Int.J.

Applied Thermodynamics, ISSN 1301-9724

Vol.5 (No.4), pp.169-175, December-2002

Lean-Burn Cogeneration Biogas Engine with Unscavenged Combustion

Prechamber: Comparison with Natural Gas

Anne ROUBAUD*, Roger RÖTHLISBERGER and Daniel FAVRAT

Laboratory of Industrial Energy Systems (LENI)
Swiss Federal Institute of Technology (EPFL)
CH-1015 Lausanne-Switzerland
Tel: +41 (0) 21 693 59 68, Fax: +41 (0) 21 693 35 02
E-mail: [email protected]

Abstract
Gaseous fuels produced, for example, by waste or agricultural by-products fermentation
(biogas) can be burned in-situ by cogeneration systems like spark-ignition internal
combustion engines. However, the more and more stringent legislation for exhaust gas
emissions requires improvement of the combustion process particularly when catalytic
after treatment is not reliable as in the case of sewage or landfill biogas. The system
proposed in this paper is the use of an unscavenged combustion prechamber instead of
direct ignition on a turbocharged 6 cylinder 150 kW gas engine. This prechamber is
used for operation with a simulated biogas (40% CO2 in natural gas). The results show
that, compared to natural gas operation for the same rated power output of 150 kW and
the same NOx emissions, the CO emissions are reduced by 15% and the HC emissions
at least by 8%. Efficiencies higher than 36% are achieved which is very promising and
the lower CO emissions give a margin to consider an increase of compression ratio.
Key words: combustion prechamber, cogeneration engine, biogas, natural gas,
efficiency, emissions

1. Introduction requirements, but they all in practice rely on a

catalytic exhaust gas after treatment for natural
As humankind becomes more and more
gas operation. The first one is based on the use of
concerned with sustainable development,
a stoichiometric mixture and a three-way
biomass represents a serious alternative to
catalyst, and another one is based on the lean
substitute part of fossil fuels in decentralized
burn mode with oxidation catalyst. Both modes
power plants. Among the available technologies,
were investigated previously (Röthlisberger et al.
cogeneration gas engines can burn most biogas
1998 and 2000). A third solution is to use
fuels such as those produced in sewage plants.
exhaust gas recirculation (EGR) and a three-way
The benefit in this case is : the CO2 balance is
catalyst (Nellen and Boulouchos 2000).
usually considered as neutral and this system
valorizes urban waste or agricultural by- However, solutions relying on a catalytic
products. Gas engines have proved able to after-treatment are not adapted to biogas since it
achieve an efficient conversion of the primary contains significant quantities of sulphur and
energy in numerous applications with natural heavy metals. Those compounds can form a
gas. However, emissions can be a limitation. deposit on the combustion chamber walls or in
With a limit of 250 mg/Nm3 (400 in the case of the catalyst and then will rapidly deactivate it. A
biogas operation) for NOX and 650 mg/Nm3 for catalytic after-treatment is therefore not reliable
CO (concentrations referred to combustion gases in the case of sewage or landfill biogas.
with 5% residual O2 (Le Conseil Fédéral Suisse Upstream filtration technologies are being
1999), Switzerland has the most stringent developed but at present they are cumbersome,
exhaust gas emissions legislation in Europe for unreliable and costly. The literature is rather
stationary combustion engines. Several engine limited when it comes to the analysis of the
operating modes enable the attaining of these influence of the biogas composition on the
*
Author to whom all correspondence should be addressed.
Int.J. Applied Thermodynamics, Vol.5 (No.4) 169
behavior of spark ignition engines considering both efficiency and
TABLE I. MAIN ENGINE SPECIFICATIONS

Manufacturer Liebherr Nozzle orifices of the prechamber 4 orifices of 2,12 mm

Type G 926 TI Prechamber volume 4540 mm3

Number of Cylinders 6 Volumetric compression ratio 12,0
Bore 122 mm Turbocharger KKK K27 3371 OLAKB
Stroke 142 mm Ignition system Fairbanks Morse IQ 250
Conrod length 228 mm Spark plugs Bosch W6DC
Total swept volume 9.96 L Number of valves 2

