Applied Thermal Engineering 22 (2002) 12171229

www.elsevier.com/locate/apthermeng

Comparison between direct and indirect (prechamber) spark ignition in the case of a cogeneration natural gas engine, part I: engine geometrical parameters
R.P. Roethlisberger, D. Favrat
*

Laboratory for Industrial Energy Systems, Swiss Federal Institute of Technology of Lausanne, CH-1015 Lausanne, Switzerland Received 3 September 2001; accepted 2 January 2002

Abstract The operation of a cogeneration internal combustion engine with unscavenged prechamber ignition was investigated. The objective was to evaluate the potential to reduce the exhaust gas emissions, particularly the CO emissions without exhaust gas after treatment. The investigation was carried out on a small size gas engine (150 kW) and required the development of cooled prechambers and the modication of the engine cylinder heads. The limit of the conventional lean burn operating mode with direct ignition is discussed and the prechamber geometrical conguration is presented. Through the generation of gas jets in the main chamber, the use of a prechamber strongly intensies and accelerates the combustion process. The prechamber operation reduces signicantly the emissions of CO and total hydrocarbon (THC) for same NOx emissions. The use of a piston generating signicantly more turbulence leads to a somewhat higher fuel conversion eciency without aecting signicantly the CO and THC emissions at low NOx emissions. Further improvement associated with the adjustment of the engine operating parameters and the turbocharger characteristics, as well as a comparison between direct and prechamber ignition operation are presented in the second part (II) of this publication. 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Gas engine; Spark ignition; Prechamber; Natural gas; Cogeneration; Emissions; Experimentation

Corresponding author. E-mail address: daniel.favrat@ep.ch (D. Favrat). URL: http://leniwww.ep.ch.

1359-4311/02/$ - see front matter 2002 Elsevier Science Ltd. All rights reserved. PII: S 1 3 5 9 - 4 3 1 1 ( 0 2 ) 0 0 0 4 0 - 6

1218

R.P. Roethlisberger, D. Favrat / Applied Thermal Engineering 22 (2002) 12171229

Nomenclature Symbols total nozzle orice area An COV coecient of variance number of nozzle orices Nn p pressure indicated mean eective pressure pmi Q heat RH relative humidity prechamber volume Vp Greek an 0 Dh Dp Du  gf k nozzle orice mean angle volumetric heating value pressure dierence crank angle interval volumetric compression ratio fuel conversion eciency relative air to fuel ratio

Subscripts ABDC after bottom dead centre ATDC after top dead centre BBDC before bottom dead centre BTDC before top dead centre cd combustion duration hr heat released id ignition delay L lower mixt fuelair mixture N p 1:013 bar, T 273:15 K 0 p 1:013 bar, T 298:15 K Abbreviations CA crank angle ST spark timing TDC top dead centre

1. Introduction Through numerous applications in Europe and in the United States, the use of small size spark ignition internal combustion engines operating on natural gas or biogas for decentralised power

