ICSET 2008
Experimental and Modeling Aspects of Producer Gas Engine
G. Sridhar
Abstract— Biomass gasification technology offers a cost
effective option of power generation for a wide variety of
applications including distributed power generation. In most
cases a reciprocating engine forms the prime mover. The work
reported here uncovers some of the misconceptions associated
with the usage of producer gas, a lower calorific gas as a
reciprocating engine fuel. This paper particularly addresses the
use of producer gas in reciprocating engines at high
compression ratio (17:1), which hitherto had been restricted to
lower compression ratio (up to 12:1). The current work clearly
indicates to the breakdown of this compression ratio (CR)
barrier and it is shown that the engine runs smoothly at CR of
17:1 without any tendency of knock or auto-ignition. These
experiments were conducted on a spark-ignition engine
converted from a diesel engine at a CR of 17. Experiments at
varying CR (17 to 11.5) have established the benefits of
operating the engine at higher CR in terms of lower de-rating
and better efficiencies. These experiments also established the
optimum parameters, namely the Minimum Advance for Brake
Torque and the corresponding equivalence ratio for maximum
power. It was also found that the exhaust emission from the
producer gas fuelled engines meet the emission norms without
using emission control devices. Furthermore, a zero-
dimensional model has been formulated using wrinkled flame
theory for flame propagation to estimate output of producer
gas engines.
I. INTRODUCTION
With engine emissions becoming stringent, gaseous fuels are
gaining prominence as cleaner fuels both for stationary and
automotive applications. LPG and CNG are establishing as
cleaner fuels for automotive applications in developing
countries like India and elsewhere. Similarly for power
generation, natural gas is being exploited largely due to
significant availability in specific locations. Likewise, gas
generated from industrial and municipal wastes such as
biogas and land-fill gas respectively are gaining prominence
in power generation. There is one another fuel, namely,
biomass derived producer gas, which is largely left
unexploited for a variety of reasons. This fuel is most
suitable for distributed power generation systems, captive
power plants for small and medium enterprises and grid
interactive independent power producer at power levels up
to 1 – 2 MWe.
Producer gas is gaseous product resulting from the thermo-
chemical conversion route (also termed gasification) of any
biomass. This can be used for fuelling a compression
Manuscript received September 30, 2008
G Sridhar is with Siemens Corporate Technology, Bangalore, India
phone: +91-80-25113634; fax: +91-90-25113666; e-mail:
sridhar.gururajarao@siemens.com
995
978-1-4244-1888-6/08/$25.00 
c 2008 IEEE
ignition (CI) engine in dual-fuel mode or a spark-ignition
(SI) engine in gas alone mode. However, SI engine is
preferred, mostly from economical point of view. Therefore
harnessing energy from biomass via gasification route is not
only proving to be economical but also environmentally
benign. Moreover, this technology being CO2 neutral
qualifies very well for carbon emission reduction
certificates.
This paper addresses the experimental and modeling work
with respect application of biomass derived producer gas in
SI engines.
II. PRODUCER GAS FUEL
A. Background
Unlike many of the gaseous fuels, producer gas has not been
adequately addressed by the scientific community because of
a variety of reasons [1]. Producer gas, a low energy density
gas is generally perceived to be inadequate in meeting the
requirement of power generation. Most of the development
work on producer gas engine has taken reported around the
World War II; subsequently there has been sporadic interest
in Europe, America and the Indian sub-continent. If one
were to summarize the findings of earlier studies, it becomes
evident that no systematic investigation has been attempted
by the earlier researchers neither in characterizing the fuel
nor its application in engines.
B. Fuel Properties
Biomass derived producer gas would typically contain 18-
20% each of H2 and CO, 2% CH4 and, rest inert like CO2
and N2. The lower calorific value would be between 4.5 –
4.9 MJ/kg, with stoichiometric air-to-fuel ratio being 1.25 +
0.05 on mass basis. It would be appropriate to compare