emissions (Henham 1998, Bucksch et al. 1999, from the combustion create strong gas jets in the
Muller 1995, Stone et al. 1993 and Huang et al. main chamber. Those jets ignite the main
1998). The main way so far to reduce emissions cylinder charge at multiple locations. The main
is to decrease the compression ratio, which engine specifications are given in TABLE I.
results in a lower efficiency than normally
achievable. Therefore, new solutions are to be
found at the level of the combustion process to
increase efficiency. Prechamber ignition is the
alternative pursued in this work. This
combustion mode has been proved to reduce the
CO and unburned hydrocarbons emissions
respectively by 40, 55% and NOx emissions
below 250 mg/Nm3 with only a slight reduction
in fuel conversion efficiency for natural gas
operation of a turbocharged 6 cylinder 150 kW
spark-ignited gas engine (Röthlisberger 2003a,
2003b, 2002a, and 2002b).
2. Experimental Setup
The engine used is a 6-cylinder in-line
Liebherr heavy-duty diesel engine type D 926 TI Figure 1. Combustion prechamber
converted for gas fuel operation with spark integrated in the cylinder head
ignition. The engine is equipped with specially
The test bed is equipped with analyzers of
modified cylinder heads and piston geometry to
O2 (paramagnetism), CO2 (IR), CO (IR), NOx
reach a volumetric compression ratio of 12.0.
(CLD) and HC (FID). The engine is fully
The piston geometry has specially been designed
instrumented to measure all needed flows,
to enhance turbulence in the combustion
temperatures, pressures, combustion air
chamber (Nellen at al. 2000). The engine is
humidity, torque and rotation speed through the
turbocharged and intercooled. Cylinder liners
dynamometer to control precisely the
allow reduction of the crevice volume at the level
experimental conditions. The engine is also
of the cylinder head gasket and hence CO and
equipped for pressure indication in the
HC exhaust gas emissions.
prechamber and the main chamber of cylinder 1.
The cylinder head is fitted with small
water-cooled combustion prechambers having a TABLE II. NATURAL GAS COMPOSITION
volume corresponding to 3% of the cylinder Volumetric %
compression volume (Figure 1). Those Nitrogen 2.036
prechambers have been optimized during the CO2 0.739
PhD thesis of R. Röthlisberger et al. (2001). Methane 91.799
The prechambers used are unscavenged. At Ethane 3.983
exhaust valve closure the prechamber is filled Propane 1.009
with burnt gases. During compression, a fresh i-butane 0.159
fuel air blend enters the prechamber and mixes n-butane 0.175
with the residual gases. At spark timing the i-pentane 0.038
prechamber mixture is ignited and the hot gases n-pentane 0.032
170 Int.J. Applied Thermodynamics, Vol.5 (No.4)
 HC with more than 6 carbons 0.031 timing was held constant at 8 CABTDC for
This testing configuration has been fully prechamber operation and 27 CABTDC for the
described in Röthlisberger et al. (2001). The test direct ignition mode.
bed fuel and air alimentation was modified to use The relative air to fuel ratio was
a natural gas/CO2 blend as fuel. As CO2 is stored 1.81<λ<1.84 for direct ignition and 1.64<λ<1.68
in liquefied form, its pressure is reduced after for prechamber ignition. The pressure traces (
preheating only and the flow is controlled by a Figure 2) show how the two ignition modes
mass flow controller before mixing with the influence the combustion.
combustion air. For the tests reported here, the Due to the important delay in spark timing
CO2 volumetric flow is adjusted to 40% of the the peak pressure is lower with prechamber
total CO2+natural gas flow. The composition of ignition. The first pressure peak corresponds to
the natural gas used in all the experiments
the end of the compression and the second one
reported here is given in TABLE II. A
results from the combustion in the cylinder after
compressed natural gas storage, which was large
enough to keep the same natural gas composition ignition by the hot gas jets issuing from the
for all tests presented here, is used,. prechamber. An earlier and much higher
combustion induced pressure peak is observed
3. Test Conditions with direct ignition.
The main experimental conditions are This difference can also be observed on the
summarized in TABLE III. The engine speed is heat release rate cycle (Figure 3) where a sharper
set to 1500 rpm, which corresponds to raise of the heat release is observed with the
cogeneration engine specifications, and the prechamber with a higher rate of combustion.
engine rated brake power output is 150 kW.
Exhaust gas emissions are expressed in 90
direct ignition
mg/Nm3 at normal conditions and corrected for 80
prechamber ignition
70 theoretical unfired
0% humidity and 5% residual oxygen (Swiss
60
standard). The relative air to fuel ratio is
P(bar)

50
calculated on the basis of the measured fuel and 40
air mass flows. 30
20

TABLE III. GENERAL EXPERIMENTAL 10

CONDITIONS 0
-100 -50 0 50 100
Crankshaft rotation speed 1500 ± 5 rpm CA

Figure 2. Comparison between direct

Rated brake mean effective 12 ± 0.1 bar ignition and prechamber ignition: main chamber
pressure pressure cycle for natural gas.
Rated brake power output 150 ± 1.3 kW
160
Intake air pressure 960 ± 5 mbar 140
direct ignition
prechamber ignition
120
Intake air temperature 23 ± 2°C
dQ/dφ kJ/Nm3