R.P. Roethlisberger, D. Favrat / Applied Thermal Engineering 22 (2002) 12171229

1219

and heat generation (cogeneration) has proven to be an eective means of simultaneously achieving an ecient and clean conversion of the primary energy. Further, the high hydrogen to carbon ratio of natural gas reduces the amount of CO2 generated per unit of energy produced. In the case of biogas, the closed carbon cycle does not aect the CO2 balance. Therefore, cogeneration with gas engines directly contributes to save energy and to preserve the global environment. However, more and more stringent exhaust gas emission legislation requires continuous improvement of the combustion process. With a limit of 250 mg/m3 N , 5% O2 (400 in the case of biogas operation) for NOx and 650 mg/ 3 mN , 5% O2 for CO [1], Switzerland has the most stringent requirement for exhaust gas emissions of stationary combustion engines in Europe. There are several engine operating modes able to full these requirements, but the two used in practice rely on a catalytic exhaust gas after treatment. The rst one is based on the use of a stoichiometric mixture and a three way catalyst. The second approach is based on the lean burn mode with oxidation catalyst. Both modes were investigated in an earlier study [3,4]. The use of a catalytic converter adds extra costs and increases the complexity of cogeneration systems, which tends to increase their maintenance costs and reduce their reliability. In the particular case of engines operating on biogas produced from sewage or landll or alternatively on biogas and natural gas, a catalyst cannot be used for exhaust gas after treatment. Sewage and landll gases contain amounts of sulfur and heavy metals, which are sucient to rapidly deactivate the catalyst. In consequence, it appears that in order to achieve high reliability and to enable alternate operation on natural gas and biogas, it is necessary to reduce the exhaust emissions below the Swiss limits for natural gas operation at the level of the combustion process. This motivates the investigation of the use of an unscavenged prechamber for application on a small size cogeneration gas engine in order to evaluate the potential to reduce simultaneously the NOx and CO, while keeping the fuel conversion eciency at a high level. The main objective of this publication is to present a comparison between direct and prechamber ignition. It has been split into two parts. The rst part, forming the present paper, describes the engine specications, the testing facilities and the experimental conditions. Further, it shows the limit of the conventional lean burn operating mode with direct ignition and discusses the prechamber geometrical conguration. Finally, it illustrates the inuence of the piston geometry in the case of prechamber ignition. The second part (II [2]) shows the inuence of the spark timing, the turbocharger characteristics and the engine load on the prechamber ignition operation. It also presents a comparison between direct and prechamber ignition.

2. Engine specications and experimental conditions 2.1. Engines specications The engine used is derived from a Liebherr heavy duty diesel engine type 926 converted for natural gas operation on the Otto principle and is intended for cogeneration applications. The main engine specications are summarised in Table 1. The engine is equipped with new cylinder heads, which enable the tting of spark plugs at the location of the diesel injectors. The combustion chamber was adapted to the spark ignition operation through a modication of the piston

1220

R.P. Roethlisberger, D. Favrat / Applied Thermal Engineering 22 (2002) 12171229

Table 1 Main engine specications Manufacturer Type Number of cyl. Bore Stroke Conrod length Total swept vol. Number of valves Firing order Liebherr G 926 TI Six in line 122 mm 142 mm 228 mm 9.96 l 2 1-5-3-6-2-4 Intake valve opening Intake valve closing Exhaust valve opening Exhaust valve closing Swirl ratio [5, Eq. (8.28)] Fuel and air mixer Turbocharger Ignition system Spark plugs 15 CABTDC 45 CAABDC 54 CABBDC 14 CAATDC 2.0 RMG 980 140/65-36-1.1 KKK K27 3371 OLAKB Fairbanks Morse IQ 250 Bosch Super W6DC

geometry and a reduction of the volumetric compression ratio. Three dierent piston geometries were tested: the original piston developed by the engine manufacturer and used as reference, called A, a piston generating a higher turbulence through an important increase of the piston front rich in a parallel project surface and developed by the Swiss Federal Institute of Technology of Zu [6], called B, both with a volumetric compression ratio of 12, and a piston with a volumetric compression ratio of 9, called C. All three pistons are represented in Fig. 1. Further, the engine is equipped with cylinder liners characterised by a reduced dead volume at the level of the cylinder head gasket to limit the carbon monoxide and hydrocarbons exhaust gas emissions [3,4,7]. The engine is turbocharged and intercooled. In order to preserve more energy for the turbocharger, the engine is equipped with an insulated exhaust manifold. More detailed information on the engine conversion is given in [8].

Fig. 1. Combustion chamber: (a) with piston B,  12:0 [6], (b) with piston C,  9:0; with the dashed outline of piston A,  12:0.

