Effects of Spark Advance, A/F Ratio and Valve Timing on Emission and Performance Characteristics of Hydrogen Internal Combustion Engine

Transcription

1 Licensed from the SAE Digital Library Copyright 29 SAE International Downloaded Monday, May 18, 29 1:6:2 AM Effects of Spark Advance, A/F Ratio and Valve Timing on Emission and Performance Characteristics of Hydrogen Internal Combstion Engine Farhad Salimi, Amir H. Shamekhi and Ali M. Porkhesalian K.N.Toosi University of Technology Copyright 29 SAE International ABSTRACT The se of hydrogen as an engine fel has great potential for redcing exhast emissions with the exception of a small amont of carbon containing emissions originating from the lbricating oil, is the only polltant emitted. In this paper by sing a trblent flame speed method, a qasi-dimensional thermodynamic model of an SI engine feled with hydrogen is developed. In this work, simlation reslts are validated by experimental data. Then the effect of spark advance, A/F ratio and valve timing on emission and performance characteristics of the modeled engine has been investigated. Hence, remarkable effects in emission and performance characteristics observed. And the behavior of the modeled engine against the above parameters has been investigated and the reason of that is discssed. Keywords: A/F ratio, Emission, Hydrogen, Performance, valve timing. INTRODUCTION Hydrogen can be prodced from many different feedstocks, inclding natral gas, coal, biomass, and water. If hydrogen is prodced from renewable sorces, the global warming potential, which nowadays is a major problem and a hot topic, of hydrogen is insignificant in comparison to hydrocarbon based fels since combstion of hydrogen prodces no carbon-based componds sch as HC, CO, and CO2 [1]. If hydrogen is prodced sing renewable energy, it is an energy carrier that redces emissions of noxios exhast gases and greenhose gases to zero or a very small fraction of the emissions fond when sing traditional fossil fels. Hydrogen has special properties so the combstion characteristics of hydrogen are very different from gasoline. The laminar flame speed of a hydrogen air mixtre at stoichiometric condition is abot 1 times that of gasoline. It has a wide flammability limit, preignition and back firing can be a problem. The flammability limits correspond to eqivalence ratios of.7 to 9 [1]. Water injection into the intake manifold is sed to mitigate preignition and provide cooling. Exhast gas recirclation and lean operation are sed to redce levels. Also the octane rating of hydrogen of 16 RON [1] allows increasing compression ratio. Hydrogen is one of the most interesting alternative fels, and is recently in the centre of attention. Hydrogen can be prodced from renewable sorces and offers lots of other benefits. One of the most practical one is its ability to rn in bi-fel conditions. Also Hydrogen internal combstion engines have the ability for an increased efficiency [2]. Comprehensive overview of hydrogen engine properties and design featres is already done by previos athors and it was conclded that hydrogen internal combstion engines have a high efficiency, are very clean and considerably cheaper than fel cells []. Limited nmbers of previos pblications have worked on hydrogen engine simlation. A two-zone qasidimensional engine model for calclating power and emission was demonstrated in Fagelson et al. pblication []. In their work laminar brning velocity is The Engineering Meetings Board has approved this paper for pblication. It has sccessflly completed SAE s peer review process nder the spervision of the session organizer. This process reqires a minimm of three () reviews by indstry experts. All rights reserved. No part of this pblication may be reprodced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, withot the prior written permission of SAE. ISSN Positions and opinions advanced in this paper are those of the athor(s) and not necessarily those of SAE. The athor is solely responsible for the content of the paper. SAE Cstomer Service: Tel: (inside USA and Canada) Tel: (otside USA) Fax: CstomerService@sae.org SAE Web Address: * * Printed in USA