properties of producer gas with methane gas because most of
the engines operating on gaseous fuels are either close to
pure methane (natural gas) or diluted methane (bio-gas,
land-fill gas). The fuel-air equivalence ratio (actual fuel to
air ratio)/(stoichiometric fuel to air ratio) at the flammability
limits compares closely for both the gases, but the laminar
burning velocity for producer gas at the lean limits is much
higher. The laminar burning velocity for producer gas (at
0.1MPa, 300K) is about 0.5 m/sec [2], which is about 30%
higher than methane, and this feature would become
important while arriving at the optimum ignition timing.
Producer gas is suitable to operate in an internal combustion
engine operation provided the gas is sufficiently clean such
that contaminant does not accumulate in the intermediary
passages to the engine cylinder. Therefore it is necessary to at a rated conditions. The intake port is of a simple design –
have a gas generator that would generate producer gas with directed type with air entering into the cylinder all around its
least amount of contaminants. Besides this being one of the periphery. The engine is designed with a valve overlap of
limitation, the fuel has largely been left unexploited due to 64° CA. The combustion chamber of the engine is formed of
certain misconceptions, namely (1) auto-ignition or a flat cylinder head and slightly offset bowl-in piston. The
knocking tendency when applied in engines of higher CR, bowl is hemispherical in shape and has a squish area of 70%
(2) large de-rating in power due to energy density being low. (percentage of piston area closely approaching the cylinder
However, these perceptions could be negated with the head). The engine was initially characterized using diesel
following viewpoints. The arguments against the classical fuel prior to conversion, which generated a net output of 24
view in favor of better knock resistivity are as follows. kW. Subsequently the above mentioned diesel engine was
Firstly, with the laminar burning velocity being high due to converted to a spark-ignition engine and the details of
the presence of hydrogen might reduce the tendency for the conversion are dealt in the earlier work [1], [4]. The engine
knock. Secondly, the presence of inert in the raw gas (CO2 was tested at varying compression ratios (CR) of 17, 14.0,
and N2) might suppress the pre-flame reactions that are 13.5 and 11.5. Load tests were conducted on the engine in
responsible for knocking on account of increased dilution. order to determine the maximum power delivered.
Also, the maximum flame temperature attainable with the Therefore, the air-to-fuel ratio and the ignition timing were
producer gas (1800 K) being lower compared to tuned in order to derive maximum output.
conventional fuels like methane (2200 K), one could expect
B. The Performance
better knock resistivity. An examination of literature clearly
indicates that producer gas has not been subjected to study 120
on knock behavior. Producer Ign Timing
o
Gas (i) 26 BTC
o
Furthermore, there is a general perception that producer gas (ii) 12 BTC
80
being a low-density energy fuel, the extent of de-rating in
power would be large when compared to high energy density
fuels like natural gas and Liquefied petroleum gas. This (i)
could be misleading because what needs to be accounted for 40
comparison is the mixture energy density and not the fuel (ii)
energy density per se. On comparison with CH4, the mixture
energy density for producer gas is lower by 23% [1], [4]. 0
Moreover, the product to reactant mole ratio for producer 340 360 380 400 420
gas is less than one. These two parameters could contribute (a)
to de-rating of engine output, to an extent of 30%. However,
it might be possible to reduce de-rating using producer gas
by working with engines of higher CR. 120
Kerosene (i) & (ii) :
Knock Random cycles
III. THE ENGINE TRIALS
80
A. Engine details
Experiments were conducted on a spark ignition (SI) engine (i)
converted from a naturally aspirated, three-cylinder, direct 40
injection diesel engine (RB 33 model) of compression ratio (ii)
(CR) 17. The reason for choosing this particular CR diesel
engine is explained as follows. It is known from
0
fundamental thermodynamics that it is beneficial to operate
an internal combustion engine at as high a CR as possible in
340 360 (b) 380 400 420