R.P. Roethlisberger, D. Favrat / Applied Thermal Engineering 22 (2002) 12171229

1221

2.2. Testing facilities The experimental investigation was carried out on a test bed especially developed for the study of gas engines (Fig. 2). The test bed is equipped with a direct natural gas supply from the network as well as from an intermediate storage tank at high pressure. The storage tank is used to keep the natural gas composition constant over several series of experiments. The natural gas was supplied from storage and the exact composition was determined by gas chromatography to enable an accurate calculation of the engine fuel conversion eciency. The composition and the lower heating value of the dierent natural gas blends used are given in Table 2. In addition to the usual pressure, temperature, ow and exhaust gas emission measurements, the cylinders 1 and 3 are instrumented for pressure indication. The pre- and main chamber of cylinder 1 as well as the main

Fig. 2. Test bed with gas engine Liebherr G 926 TI.

Table 2 Chemical composition and properties of the natural gas blends used NG1 CH4 (vol%) C2 H6 (vol%) C3 H8 (vol%) i-C4 H10 (vol%) n-C4 H10 (vol%) i-C5 H12 (vol%) n-C5 H12 (vol%) C 6 (vol%) N2 (vol%) CO2 (vol%) 0;L a (MJ/m3 ) Dh N
a

NG2 95.90 1.77 0.46 0.07 0.08 0.02 0.01 0.00 1.42 0.27 36.10

NG3 97.45 0.88 0.26 0.04 0.05 0.02 0.01 0.00 1.30 0.00 35.82

NG5 95.68 1.76 0.50 0.08 0.09 0.02 0.02 0.00 1.54 0.31 36.08

NG7 95.78 1.95 0.49 0.08 0.07 0.02 0.02 0.00 1.34 0.25 36.20

92.17 3.48 0.86 0.13 0.15 0.03 0.00 0.05 2.35 0.79 36.44

C 6 considered as n-C6 H14 .

1222

R.P. Roethlisberger, D. Favrat / Applied Thermal Engineering 22 (2002) 12171229

Table 3 General experimental conditions Crankshaft rotational speed Maximum brake mean eective pressure Maximum brake power output Engine room temperature Intake air pressure Intake air temperature Intake air relative humiditya Natural gas pressureb Natural gas temperature Fuelair mixture temperature after intercooler Exhaust gas pressure after turbocharger Cooling water inlet temperature Cooling water mass ow Lubricating oil inlet temperature Lubricating oil mass ow
a b

1500 5 rpm 12 0:1 bar 150 1:3 kW 2936 C 980 5 mbar 25 0:5 C 50 0:5% 980 5 mbar 1427 C 90 1 C 1050 5 mbar 80 0:5 C 2:45 0:03 kg/s 90 0:5 C 1:21 0:01 kg/s

When not otherwise specied. Through zero pressure regulator controlled by intake air pressure.

chamber of cylinder 3 are equipped with high pressure transducers. Further, the intake and exhaust ports of cylinder 1 are tted with low pressure transducers. The angular resolution is 0.5 and the presented results are averaged over 100 engine cycles. The testing facilities and the complete instrumentation have already been described thoroughly elsewhere [3,8]. 2.3. Experimental conditions The experimental conditions are summarised in Table 3. The emission concentration is expressed in mg/m3 N at normal conditions (1013 mbar, 0 C) and corrected for humidity to 0% RH (dry) and an oxygen (O2 ) residual concentration of 5%, corresponding to the Swiss standard. NOx and total hydrocarbon (THC) are expressed in equivalent NO2 and CH4 , respectively. The ignition delay (Duid ) is dened as the dierence between the average angular positions where 5% of the heat has been released (start of combustion) and of the spark timing. The combustion duration (Ducd ) is dened as the dierence between the average angular positions where 90% (end of combustion) and where 5% of the heat has been released. The emissions of NOx , CO, THC, O2 and CO2 were measured and the relative air to fuel ratio is calculated on the basis of the exhaust gas composition. 3. Results and discussion 3.1. Direct ignition operation This section reviews the principal results obtained with direct spark ignition in early phases of the gas engine study [3,4,7,8] and illustrates the inuence of the main engine parameters aecting the principal exhaust gas emissions.