2 calclated from a second order reaction with estimated activation energy. In another literatre [] the same model was sed in order to predict the performance of a spercharged hydrogen engine. They report an overestimation of the rate of pressre rise that can be a conseqence of an overestimation in brning velocity. Another pblication [6] se a Wiebe s law in a zero-dimensional model. And for a varying compression ratio and ignition timing the optimm cylinder diameter for the minimm emission and maximm power for a fixed eqivalence ratio calclated. In Verhelst s literatre [7] qasi-dimensional model was preferred to mlti-dimensional one becase of a reasonable accracy and fast comptation on PC system. And they developed a complete cycle simlation code for SI engine and they looked thoroghly at the trblent combstion in hydrogen felled engine. In another pblication [8] a simlation and design tool applicable to hydrogen powered spark ignition engine systems is introdced. The code is applicable to single and mlti-cylinder engines nder steady state or transient operating conditions the model is validated against experimental data for the intake flow model. In this paper, first of all a qasi-dimensional code for the for strokes of SI hydrogen engine is developed. In this simlation the trblent flame speed is modeled based on previos literatre and also some modification was applied to the flame speed method. The model then calibrated matching the data obtained in the previos experiment [9]. A combination of valve timing, spark advance (SA) and air to fel ratios variations on engine emission and performance is also stdied. ENGINE MODEL The engine model is a qasi-dimensional two-zone model which solves the basic differential eqations for the intake, compression, power and exhast strokes. In this model, the combstion chamber is divided into two zones, which one shold imagine as being divided by flame front, zone 1 contains nbrned mixtre and zone 2 contains brned mixtre. Thermal formation also takes place in brned zone, which is described by the extended Zeldovich mechanism [1]. The flame front is assmed to travel by a speed called trblent flame speed which is a fnction of laminar flame speed that is compted from previos stdies [11]. The engine model ses Woschni correlation [12] to estimate engine heat transfer this correlation is not accrate for hydrogen, considering the lack of any accrate validated alternative the error of the Woschni correlation for hydrogen engines is inevitably accepted in the model. Brned an nbrned zones are calclated by assming that the flame travels in a sphere like shape. The engine model also incldes a friction model Licensed from the SAE Digital Library Copyright 29 SAE International Downloaded Monday, May 18, 29 1:6:2 AM to predict brake mean effective pressre. The frictional processes in an internal combstion engine can be categorized into three main components: (1) the mechanical friction, (2) the pmping work, and () the accessory work. For this work a method, which calclates the total friction work accrately, was sed [1]. The applied friction model spontaneosly predicts all the categories above. The composition of the reaction prodcts is calclated from the chemical eqilibrim at a given pressre and temperatre of the 12 species N 2, NO, N, CO 2, CO, OH, H, O 2, O, H 2 O, H 2, Ar. An optimization calclation procedre is sed to calclate the mole fraction of each species and the total mole fraction [1]. The engine model is validated by comparing the simlated reslt with the experimental data taken from a previos engine experiment [9]. The engine sed in the experimental evalation is a dedicated hydrogen-feled engine. The dedicated hydrogen-feled engine is the descendant of a gasoline-feled engine, which was later converted to hydrogen-feled one. LAMINAR FLAME SPEED - Flame speed is a critical effective parameter in model reslts, so sing an accrate formlation is essential. Almost all of the trblent combstion models assme that the combstion happens in flamelet regime. It is then assmed to travels locally at the laminar flame speed therefore it is necessary to know the laminar flame speed of the hydrogen/air mixtre. First a short overview over hydrogen brning velocity is given. Li and MacFarlane In their pblication [1] laminar brning velocity of hydrogen/air mixtre was measred and their measrements reslted in a formla as a fnction of fel/air eqivalence ratio and residal gas mole fraction. Milton and Keck Milton and Keck [1] took ot the laminar brning velocity of stoichiometric hydrogen/air mixtre from some experiments. Then they fitted a correlation to the experimental data. Ijima and Takeno - Ijima and Takeno [11] described the laminar brning velocity of hydrogen/air mixtre by a zero-dimensional analysis. Their experiments reslted in a formla which will be shown. Other Kinds of formla has been developed by some other athors too, inclding Koroll, Kmar and Bowels [16], Taylor et al. [17], Law et al [18], Kobayashi et al [19]. All of the above formla can be seen in their pblications. From all the correlations developed by the athors above only Iljima and Takeno s [11] formla is a fnction of the three of the fel/air eqivalence ratio, temperatre and pressre. The other s formlation doesn t inclde one of them so their eqation has the enogh integrity

