increased or decreased. This “load board” consisted of 16 different light bulbs of 100-200 W each, wired in parallel, with every two bulbs sharing a switch, to allow easy load variation for flexible testing. To measure the power the engine’s electric generator produced, a Wattsup pro power meter (Model 99333) was used to continuously monitor the power output of the engine during testing. The engine was tested under loads from 0.2 to 0.8 kW to enable direct comparison across all tests. Emission analysis was conducted with an ENERAC (M700) continuous monitoring unit. The instrument’s probe was inserted into the exhaust flow. A pump located inside the device drew a small amount of sample from the main exhaust gas. The sample was dried before entering the analyzer, via an onboard water trap and filtered for particulate matter. The emissions are reported on a volumetric basis. The overall schematic of the test system and specifications of instruments are shown in Figure 2 and Table 2, respectively.

3. RESULTS AND DISCUSSION The composition of LFG, specifically the CH4 and CO2 content, can vary due to many factors including waste composition and age of the landfill. Therefore, the engine was tested over a range of CH4 and CO2 concentrations for CO, UHC, and NOx emissions. The engine was then tested with the addition of syngas to the landfill gas for a range of CH4, CO2, H2, and CO concentrations. The catalytic reactor tests provided data on syngas concentration and H2/CO ratios that can be achieved while reforming landfill gas for various reactor conditions. While there was extensive testing on the catalytic reactor,14 only the pertinent results related to the engine tests will be discussed. 3572



ARTICLE

Figure 2. Schematic of LFGTE test system.

Table 2. Specification of Instruments measured parameter emission analyzer

range

CO UHC

0-2000 ppm 0-2000 ppm

accuracy (2% (3%

NO

0-300 ppm

(2%

NO2

0-300 ppm

(2%

power meter

true power

0-1500 W

(1.5%

rotameter

flow rate of CH4, CO2, CO, and H2

0-20 dm3/min

(2%

flow rate of air

0-200 dm3/min

(2%

Figure 3. Adiabatic flame temperature calculation and equilibrium calculation for various CH4 and CO2 mixtures.

3.1. Effect of LFG CO2 Content on Emissions. A primary consideration when investigating the source of emissions from

combustion systems is the temperature at which the combustion occurs. While measurements are ideal, they are not possible to 3573



Figure 4. CO emissions measurements as a function of CO2 fraction and engine load.

obtain without extreme modifications to the engine which also will likely impact the performance. Therefore, the adiabatic flame temperature was calculated to ensure that changes in gas composition did not change the maximum possible combustion temperature. Figure 3 shows the results of those calculations as a function of engine load. The flame temperature was calculated on the basis of the inlet gas composition at a temperature of 300 K, the outlet gas composition, and the total flow rate which had to be adjusted for each composition in order to achieve the same engine load. Figure 3 shows the result of the flame temperature calculations as points and the result of equilibrium calculations based on the inlet gas composition as lines. Figure 3 shows that the adiabatic flame temperature increases with engine load. As the CO2 content of the gas mixture increased, the flame temperature remained nearly constant for the mixtures containing 0-25% CO2. The mixture of 50% CH4 and 50% CO2 had a slightly lower flame temperature with a maximum temperature difference of 40 K from the other mixtures at 0.4 kW load. This indicates that, for 0-25% CO2 fuel mixtures, observed changes in emissions can be attributed to impacts of gas composition and the resulting chemistry of combustion rather than temperature effects. However, for 50% CO2 fuel mixtures, temperature can have a small effect on emissions. The emissions of CO and UHC are indicative of combustion performance. Since CO and UHC are produced when the combustion is not complete, more CO and UHC emissions typically result when there is poor mixing of the fuel and air, not enough residence time, or low combustion temperature (although very high temperature will also result in high CO, UHC emissions will be very low). In Figure 4, CO emission measurements are presented as a function of the CO2 fraction in the gas mixture. For each concentration of CO2, the engine load was changed by adjusting the load board. Figure 4 shows that, as the CO2 fraction increases, CO emissions increase by 80% for the 0.2 kW load condition and double for the 0.8 kW load condition. Increasing CO2 may result in higher CO emissions due to dilution of the fuel leading to less complete combustion. The increase in CO2 may also promote CO2 dissociation, leading to greater amounts of CO. Higher engine load results in higher flame temperatures, as shown in Figure 3, which also promotes CO2 dissociation beginning at approximately 1800 K. This may explain why at higher loads the