R.P. Roethlisberger, D. Favrat / Applied Thermal Engineering 22 (2002) 12171229

1223

The experimental investigations carried out on the basis of the original combustion chamber geometry, formed by piston A (Fig. 1), have shown that the combustion chamber crevice volume located between the cylinder liner, the cylinder head and the cylinder gasket represents a major source of CO and THC emissions. This nding led to a modication of the cylinder liner and gasket to reduce as much as possible this crevice volume. The emissions and performance after modication are represented in Fig. 3. The results are generally represented as a function of NOx emissions in order to clearly illustrate the eect of the dierent combustion chamber congurations on this determining pollutant. The evaluation of piston B (Fig. 1a) has shown that the higher turbulence level generated leads to a shortening of the combustion process (Fig. 3a) and an improvement of the combustion stability, expressed by the coecient of variance of the indicated mean eective pressure (COVpmi ) (Fig. 3b), which results in an important reduction of the THC emissions (Fig. 3c). However, the CO emissions are not signicantly aected. The faster combustion process yields, at constant NOx emissions, higher fuel conversion eciency (Fig. 3b). The lower methane concentration of the natural gas blend used for the test of piston B (NG1: 92.2%) compared to the blend used for piston A (NG2: 95.9%) accounts for a part of the combustion acceleration [9].

Fig. 3. Direct ignition, performance and emissions of the combustion chambers A and B,  12:0, and C,  9:0, as function of NOx emissions: mean heat-release rate and integral (a), fuel conversion eciency and coecient of variance of pmi (b), CO and THC emissions (c) and mean pressure cycle (d); ST 26:7 CABTDC , Tmixt 90 50 C, RH 30 45%.

1224

R.P. Roethlisberger, D. Favrat / Applied Thermal Engineering 22 (2002) 12171229

However, the eect of fuel composition is somewhat attenuated by the lower fuelair mixture temperature used with piston B. The much better combustion stability enables a further increase in relative air to fuel, which results in a further reduction of NOx emissions, far below the Swiss limit. However, the satisfaction of the Swiss limit for CO requires the use of an oxidation catalyst. A complementary study of the inuence of specic engine parameters was undertaken in order to evaluate the potential to reduce the relatively high CO emissions. The parameters investigated were the volumetric compression ratio, the intake and exhaust valve overlap duration and the location of the rst piston compression ring. The results are presented and discussed in detail in [8]. This complementary study indicates that the volumetric compression ratio represents the major parameter inuencing the engine exhaust emissions. Its reduction results in a signicant decrease of the CO emissions. This is illustrated by the experimental results obtained with piston C designed to reduce the compression ratio to 9.0 and to increase as much as possible the squish surface in order to promote the generation of turbulence (Fig. 1b). The decrease of cylinder pressure resulting from the reduction of volumetric compression ratio from 12.0 to 9.0 (Fig. 3d) leads to a slowing down of the combustion process (Fig. 3a), which yields an important decrease of fuel conversion eciency (Fig. 3b). In comparison with piston A, the decrease in peak combustion temperature resulting from the reduction of peak cylinder pressure enables a further reduction of the NOx emissions and a decrease in relative air to fuel ratio, which improves the combustion stability. The much lower cylinder pressure reduces the amount of unburned mixture stored in the combustion chamber crevices at the end of the primary combustion process and therefore the quantity of CO emissions generated through hydrocarbon partial secondary oxidation. On the other hand, the slower combustion process results in an increase of the average temperature during expansion, which promotes a more complete oxidation of the unburnt hydrocarbons emerging from the crevices. In comparison to piston A, both of these eects result in a simultaneous and important reduction of CO and THC emissions (Fig. 3c). From the results presented, two distinct methods to satisfy the requirements of the Swiss Clean Air Act are emerging. The rst method, only applicable to natural gas, consists in using a high turbulence combustion chamber to maximise the fuel conversion eciency and reduce the NOx emissions, while abating the CO emissions by exhaust gas after treatment with an oxidation catalyst. The second method, destined for biogas, is based on the reduction of the volumetric compression ratio to decrease the raw exhaust emissions under the Swiss limit. This approach, however, leads to an important reduction of the fuel conversion eciency. These two methods clearly illustrate the actual limits of the lean burn operation with direct spark ignition. The reduction of the CO emissions without catalytic treatment, in order to decrease the investment and maintenance cost and to improve the engine reliability to satisfy the emissions regulation, leads to an increase of the operating cost through an augmentation of the specic fuel consumption. The simultaneous requirement of low exhaust gas emissions, high primary energy conversion eciency and reliable engine operation (therefore without exhaust gas after treatment) motivated the search for a new degree of freedom related to the combustion chamber conguration. This was realised through the addition of a small combustion prechamber and the transfer of the ignition location into this prechamber.

