UCRL-JC-124094
Analysis of Experimental Hydrogen Engine
Data and Hydrogen Vehicle Performance and
Emissions Simulation
Salvador Aceves
This paper was prepared for submittal to
Hydrogen Program Review
A p d 29-31,1996
Miami, Florida
Thisisapreprintof apaperintendedforpublicationin ajoumalorproceedings. Since
changes may be made before publication, this preprint is made available with the
understandingthat it will not be cited or reproducedwithout the permission of the
author.
DISCLAIMER
This document was prepared as an account of work sponsored by an agency of
the United States Government. Neither the United States Government nor the
University of California nor any of their employees, makes any warranty, express
or implied, or assumes any legal liability or responsibility for the accuracy,
completeness, or usefulness of any information, apparatus, product, or process
disclosed, or represents that its use would not infringe privately owned rights.
Reference herein to any specific commercial product, process, or service by trade
name, trademark, manufacturer, or otherwise, does not necessarily constitute or
imply its endorsement, recommendation, or favoring by the United States
Government or the University of California The views and opinions of authors
expressed herein do not necessarily state or reflect those of the United States
Government or the UNversity of California, and shall not be used for advertising
or product endorsement purposes.
DISCLAIMER
Portions of this document may be illegible
in electronic image products. Images are
produced from the best available original
document.
.
ANALYSIS OF EXPERIMENTAL HYDROGEN ENGINE DATA AND
HYDROGEN VEHICLE PERFORMANCE AND EMISSIONS
SI MULATlON*
Salvador M. Aceves
Lawrence Livermore National Laboratory
Livermore, CA 94550
Abstract
This paper reports the engine and vehicle simulation and analysis done at Lawrence
Livermore (LLNL,) as a part of a joint optimized hydrogen engine development effort.
Project participants are: Sandia National Laboratory, California (SNLC),responsible for
experimental evaluation; Los Alamos National Laboratory (LANL),responsible for detailed
fluid mechanics engine evaluations, and the University of Miami, responsible for engine
friction reduction.
Fuel cells are considered as the ideal power source for future vehicles, due to their high
efficiency and low emissions. However, extensive use of fuel cells in lightduty vehicles is
likely to be years away, due to their high manufaduring cost.
Hydrogen-fueled, spark-ignited, homogeneous-charge engines offer a near-term alternative to
fuel cells. Hydrogen in a spark-ignited engine can be burned at very low equivalence ratios, so
that NO, emissions can be reduced to less than 10 ppm without catalyst. HC and CO
emissions may result from oxidation of engine oil, but by proper design are negligible (a few
ppm). Lean operation also results in increased indicated efficiency due to the thermodynamic
properties of the gaseous mixture contained in the cylinder. The high effective octane
number of hydrogen allows the use of a high compression ratio, further increasing engine
efficiency.
*Workperformed under the auspicesof the U.S.Departmentof Energy by Lawrence Livermore National
L~~~IW
~ndeaC~ntract
XY
W-7405-ENG-48.
In this paper, a simplified engine model is used for predicting hydrogen engine efficiency and
emissions. The model uses basic thermodynamic equations for the compression and
expansion processes, along with an empirical correlation for heat transfer, to predict engine
indicated efficiency. A friction correlation and a superchargerhrbocharger model are then
used to calculate brake thermal efficiency. The model is validated with many experimental
points obtained in a recent evaluation of a hydrogen research engine. The experimental data
are used to adjust the empirical constants in the heat release rate and heat transfer
correlation. The adjusted engine model predicts pressure traces, indicated efficiency and NO,
emissions with good accuracy over the range of speed, equivalence ratio and manifold
pressure experimentally covered.
The validated model is applied to conditions that are considered to be of interest to vehicular
applications in hybrids as well as conventional cars. It is recognized that using the engine
model for conditions far from the experimental points may result in inaccuracies. Therefore,
cylinder geometry is kept constant in the analysis, and only engine speed is varied beyond the
range for which experimental data points are available. It is expected that such an
extrapolation does not introduce large errors.
The results present information that can be used to predict engine performance for vehicular
applications, and are expected to seme as a first-order guide for engine sizing (number of
cylinders) and control strategy selection. The results indicate that hydrogen lean-bum sparkignited engines can provide Equivalent Zero Emission Vehicle (EZEV)levels in either a series
hybrid or a conventional automobile.