order to attain high overall efficiencies. But the gain in

efficiency beyond a certain CR can be expected to be Fig 1 (a) Normal combustion under wide open throttle condition with
marginal due to the other counter influencing factors such as producer gas at different ignition timings at CR=17, (b) Incipient knock
with kerosene at no-load condition with ignition timing of 30° BTC at
increased heat loss and friction. But in the case of a SI CR=16.5.
engine the limita- tion of CR comes from a supplementary
factor namely, ‘knock’. It is to be noted that knock
The first and the foremost result of these tests was that the
sensitivity identifies the Highest Useful Compression Ratio
engine worked smoothly without any sign of knock at the
(HUCR) for most of the fuels. However, for higher-octane
CR of 17. There was no sign of audible knock during the
fuels, it has been experimentally established that the upper
entire load range. Moreover, the absence of knock is clear
limit of CR is 17 beyond which there is a fall in efficiency
from the pressure-crank angle (p-ș), which does not show
[3]. The chosen engine for the study was a naturally
any pressure oscillations, either at part load or at full load
aspirated engine of 3.3 litre capacity is designed at a CR
(wide open throttle) conditions. A comparison of normal and
=17, to operate at an air-to-fuel ratio of 20 - 21 (with diesel)

996
 abnormal (due to knock) combustion is shown in Fig. 1. The
normal performance (without pressure oscillations) shown as
(i) and (ii) in Fig. 1 (a) corresponds to two firing cycles at
ignition advance of 26° and 12° Before Top Centre (BTC)
respectively. These correspond to operations with producer
gas at CR=17 under wide open throttle conditions. Whereas,
performance with incipient knock (with pressure
oscillations) shown as (i) and (ii) in Fig. 1 (b) corresponds to
two random firing cycles. These correspond to operations
with kerosene fuel on a single cylinder SI engine (converted
from a diesel engine) at CR=16.5 under no-load conditions.
80
o
17 CA
Black : Firing
Grey : Motoring
o
12 CA
Cylinder Pressure, bar
brake output of 20 kW (17.5 kWe) at an efficiency of 30.7%
compared to 24 kW (21 kWe) brake output at 33%
efficiency with diesel (compression ignition mode). The
efficiency calculation is based on the ratio of net brake
output to the energy content of the air and gas mixture. The
useful output and efficiency decreased with the lowering of
CR. A maximum net brake output of 17.6 kW (15.3 kWe) at
an efficiency of 27.5% was obtained at CR of 11.5. The
power output at intermediate CR of 14.5and 13.5 were 18.8
and 18.6 kW respectively and with efficiencies around 29%.
The efficiency at CR = 13.5 was comparable to that at 14.5
probably due to relatively leaner operation. The extent of de-
rating in brake power was about 16.7% at CR = 17 and
increased to as high as 26% at CR = 11.5 compared with
baseline operations in diesel mode.

60 o
22 CA
o
33 CA
o
26 CA
o
6 CA TABLE 1
40 MAXIMUM NET ENGINE OUTPUT AT VARYING CR
CR IGN, Φ Net Net Brake Efficiency :
BTC Elec. Power Gas-to-Shaft,
20
Power, (BPNet) , kW %
kWe
17.0 06 1.10 17.5 20.0 30.7
0
300 320 340 360 380 400 420
14.5 10 1.10 16.4 18.8 29.0
Crank Angle, degree
13.5 14 1.06 16.2 18.6 29.3
(a)
11.5 15, 17 1.07 15.3 17.6 27.5
60
ĭ =Equivalence Ratio: (Actual fuel- to- air ratio)/(Stoichiometric fuel- to- air ratio)
Black : Firing
Grey : Motoring
The Minimum Advance for Brake Torque (MBT) for
Cylinder Pressure, bar

40
o
27 CA o
producer gas fuel is identified to be in the following crank
16 CA
o
38 CA angle intervals. 6-10° BTC for CR=17.0, 10-12° BTC for
CR=14.5, 12-14° BTC for CR=13.5 and 15 –17° BTC for
20
o
6 CA
CR=11.5. This range is identified keeping in view the
modest variation in the net power output.

