Experimental Study of the Performance of a Stationary Diesel Engine

---

Energy Fuels 2009, 23, 5062–5072 Published on Web 09/16/2009

: DOI:10.1021/ef900311w

Experimental Study of the Performance of a Stationary Diesel Engine Generator with Hydrogen Supplementation Vamshi K. Avadhanula,† Chuen-Sen Lin,*,† Dennis Witmer,‡ Jack Schmid,‡ and Praveen Kandulapati† †

Department of Mechanical Engineering, and ‡Alaska Center for Energy and Power, University of Alaska Fairbanks Fairbanks, Alaska, USA Received April 8, 2009. Revised Manuscript Received August 29, 2009

Recent claims of engine efficiency improvement via hydrogen generated by small electrolyzers have been made by numerous businesses selling these devices. The technical literature indicates that using hydrogen in gasoline internal combustion results in some improvement in engine emissions. Literature with detailed experimental results with hydrogen supplementation in a stationary diesel engine generator set was not found. This paper presents the results of the experiment that was conducted on a 125 kW stationary diesel engine generator set by addition of supplementary hydrogen in the intake air stream. The hydrogen was from a compressed hydrogen tank, and the testing was completed in two separate runs, the first with H2/air ratios by volume between 0 and 0.7% and the second with H2/air ratios by volume between 0 and 2.3%. The upper limits of these ranges are far above the levels provided by the small electrolyzers currently being sold. The experiments were conducted at a constant load of 56 kW and 1200 rpm. It was observed that the amount of fuel energy consumed (the sum of the diesel and H2 fuel values) remained constant as the engine was run at constant load and that no increase in thermal efficiency occurred. Net indicated work per cycle per cylinder was measured using an in-cylinder pressure sensor. The net indicated work per cycle per cylinder was nearly constant throughout each experiment. Some combustion parameters, such as start of ignition and end of premixed combustion, have been checked and showed no significant variations due to the added hydrogen in the intake air stream. O2, SO2, and CO emissions decreased as the hydrogen in intake stream increased. As the hydrogen flow rate in intake air increased NO2 emissions increased, NO emissions and NOx emissions decreased up to 50 dm3/min of H2 and then increased up to 150 dm3/min of H2. All the above-mentioned changes are relatively small amounts. These results indicate that there is no change in engine efficiency with the addition of hydrogen and only minor changes in emissions. However, the efficiency of the engine system (the engine and its auxiliary components) may increase with the added hydrogen, if the hydrogen is generated from the engine waste energy, such as engine exhaust waste heat.

manifold. Since the amount of oxygen (O2) produced in the hydrogen electrolyzer is negligible when compared with the oxygen present in the intake air, our experiment was conducted using commercial high-purity bottled hydrogen, which allowed introduction of larger flows of hydrogen into the engine without the need for a large (and expensive) electrolyzer. The experiment was conducted on a Detroit Diesel 50 series engine, with data acquisition using National Instruments (NI) hardware and LABVIEW software. The performance characteristics of the engine, such as fuel energy consumption (Q•T), net indicated work per cycle (Wc,in), and emissions data were measured for different flow rates of hydrogen and compared with zero hydrogen flow rate (i.e., normal engine performance data). Lower heating value and density of hydrogen5,6 and No. 2 diesel fuel7 used for calculations are listed in Table 1. A literature review was conducted to find previous studies on hydrogen supplementation tests that were conducted on the internal combustion engines, both spark ignition (SI) and compression ignition (CI) engines.

Introduction and Literature Review In recent years, numerous companies selling small-scale hydrogen electrolyzers claimed that considerable engine efficiency improvement has been observed by using these devices.1-3 Hydrogen electrolyzers dissociate pure water (H2O) into hydrogen (H2) and oxygen (O2) as shown in eq 1.4 In these devices, the electric power for this process is from the engine electrical circuit such that hydrogen is produced only when the engine is running. The products (H2 and O2) of the process are then mixed with the fuel-air mixture in gasoline engines and the air intake manifold in diesel engines. Most of the claims were for gasoline engines,2,3 but a few were for diesel engines.1,2 2H2 OðlÞ f 2H2 ðgÞ þ O2 ðgÞ E O ¼ 1:229 V

ð1Þ

This work studies the effects on a diesel engine by introducing different volume flow rates of hydrogen into the intake *To whom correspondence should be addressed. E-mail: clin@ alaska.edu. Telephone: 907-474-5126. (1) URL: http://www.hydrorunner.com/index.php (accessed on February, 15, 2009). (2) URL: http://www.greenfuturetechnology.com/index.html (accessed on February, 15, 2009). (3) URL: http://www.punchhho.com/index.php (accessed on February, 15, 2009). (4) Lide, D. R. CRC Handbook of Chemistry and Physics; CRC Press, Inc.: Boca Raton, FL, 1995. ISBN 0-8493-0596-9. r 2009 American Chemical Society

(5) Yuksel, F.; Ceviz, M. A. J. Energy 2003, 28, 1069–1080. (6) National Institute of Standards and Technology. URL: http:// webbook.nist.gov (accessed on December, 10, 2008). (7) Sastry, K. Properties and Performance Evaluation of Syntroleum Synthetic Diesel Fuels; Master's Thesis, University of Alaska Fairbanks: 2005.