3 that was needed for this work. Iljima and Takeno s correlation is not a fnction of residal gas, therefore in this paper their correlation is modified by a residal gas effect term which is taken from Verhelst s correlation [2]. Iljima and Takeno s [11] formla is as follows:. T 1 + = T l l ( ) p Log T P Flame speed is in (m/s). Licensed from the SAE Digital Library Copyright 29 SAE International Downloaded Monday, May 18, 29 1:6:2 AM P (1) θ 6 = TDC (1. ) =.7U TDC P θ is the crank angle (6 at TDC of compression). The above correlation has been validated for specific engine geometry. In this paper the method above was calibrated by some calibration factors to fit the crrent engine geometry and hence prodcing accrate reslts. (7) The modified correlation is: l T P = T l ( ) 1 + p Log (1 f) (2) T P Where; P is the pressre, T is the nbrned temperatre, f is the residal gas volme fraction, T =291(K), P =112(Pa). α T and β P are as follows: T p = 1. = ( φ 1) ( φ 1) Where, φ is the fel air eqivalence ratio. The parameter γ expressing the effect of residal gases is given by: γ φ = () () MODEL VALIDATION The engine sed is a dedicated hydrogen-feled engine which was converted from a gasoline engine to bi-fel operation and later sed when operated by hydrogen. By sing the modified Ijima and Takeno method a two zone engine model in Matlab environment is developed. Engine specification, which can be observed in Table 1, was imported to the model. The calclated data was validated by previos experimental work [9]. The calclated and experimental in cylinder pressre verss the crank angle is really near to each other, which can be observed in Fig1. This shows that, sing the modified Iljima and Takeno flame speed method was sitable. The fel air ratio for Fig1 is not a practical one bt only on this particlar condition, P-theta diagram was presented in the related literatre [9] so this was sed for validation inevitably. Table 1 Engine specification And l which is the laminar brning velocity of hydrogen at 291(K) and 1(atm), in (m/s), given by: 2 l = 2.98 ( 1) +.2 ( φ 1.7 ) φ () Displacement Bore.8 cc 86. mm By sing the above vales the simlation reslts was accrate for different engine speed and conditions. It shold be considered that the method above does not prodce a stable and stretch-free laminar brning velocity. TURBULENT FLAME SPEED Many methods for describing and calclating the trblent flame speed have been developed by previos athors. In this paper the so-called Damkohler and derivatives method [2] is sed according to this model trblent flame speed is as follows: + t = l (6) Stroke 7.2 mm Compression ratio 9.7 Intake valve diameter 6. mm Exhast valve diameter 2. mm Intake valve lift 8. mm Exhast valve lift.7 mm Where [2, 21]:

4 6 Experimental Simlated Licensed from the SAE Digital Library Copyright 29 SAE International Downloaded Monday, May 18, 29 1:6:2 AM fel/air eqivalence ratio is shown in Fig..The reslts are near enogh with a little difference which can be a reslt of heat transfer and flame speed correlations error, this shows that the engine model works good enogh. 7 6 P (bar) Experimental Simlated CA (degree) φ A) 6 Experimental Simlated Figre 2 concentration verss fel/air eqivalence ratio at 28 (rpm) 8 7 Experimental Simlated P (bar) BMEP (Kpa) 6 B) CA (degree) Figre 1 A) In cylinder pressre verss crank angle at 28 (rpm), φ=1.6 B) Combstion part The reslts for the experimental and simlated concentration, which is calclated from extended Zeldovich mechanism [1], verss eqivalence fel/air ratio for some points, can be seen in Fig.2. As it is seen that the simlated and experimental reslts for concentration is pretty good for low fel air eqivalence ratio bt the model nder-predicts concentration at high fel air eqivalence ratio, which is not a practical operating condition. And this difference from experiment can be probably becase of the error of sing Woschni s heat transfer correlation and the error of the flame speed correlation sed. As this error occrs in a non practical operating condition, it can be neglected. Also the experimental and calclated reslts for BMEP verss φ Figre BMEP verss fel/air eqivalence ratio at 28 (rpm) AIR TO FUEL RATIO A/F ratio plays a very important role in engine performance and emissions characteristics. Hydrogen has special properties so the combstion characteristics of hydrogen are very different from gasoline. The laminar flame speed of a hydrogen air mixtre at stoichiometric condition is abot 1 times that of gasoline. The wide flammability limit of hydrogen allows the se of very lean fel/air eqivalence ratios, as low as.2, which reslt in redcing emissions.