100

Intake air relative humidity 50 ± 0.5 % 80

60
Exhaust gas pressure 1050 ± 5 mbar 40

after turbocharger 20
0
CO2 percentage in fuel 40 ± 0.5 % -20
-30 -10 10 30 50 70
°V
4. Results
Figure 3. Comparison between direct
The objectives of this work are to compare ignition and prechamber ignition: main chamber
the benefits, in terms of engine efficiency, heat release rate cycle for natural gas.
stability and exhaust gas emissions, of using a
prechamber with either synthetic biogas or This more rapid combustion leads to a more
natural gas. stable operation and then to a lower coefficient
of variance of pmi (Figure 4). But the late spark
4.1. Natural gas performances timing ignition in the prechamber shifts the
A new set of tests with natural gas has been combustion process in the expansion phase.
made both with direct ignition and prechamber Following this, the maximum peak pressure is
ignition. For these preliminary tests, the spark

Int.J. Applied Thermodynamics, Vol.5 (No.4) 171

lower but all the heat released produces a (Figure 7) for the same NOx emissions. On the
positive mechanical work. one hand, the hot gas jets issuing from the
5 prechamber to ignite the main combustion
prechamber ignition
4.5
Swiss limit direct ignition
chamber create a more homogeneous and a faster
4 combustion so that partial oxidation has less time
3.5
to occur. On the other hand, less fuel air mixture
COVpmi in %

3
2.5 is going into crevices volumes thanks to a lower
2 pressure in the cylinder.
1.5
1 4.2. Synthetic biogas performances
0.5 compared to natural gas
0
0 100 200 300 400 The first set of experiments made with
NOx mg/Nm3, 5%O2, O%RH synthetic biogas was to see the lower limit in
Figure 4. Coefficient of variance of pmi in stability and in NOx emissions with this fuel.
% for pure natural gas The spark timing in the prechamber was 8
CABTDC and 1.54<λ<1.55. The second set of
0.5 experiments objective was to observe the
0.48 prechamber ignition
direct ignition
influence of the spark timing on the
0.46
0.44 performances from 8 to 13 CABTDC.
0.42
0.4
Of

70
0.38 biogas
0.36 60 natural gas
0.34 theoretical unfired
50
0.32
40
P(bar)

0.3
0 100 200 300 400
30
NOx mg/Nm3, 5%O2,0% RH
20
Figure 5. Fuel conversion efficiency for
natural gas 10

0
-100 -50 0 50 100
1400
prechamber ignition CA
CO mg/Nm3, 5%O2, 0% RH

1200 direct ignition

Figure 8. Comparison between natural gas
1000
and synthetic biogas pressure cycle
800

600
180
biogas
400 160
natural gas
140
200
120
dQ/df kJ/Nm3

0 100
0 100 200 300 400
80
NOx mg/Nm3, 5%O2, 0% RH
60
Figure 6. CO emissions for natural gas 40
20
0
2500
-20
prechamber ignition
-30 -20 -10 0 10 20 30 40 50 60 70 80
HC mg/Nm3, 5% O2, 0% RH

direct ignition
2000 CA

1500 Figure 9. Comparison between natural gas

and synthetic biogas heat release rate cycles
1000
The main cylinder pressure trace (Figure 8)
500 shows that the operation with synthetic biogas is
0
similar to the one with natural gas but with a
0 100 200 300 400 lower peak pressure due to a higher heat capacity
NOx mg/Nm3, 5% O2, 0% RH of the gas mixture with the CO2. The heat release
Figure 7. HC emissions for natural gas rate is very similar to the one for natural gas
(Figure 9). The prechamber seems to operate in
The balance of those two phenomena
the same way with synthetic biogas as for natural
explains the lower fuel conversion efficiency of
gas.
2% (Figure 5). The prechamber allows a
reduction of at least 40% of the CO emissions
(Figure 6) and 55% for the HC emissions
172 Int.J. Applied Thermodynamics, Vol.5 (No.4)
 As the combustion process is the same, the 2000

biogas ST -8 CA
engine stability in terms of variance coefficient is 1800

natural gas ST -8 CA
similar to the natural gas operation (Figure 10). 1600

HC mg/ Nm3, 5%O2, 0% RH

5 biogas with variable
1400
biogas ST -8 CA ST advance
natural gas ST -8 CA 1200
4 biogas with variable ST advance
1000
COV pmi %