ARTICLE

Figure 5. UHC emissions measurements as a function of CO2 fraction and engine load.

rate of CO production increases faster with increasing CO2 fraction. For CO2 concentrations less than 25%, the emissions decrease as the load increases for a given CO2 fraction. This is likely because an increase in load results in slower engine speed and therefore more residence time in the combustion chamber to fully oxidize CO to CO2. The UHC emission data are shown in Figure 5 as a function of CO2 concentration in the inlet fuel mixture. The data follow the same trend as CO emissions. Again, as the CO2 in the fuel mixture increases to 50%, the emissions rise from below 80 ppm to 150 ppm for 0.2 kW load condition. Similar to the CO emissions, the increasing CO2 fraction may dilute the fuel, leading to incomplete combustion and higher UHC emissions, even though the flame temperature does not change significantly as CO2 fraction increases. The last major emissions component examined was NOx. Since the testing utilized simulated landfill gas composed of pure CH4 and CO2, there were no contaminants such as H2S; therefore, SOx emissions were not produced. In addition, in the simulated LFG there were no nitrogen components other than air, leaving three main mechanisms for NOx formation. The primary mechanism is thermally initiated and controlled, and is also called the Zeldovich mechanism. High temperatures cause the nitrogen from the air to dissociate into nitrogen radicals, thus becoming very reactive, and combine with oxygen to form NO. Further reactions in the combustion chamber and downstream in the exhaust convert some of the NO to NO2. While this mechanism does not become significant until temperatures exceed 1800 K, NOx increases exponentially as temperature increases above 1800 K, so once engaged it quickly becomes the biggest contributor. Another mechanism is the Fenimore or prompt NOx route. Under the Fenimore mechanism, fuel radicals such as CH3 and others react with diatomic nitrogen from the air to form a CHN molecule and a nitrogen radical. This chemically produced nitrogen radical then can react with oxygen the same way as occurs in the Zeldovich mechanism. Finally, NOx can be produced via the nitrous mechanism which is initiated at high pressures found in combustion cylinders. Here diatomic nitrogen reacts with an oxygen radical to form N2O, which can then reduce to NO and participate in the NO to NO2 mechanism in the system. 3574



ARTICLE

Shown in Figure 6 are NOx emissions measured as a function of CO2 concentration. For each load condition, as CO2 fraction increases, NOx emissions slightly decrease. The largest reductions were at the 0.2 kW load condition where the NOx concentration decreased from 24.8 ppm at 0% CO2 to 12.4 ppm at 50% CO2 and at the 0.8 kW load condition where the NOx decreased from 125 to 100 ppm. This reduction could be caused by the higher CO and UHC emissions as CO2 fraction increases, shown in Figures 4 and 5. These emissions produce a reducing atmosphere that inhibits the production of NOx. The CO may participate in a CO þ N route to form CON, which upon further reaction forms CO2 and N2, effectively preventing the N2 radicals from forming CHN radicals. For all CO2 concentrations, as the load increases, resulting in higher residence time and higher