5062

pubs.acs.org/EF

Energy Fuels 2009, 23, 5062–5072

: DOI:10.1021/ef900311w

Avadhanula et al.

model was studied at 2500 rpm and variable torque ranging from 15 to 75 N 3 m. The engine was tested for H2-to-gasoline ratios by mass from 15 to 201%. For lean hydrogengasoline-air mixtures at lower torques, NOx emissions were less compared to pure gasoline results. For rich mixtures at higher torques (60 N 3 m and higher), NOx emissions were higher compared to pure gasoline mixtures. CO2 emissions decreased with increase in the quantity of hydrogen at a torque value. On the basis of the above literature studies of adding H2 into gasoline engines, lean air-gasoline mixtures are possible, which reduce the possibility of engine misfiring.5,8,10 For lean mixtures, NOx emissions9-11 and CO emissions10 decreased and HC emissions increased.10 For rich mixtures, a reduction in CO and HC emissions and an increase in NOx emissions were observed.8,9 There was an increase in thermal efficiency or decrease in specific fuel consumption by addition of hydrogen to the air-gasoline mixture.5,8-10 However, does not conclusively prove that the supplementary hydrogen contributes to the reduced specific fuel consumption. On the basis of our estimates, by considering the contribution of energy in both gasoline and hydrogen, specific fuel energy (energy from both gasoline and hydrogen) consumption improvement was negligible. For a given power output, the addition of hydrogen results in a corresponding reduction in gasoline consumption.5,8,10,11 In all the above experiments of hydrogen supplementation in gasoline engines, hydrogen was from a compressed hydrogen tank. Compression Ignition Engines. S.O. Bade Shrestha et al.12 conducted experiments on a Chevrolet Silverado 6.5 L turbocharged V8 diesel engine. A hydrogen generation system (HGS), which produces hydrogen and oxygen by electrolysis of water, was used as hydrogen source. The hydrogen-to-oxygen flow rate into the engine was 690 cm3/min per unit of HGS. One to three units were used. Hydrogen-to-air by volume was estimated to be about 0.074-0.235%. The general results showed that there was enhancement in combustion process and reduced exhaust emissions as the amount of hydrogen flow rate increased by switching on additional HGS units. The maximum reductions of particulate matter (PM), CO, and NOx are up to 60, 30, and 19%, respectively, when compared to the pure diesel experiment. HC emissions did not show any regular trend. A. Tsolakis et al.13 conducted experiments on a Lister Petter TR1 single-cylinder, naturally aspirated, direct-injection diesel engine. The engine has a rated power output of 8.6 kW (6.5 bar IMEP) at 2500 rpm. The experiment was conducted at 5.5 bar IMEP and 1500 rpm. The experiment was conducted without exhaust gas recirculation (EGR) for H2-to-intake charge by volume varied between 0.5 and 1% and with EGR for hydrogen in intake charge ratio between 0.5 and 2%. For both cases, NOx emissions and smoke decreased and no significant change in engine efficiency was observed. According to these two studies of supplementation of hydrogen in intake air stream of diesel engine, NOx emissions decreased.12,13 CO emissions decreased from Shrestha et al.12 HC emissions did not show any trend according to Shrestha et al.12 and reduced according to Tsolakis et al.13

Table 1. Properties of Hydrogen and No. 2 Diesel Fuel Used for Calculations5-7 property

hydrogen

density (kg/m3) low heating value (kJ/kg)

0.080 993 @ 30 °C 120 000

No. 2 diesel fuel 784.2-835.3 43 654.355

Spark Ignition Engines. F. Yuksel and M.A. Ceviz5 conducted experiments on a 4-cylinder 75 kW gasoline engine with hydrogen flow rates of 0.129, 0.168, and 0.208 kg h-1, corresponding to hydrogen to air ratios by volume between 1.92 and 4.76%. A small decrease in brake horsepower was observed as the energy density of the mixture enriched with hydrogen was less. Thermal efficiency was reported to increase to nearly 44.0% at 0.208 kg h-1 of hydrogen when compared to pure gasoline, for which thermal efficiency was 38.1%; and we know that the heating value of hydrogen is nearly 2.72 times greater than that of gasoline. Hydrogenenriched fuel contributed to leaner air-fuel mixtures. In average, consumption of gasoline decreased by about 11.5%, and heat loss to cooling water decreased by 36%. Engine performance parameters deteriorated for hydrogen-to-gasoline ratios greater than 5% by mass due to the intake air limitations. Petkov and Barzev8 have conducted experiments on a 4-cylinder VAZ-2103 gasoline engine. They conducted experiments at varying loads from 3 to 27 kW at 2400 rpm. The engine was tested at hydrogen flow rates of 0.18 and 0.29 kg h-1. The mass percentage of hydrogen in total mass of gasoline and hydrogen is in the limits of 2-6% at 0.18 kg h-1 and 4-10% at 0.29 kg h-1 of hydrogen flow rates. The engine was tested at full load only on gasoline and at part loads with both hydrogen and gasoline. With hydrogen supplementation in the air-fuel mixture, results show that CO decreases between 25 and 96%, HC reduces between 15 and 64%, NO emissions increased by 70%, and gasoline consumption decreases between 23 and 25%. Results depend on engine load. Nicolae Apostolescu and Radu Chiriac9 conducted experiments on a single-cylinder, 4-stroke 1.3 L gasoline engine with rated power output of 39.7 kW at 5250 rpm. The experiment was conducted at 45, 30, and 15% load and 1800 rpm. Hydrogen was introduced into the intake manifold. Hydrogen-to-air ratios by volume were between 1.44 and 9.17%. Brake thermal efficiency was reported to increase by 10%. A reduction in HC by up to 35% for lean mixtures and an increase in NOx, except for lean mixtures, were observed. Hoehn and Dowdy10 conducted experiments on a road vehicle fueled with hydrogen-enriched modified gasoline engine. The numerical results of the experiment were given at 40 BHP and 2000 rpm. The ratio of H2 to air by volume is about 11.9% and H2 to gasoline by mass is 15%. When compared with pure gasoline results, for hydrogen-enriched gasoline the thermal efficiency was reported to increase from 32.5 to 37.5%, NOx decreased by 97.7%, CO decreased by 63.3%, and HC increased by 1000%. G. Fontana et al.11 had done a numerical simulation study of a gasoline engine using modified KIVA-3 V code. The (8) Petkov, T.; Barzev, K. Int. J. Hydrogen Energy 1987, 12 (10), 701– 704. (9) Nicolae, A.; Radu, C. SAE Tech. Pap. Ser. No. 960603, 1996. (10) Hoehn, F. W.; Dowdy, M. W. Feasibility Demonstration of a Road Vehicle Fueled with Hydrogen-Enriched Gasoline. 9th Intersociety Energy Conversion Engineering Conference, San Francisco, CA, 1974. (11) Fontana, G.; Galloni, E.; Jannelli, E.; Minutillo, M. SAE Tech. Pap. Ser. No. 2002-01-2196, 2002.