3 800
ST=- 8 CA
ST=- 8 CA 600 ST=- 13 CA
2
400
0 100 200 300 400 500
1 ST=- 13 CA NOx mg /Nm3, 5% O2, 0% RH

0
Figure 13. Comparison between natural
0 100 200 300 400 500 gas and synthetic biogas, HC emissions
mg NOx/Nm3, 5% O2, 0% RH
With those conditions (ST 8 CABTDC) the
Figure 10. Comparison between natural CO and NOx emissions always fulfill the
gas and synthetic biogas, coefficient of variance legislation requirements but with a reduction in
of pmi efficiency of 1%, 0.355<ηf<0.366. One way of
recovering this difference in efficiency is to
0.4
increase the spark-timing advance (Figure 11).
ST= -13 CA
0.38
4.3 Influence of spark timing advance

0.36 80
ηf

-8 CA
ST= -8 CA Biogas ST -8 ca 70
0.34 -9,5 CA
60 -10,5 CA
natural gas ST -8 ca
-11,5 CA
0.32 50
biogas with variable -13 CA
P(bar)

ST advance 40
0.3
0 100 200 300 400 500 30
mg NOx/Nm3, 5% O2, 0% RH
20
Figure 11. Comparison between natural 10
gas and synthetic biogas, fuel conversion 0
efficiency -100 -50 0 50 100
CA

1000
biogas ST -8 CA
Figure 14. Influence of spark timing on
main cylinder pressure cycle
CO mg/Nm3, 5%O2, 0% RH

900
natural gas ST -8CA

800 biogas with variable When the spark timing advance increases,
ST advance
700
the maximum pressure of the cycle is higher
(Figure 14) and comes sooner but the entire
600
ST=-8 CA
work is still during the expansion phase. The
ST=-13 CA
500 higher peak pressure induces a higher
400
combustion temperature in the main combustion
0 100 200 300 400 500 chamber and also promotes the NOx thermal
NOx mg /Nm3, 5%O2, 0% RH
formation (Figure 15). Due to this higher
Figure 12. Comparison between natural pressure, more fresh mixture is trapped in the
gas and synthetic biogas, CO emissions crevice volume during the early stage of the
Due to the lower peak pressure, the fuel combustion process. This leads to an increase of
conversion efficiency is 1% lower at constant the CO emissions throw secondary oxidation.
spark timing (Figure 11). But on the other hand Moreover, the fresh fuel-air mixture has less time
the CO emissions (Figure 12) and the HC to penetrate in the prechamber and the hot
emissions (Figure 13) are 15% and 8% lower combustion gas jets from the prechamber are
respectively. weaker. Following this, the jets are less efficient
to oxidize hydrocarbons trapped in crevices.With
an earlier spark timing, the COV of pmi is
decreasing (Figure 10), giving a good general
stability. As the stability of the end of
combustion in the expansion phase is better, the
HC emissions are lower.

Int.J. Applied Thermodynamics, Vol.5 (No.4) 173

 emissions mg/Nm3, 5%O2, 0% 800 perspectives for prechamber ignition as a way to
700 a cleaner and more rational use of biogas without
600 any major economic burden.
500
Nomenclature
RH

400
300 CA crank angle
200 NOx CLD Chemi-luminescence Detector
CO
100 THC
CO carbon monoxide
0 CO2 carbon dioxide
-14 -12 -10 -8 -6 -4 COV coefficient of variance
ca
FID Flame Ionisation Detector
Figure 15. Influence of spark timing on the HC Hydrocarbons
engine exhaust gas emissions IR Infra-Red Detector
The positive aspect of a larger ignition NOx nitrogen oxides
advance is a higher fuel conversion efficiency Nm3 normal m3 (1,013 bar and 273.15 K)
(Figure 16) due to a higher peak pressure. To O2 oxygen
maintain a value of ηf higher than 0.365, one has Q heat
to use a spark timing between 10.5 and 12 RH relative humidity
CABTDC and the emission limit for NOx and CO ST spark timing
will still be fulfilled. pmi indicated mean effective pressure
rpm rotations per minute
0.385
λ relative air to fuel ratio
0.38
ηf fuel conversion efficiency
0.375 BTDC Before Top Dead Center
0.37
Acknowledgements
ηf