temperatures, NOx emissions increase. This suggests that the dominant NOx generation mechanism is likely the thermally controlled Zeldovich mechanism. In summary, as the CO2 fraction of the landfill gas increases, the CO and UHC emissions increase by nearly 100%. However, the NOx emissions decrease by 50% for the 0.2 kW load condition and by 20% for the 0.4, 0.6, and 0.8 kW load conditions. To assess the industrial relevance of these tests, emissions from the Honda tested engine were compared to emissions from commercially operating LFGTE systems.20 Figure 7 presents the emissions comparison of CO, UHC, and NOx between the various LFGTE systems and the Honda GC160E engine results detailed in this study. In Figure 7, emissions are expressed as the mass of emitted pollutant per supplied power. Table 3 gives the specifications of the four combustions systems. Compared to the small Honda engine used in this study, the three LFGTE systems were much larger, operating at an industrial scale on real landfills. The two turbines (Solar and Capstone) emitted much less CO, UHC, and NOx than the Caterpillar engine. For all engines except the Caterpillar, CO was the highest emissions value. These emissions differences between the Caterpillar system and the other two industrial LFGTE systems are likely due to the inherent differences between IC engines and turbines: generally turbines use more air for combustion than IC engines (i.e., have a lower equivalence ratio), allowing more complete combustion. Also because the combustor is separate from the power system in a turbine, more uniform combustion can be achieved. CO and NOx emissions of the Honda engine were similar to those of the Capstone CR200 microturbine, but UHC emission was considerably less than the others. Also, the Honda’s emissions were very different from the Caterpillar’s, especially the UHC, even though both of them were IC engines. This was, in part, because the Honda engine was far smaller than the other three systems, producing less power and consuming less fuel. More importantly, the three field LFGTE systems were fueled on real LFG while the Honda engine used a simulated LFG as a fuel. While composition of a real LFG can vary according to the type of MSW and landfill conditions, composition of the simulated LFG was constant throughout each experiment. Also, the real LFG almost certainly contained some contaminants such as H2S whereas the simulated LFG did not. These differences in composition and purity of fuel and operating conditions may have contributed to the difference in emissions. However, in spite of these differences among the four systems, the emissions are comparable for the Honda engine used in this study since the field engines are orders of magnitude larger yet the emissions are different by a factor of 5 at the most. Therefore, results from this study can be considered representative of what would be expected in field operation. 3.2. Syngas Production in Catalytic Reactor. The addition of syngas was investigated as a method of reducing CO, UHC, and NOx emissions in landfill gas mixtures with high CO2

Figure 6. NOx emissions measurements as a function of CO2 fraction and engine load.

Figure 7. Emissions of four different combustion systems.

Table 3. Specification of Four Different Combustion Systems Caterpillar G3516LE

Solar Centaur 40

Capstone CR 200

Honda GC160E

type

IC engine

gas turbine

microturbine

IC engine

dimensions (L W H, mm) weight (kg)

3327 1703 1859 8022

8763 2440 2591 26015

3660 1700 2490 2270

337 369 331 11.5

power output (kW)

815

3515

200

2

3575



ARTICLE

content. Dry and autothermally reforming landfill gas were explored as a way to supply the syngas to the engine. The catalytic reactor tests were performed to supply actual reactor output data on syngas composition as a result of dry and autothermally reforming landfill gas. Figure 8 shows a schematic of a potential LFGTE system in which a portion of the LFG is routed to the catalyst and reformed to H2 and CO, which is then mixed back into the LFG stream. The result is a more reactive mixture of primarily CH4, CO2, H2, and CO. Preliminary calculations indicate the slip stream of LFG will be between 10 and 50% depending on the composition of the landfill gas. The amount of syngas produced in the reformer and the H2/ CO ratio of the syngas are functions of the catalyst and reactor operating conditions for a given LFG composition. Figure 9 shows the results of autothermally reforming a gas mixture of 8% CH4, 8% CO2 , 5% O2 , and 79% N2 using a Rh/γ-Al2O3 monolith. Product gas composition is shown as a function of monolith reactor outlet temperature. As the monolith is heated, CH4, CO2, and O2 are consumed to produce H2, CO, and H2O as a result of the following global reactions:14 methane combustion: CH4 þ2O2 f2H2 OþCO2 methane partial oxidation: 1 CH4 þ O2 f 2H2 þ CO 2

ΔH ¼ -802 kJ mol-1

ð1Þ

ΔH ¼ -36 kJ mol-1

ð2Þ

Figure 8. Schematic of potential LFGTE system.