0 The emissions measured are Nitric oxide (NO) and Carbon

300 320 340 360 380 400 420

Crank Angle, degree

(b)
Fig. 2 p-ș Recording corresponding to maximum brake output at varying
ignition advance (a) CR=17. (b) CR=11.5 ensemble-averaged data over 30
consecutive cycles.
The pressure-crank angle (p-ș) recording is shown in Fig 2
CRs of 17 and 11.5, none of them showed any trace of
knock for all ranges of load inclusive of peak. It is clear
from these curves that smooth and normal combustion
seemed to occur even at advanced ignition timing of 33° CA
corresponding to CR of 17. Faster burn rate due to presence
of hydrogen in the fuel gas could be the principal factor for
the no-knock performance. The effect of the ignition
advance on the pressure history is evident from the above
curves. There are substantial differences in the maximum
cylinder pressure and their point of occurrence. The ratio of
maximum cylinder pressure between 17 and 11.5 CR at
corresponding ignition timing is about 2.
The results of the power output with producer gas are shown
in Table 1. At CR=17, the engine delivered a maximum net
Monoxide (CO). The variation of NO in gas mode at varying
CR with ignition advance is shown in Fig. 3. The NO level
has been represented as an emission index in units of gram
per unit MJ of input energy. These results are compared with
the CPCB norms (equivalent to EURO I norm) meant for
diesel engine powered vehicle, as there are no existing
CPCB norms for stationary engines. Also the data is
compared with the Swiss emission norm, as it is generally
understood that their norms are stringent. The NO level
reduced with the retardation of ignition timing and this
feature is observed for all CRs. The NO level was observed
to be maximum at the highest CR corresponding to advanced
ignition timings, whereas in the MBT range of 6 to 20° BTC
the NO was lower and comparable in almost all the cases. It
is well known that NO generation is strongly dependent on
the temperature, oxygen availability and residence time in
the combustion chamber. With the flame speed of the gas
mixture being high, the ignition setting is retarded whereby
the residence time in the high temperature combustion
chamber is automatically reduced. Therefore the low NO
levels at retarded ignition setting are an expected and
consistent behavior. The NO emissions are lower than the
CPCB and Swiss norm for all cases around MBT.