0.365
0.36
This work is financially supported by the
0.355
0.35
Swiss Federal Office of Energy, in collaboration
0.345 with the engine manufacturer Liebherr Machines
-14 -12 -10
CA
-8 -6 -4 Bulle S.A. and the cogeneration group
manufacturer DIMAG S.A.
Figure 16. Influence of spark timing on the
engine efficiency References
Bucksch, S., Egebäck K.E., 1999, “The Swedish
5. Conclusion
program for investigations concerning biofuels”,
The results presented in this paper indicate The science of the total environment, 235, 1999,
that the cogeneration engine fitted with pp. 293-303.
unscavenged combustion prechamber operates
Henham, M. K., 1998, “Assessment of a
well with a synthetic biogas fuel as the advantage
simulated biogas as fuel for spark ignition
of prechamber ignition compared to direct
ignition is kept. The use of simulated biogas for engine”, Energy Conversion & Management,
the same rated brake power output of 150 kW Vol. 39, 1998, pp. 293-303.
and the same NOx emissions, reduces the CO Huang, J., Crookes R. J., 1998, “Spark ignition
emissions by 15%, and the HC emissions by 8%. performance with simulated biogas – a
The fuel conversion efficiency varies between comparison with gasoline and natural gas”,
0.36 and 0.37 (with the given volumetric Journal of Inst. of Energy, Vol. 71, 1998, pp.
compression ratio of 12). The combustion 197-203.
process is essentially unchanged. The velocity
and the heat release rate is the same but the peak Le Conseil Fédéral Suisse 1999, “Ordonnance
pressure is lower due to a higher heat capacity of sur la protection de l'air (OPair) du 16 décembre
the gas mixture. With a spark timing of 8 1985” (Etat au 12 octobre 1999).
CABTDC, the emissions of CO and NOx are well Muller, G. P., 1995, “Landfill gas application
below the Swiss limits. Increased efficiencies in development of the Caterpillar G3600 spark
a range of 0.365 to 0.375 for an ST varying ignition engine”, Journal of Engineering for Gas
between 10.5 and 12 CABTDC can be achieved Turbine and Power, Vol. 117, 1995, pp. 820-
while remaining within the Swiss limits but with 825.
an increase in CO and NOx emissions. Another
solution could be to increase the volumetric Nellen, C., and Boulouchos, K., 2000, “Natural
compression ratio. These results open significant gas engines for cogeneration: highest efficiency
and near zero emissions through turbocharging,
174 Int.J. Applied Thermodynamics, Vol.5 (No.4)
EGR and 3-way catalytic converter”, SAE geometrical parameters”, Applied Thermal
technical paper, 2000-01-2825, 2000 Engineering, Vol. 22, 2002, pp. 1217-1229.
Röthlisberger, R. P., Leyland, G., Favrat, D., and Röthlisberger, R. P., Favrat, D., 2002b
Raine, R. R., 1998, “Study of a Small Size “Comparison between direct and indirect
Cogeneration Gas Engine in Stoichiometric and (prechamber) spark ignition in the case of a
Lean Burn Modes: Experimentation and cogeneration natural gas engine, part. II: engine
Simulation”, SAE Paper, SP-1391, 982451, 1998 operating parameters and turbocharger
characteristics”, Applied Thermal Engineering,
Röthlisberger, R. P., Raine, R. R., Kleemann, R., Vol. 22, 2002, pp. 1231-1243.
and Favrat D., 2000, “Experimental Results and
Modelling of Carbon Monoxide Emissions from Röthlisberger, R. P., Favrat, D., 2003a,
a Natural Gas Fuelled Spark-Ignition “Investigation of the prechamber geometrical
Cogeneration Engine”, ImechE Int. Conf. on configuration of a natural gas spark ignition
Computational and Experimental Methods in engine for cogeneration; part I, numerical
Reciprocating Engines, IMechE Conference simulation”, accepted for publication in the Int. J.
Th. Sc., 2003.
Transactions, ISBN 1-86058-275-3, pp. 127-138,
2000. Röthlisberger, R. P., Favrat, D., 2003b
Röthlisberger, R. P., 2001, “An Experimental “Investigation of the prechamber geometrical
Investigation of a Lean Burn Natural Gas configuration of a natural gas spark ignition
Prechamber Spark Ignition Engine for engine for cogeneration; part II,
Cogeneration”, Swiss Federal Institute of experimentation”, accepted for publication in the
Technology of Lausanne, Thesis No. 2346, 2001. Int. J. Th. Sc., 2003.
Röthlisberger, R. P., Favrat, D., 2002a, Stone, C. R., Gould, J., 1993, “Analysis of
“Comparison between direct and indirect biogas combustion in spark ignition engines by
(prechamber) spark ignition in the case of a means of experimental data and computer
cogeneration natural gas engine, part I: engine simulation”, Journal of Inst. of Energy, Vol. 66,
1993, pp. 180-187.

Int.J. Applied Thermodynamics, Vol.5 (No.4) 175