dry reforming: CH4 þCO2 f 2H2 þ2CO

ΔH ¼ 247 kJ mol-1

ð3Þ

steam reforming: CH4 þH2 O f 3H2 þCO

ΔH ¼ 206 kJ mol-1

ð4Þ

water gas shift: H2 OþCO H CO2 þH2

ΔH ¼ -42 kJ mol-1

ð5Þ

In Figure 9, the O2 is consumed rapidly at low monolith temperature due to methane combustion and partial oxidation to produce H2O, CO2, H2 and CO. As the temperature increases and the oxygen is consumed, a combination of dry reforming, steam reforming, and water gas shift occurs to fully convert the CH4 to produce primarily syngas. The mixture also contains a small amount of H2O produced from the combustion reaction and residual CO2 that is unreacted. After the CH4 is consumed, above 923 K, the water gas shift reaction is still active, consuming CO2 and H2 to produce H2O and CO. Figure 10 shows the amount of syngas production and the H2/ CO ratios obtained while autothermally and dry reforming a simulated landfill gas. Data for dry reforming a 8% CH4, 8% CO2, and 84% N2 mixture is shown to compare to autothermally reforming a 8% CH4, 8% CO2, 5% O2, and 79% N2 mixture. H2/ CO ratio is shown on the primary ordinate and syngas production is shown on the secondary ordinate as a function of monolith outlet temperature. Syngas production is calculated as the sum of H2 and CO divided by the sum of all non-N2 products. Figure 10 shows that as the monolith outlet temperature increases to 973 K, at full conversion of CH4, the syngas production reaches 70% for ATR and 90% for dry reforming. For ATR, the H2/CO ratio decreases from 4.7 at 623 K to 1.2 at 973 K, and for dry reforming, the H2/CO ratio increases from 0.5 at 673 K to 0.97 at 973 K. Figures 9 and 10 show that LFG can be reformed to produce syngas at various H2/CO ratios and concentrations as a function of the reforming temperature.

Figure 9. Autothermal reforming of a 1:1 CH4:CO2 gas mixture using a Rh/γ-Al2O3 monolith. Gas composition is a function of monolith outlet temperature. 3576



ARTICLE

Figure 10. Syngas production and H2/CO ratios obtained while dry and autothermally reforming a 1:1 CH4:CO2 gas mixture as a function of monolith outlet temperature.

ATR provides a wider range of H2/CO ratios due to the multiple reactions occurring in the system, discussed previously. The benefit of ATR compared to dry reforming is that ATR includes the exothermic reactions of methane combustion and partial oxidation that supply heat for the endothermic reforming reactions, resulting in lower heat input, or in autothermal operation, zero external heating. 3.3. Syngas Addition to LFG. As previously discussed, the composition of syngas produced in an autothermal reactor varies depending on factors such as catalyst formulation and reactor operation conditions. In this study the ratios of syngas used (H2/ CO = 0.5, 1.0, and 2.0) were chosen based on the catalytic work discussed in section 3.2. These fuel mixtures were tested in the Honda engine to compare to the emissions output from the CH4 and CO2 mixtures (i.e., 0% syngas addition). The adiabatic flame temperature was calculated for each of the syngas cases and compared to the simulated landfill gas without syngas addition to understand how the change in gas composition affected the combustion temperature. These values are shown in Table 1. With syngas addition, the adiabatic flame temperature dropped by 95-121 K depending on the amount of syngas added. Because the flame temperature varies as a result of syngas addition, this likely has an effect on the emissions results. However, since the temperature trends are known, useful observations can still be made about the testing. Figure 11 shows the measurements of CO emissions at 0.8 kW load condition as a function of the mixture of syngas with H2/CO ratios of 0.5, 1.0, and 2.0 added to the simulated LFG fuel. The 0% syngas condition is the same data shown in Figure 4 at the 50% CO2 condition. Figure 11 shows that the addition of as little as 5% syngas with the simulated LFG reduced the CO emissions by nearly a factor of 4. This reduction is due to the H2 and CO in the syngas making the LFG fuel much more reactive. H2 has extremely high laminar flame speed, high flammability limits, and low ignition energy caused by its low dissociation energy; further, CO is a highly ignitable gas, and its burning rate is also very fast.21 Therefore, the addition of syngas to LFG fuel changes both the chemical and