R.P. Roethlisberger, D. Favrat / Applied Thermal Engineering 22 (2002) 12171229

1225

3.2. Prechamber ignition operation The inuence of the main combustion chamber geometry is presented and discussed for a prechamber geometrical conguration close to optimum. 3.2.1. Prechamber design A small water cooled prechamber, representing about 3% of the main chamber compression volume, was designed to be housed in the boss of the existing cylinder head used for the spark plug well. The prechamber integrated to the modied cylinder head is represented in Fig. 4a. It was designed on a modular basis formed by three parts (nose, body and xing clamp), in order to limit the machining work for the experimental parametric study. A picture of the whole prechamber is given in Fig. 4b and a detail of the body with spark plug electrodes and positioning pin is showed in Fig. 4c. The part of the prechamber nose protruding in the main chamber can be seen in Fig. 4d. For later industrial application, the three dierent parts could be welded together. The limited space in the prechamber body (limited by the size of the original cylinder head boss) required the modication of the spark plug tightening hexagon. In order to limit the prechamber and spark plug operating temperature, part of the engine water owing in the cylinder head is used to cool the nose and the bottom of the body. Several prechamber congurations were realised in order to perform variations of the cross sectional area, number, orientation and distribution of the nozzle orices, as well as of the prechamber internal volume and shape. The congurations tested were selected on the basis of the results of a CFD numerical simulation with KIVA-3V. The numerical and experimental results are

Fig. 4. (a) Combustion prechamber integrated in the cylinder head, (b) complete prechamber with xing clamp and copper joints, (c) detail of the prechamber body with spark plug electrodes and positioning pin, (d) detail of the prechamber nose protruding into the main combustion chamber.