997

Time, millisecond
0 0.4 0.8 1.2 1.6 2 2.4 2.8
1.0
Euro Stage II norm : 1.67 g/MJ
CR = 17.0
CPCB (India) norm : 2.22 g/MJ
0.8 CR = 14.5
CR = 13.5
CR = 11.5
NO, g/MJ
0.6
0.4
0.2
0.0
0 5 10 15 20 25
MBT Timing
Ignition advance, oCA
Fig. 3 Variation of NO with ignition advance at various CRs
IV. THE ENGINE MODELING
To understand the in-cylinder behavior, a zero-dimensional
(0-D) modeling study has been conducted using wrinkled
flame theory for flame propagation. These are compared
with the earlier presented experimental results. The heat
release part, which is key to the modeling, is based on the
well-postulated Eddy Entrainment and Laminar Burn-up
(EELB) meant for turbulent pre-mixed wrinkled flames. The
required inputs on the laminar burning velocity and
turbulence parameters are obtained from separate studies.
The data related to laminar burning velocity for producer gas
+ air mixture at thermodynamic conditions typical of
unburned mixture in an engine cylinder was obtained from
one-dimensional flame calculations. The turbulence
parameters were obtained by conducting a 3-D CFD study
on a bowl-in-piston geometry to simulate motored or non-
firing conditions. The above mentioned data was used in the
0-D model to make pressure-time (p-θ) computation over the
complete engine cycle, for a range of test cases at varying
compression ratio (CR) and ignition timing.
The model comprises of sub-models to simulate the four
processes of reciprocating engine cycle namely intake,
compression, heat release followed by expansion and
exhaust. The various sub-models used in the above
simulation are (a) the filling and emptying technique for
intake and exhaust processes as outlined in Heywood [5], (b)
Eddy Entrainment and Laminar Burn-up (EELB) model for
simulation of heat release as derived by Keck [6] and (c) the
heat loss due to convection based on Annand’s convective
heat transfer correlations. The flame propagation (or heat
release) is modeled as a two-zone model, where a thin
wrinkled multi-connected laminar flame separates the
burned and the unburned mixture. The EELB model as
formulated by Keck [6] is represented by two equations
namely,
dmb μ
= ρu Af Sl +
dt τb
dμ μ
= ρ u A f uT − (2)
3.2 3.6 4 4.4 dt τb
where μ = ρ u A T ( AA − A f ) . Eq. 1 represents the mass
burn rate, whereas Eq. 2 represents the rate of change of
unburned mixture within the flame front. In these equations,
there are two quantities, namely the characteristic speed and
length, which could be related to the turbulence parameters
'
namely u and A T as identified in Heywood [5]. These
turbulence parameters can either be determined using
empirical correlations or from fundamental studies [5].
Keck [6] has provided empirical relationship for obtaining
characteristic speed and length by coupling with few other
30 35 40 variables related to engine geometry. But these correlations
have been formulated for simpler geometries where
turbulence is more dependent on the intake port design and
less on the fluid-piston interaction inside the cylinder.
Extending the same correlation is not appropriate for
complex geometries where the flow is known to get
substantially modified in the compression process. Therefore
a CFD study was conducted on a cylinder – piston geometry
(with bowl) simulating motoring condition for the complete
engine cycle in one of the previous work by the above
authors [8]. Turbulence parameters namely turbulence
intensity and length scales are extracted from the CFD
results and used as a 0-D model input. It was further
assumed that during combustion, the unburned mixture
undergoes isentropic compression sufficiently rapidly that
simple distortion process occurs as outlined in Heywood [5].
Laminar burning velocity is another input parameter
required for the heat release model. It was computed using a
code called ‘FLAME CODE’ [9] that uses a transient
calculation procedure to extract the properties of one
dimensional laminar premixed flame. The code has been
validated with experimental results at ambient pressure and
initial temperatures [10]. The fuel considered was of
nominal composition with 20 % each of H2 and CO, 2%
CH4, 12% CO2 and rest N2. Therefore theoretical burning
velocity calculation has been done at varying equivalence
ratio, initial temperature and pressure and with varying
amounts of recycled gas ranging from 0 to 10%. The detail
of this part of the work is referred in earlier works of the
authors [8]. An expression derived from this analysis for
fuel-air equivalence ratio between 0.9 to 1.3 is as follows:
Sl (cm/s) =
0.2744
§ p ·
94.35 ¨¨ ¸¸ (0.96 + 1.2(Φ − 1))(1 − 2.4ψ )
© p0 ¹
(1)
998
where, p is the actual pressure during combustion in bar
(abs), po the reference pressure (1.0 bar), Ȍ the external EGR
fraction and ĭ the fuel-air equivalence ratio [8].
One experimental firing case is used as a test case in tuning
the 0-D model in terms of constants required for
computation of heat loss. Subsequently, keeping these
parameters fixed 0-D predications are made for twelve test
cases at varying CRs and ignition settings. These are further
categorized under two sub-groups, namely Simple and
Complex depending upon the intricacy involved in the
predictions. Under the Simple sub-group the predictions
match very well with the experimental p-ș curve. These
bar and as a consequence the expansion pressures are lower.
However, computation at 18° CA ignition advance matched
reasonably well with the experimental data even though the
computed pressure deviates after about 356° CA. The
predicted duration of combustion is shorter at retarded
ignition setting implying faster burn rate. Similar is the case
with computations at CR=11.5 at an ignition advance of 27°
CA.
70
Dotted: Experimental
60
Black : Predictions
Grey : Motoring

Cylinder Pressure, bar

50
computations are attempted assuming simple spherical flame
propagation model. Whereas, the cylinder pressure was 40
under-predicted under Complex sub-group using spherical
flame propagation model. The Simple cases corresponded to 30

advance ignition setting, whereas the Complex cases

20
corresponded to retarded ignition setting, which is close to o
31.4 CA
the MBT. Some of the sample test cases are presented in 10
this paper. o
Ign - 334 CA
0
300 320 340 360 380 400 420
The Simple four cases are computed with the well-observed
Crank Angle, degree
phenomenon of a spherical flame propagating into the Fig. 4 Simple case: p-ș computations at 26°
unburned mixture. With the ignition occurring at the pre-set CA ignition advance @ CR=17.0. Combustion duration is 31.4° CA.
time, a flame kernel forms at the ignition site. During the
ignition delay period, the flame kernel is assumed to move 40