5 Licensed from the SAE Digital Library Copyright 29 SAE International Downloaded Monday, May 18, 29 1:6:2 AM rpm 1 rpm 18 rpm 2 rpm rpm rpm rpm φ Figre emission verss fel/air eqivalence ratio at different engine speeds for MBT ignition timing A/F ratio has a major effect on. This effect is shown on Fig.. As it is shown the vale of is varied from near to near 1 bar respecting to fel/air eqivalence ratio. Maximm amont of occrs when the fel/air eqivalence ratio is abot 1, stoichiometric condition, and this repeats in a wide range of engine speeds. SPARK ADVANCE Spark advance (SA) is another parameter that has a major effect on engine performance and emission. Fig.6 shows the verss spark advance (SA) for different engine speeds. As it is seen in low engine speeds, 1-18 (rpm), the maximm happens when the spark advance range is [- ]. Bt as the engine speed increases the maximm happens in more advanced spark and in (rpm) this point is near 2 degrees before TDC. 6. The major effect of A/F ratio on engine emission for different engine speeds is shown in Fig.. As it is shown the vale of emission is varied from 16 (ppm) to near zero respecting to fel/air eqivalence ratio. Maximm amont of emission occrs when the fel/air eqivalence ratio is abot.8 and this happens in a wide range of engine speeds. As it is shown in Fig. the concentration peak at near.8 fel/air eqivalence ratio, bt as the mixtre becomes leaner the concentration falls dramatically. One of the most important parameter in determining SI engine emissions is the fel/air eqivalence ratio and in hydrogen engines it is mch more important and controlling this parameter is really important and effective rpm 1 rpm 18 rpm 2 rpm rpm rpm rpm 1 2 SA (CA BTDC) rpm 1 rpm 18 rpm 2 rpm rpm rpm rpm Figre verss fel/air eqivalence ratio at different engine speeds for MBT ignition timing φ Figre 6 verss spark advance (SA) at different engine speeds, φ=. Spark timing significantly affects emission levels. Advancing the timing so that combstion occrs earlier in the cycle increases the peak cylinder pressre, retarding the timing decreases the peak cylinder pressre. Higher peak cylinder pressres reslt in higher peak brned gas temperatres, and hence higher formation rates. For lower peak cylinder pressres, lower formation rates reslt [1]. This matter is shown on Fig.7 and it can be seen that the varies between abot 6 (ppm) to near zero respectively to spark advance. This shows the major importance of spark advance on hydrogen engine s characteristics.

6 Licensed from the SAE Digital Library Copyright 29 SAE International Downloaded Monday, May 18, 29 1:6:2 AM SA (CA BTDC) EVO advance (CA) Figre 7 emission verss spark advance (SA) for 1 (rpm), φ=. VALVE TIMING In internal combstion engine valves behavior (lift and timing) is one of the most important parameters which have a major effect on the engine operation and emission. By sing VVT technology we are able to control engine behavior in any conditions with the prpose of decreasing emission and optimizing engine operating characteristics IVO advance (CA) Figre 8 emission and verss IVO for 1 (rpm), φ=. In this paper the model calclates the volmetric efficiency according to valve timing and also the in cylinder residal gas is calclated according to valve lift and timing based on Senecal et al. literatre [22]. Figre 9 emission and verss EVO for 1 (rpm), φ=. As it is shown on Fig.8 as the IVO advances concentration and increases de to volmetric efficiency increase, and conseqently increase of incylinder temperatre bt when the volmetric efficiency approaches its maximm the valves overlap effect shows p. Valves overlap factor increase the amont of residal gas. Any brned gas in the nbrned mixtre redces the heating vale per nit mass of mixtre and, ths, redce the adiabatic flame temperatre [1]. Therefore as the residal gas mass fraction increases, decreases and as a reslt of redction of incylinder temperatre, concentration falls. As the EVO advances the exhast valve opens in higher in-cylinder pressre therefore more brned gas leaves the cylinder and there is less residal gas so the concentration and increase becase more fresh mixtre enters the cylinder bt as the EVO advances more and more in one point the valves overlap factor overcomes the first effect and therefore residal gas increases and this cases an decrease in and. These trends can be seen on Fig.9. One of the other effective variables which can be controlled by sing VVT mechanism is valves (Intake and exhast) lift. By increasing the intake valve lift the volmetric efficiency and therefore in-cylinder temperatre increases and this case an increase in and concentration, intake valve lift increase also cases an increase in valves overlap factor so as the intake valve lift increases the valves overlap effect cases an increase in residal gas and therefore decrease in concentration, the trends mentioned can be seen on Fig.1. The moment of intake and exhast valves opening has a great effect on engine emission and operation which can be seen on Figs.8 and.9.