1226

R.P. Roethlisberger, D. Favrat / Applied Thermal Engineering 22 (2002) 12171229

detailed in [8]. The nal conguration combining all the positive features results in a prechamber with an internal volume of 4540 mm3 (2.9% of the compression volume) and four nozzle orices of diameter 2.12 mm (An 14:10 mm2 ), non-evenly distributed and oriented at an angle of %78 from the prechamber axis. Although the selection of this combination is based on a limited number of variations performed for each geometrical parameter investigated, the nal prechamber conguration obtained should be relatively close to the global optimum. 3.2.2. Main combustion chamber geometry In order to perform a rst evaluation of the inuence of the main combustion chamber geometry on the combustion process and exhaust gas emissions, two dierent piston bowl congurations were investigated: the original (A) and the high turbulence (B) (Fig. 1a). Both pistons are characterised by a volumetric compression ratio of 12. The engine was operated at a constant spark timing of 8.3 CABTDC . The results corresponding to the two dierent piston geometries are presented together in Fig. 5. At a constant relative air to fuel ratio, the transition from piston A to B reduces slightly the intensity and duration of the prechamber pressure pulse, which is illustrated by the dierence between the prechamber and the main chamber pressure (Fig. 5c). This is expected to decrease the penetration of the gas jets into the main combustion chamber. However, the higher turbulence generated by piston B strongly intensies and accelerates the main chamber combustion process, particularly in the late stage (Fig. 5b). In turn, the faster combustion process induced by piston B causes an important increase in peak cylinder pressure (Fig. 5a). At equal NOx emissions, the signicantly shorter combustion duration characterising piston B (Fig. 5d) causes an increase in fuel conversion eciency, particularly for the highest values of relative air to fuel ratio (Fig. 5f). Further, the more rapid late stage of the combustion process reduces slightly the cycle-by-cycle variability in cylinder 1. This is indicated by a somewhat lower coecient of variance of pmi . The opposite trend observed in cylinder 3 was identied later and originates in a slight leakage of water between the prechamber cooling jacket and the main chamber of cylinder 3 due to an inadequate match of machining tolerances. Because no change was made at the level of prechamber 3 during the experimentation of the two dierent piston geometries, the results obtained remain consistent (further experiments after improvement of the water-tightness of prechamber 3 have shown a similar combustion stability in cylinder 1 and 3). When considering the same NOx emissions, the transition from piston A to B causes a moderate reduction of CO and THC emissions (Fig. 5e). Further, the dierence between both pistons progressively decreases when increasing the relative air to fuel ratio in order to reduce the NOx emissions. The dierent behaviour of the CO and THC emissions when using piston B seems to indicate a fundamental change in the interaction between the gas jets and the main chamber ows. This no longer gives a reduction in CO emissions while increasing the relative air to fuel ratio. Piston B leads to a higher peak cylinder pressure during combustion and a lower mean combustion temperature during late expansion. The former eect should increase the amount of unburnt mixture located in the combustion chamber crevices during the primary oxidation process. The latter eect tends to lower the extent of the secondary oxidation of the unburnt hydrocarbons emerging from these crevices. However, piston B tends also to promote a more rapid propagation of the ame front towards the squish region and therefore should reduce the amount of unburnt mixture owing into the piston top land and cylinder head gasket crevices. The lower

R.P. Roethlisberger, D. Favrat / Applied Thermal Engineering 22 (2002) 12171229

1227

Fig. 5. Prechamber ignition, inuence of the main combustion chamber on the engine performance and emissions; main chamber pressure (a), heat-release rate and integral (b) and pressure dierence between pre- and main chamber (c) at constant relative air to fuel ratio; ignition delay and combustion duration (d) CO and THC emissions (e) and fuel conversion eciency and coecient of variance of pmi (e) as function of NOx emissions; Vp 4540 mm3 , An 14:10 mm2 , Nn 4, an % 78, ST 8:3 CABTDC , NG7.

CO and THC emissions associated with piston B tend to demonstrate that this last eect is determinant. The increase in peak cylinder pressure induced by piston B results in a higher maximum combustion temperature. Consequently, this requires a moderate increase in relative air to fuel ratio in order to achieve the same NOx emissions (Fig. 5e). In the case of both pistons, A and B, the reduction of the NOx emissions is essentially limited by the particularly high cycle-by-cycle variability in cylinder 3.

1228

R.P. Roethlisberger, D. Favrat / Applied Thermal Engineering 22 (2002) 12171229

Compared to piston A, piston B achieves a somewhat better fuel conversion eciency. The dierence between the pistons increases with the reduction of NOx emissions to reach a maximum value of approximately 0.5% point. On the other hand, the behaviour of the CO and THC emissions with the increase of relative air to fuel ratio seems to indicate no signicant advantage of piston B over piston A at a low level of NOx emissions. Due to its capacity to achieve a higher fuel conversion eciency, piston B is used for the investigations of the turbocharger turbine and engine load.