Figure 11. CO emissions measurement as a function of syngas fraction mixed with simulated LFG (50% CH4-50% CO2) at 0.8 kW load condition.

physical processes in combustion, resulting in a more reactive fuel mixture allowing more complete combustion than pure LFG fuel. Furthermore, adiabatic flame temperature calculations indicate that the addition of 5% syngas would decrease the flame temperature by approximately 121 K, which typically would increase CO emissions. However, Figure 11 shows that even with a 121 K decrease in adiabatic flame temperature the syngas still reduces the CO emissions by a factor of 4. Interestingly, more syngas (10 and 15%) does not measurably reduce the emissions any further. This may be an indication of the fluid dynamics within the cylinder governing the emissions production. The CO emissions generally decreased as the H2/ CO ratio increased from 0.5 to 2.0 for all syngas (5-15%) added mixtures. Therefore, increasing the H2 content of the syngas is more effective in reducing emissions than increasing the syngas content in the fuel. A similar trend is shown for UHC emissions production with syngas addition. Figure 12 shows the measured UHC emissions 3577



Figure 12. UHC emissions measurement as a function of syngas fraction mixed with simulated LFG (50% CH4-50% CO2) at 0.8 kW load condition.

Figure 13. NOx emissions measurement as a function of syngas fraction mixed with simulated LFG (50% CH4-50% CO2) at 0.8 kW load condition.

as function of syngas mixed with the simulated LFG at 0.8 kW load condition. Similar to CO, UHC is also produced when the combustion is not complete; accordingly, more reactive LFG-syngas fuel mixtures or higher flame temperatures reduce UHC emissions. Here again, the UHC emissions are reduced by approximately a factor of 5 when syngas is added to the simulated LFG, even when the flame temperature was decreased by 121 K. Similar to the CO emissions data, an increase of syngas beyond 5% does not have a noticeable effect on reducing the emissions output. Finally, the emissions for NOx are shown in Figure 13 as a function of syngas added to the simulated LFG fuel at 0.8 kW load condition. The addition of syngas to LFG also decreased NOx emissions because even though the fuel mixture was more reactive (due to the already mentioned high flame speed, high flame stability and low ignition energy of syngas), the syngas allowed a more stable combustion of the fuel at lower temperature, as indicated by flame temperature calculations. Figure 13 shows that NOx is reduced by approximately 40% from 0 to 5% syngas fraction. Again, the major emissions reduction occurs with

ARTICLE

Figure 14. Brake specific fuel efficiency as a function of syngas fraction mixed with simulated LFG (50% CH4-50% CO2) at 0.8 kW load condition.