(12) Bade Shrestha, S. O.; LeBlanc, G.; Balan, G.; de Souza, M. SAE Tech. Pap. Ser. No. 2000-01-2791, 2000. (13) Tsolakis, A.; Megaritis, A.; Wyszynski, M. L. J. Energy Fuels 2003, 17, 1464–1473.

5063

Energy Fuels 2009, 23, 5062–5072

: DOI:10.1021/ef900311w

Avadhanula et al.

Engine efficiency also did not improve much by hydrogen supplementation.13 Literature with detailed experimental results (engine thermal efficiency, emissions, and net indicated work) of H2 test on a midsize turbo-charged stationary diesel engine generator set, which is the type of diesel engine generator applied in numerous rural villages, was not available. This paper may help in getting some useful experimental results in this regard.

Table 2. Diesel Engine Specifications engine type aspiration number of cylinders number of valves compression ratio bore (mm) stroke (mm) connecting rod to crank radius ratio displacement volume (dm3) rated power cooling system fuel injection

Experimental Setup This experiment was conducted to evaluate the impact of hydrogen additions on both engine efficiency and emissions. This experiment was conducted on a Detroit Diesel series 50 conventional diesel engine-generator set and has no EGR or exhaust after-treatment equipment. The diesel engine specifications are given in Table 2. The schematic line diagram of the experimental setup is shown in Figure 1. The engine was coupled to a 125 kW electric generator. A 250 kW resistive/reactive load bank was used for load control. The main components of the experiment are diesel engine, compressed hydrogen tank, MKS hydrogen flow controller, National Instruments data acquisition and control (DAQ) system, exhaust gas analyzer, and in-cylinder pressure measuring sensor. DAQ functions were performed using a LabView virtual instrument program (VI) operating on a National Instruments PXI system. The hydrogen gas from a compressed hydrogen tank was supplied via a separate hydrogen line, and flow was measured and controlled by a MKS hydrogen flow controller supplied by MKS Instruments. The exhaust gas analyzer was a TESTO 350-XL. A Kistler 6124B21 noncooled-type combustion pressure transducer and Kistler 5010B dual mode charge amplifier were used for measuring in-cylinder pressure of the No. 3 cylinder. An electronic circuit was used to tap the injection timing signal from the wires running from the engine’s electronic control module (ECM) to the electronic unit injector (EUI). Temperature was measured at different points of the engine using type-K (Cr-Al) thermocouples. For measuring the intake air flow, a Meriam Instruments laminar flow element was placed in the intake air line between the air filter and turbo charger. The hydrogen flow was introduced into the air filter, resulting in the laminar flow element measuring the flow rate of both air and hydrogen combined. For measuring the diesel fuel flow rate, CANbus data was used. Numerous other installed sensors (e.g., thermocouples, pressure gauges, etc.) enabled data collection for monitoring general performance parameters.

Detroit Diesel series 50 engine turbocharged 4 4 (2-intake, 2-exhaust) 15:1 130 160 4.1423 8.5 125 kW at 1200 rpm water cooled electronic unit injector