7 Intake valve lift (mm) Figre 1 emission and verss intake valve lift for 1 (rpm), φ=. As the exhast valve lift increases more brned gas leave the cylinder and therefore there is less residal gas in the cylinder and this cases an increase in concentration, as it is shown infig11, as the residal gas decreases more fresh mixtre enters the cylinder and increases Exhast valve lift (mm) Figre 11 emission and verss exhast valve lift for 1 (rpm), φ=. Valves opening dration which is the time between valves (intake or exhast) opening till closing can be controlled by sing VVT mechanism at different conditions. Licensed from the SAE Digital Library Copyright 29 SAE International Downloaded Monday, May 18, 29 1:6:2 AM... The mentioned series of events also affects the. By increasing the intake valve opening dration, increases and by frther increase in dration, declines becase of an increase in residal gas mass fraction Intake valve opening dration (CA) Figre 12 emission and verss intake valve opening dration for 1 (rpm), φ=. As the exhast valve opening dration increases more brned gas leaves the cylinder so more fresh mixtre fills the cylinder in the intake stroke this case a sharp increase in and in-cylinder temperatre. Therefore, as it is shown on Fig.1 concentration rises too. By increasing dration more, the overlap factor increases and conseqently residal gas increases, so the concentration and falls slightly When the intake valve opening dration increases volmetric efficiency increases too and this case a rise of in-cylinder temperatre and conseqently concentration grows. Bt when the dration increases more and more valves overlap factor increases too and therefore residal gas grows and eventally concentration redces. Exhast valve opening dration (CA) Figre 1 emission and verss exhast valve opening dration for 1 (rpm), φ=. CONCLUSION The prpose of this work was to develop a hydrogen engine model which can be sed as an engine simlator

8 to predict engine emission and performance characteristics with accrate reslts. The engine simlator was validated by experimental data. In this paper also sensitivity of hydrogen engine to spark advance, A/F ratio and valve timing stdied and their importance were shown. It was shown that the hydrogen engine has its most concentration at the point near to φ =.8. Also SA effect was stdied and the major effect of this parameter considered and discssed. Variation of valves lift, opening time and dration was stdied and the reason of their effects flly discssed. The valve timing stdies can also be applied to VVT mechanism. REFERENCES 1. Fergson, Colin R, Kirkpatrick, Allan T., Internal Combstion Engines: Applied Thermoscience, 2 nd Edition, John Wiley & Sons Inc, Tang X. et al., Ford P2 hydrogen engine dynamometer development, presented at SAE World Congress, , 22.. Verhelst, Sebastian, Sierens, Roger, Stefaan, Verstraeten, A critical review of experimental research on hydrogen feled SI engines, presented at SAE World Congress, 26-1-, 26.. Fagelson, J.J., McLean, W.J, de Boer, P.C.T (1978). Performance and emissions for spark-ignited Combstion engines sing alternative fels qasi One-dimensional modeling. I. hydrogen feled engines. Combstion Science and Technology, 18(pp7-7).. Prabh-Kmar, G.P, Nagalingam, G., Gopalakrishnan, K.V (2). Theoretical stdies of a spark- Ignited spercharged hydrogen engine. Int l. Jornal of Hydrogen Energy, 28(pp77-8). 6. Ma, J., S, Y., Zho, Y., Zhang, Z. (2). Simlation and Prediction on the performance of a vehicle s hydrogen engine. Int l Jornal of Hydrogen Energy, 28(pp77-8). 