the addition of 5% syngas with little improvement as more syngas is added. Moreover, there is a slight increase in emissions as the syngas fraction is increased from 5 to 15%. The flame temperatures calculated for syngas addition exceed 2100 K, so the dominant NOx formation mechanism is the thermal NOx mechanism. As the syngas fraction was increased from 5 to 15%, adiabatic temperature increased by 26 K, causing NOx emissions to increase. For the syngas addition, the highest NOx values were found at a syngas fraction of 15% with a H2/CO ratio of 1. Preliminary calculations using the GRI 3.0 mechanism model22,23 at constant flame temperature confirm the emissions trends for CO and UHC, particularly the emissions decrease with addition of 5% syngas but an increase for 10 and 15% syngas. For NOx emissions, however, the GRI model predicts an increase in NOx at 5% syngas with a decrease in NOx for 10 and 15% syngas. However, the engine data show a decrease in NOx at 5% syngas with an increase for 10 and 15% syngas which is attributed to temperature effects. In summary, the CO and UHC emissions trends are real but the NOx is more sensitive to temperature effects and is the subject of further quantification. To demonstrate that the effects of syngas addition are truly related to improvements in combustion stability and robustness, a comparison of brake specific fuel efficiency was calculated. The fuel efficiency is defined as the ratio of engine power output to the heat release rate of the fuel: P 3600 ¼ ð6Þ ηf ¼ m_ f QHV sfcQHV QHV is the lower heating value of the fuel [MJ/kg]. The brake specific fuel consumption (bsfc) is defined as follows: m_ f ð7Þ bsfc ¼ P m_ f is the mass flow rate of fuel [g/h], and P is the engine power output [kW].24 Figure 14 shows the results of the fuel efficiencies as a function of the syngas percentage in the fuel mixture. Figure 14 shows that the engine efficiency increases by approximately 10% due to 5% syngas addition at 0.8 kW load condition. The fuel conversion efficiency decreases slightly as syngas fraction increases. This is because the fuel consumption 3578


Industrial & Engineering Chemistry Research rate increases as more syngas is added. Therefore, the reductions in emissions can be attributed to H2 and CO creating a more reactive mixture allowing for more complete combustion.

4. CONCLUSIONS This work has shown that as landfill gas quality decreases due to increasing amounts of CO2, when combusted the CO and UHC emissions significantly increase while NOx emissions slightly decrease. As the engine load increases, resulting in a rise in adiabatic flame temperature, the CO and UHC emissions decrease while NOx emissions increase. The emissions data obtained from these experiments were compared to commercial LFGTE systems operating in the field, showing that the data from this study can be considered representative of a fieldoperated IC engine fueled with LFG. For effective use of LFG fuel for IC engines, pollutant emissions such as CO, UHC, and NOx should be reduced. This study shows that syngas addition is effective for emissions reduction and may be easier to obtain via a compact, on-stream reformer. Syngas addition to the LFG fuel not only also reduced pollutant emissions but also improved the fuel conversion efficiency due to the H2 and CO in syngas allowing more complete combustion in the engine cylinder. It was found that 5% syngas addition was most effective in reducing pollutant emissions and improving engine efficiency, and a higher addition of syngas did not significantly decrease the emissions further. Furthermore, greater concentrations of H2 in the syngas fraction (i.e., higher ratios of H2/CO) resulted in greater reduction of pollutant emissions compared to lower H2/CO ratios. Therefore, the addition of syngas is effective in reducing engine emissions and improving fuel conversion efficiency, even for landfill gases containing very low CH4/CO2 ratios. ’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT The authors gratefully acknowledge the Waste-to-Energy Research Technology Council, Sustainable Utilization of Resources, and the Earth Engineering Center at Columbia University for partial support of this project. ’ REFERENCES (1) LFG Energy Project Development Handbook; U.S. EPA Landfill Methane Outreach Program (LMOP); 2010. (2) An Overview of Landfill Gas Energy in the United States; U.S. EPA Landfill Methane Outreach Program (LMOP); 2010. (3) Tapping landfill’s energy potential. Milwaukee Journal Sentinel, Eastern Racine County Section; 2003. (4) Marcari, N. C.; Richardson, R. D. Operation of a caterpillar 3516 spark-ignited engine on low-btu fuel. J. Eng. Gas Turbines Power 1987, 109 (4), 443. (5) Mueller, G. P. Landfill gas application development of the caterpillar G3600 spark-ignited gas engine. J. Eng. Gas Turbines Power 1995, 117 (4), 820. (6) Karim, G. A.; Wierzba, I. Methane-Carbon Dioxide Mixtures as a Fuel. SAE Tech. Pap. 1992, Paper No. 921557; DOI: 10.4271/ 921557. (7) Huang, J.; Crookes, R. J. Assessment of simulated biogas as a fuel for the spark ignition engine. Fuel 1998, 77, 1793.