air in a stepwise manner. For example, hydrogen was supplied at 1.0 dm3/min for 5 min and then dropped to 0.0 dm3/min for the next 5 min and again increased to 2.0 dm3/min for the next 5 min and dropped to 0.0 dm3/min for the next 5 min, and so on. Hydrogen flow was controlled by the MKS flow controller, via control signals initiated by the LabView VI and sent to the NI SCXI-1124 analog output board of the DAQ system. The hydrogen flow rate schedule, which contains the information of the desired flow rate with respect to time, was stored in the LabView VI. The NI SCXI-1120 analog input board of the DAQ system provided feedback reading the actual flow rate of hydrogen through the flow controller inducted into the intake air stream. During the test, for each hydrogen flow rate, direct measurements were taken for following performance parameters: ambient temperature; CANbus diesel fuel flow rate; CANbus diesel fuel temperature; intake air flow rate; intake manifold temperature and pressure (after intercooler); in-cylinder pressure versus crank position; and O2, CO, SO2, NO, NO2, and NOx concentrations in the exhaust. Laminar flow element data, ambient temperature, CANbus fuel flow rate data, CANbus fuel temperature, intake manifold temperature and pressure, and hydrogen flow rate data samples were taken at a frequency of 0.1 Hz. For each flow rate of hydrogen, 30 readings of all the above parameters were collected. The average of these 30 readings is used for further analyses. The in-cylinder pressure measurement device was calibrated using a procedure described by Kandulapati.14 In-cylinder pressure data was sampled using a NI PXI-4472 high-speed dynamic acquisition module of the DAQ system at a frequency of 70.1 kHz. Five 0.2 s (two thermodynamic cycles) in-cylinder pressure samples were taken with respect to the crank angle for each hydrogen flow rate. The pressure at a crank angle for each hydrogen flow rate was taken as the average of 5 readings. The pressure profiles measured using these pressure sensors will have negative baseline drift of the cylinder pressure during intake and exhaust strokes. This is adjusted by using the cylinder pressure at inlet valve close (IVC), which is equal to the intake manifold pressure. Exhaust gas was analyzed for H2 flow rates of 0, 10, 50, 100, and 150 dm3/min in August. Oxygen (O2) was measured as a volume percent in exhaust. Carbon monoxide (CO), sulfur dioxide (SO2), nitric oxide (NO), nitrogen dioxide (NO2), and oxides of nitrogen (NOx) are measured in parts per million (ppm). The readings are directly taken in respective units from the memory of the exhaust gas analyzer. The emissions analyzer is not a continuous measuring device. As the hydrogen was supplied for 5 min at each flow rate, a sample of exhaust gas was sent through the analyzer for 3 min for emissions analysis. The analysis results of emissions readings were stored in the analyzer memory. For the next few minutes, ambient air was sent through the analyzer to rinse the sensors to make them ready for the next sample. The calibration of the analyzer was verified with calibration gases prior to the experimental runs.

Experimental Procedure The experiment was conducted on two days. The first day was in June to test only for engine performance with H2 injection rates ranging between 0 and 50 dm3/min (H2/air ratio by volume from 0 to 0.7%) and emissions data was taken, but the emissions equipment may not be properly calibrated. However other engine performance data are believed to be reliable. The second day was in August, to test for both engine performance and emissions with carefully calibrated emissions equipment and an extended hydrogen flow rate, which ranged from 0 to 150 dm3/min (H2/air ratio by volume from 0 to 2.3%). For each case, all the data was obtained in one day between 10:00 a.m. and 5:00 p.m. The engine-generator set was located inside a converted shipping Conex. Intake air for the diesel is taken from within the Conex and the state of the intake air, temperature, and relative humidity were observed to be consistent during the course of an experimental run on a single day; however, intake air conditions in June and August were observed to be different from each other. For each experimental run the engine was operated at 56 kW for approximately 3 h to attain near steady-state conditions prior to the experimental run. The experiment was conducted at a load of 56 kW and 1200 rpm. Hydrogen was introduced into the intake

(14) Kandulapati, P. K. In-Cylinder Based Combustion Performance Evaluation of Syntroleum Synthetic and Conventional Diesel Fuels; Master's Thesis, University of Alaska Fairbanks: 2006.

5064

Energy Fuels 2009, 23, 5062–5072

: DOI:10.1021/ef900311w

Avadhanula et al.

Figure 1. Schematic line diagram of experimental apparatus.

Computation Method

Table 3. Direct Data Obtained for Different Hydrogen Flow Rates

For this experiment, the engine was run at a constant load of 56 kW and 1200 rpm. This section discusses the method used to evaluate, based on experimental data, the volume ratio of hydrogen to intake air flow rates, energy supplied from both diesel fuel and hydrogen, the amount of diesel fuel saved for each of the hydrogen injection flow rates, and the thermal efficiency. Other computations, which include mean in-cylinder gas temperature and net heat release rate, are also discussed in this section. Fuel Energy Supplied. From laminar flow element data we get the volume flow rate of air and hydrogen. The calibration of the laminar flow element was described by Sastry.7 Hydrogen flow rates were feedback controlled, and the deviation between the measured and desired hydrogen flow data was within 1%. The volume flow rate of air is calculated by subtracting hydrogen volume flow rate from the total volume rate measured from laminar flow element. Volume percent of hydrogen in air can be evaluated using the hydrogen and air flow rates. In general, the total volume flow rate is much greater than the hydrogen volume flow rate. The mass flow rate of hydrogen can be calculated from the volume flow rate of hydrogen and density of hydrogen, which is listed in Table 1. The average ambient temperature throughout the experiment was 30 °C, and the hydrogen mass flow controller was placed in ambient conditions.

CANbus intake VH2• fuel V•T ambient manifold intake 3 3 • (dm / (dm / tempera- temperaVf temperature manifold min) min) ture (°C) ture (°C) (L/h) (°C) pressure (kPa)

0 1 2 3 4 6 8 10 12 16 20 30 50

6761.24 6693.02 6734.77 6747.02 6747.16 6719.83 6728.31 6736.33 6761.68 6759.38 6753.76 6756.96 6743.74

31.0 31.0 29.7 30.0 31.4 30.2 30.6 30.4 31.1 31.4 31.9 32.6 33.3

A: June 25.13 15.489 25.94 15.471 23.74 15.361 24.30 15.337 26.02 15.400 25.08 15.232 24.99 15.203 25.02 15.195 25.28 15.152 25.62 15.085 25.99 14.955 26.27 14.744 27.68 14.413

39.0 37.0 38.0 38.9 39.0 36.0 37.5 38.0 39.0 40.0 39.7 40.0 42.0

127.18 126.62 127.33 127.26 127.73 127.00 126.81 126.81 127.00 126.87 127.33 126.94 126.68

0 10 50 100 150

6652.20 6662.09 6653.22 6598.11 6574.17

29.4 27.9 29.2 29.1 30.1

B: August 23.26 15.293 22.35 15.026 23.47 14.235 23.21 13.411 23.90 12.833

38.7 35.5 38.6 38.2 40.4

129.16 128.95 128.53 128.37 128.06

From CANbus data we get the volumetric diesel fuel flow rate. Volumetric diesel fuel flow rate was converted into mass flow rate using the calibration curve that was derived by comparing CANbus data to readings obtained 5065

Energy Fuels 2009, 23, 5062–5072

: DOI:10.1021/ef900311w

Avadhanula et al.