vertically downward due to the surrounding turbulent fluid 35 Dotted: Experimental

flow. Subsequent to the ignition delay period, the EELB Black : Predictions
30 Grey : Motoring
model of flame propagation is invoked wherein a spherical
Cylinder Pressure, bar

flame is assumed to propagate into the unburned mixture, 25

with continued movement of the flame due to local fluid
20
velocities. This spherical flame propagation continues till the
flame encounters a wall, further the entrained unburned 15
mixture is assumed to burn exponentially [6]. During the
10
quasi-steady flame propagation, typical turbulent burning 59 CA
o

velocities are of the order of 7-9 m/s (at CR = 17.0) and time
scale of the order of 0.5– 0.6 ms during the initial stages of
flame propagation, and once the flame reaches the wall, time
scale for exponential burning is of the order of 0.8 to 1.0 ms
and somewhat similar to the value (0.6 to 1.0 ms) reported
by Keck [6]. Using the above theory, computations for four
test cases have been attempted. The simulation scheme is
organized to calculate the complete engine cycle, till
successive cycles converge. The first cycle simulation is
done assuming a recycled gas of 12% H2O, 14% CO2 and N2
as balance. Two more cycles are subsequently attempted
using the product gases obtained from the equilibrium
calculations as the recycled gas. These computations are
attempted with a time step of 0.2° CA (0.022 ms) and
confirmed to be time step independent. The computations for
two of the four test cases are shown in Fig. 4 and 5. The
computed results at CR=17 corresponding to an ignition
advance of 22° CA match excellently with the experimental
data, with marginal variation in the peak pressure and its
point of occurrence. In the computations at CR=13.5 at an
ignition advance of 25° CA, the pressures are marginally
higher at the commencement of heat release, with peak
pressure falling short of the experimental value by about 1.5
5
o
Ign - 333 CA
0
300 320 340 360 380 400 420
Crank Angle, degree
Fig. 5 Simple case: p-ș computations at 27° CA ignition advance @
CR=11.5. Combustion duration is 59° CA.
However, the computational results at retarded ignition
timing seemed to deviate largely from the experimental
results. This becomes evident from the two test cases
presented in Fig. 6 and 7. From Fig. 6 it is evident that there
is deviation in the computed pressures beyond a crank angle
(CA). At careful look at the experimental curve shows that
there is a steep rise in the cylinder pressure, giving an
impression of auto-ignition. However, this not true because
if auto ignition were to occur then it should have occurred
even at advanced ignition timing, wherein the
thermodynamic condition is of the same severity. The
possible cause for the discrepancy between the experimental
and computation results at less advanced or retarded ignition
timing is brought out in the following paragraphs.