ARTICLE

(8) Crookes, R. J. Comparative bio-fuel performance in internal combustion engines. Biomass Bioenergy 2006, 30 (5), 461. (9) Shrestha, S. O.; Narayanan, G. Landfill gas with hydrogen addition—A Fuel for SI engines. Fuel 2008, 87, 3616. (10) Porpatham, E.; Ramesh, A.; Nagalingam, B. Investigation on the effect of concentration of methane in biogas when used as a fuel or a spark ignition engine. Fuel 2008, 87, 1651. (11) Akansu, S.; Dulger, Z.; Kahraman, N.; Veziroglu, T. Internal combustion engines fueled by natural gas-hydrogen mixtures. Int. J. Hydrogen Energy 2004, 29, 1527. (12) Rakapoulos, C. D.; Kyritsis, D. C. Hydrogen enrichment effects on the second law analysis of natural and landfill gas combustion in engine cylinders. Int. J. Hydrogen Energy 2006, 31 (10), 1384. (13) Rakapoulos, C. D.; Michos, C. N. Generation of combustion irreversibilities in a spark ignition engine under biogas-hydrogen mixtures fueling. Int. J. Hydrogen Energy 2009, 34 (10), 4422. (14) Kohn, M. P.; Castaldi, M. J.; Farrauto, R. J. Auto-thermal and dry reforming of landfill gas over a Rh/γAl2O3 monolith catalyst. Appl. Catal., B 2010, 94, 125. (15) Sobyanin, V.; Sadykov, V.; Kirillov, V.; Kuzmin, V.; Kuzin, N.; Vostrikov, Z.; Smirnov, E.; Sorokin, A.; Brizitzkiy, O.; Terentyev, V.; Khristolyubov, A.; Luksho, V.; Afanasiev, A. Syngas as a fuel for IC and diesel engines: efficiency and harmful emissions cut off. Proceedings of International Hydrogen Energy Congress and Exhibition; Istanbul, Turkey, 2005. (16) Mustafi, N. N.; Miraglia, Y. C.; Raine, R. R.; Bansal, P. K.; Elder, S. T. Spark-ignition engine performance with `Powergas’ fuel (mixture of CO/H2): A comparison with gasoline and natural gas. Fuel 2006, 85, 1605. (17) Shah, A.; Srinivasan, R.; Filip To, S.; Columbus, E. Performance and emission of a spark-ignited engine driven generator on biomass based syngas. Bioresour. Technol. 2010, 101, 4656. (18) Papagiannakis, R. G.; Rakopoulos, C. C.; Hountalas, D. T.; Giakoumis, E. G. Study of the performance and exhaust emissions of a spark-ignited engine operating on syngas fuel. Int. J. Altern. Propul. 2007, 1 (2/3), 190. (19) Smith, J.; Barley, G. Stoichiometric Operation of a Gas Engine Utilizing Synthesis Gas and EGR for NOX Control. J. Eng. Gas Turbines Power 2000, 122, 617. (20) Westby, K. J.; Castaldi, M. J. A Comparison of Landfill Gas to Energy Technologies. Presented at IT3; 2008; Paper No. 120. (21) Dong, C.; Zhou, Q.; Zhao, Q.; Zhang, Y.; Xu, T.; Hui, S. Experimental study on the laminar flame speed of hydrogen/carbon monoxide/air mixtures. Fuel 2009, 88, 1858. (22) Smith, G.; Golden, D.; Frenklach, M.; Moriarty, N.; Eiteneer, B.; Goldenberg, M.; Bowman, T.; Hanson, R.; Song, S.; Garinder, W.; Lissianski, V.; Qin, Z. GRI Mechanism 3.0; http://www.me.berkelye. edu/gri_mech/. (23) Ianni, J. C. Kintecus, Windows Version 4.00; www.kintecus. com; 2010. (24) Heywood, J. B. Internal Combustion Engine Fundamentals; McGraw-Hill: New York, 1988.

3579


Performance of an Internal Combustion Engine ... - ACS Publications

Recommend Documents