Nomenclature
A
B
c1, c2
h
n
P
9
s,
T
V
W
X
8
cylinder surface area
cylinder bore
constants in Equation 4.
heat transfer coefficient
shape parameter for bum fraction curve
pressure
.
heat transfer rate
mean piston speed
temperature
volume
gas velocity parameter
fraction of bum
crank angle
Subscripts
d
m
r
displacement
motored
reference
W
Wall
Introduction
Fuel cells have been recognized as the optimum power source for the lightduty fleet, due to
their high efficiency and low (near zero) emissions (DeLuchi, 1992). Fuel cells provide the
near-zeroemission benefits of electric vehicles, with the potential for long range and
performance comparable to that of conventional cars. The major obstacles in the way of
generalized use of vehicular fuel cells are their high cost, and the lack of an adequate fueling
infrastructure. The need for a new fueling infrastructure can be reduced if a fuel reformer is
installed onboard the vehicle, to convert a liquid fuel (e.g. methanol) to the high-purity
hydrogen fuel required in PEM fuel cells (Appleby, 1993).
Hydrogen can also be used in spark-ignited piston engines. Hydrogen has very special
properties, including a very high laminar flame speed, a high effective octane number, and no
toxicity or ozone-forming potential (Smith, 1994). Homogeneous-charge spark-ignited
piston engines can be designed to take advantage of these characteristics. The high laminar
flame speed allows the use of very low equivalence ratios (as low as 0.2), reducing NO,
emissions to near-zero levels without requiring a catalytic converter, that may deteriorate
with time. The use of low equivalence ratios also increases the indicated efficiency, and
reduces the need for throttled operation (Heywood, 1988). The engine can have a high
compression ratio, due to the high hydrogen octane number. Piston engines do not require a
high-purity fuel, and can burn mixtures of hydrogen, carbon monoxide, and other gases.
Hydrogen piston engines can therefore be optimized to yield a high efficiency and near-zero
emissions. While fuel cells have the potential for greater improvements in efficiency as well
as emissions, piston engines are an inexpensive and well-known technology that can be
applied in the short-term for obtaining basically the same benefits as fuel cell utiIization.
This paper analyzes the applicability of hydrogen homogeneous-charge spark-ignited piston
engines to Equivalent Zero Emission Vehicles (EZEV).' The analysis uses an engine model
that is calibrated to match the data obtained in a recent experiment (Van Blarigan, 1996).
The model is then used to generate engine performance maps for supercharged and
turbocharged operation. These maps are applied for predicting fuel economy and emissions
for conventional and series hybrid vehicles.
Ford Motor Company has publicly stated that piston engines can be produced for $20/kW,
which provides the economic incentive to evaluate piston engine technology for hydrogen
utilization.
Based on our engine modeling and vehicle simulation work, we believe that we can develop the
design rules for building 40% efficient hydrogen engines with emissions that are lower than
those from electric vehicles when power plant emissions are accounted for.
This research work has two major parts:
'
Equivalent Zero Emission Vehicles are defined here as those that generate less emissions when operating
inside the Los Angeles Basin thanthe power plant emissionsgenerated as a result of electric car operation.
These emission levels have been estimated as being one tenth of CARB ULEV standards, and are b e i
consideredfor approvalby CARE (CARB, 1995).
1. Analysis of hydrogen engine experimental data: This task consists of using a simplified
engine model to evaluate the experimental results obtained by Sandia National
LaboratoryKalifomia, for the current generation of optimized hydrogen spark-ignited
engine. The simplified engine model is used in conjunction with an optimizer, to obtain
optimum fits for the pressure traces obtained by the SNLC researchers. Bum duration and
bum fixtion parameters are then obtained from the optimum pressure trace fits. Pressure
traces are always fitted to within a very good approximation, indicating that the heat
release parameters obtained from the simplified engine model can be considered reliable.
2. The results of the analysis of the combustion obtained in (1) are then used to develop an
engine model that can predict hydrogen engine performance and emissions for conditions
not tested in the experimental work. Friction, supercharger and turbocharger models are
incorporated into the engine model, to predict brake thermal efficiency (ratio of net output
work to chemical energy input). The engine model can then be incorporated into a hybrid
and conventional vehicle simulation code, to predict hydrogen vehicle fuel economy and
emissions. Hydrogen vehicle fuel economy is a necessary input to infhstructure studies and
fuel storage design.
The purpose of this work is to generate a design guide for hydrogen spark-ignited engines that
a user may consult for selecting a hydrogen engine (especially for sizing the engine) for a
particular application and for the efficiency and emissions constraints of the given application.
The guide will indicate the performance of vehicles built with an existing engine, and point to
engine designs that will result in improved performance and emissions.