7. Verhelst, S., Verstraeten, S., Sierens, R., Development of a simlation code for hydrogen Felled SI engines, presented at Proceedings ASME Spring Technical Conference, 26 ICES26-117, Drew, Alan N., Timoney, David J., Smith, William J. (27). A simlation and design tool for hydrogen SI engine Systems-Validation of the intake hydrogen flow model. Int l Jornal of Hydrogen Energy, 2 (pp8-92). 9. Swain, Michael R., Schade, Gregory J., Swain, Matthew, Design and testing of a Dedicated Hydrogen-Feled Engine, presented at SAE World Congress, 96177, Heywood, J.B., Internal Combstion Engine Fndamentals, McGraw-Hill Book Company, Licensed from the SAE Digital Library Copyright 29 SAE International Downloaded Monday, May 18, 29 1:6:2 AM 11. Iljima, T, Takeno, T. (1986). Effects of temperatre and pressre on brning velocity. Combstion and Flame, 6(pp-). 12. Woschni, G., A niversally applicable eqation for the instantaneos heat transfer coefficient in the in the internal combstion engine, presented at SAE World Congress, 6791, Shamekhi, A.H., Simlation and Fzzy Spark Advance Control in SI engines by Ion Crrent Sensing, PhD Thesis, Department of Mechanical Engineering, K. N. Toosi University of Technology, September, Li, D.D.S, MacFarlane, R. (198). Laminar brning Velocities of hydrogen-air and hydrogen-air-steam Flames. Combstion and Flame, 9(9-71). 1. Milton, B, Keck, J. (198). Laminar brning velocities in stoichiometric hydrogen and hydrogenhydrocarbon gas mixtres. Combstion and Flame, 8(pp1-22). 16. Koroll, G.W., kmar, R. K., Bowels, E.M. (198). Brning Velocities of hydrogen-air mixtres. Combstion and Flame, 9(pp-). 17. Taylor, S.C, Brning velocity and the inflence of flame stretch, PhD thesis, Leeds University, Tse, S.D, Zh, D.L, Law, C.K., Morphology and brning rates of expanding spherical flames in H2/O2/inert mixtre p to 6 atmospheres, presented at 28 th Symp. (Int.) On Combstion, pages , Ogami, Y., Kobayashi, H., A stdy of laminar brning velocity for H2/O2/He premixed hydrogen/air flames at high pressre and high temperatre, presented at 6 th ASME-JSME Thermal Engineering Joint Conference, TED-AJ-7, Verhelst, S., A stdy of the combstion in hydrogenfelled internal combstion engines, PhD thesis, Ghent University, Gent, Belgim, Hall MJ, Bracco FV., A stdy of velocities and trblence intensities measred in firing and motored engines, presented at SAE World Congress, 87, Senecal, P.K., Xin, J., Reitz, R.D., Prediction of Residal Gas Fraction in IC Engines, presented at SAE World Congress, 9622, ADDITIONAL SOURCES 1. RAMOS, J.I., Internal combstion engine Modeling, Hemisphere Pblishing Corporation, Merker, Gnter P, Schwarz, Christian, Stiesch, Gnnar, Otto, Frank, Simlating Combstion, Springer, 2.

9 DEFINITIONS, ACRONYMS, ABBREVIATIONS Licensed from the SAE Digital Library Copyright 29 SAE International Downloaded Monday, May 18, 29 1:6:2 AM SA: Spark advance A/F: Air to fel atm: atmosphere BMEP: Brake mean effective pressre BTDC: Before top dead centre CA, θ: Crank angle CI: Compression ignition EVO: Exhast valve opening f: Residal gas volme fraction : Indicated mean effective pressre IVO: Intake valve opening MBT: Maximm brake torqe : Oxides of nitrogen P: Pressre ppm: Parts per million RON: Research octane nmber SI: Spark ignition T: Temperatre TDC: Top dead centre : Root mean sqare trblent velocity U P : Mean piston speed α T : Temperatre exponent β P : Pressre exponent γ: Residal gas coefficient φ : Fel to air eqivalence ratio VVT: Variable valve timing Sbscripts: : Reference condition t: trblent : Unbrned rpm: revoltion per minte