Table 4. Volume Percent of H2 in Air, Total Fuel Energy Supplied, and Percent of Hydrogen Energy in Total Fuel Energy with Respect to Different Flow Rates of Hydrogen in June and August VH2• (dm3/min)

VH2 (%)

MH2• (kg/h)

M•f (kg/h)

0 1 2 3 4 6 8 10 12 16 20 30 50

0.000 0.015 0.030 0.044 0.059 0.089 0.119 0.149 0.178 0.237 0.297 0.446 0.747

0 4.860  10-03 9.719  10-03 1.458  10-02 1.944  10-02 2.916  10-02 3.888  10-02 4.860  10-02 5.831  10-02 7.775  10-02 9.719  10-02 1.458  10-01 2.430  10-01

12.546 12.553 12.453 12.424 12.474 12.370 12.330 12.319 12.273 12.209 12.106 11.932 11.644

0 10 50 100 150

0.000 0.150 0.757 1.539 2.335

0 4.860  10-02 2.430  10-01 4.860  10-01 7.289  10-01

12.391 12.208 11.535 10.870 10.383

QH2• (kW)

Q•f (kW)

QH2/T (%)

42.84 42.38 42.98 43.16 42.98 43.15 43.33 43.41 43.72 43.91 44.22 44.82 45.70

0.000 0.162 0.324 0.486 0.648 0.972 1.296 1.620 1.944 2.592 3.240 4.860 8.099

152.135 152.216 151.012 150.653 151.263 150.000 149.520 149.379 148.823 148.046 146.801 144.687 141.203

0.000 0.106 0.214 0.322 0.427 0.644 0.859 1.073 1.289 1.721 2.159 3.250 5.425

42.68 43.32 45.51 47.53 49.19

0.000 1.620 8.099 16.199 24.298

150.250 148.035 139.874 131.811 125.903

0.000 1.082 5.474 10.944 16.177

air/diesel fuel ratio A: June

B: August

from a gravimetric device during the previous experiment.7 The energy in diesel fuel stream and energy in hydrogen stream can then be evaluated using the mass flow rates and heating values of respective fuels from Table 1. The total fuel energy supplied is the sum of energy in diesel fuel and energy in hydrogen fuel. The percentage contribution of hydrogen energy to the total energy is calculated as the ratio of hydrogen energy and the total fuel energy. As the engine was run at constant load, determining the reduction in diesel fuel use when hydrogen is supplied is straightforward. The amount of diesel fuel saved for a given hydrogen flow rate was calculated. Q•T is the total fuel energy present in both diesel fuel and hydrogen. The amount of diesel fuel required (M•f,req) to produce Q•T amount of energy is given by: Q•T ð2Þ Mf•, req ¼ HVf

engine, the in-cylinder volume (V) for a given crank angle can be determined by knowing the clearance volume, bore diameter, stroke length, and connecting rod length to crank radius ratio.15 From cylinder pressure data and the corresponding cylinder volume data throughout the engine cycle can be plotted on a P-V diagram. The net indicated work per cycle (Wc,in) is the work delivered to the piston over the entire four-stroke cycle,15 which is obtained by subtracting pumping work from compression and expansion work. In this study, the area under the curve (for a P-V diagram) was found by using the MATLAB function trapz(P, V), where P and V are the pressure and volume data, respectively. Mean In-Cylinder Gas Temperature (TGas,m). The mean cylinder gas temperatures were calculated using the ideal gas law.16 The mass inside the cylinder is equal to the mass of air and hydrogen until the point of fuel injection. Once the injection starts, the total mass will be equal to the mass of air, hydrogen, and diesel fuel injected. The residual mass from the previous cycle that is trapped in the clearance volume is neglected in this study. Temperature calculations were performed only for timing between inlet valve close (IVC) to exhaust valve open (EVO). Net Heat Release Rate (dQnet/dθ). Net heat release rate was estimated using the first law of thermodynamics single zone model.15,17 Calculations were based on measured in-cylinder pressure versus crank angle. The specific heat ratio depends on temperature and is given by Brunt et al.,18 in this work it was calculated using mean in-cylinder gas temperatures.

As we are supplying hydrogen, actual diesel mass flow rate (M•f ) is less than the required diesel mass flow rate (M•f,req) to generate same amount of power output. Therefore the amount of diesel fuel saved (Mf,saved%) for a given amount of hydrogen injection is given by: ! Mf•, req - Mf•  100 ð3Þ Mf , saved % ¼ Mf•, req Finally, the thermal efficiency (ηth), now called fuel conversion efficiency,15 can be calculated by the relation: ! Pb  100 ð4Þ ηth ¼ Q•T

Results and Discussion Here the results of June and August are presented in separate tables and figures with the differences between the two results discussed separately.

where Pb is the engine load, which is equal to 56 kW, and Q•T is total fuel energy from both diesel and hydrogen fuels. Net Indicated Work Per Cycle (Wc,in). In-cylinder pressure data was sampled with respect to crank angle for each hydrogen flow rate. In this analysis, 0° crank angle represents the beginning of the intake stroke. For a common diesel

(16) Zeng, P.; Assainis, D. N. SAE Tech. Pap. Ser. No. 2004-010922, 2004. (17) Stone, R. Introduction to Internal Combustion Engines; Society of Automotive Engineers, Inc.: Warrendale, PA, 1994. ISBN 1-56091390-8. (18) Brunt, M. F.; Rai, H.; Emtage, A. L. SAE Tech. Pap. Ser. No. 981052, 1998.