4. Conclusions In the context of cogeneration with internal combustion engines, the operation with prechamber ignition on natural gas was investigated in order to evaluate the potential to reduce the exhaust gas emissions, particularly the CO emissions, below the Swiss limits (NOx and CO emissions: 250 and 650 mg/m3 N , 5% O2 , respectively), without exhaust gas after treatment. This required the development of cooled prechambers and the modication of the engine cylinder heads. The nal prechamber conguration combining all the positive features established through a parametric study was used to carry out variations of the piston geometry, the turbocharger characteristics and the engine operating parameters in order to determine their inuence on the engine performance and emissions. The main experimental results concerning the transition from direct to prechamber ignition and the inuence of the piston geometry are the following: In comparison with the direct ignition, the prechamber ignition strongly intensies and accelerates the combustion process; this results from the distribution of ignition sources in the main combustion chamber by means of jets of combustion products issuing from the prechamber; the penetration of the gas jets into the main chamber dramatically increases the initial ame front surface. In comparison to the original piston, the use of a piston generating signicantly more turbulence leads to an approximately 0.5%-point higher fuel conversion eciency; however, it induces no signicant change in the CO and THC emissions at low NOx emissions. The results regarding the inuence of the engine operating parameters and the turbocharger characteristics as well as a comparison between direct and prechamber ignition are presented in the second part of this publication (II, [2]).

Acknowledgements This research work was carried out at the Laboratory for Industrial Energy Systems (LENI) of the Swiss Federal Institute of Technology of Lausanne (EPFL). It was nancially supported by the Swiss Federal Oce of Energy (OFEN), under the grant no. 69801, the engine manufacturer Liebherr Machines Bulle S.A., the cogeneration group manufacturer Dimag S.A., the Research Fund of the Gas Industry (FOGA), as well as by the EPFL, who are all gratefully acknowledged.

R.P. Roethlisberger, D. Favrat / Applied Thermal Engineering 22 (2002) 12171229

1229

References
[1] Le Conseil f ed eral suisse, Ordonnance sur la protection de lair (OPair), du 16 d ecembre 1985 (Etat au 12 octobre 1999). [2] R.P. Roethlisberger, D. Favrat, Comparison between direct and indirect (prechamber) spark ignition in the case of a cogeneration natural gas engine, part II: engine operating parameters and turbocharger characteristics, Applied Thermal Engineering, 2002. [3] R.P. Roethlisberger, G. Leyland, C.-A. Paschoud, D. Favrat, Swiss Motor, Modication dun moteur Diesel pour le fonctionnement avec m elange stchiom etrique (k 1) et pauvre (k ) 1), Swiss Federal Oce of Energy, Final Report, 1998. [4] R.P. Roethlisberger, G. Leyland, D. Favrat, R.R. Raine, Study of a Small Size Cogeneration Gas Engine in Stoichiometric and Lean Burn Modes: Experimentation and Simulation, SAE Paper, 982451 and SP-1391, 1998. [5] J.B. Heywood, Internal Combustion Engine Fundamentals, McGraw-Hill International Editions, Automotive Technology Series, 1988. [6] K. Boulouchos, A. Zbiorczyk, J. Inhelder, C. Frouzakis, Optimale Auslegung umweltfreundlicher Gasmotoren mit Hilfe von Simulationswerkzeugen, MTZ Motortechnische Zeitschrift 58 (1997) 486499. [7] R.P. Roethlisberger, R.R. Raine, R. Kleemann, D. Favrat, Experimental Results and Modelling of Carbon Monoxide Emissions from a Natural Gas Fuelled Spark-Ignition Cogeneration Engine, IMechE International Conference on Computational and Experimental Methods in Reciprocating Engines, IMechE Conference Transactions, ISBN 1-86058-275-3, 2000, pp. 127138. [8] R.P. Roethlisberger, An Experimental Investigation of a Lean Burn Natural Gas Prechamber Spark Ignition Engine for Cogeneration, Swiss Federal Institute of Technology of Lausanne, thesis no. 2346, 2001. Also published under Swiss Motor, Modication dun moteur diesel pour le fonctionnement au gaz naturel en cogeneration, fonctionnement avec pr echambre de combustion, Swiss Federal Oce of Energy, Final Report (in English), 2001. [9] C. Nellen, K. Boulouchos, Natural Gas Engines for Cogeneration: Highest Eciency and Near-Zero-Emissions through Turbocharging, EGR and 3-Way Catalytic Converter, SAE Paper, 2000-01-2825, 2000.