999

Coming back to the computation results, it is not
immediately obvious as to why there is a deviation beyond a
certain point. The deviation in pressure might be occurring
as a result of increased burn rate either due to enhanced
surface burning area or the enhanced entrainment. Higher
surface area for burning is possible provided the flame does
not quench at the surrounding walls and the flame remains
floating. This calls for the displacement of the spherical
flame in a more complex manner than what has been
accounted in this analysis. However, considering this effect
had marginal effect on the maximum pressure rise. This
behavior was found to be similar with all the six test cases.
Therefore at less advanced ignition setting the enhanced
fluid dynamics due to reverse squish flow could be
modifying the burn rate to such an extent that there is a steep
rise in cylinder pressure. The cold flow CFD studies [7]
clearly indicated high velocity jets coming out of the bowl
and reentering the cylinder during reverse squish period.
And whenever major part of the combustion occurred during
this time there appears to be abrupt increase in the burn rate
leading to steep pressure rise. This effect appears to be more
severe in for higher CR of 17 and 13.5 as compared to
CR=11.5. This could true because the peak reverse squish
velocities with CR=17 is of the order of 31 m/s as against 6
m/s for CR=11.5 [7]. Therefore the deviations in the
computed results at CR = 11.5 is lower compared to CR= 17
and 13.5. The 0-D model has the drawback that it cannot
take into the fluid flow; therefore a detailed 3-D CFD
modeling would be essential to understand this phenomenon
60
Dotted: Experimental
50 Black : Predictions
Grey : Motoring
Cylinder Pressure, bar
40
30
20
10
o of premixed gases”, Ph.D. thesis, Indian Institute of Science, 1991.
Ign - 354 CA
0
300 320 340
Crank Angle, degree
Fig. 6 Complex case: p-școmputation at 6° CA ignition advance @
CR=17.0. Combustion duration is 29.4° CA.
V. CONCLUSION
The first principal contribution of this work is the revelation
of non-knock performance of producer gas in high
compression ratio (CR) spark-ignited engine. The engine
operation was found to be smooth at the highest CR of 17.0
without any undesirable effects of knock as discerned from
the pressure-crank angle curve, which showed smooth rise in
pressure without any pressure oscillations. However, there
was de-rating of the engine power, which is about 16.7% at a
CR of 17.0 and increased to 26% at a CR of 11.5. The
reduction in power output with producer gas is attributed to
reduction in the net calorific value of the fuel gas and air
mixture, and to the net reduction in number of molecules as
the mixture goes to product gases. In terms of emissions, the
NOx levels are well below the emission norms in India and
abroad. In the modeling part, a zero-dimensional (0-D)
model is constructed to simulate the processes of a spark-
ignition engine operation. The indicated power (IP) was
calculated by integrating the pressure-volume data over a
complete cycle. At advanced ignition timing, the 0-D model
is able to make reasonably good predictions. The accuracy in
IP prediction is around 3% for CR=17.0 and 11.5, and
between 5 and 8% for the other two cases at CR=13.5. But
at retarded ignition setting the error was large. This is
attributed to the inability of 0-D model to capture the fluid
flow; however, it might be possible to uncover this
phenomenon by conducting a detailed 3-D CFD modeling.
REFERENCES
[1] G. Sridhar, P.J. Paul, and H.S. Mukunda, “Biomass derived producer
gas as a reciprocating engine fuel – an experimental analysis”,
Biomass & Bioenergy, 21, 2001, pp 61-72
[2] S. Kanitkar, P. Chakravarty, P.J. Paul and H.S.Mukunda,, “The Flame
Speeds, Temperature and Limits of Flame Propagation for Producer
Gas-Air Mixtures,” Proceedings of 4th National Meet on Biomass
Gasification and Combustion, Mysore, Vol. 4, 1993, pp. 50-62.
[3] D.F. Caris and E.E. Nelson, “A New look at high compression
engines”, SAE Transactions, Vol 67, pp. 112-124.
[4] G. Sridhar, “Experimental and Modelling studies of producer gas
based spark-ignited reciprocating engines”, Ph.D. thesis, Indian
Institute of Science, 2003
[5] J.B.Heywood, Internal Combustion Engine Fundamentals,
International edition: McGraw-Hill, 1988.
[6] C.J. Keck, “Turbulent flame structure and speed in SI engines”,
Proceedings of 19th international symposium on combustion, 1982, pp.
1451-1466.
[7] G. Sridhar, P.J. Paul, and H.S. Mukunda, “ Simulation of fluid flow in
high compression ratio reciprocating IC engine”, Proceedings of
Institution of Mechanical Engineers, Part A, Journal of Power &
o
șb:29.4 CA
energy, Vol. 218, 2004, pp. 403-416.
[8] G. Sridhar, P.J. Paul, and H.S. Mukunda, “Computational Studies of
Laminar Burning Velocity of a Producer Gas and Air Mixture at
typical Engine Conditions”, Proceedings of Institution of Mechanical
Engineers, Part A, Journal of Power & energy, Vol. , 219[3], 2005,
pp. 195- 202.
[9] K.N. Lakshmisha, “Computational studies on the flammability limits
[10] P. Chakravarty, D.P. Mishra, P.J. Paul and H.S.Mukunda,, “The
360 380 400 420
theoretical calculations of limits of flame propagation for producer
gas-air mixture”, Proceedings of 4th National Meet on Biomass
Gasification and Combustion, Mysore, Vol. 4, 1993, pp. 28-37.
1000