(15) Heywood, J. B. Internal Combustion Engines Fundamentals; McGraw Hill: New York, 1988. ISBN 0-070100499-8.

5066

Energy Fuels 2009, 23, 5062–5072

: DOI:10.1021/ef900311w

Avadhanula et al.

Figure 2. Total fuel energy (both diesel fuel and Hydrogen) vs hydrogen flow rate from (a) June and (b) August data.

The experimental results have three different types of data: the first type is obtained with a sample rate of 0.1 Hz, the second type with a sample rate of 70.1 kHz, and the third type is the emissions data directly read from emissions equipment for each hydrogen flow rate. The first type of data, which is listed in Table 3, includes hydrogen flow rates (VH2•) and the corresponding total air flow rates (V•T), ambient temperature, CANbus fuel flow rate (V•f ), CANbus fuel temperature, and intake manifold temperature and pressure. These data are used to calculate the volumetric ratios of hydrogen to air, total fuel energy supplied to the engine from both hydrogen and diesel fuel, and percentage of hydrogen energy in total fuel energy supplied. The second type includes measured in-cylinder pressures and the corresponding crank positions of the third cylinder, which is equipped with an in-cylinder pressure sensor. This data is used to calculate the mean in-cylinder gas temperatures, net heat release rate, and the net work produced per cycle in the cylinder to detect if indicated works are about the same for different hydrogen flow rates at the given generator load of 56 kW and 1200 rpm. The third type of

measurement includes the emissions data of O2, CO, SO2, NO, NO2, and NOx, which are shown later in Figures 7 and 8. Fuel Energy Supplied. Table 4 gives the flow rate of hydrogen (VH2•), the corresponding percentage of hydrogen in intake air by volume (VH2%), mass flow rate of hydrogen (MH2•), the diesel fuel flow rate (M•f ), air to diesel fuel ratio, fuel energy supplied by hydrogen (QH2•) and diesel fuel (Q•f ), and percent of hydrogen energy in total fuel energy (QH2/T%) with respect to each hydrogen flow rate for respective experiments during June and August. Figure 2 gives a bar graph of total fuel energy supplied with respect to different flow rates of hydrogen. Here we can observe that for both the experiments, total fuel energy supplied remains almost constant as the engine is run at constant load and speed (56 kW and 1200 rpm) for different flow rates of H2. At 150 dm3/min of hydrogen flow rate, 16.2% of the total fuel energy supplied comes from hydrogen energy, in other words, at 150 dm3/min (i.e., 2.3% of hydrogen by volume in intake air) of hydrogen flow rate there is diesel fuel savings of 16.2% by mass as shown by linear curve fit (eq 6) in Figure 3B. The linear curve fit in eq 6, which correlates the 5067

Energy Fuels 2009, 23, 5062–5072

: DOI:10.1021/ef900311w

Avadhanula et al.

Figure 3. Linear curve-fit between mass percent of diesel fuel saved (Mf,saved%) and volume percent of hydrogen in intake air (VH2%) for (a) June and (b) August data.

diesel fuel savings to measured hydrogen flow rate, has a coefficient of determination (R2) value of 0.99. A similar observation can be made for June data, and a linear curve fit (eq 5) with R2 = 1.0 is shown in Figure 3. ð5Þ Mf , saved % ¼ 7:2725VH2 % - 0:0033

where Mf,saved% is the mass percent of diesel fuel saved. We can observe that for same H2 flow rates of 0, 10, and 50 dm3/min, the diesel fuel flow rate is slightly less in August than in June by 1.23, 0.9, and 0.94% respectively. This may be attributed to the ambient air conditions during respective seasons.19

The equation for Mf,saved% (eqs 5 and 6) is different for June and August because of a small difference that is observed in diesel fuel flow rate in two seasons and higher hydrogen flow rate range in August. It may be observed from Table 4 that the percent of hydrogen energy (QT,H2%) in total fuel energy (Q•T) remains constant for 10 and 50 dm3/min of hydrogen irrespective of the season. According to the above discussion, the application of supplementary hydrogen may not improve the thermal efficiency of the engine. Based on data from NREL report20 on commercial electrolyzers, some basic calculations have been done; ideally, 39 kWh (140 400 kJ) of electric energy is required to produce 1 kg of hydrogen from electrolysis of

(19) URLhttp://www.alaska.edu/uaf/cem/ine/aetdl/conferences/2008Presentations/SmallHydroElectrolyzers.J.Schmid.pdf (accessed on Jan. 10, 2008).

(20) Kroposki, B.; Levene, J.; Harrison, K.; Sen, P. K.; Novachek, F. National Renewable Energy Laboratory (NREL) Report No. TP581-40605, Sept. 2006.

Mf , saved % ¼ 6:9507VH2 % þ 0:0887

ð6Þ

5068

Energy Fuels 2009, 23, 5062–5072

: DOI:10.1021/ef900311w

Avadhanula et al.

water. Typically, the efficiency of commercial electrolyzers is in the range between 56 and 73%,20 which corresponds to 252 360 kJ/kg and 192 240 kJ/kg of electric energy, respectively. Furthermore, if hydrogen is generated from electrical power, the thermal efficiency will become worse because the energy used for hydrogen generation (192 240 kJ/kg at 73% efficiency of electrolyzer) is higher than the amount of heat contained in the generated hydrogen (heating value of hydrogen is 120 000 kJ/kg), which in turn is higher than the amount of engine work created through the hydrogen combustion in the engine. However, if the hydrogen is generated from the engine waste energy (e.g., engine exhaust heat), the application of supplementary hydrogen will likely increase the efficiency of the engine system, which comprises the engine and all its auxiliary components, including the waste energy hydrogen generator. Net Indicated Work Per Cycle (Wc,in). The experiment was conducted with the diesel generator setting at a constant generator load of 56 kW. To confirm that the engine was also having a constant power for tests at different hydrogen flow rates, the net indicated work per cycle of the third cylinder was used. The net indicated work (Wc,in) applied to the third cylinder per thermal cycle, which provides about one-fourth of the engine output, was calculated using the cylinder P-V diagram, which was derived from the average of five consecutive measured in-cylinder pressure versus crank position data. Table 5 gives the area under P-V curve for different hydrogen flow rates for June and August. The area under the P-V curve indicates the net indicated work (Wc,in) done per cycle on the piston in the third cylinder. From Table 5A we can observe that Wc,in remain almost constant with maximum occurring at 10 dm3/min of H2 (1.94 kJ/cycle) and minimum occurring at 50 dm3/min of H2 (1.83 kJ/cycle); and the variation from its mean value of all calculated indicated work is ( 3%, which is within the uncertainty range of the equipment. Similarly from Table 5B for August data, Wc,in occurs maximum at 150 dm3/min of H2 (1.99 kJ/cycle) and minimum at 10 dm3/min of H2 (1.95 kJ/cycle); and the variation from its mean value of all calculated indicated work is ( 1.1%, which is also well within the uncertainty of equipment. Data in Table 5 indicate that no significant difference in net indicated work has been observed for the engine running at different hydrogen flow rates. The Wc,in obtained from the August data was observed consistently a bit higher than that of Wc,in obtained from the June data for same hydrogen flow rates (i.e., 0, 10, and 50 dm3/min). This may be attributed to the difference of the intake air conditions during the experiments in June and August. The effect of intake air condition on engine performance was studied by Jack.19 For studying the ignition timing, mean in-cylinder gas temperatures, and net heat release rate, we need in-cylinder pressure versus crank angle data (Figure 4). Due to the similarity of the June and August pressure versus crank angle data, only the data for August experiment are presented, which include both pressure and emissions data and are useful for further discussions. The crank angle at which there is rapid raise in cylinder pressure corresponds to the start of ignition.21 In Figure 4, the change in the start of

Table 5. Indicated Work Per Cycle for Different Flow Rates of Hydrogen for June and August Data VH2• (dm3/min)

Wc,in (kJ/cycle) A: June

0 1 2 3 4 6 8 10 12 16 20 30 50

1.8602 1.8771 1.8363 1.8643 1.849 1.9122 1.9379 1.9434 1.8827 1.8328 1.849 1.8377 1.8325 B: August

0 10 50 100 150

1.9552 1.9511 1.9845 1.9814 1.9941

ignition for different hydrogen flow rates is less than 1° crank angle, as can be observed. The hydrogen flow rate is considered to have insignificant effect on timing of start of ignition as observed. Mean In-Cylinder Gas Temperatures (TGas,m). Figure 5 give the mean in-cylinder gas temperatures calculated from the pressure data using ideal gas law. The general trend is, the higher the hydrogen flow rate, the higher the peak TGas,m and residence time of the high-temperature gases in the cylinder. Net Heat Release Rate (dQnet/dθ). Figure 6 gives the calculated net heat release rate with respect to crank angle for different hydrogen flow rates. Based on Figure 6, the ends of premixed combustions are estimated using the first minimum values of the corresponding heat release curves.15 The variation in estimated end of premixed combustions is within 1° crank angle, which is considered to have insignificant effect on premixed combustion. Emissions. The emissions data in August are used in this discussion. Figures 7 and Figure 8 give the exhaust gas emissions data and emissions plot at different hydrogen flow rates. Oxygen (O2) emissions decreased as the hydrogen flow rate in intake air increased, as shown in Figure 7. O2 emissions decreased by 1.82% at 150 dm3/min of hydrogen flow rate compared to 0 dm3/min of hydrogen. This was expected due to the combustion with increased H2 fuel. Carbon monoxide (CO) emissions did not show any regular trend as the hydrogen flow rate increased in the intake manifold. But on overall observation of data, when compared to 0 dm3/min of hydrogen flow rate, CO emissions decreases with the increase of the hydrogen flow rate, except at 10 dm3/min flow rate. The reduced CO emissions at higher hydrogen flow rate may be attributed to the corresponding higher air/diesel fuel ratio (Table 4B). In general, higher air to diesel fuel ratio is considered to associate with lower CO emissions.12,15,17 However, the effect of higher in-cylinder gas temperature (Figure 5) (corresponding to the higher hydrogen flow rate), which reverses the dissociation reaction back to CO,15,17 may help in increasing CO emissions and also needs to be considered.

(21) Huang, Y.; Zhou, L.; Wang, S.; Liu, S. J. Automobile Eng. 2006, 20 (6), 827–835.

5069

Energy Fuels 2009, 23, 5062–5072

: DOI:10.1021/ef900311w

Avadhanula et al.

Figure 4. In-cylinder pressure vs crank angle for August data.

Figure 5. Mean cylinder gas temperature vs crank angle from August data.

As we know from the above discussion of fuel energy supplied, at 150 dm3/min of hydrogen flow rate there is a decrease in fuel consumption by 16.2%; and we also know that sulfur emissions come from diesel fuel properties, so we expect a decrease in total sulfur emissions by the same amount. As shown in Figure 7 sulfur dioxide (SO2) emissions decreased at a steady rate as the hydrogen flow rate increased. At 150 dm3/min of hydrogen flow rate, fuel consumption decreased by 16.2% but SO2 emissions decreased by 73% when compared to 0 dm3/min of hydrogen flow rate. The decrease in SO2 emissions may also be attributed to the higher air/diesel fuel ratios at higher hydrogen flow rate22

(Table 4B) and the possibility of formation of other sulfur components such as sulfur trioxide (SO3) and simply sulfur (S). Since SO2 can further react with water (H2O) to form sulfurous acid (H2SO3), which is considered the most influential sulfur compound in limiting the capacity of exhaust heat recovery systems due to its high dew point and high corrosive property, the reduction in SO2 concentrations in the exhaust is desirable. This may also help in the reduction of corrosion in exhaust after treatment equipment. Figure 8 shows the NO and NO2 emissions with respect to different hydrogen flow rates, and the emissions analyzer measures NOx as the sum of NO and NO2. The overall NOx decreased up to 50 dm3/min hydrogen and then increased for higher hydrogen flow rates when compared to pure diesel

(22) Corro, G. J. React. Kinet. Catal. Lett. 2002, 75 (1), 89–106.

5070

Energy Fuels 2009, 23, 5062–5072

: DOI:10.1021/ef900311w

Avadhanula et al.

Figure 6. Net heat release rate vs crank angle for different hydrogen flow rates for August data.

Figure 7. O2, CO, and SO2 emissions at different hydrogen flow rates.

NOx emissions. NOx formation is strongly dependent on burned gas temperature and air/fuel ratio.15,17 NOx emissions also depend on the fuel nitrogen content. Higher gas temperature, lower air to fuel ratio, and more diesel fuel combusted result in higher NOx emissions. When the gas temperature exceeds a critical level, NOx will be generated much faster. In the presence of hydrogen flow, the diesel fuel into the cylinder is less, the air to fuel ratio is higher (Table 4B), and the gas temperature (and residence time) is higher. These first two parameters are expected to reduce the NOx emissions, but the third parameter is expected to increase the NOx emissions. In addition, hydrogen property

of faster burning, lower activation energy, higher heating value, and more molecular collisions than conventional liquid and gaseous fuels12 may also need to be considered for the evaluation of the effect of the supplementation hydrogen on NOx emissions. NOx emissions decreased from Shrestha et al.12 and Tsolakis et al.13 for H2 to air ratio between 0.074-0.235% and 0.5-1%, respectively, thus matching the results for compression ignition engines up to the H2 to air volume percent mentioned in this study. The decreasing trend of CO emissions in the presence of H2 also matches with literature for compression ignition engines.12 5071

Energy Fuels 2009, 23, 5062–5072

: DOI:10.1021/ef900311w

Avadhanula et al.

Figure 8. Concentrations (ppm) of NOx, NO, and NO2 in exhaust at different hydrogen flow rates.

emissions results, O2 emissions were observed to decrease monotonically with the increase of hydrogen flow rate. (5) The decrease of SO2 content in the exhaust was proportional to the hydrogen flow rate. (6) CO emissions showed no simple and clear trend with respect to the hydrogen flow rate. But in general the presence of hydrogen (up to 2.3% of H2 in air by volume) seemed to decrease the CO emissions except at very low hydrogen flow rate (i.e., less than 0.5% of H2 in air by volume). (7) NOx emissions decreased up to 50 dm3/min of hydrogen flow rate and then increased with a slow rate when the hydrogen flow rate increased up to 150 dm3/min. On the basis of the data obtained in the experiment, NOx emissions reduced in the presence of hydrogen in intake air. (8) The effect of the hydrogen flow rate, of less than 2.3% by volume in air, on the combustion parameters of in-cylinder pressures, calculated mean in-cylinder gas temperatures, start of ignition, end of premixed combustion, calculated net heat release rate are observed to be insignificant. (9) The application of supplementary hydrogen may increase the efficiency of the engine system (the engine and its auxiliary components including the hydrogen generator), if the hydrogen is generated by engine waste energy, such as engine exhaust waste heat.

Conclusions Experiments on a midsize diesel engine-generator set have been conducted with supplementary hydrogen in intake manifold. The engine was tested for hydrogen to air ratio by volume of 0-0.7% in June 2008 and 0-2.3% in August 2008. The following conclusions have been arrived at from the results obtained: (1) No considerable improvement in thermal efficiency or reduction in total fuel energy consumption (sum of diesel fuel and hydrogen fuel) has been observed. (2) As the hydrogen flow rate (or hydrogen energy supplied) increased in the intake manifold, there was decrease in diesel fuel energy consumed by the engine. According to the experimental data obtained from both June and August, the decrease in diesel fuel consumption was strongly proportional to the increase in the hydrogen flow rate. (3) For each of the June and August experiments, the calculated results of net indicated work per cycle per cylinder for different hydrogen flow rates remained almost constant with small variation, which is within the uncertainty limit of the measuring equipment. In other words, the change in H2 flow rate did not cause any significant change in engine net indicated work. (4) From

5072