This project complements the ongoing experimental work at Sandia National
Laboratory/California (SNLC), the detailed fluid mechanics and combustion analysis done at
Los Alamos National Laboratory (LANL), and the friction reduction work at University of
Miami, without duplicating any of the work.
Past Results
This project has resulted in a set of design guidelines for high efficiency and low emission
hydrogen engines and vehicles. These guidelines include: extremely lean engine operation for
high indicated efficiency and very low nitric oxide (NO,) emissions; high compression ratio for
high efficiency; long piston stroke, controlled turbulence, and big cylinders for reduced engine
heat transfer losses; relatively low engine speed for reduced friction; supercharged operation for
improving power output and efficiency; and the use of a series hybrid vehicle configuration for
maximum he1 economy, thus reducing the problem of on-board fuel storage for vehicular use.
This project has also demonstrated that an optimized hydrogen spark-ignited engine built
according to the guidelines listed above, offers most of the emissions and efficiency benefits of
using fuel cells, without the cost and packaging restrictions of current PEM fie1 cells.
Optimized hydrogen spark-ignited engines are a current technology that can be used in
approaching the 80 mpg goal of the Partnership for a New Generation of Vehicles (PNGV), as
well as the Equivalent Zero Emission Vehicle (EZEV) standards being considered by the
California Air Resources Board (CAW).
The results of this project have been reported in 6 publications in the open literature. One of
these papers was written jointly with SNLC, and another was written jointly with LANL.
Experimental Engine Evaluation
The engine used in the experimental evaluation is an Onan engine which was modified by
incorporating a head containing two spark plugs, along with the original two valves. ?he
combustion chamber is a simple right circular cylinder with no squish and a flat top piston.
This geometry has been shown to be the most efficient shape for reducing heat transfer losses
in lean-bum engines (Olsson and Johansson, 1995). Engine chamteristics are listed in Table
1, along with the range of conditions used in the experiment. The design of the Onan engine
is based on the arguments set forth by Smith et al., 1995, to obtain a high efficiency, low
emission engine. These guidelines include: extremely lean engine operation for high indicated
efficiency and very low NO, emissions; high compression ratio for high efficiency; long
piston stroke, controlled turbulence, and big cylinders (low surface area to volume ratio) for
reduced engine heat transfer losses; relatively low engine speed for reduced friction; and
supercharged or turbocharged operation for improving specific power output and efficiency.
Figures 1 and 2 show the most important experimental results. Figure 1 shows indicated
efficiency as a function of equivalence ratio, for all the experimental points at MBT timing
obtained in the analysis. Engine speeds and supercharged operation are indicated with
different symbols. A 0.39 equivalence ratio was selected for most supercharged runs.The
figure shows that indicated efficiency increases as a fhction of engine speed, as a
consequence of reduced heat transfer losses. 'The variation of indicated efficiency with
equivalence ratio is best observed for 1200 rpm operation, for which the greatest fuellair
range was used. Indicated efficiency reaches a maximum near a 0.40 equivalence ratio.
Increasing the equivalence ratio from this point results in a decreased indicated efficiency, due
to a decreased specific heat ratio (y=c,,/~,) for the gas inside the cylinder (Heywood, 1988).
Decreasing the equivalence ratio from the optimum point increases the timing losses, due to
slower heat release, thereby reducing the indicated efficiency. Supercharged operation results
in small indicated efficiency gains due to slightly lower heat transfer losses per unit mass of
fuel at the higher densities. Supercharged operation has a larger effect on brake thermal
efficiency by increasing the output work relative to the frictional work.
Figure 2 shows NO, emissions as a finction of equivalence ratio. The figure shows that NO,
emissions are very insensitive to engine speed and supercharged operation, and correlate very
well with equivalence ratio. It is concluded that the modified Zeldovich mechanism
(Heywood, 1988) describes NO, production well. Spatial simulations using a 2-dimensional
version of KIVA (Amsden, 1993) indicate that 80% of the NOx is produced by the first 20%
of the burned gas when the last-bumed gas recompresses the first-burned gas.
Engine Model
The engine model uses first principles and correlations to predict piston engine efficiency and
power output. The engine model is a lumped (zerodimensional), timedependent model
which solves the basic differential equations for the compression and power strokes. The
following empirical expression is specified for the shape of the heat release cuwe (Ferguson,
1986):
where 8 is the instantaneous crank angle, 8, is the ignition angle, and 8 b and n are shape
parameters for the heat release curve. The values of e,, ob, and n are determined for each
experimental run by using an optimizer (Haney et al., 1992) to find the combination of the
three parameters that minimizes the differences between the experimental pressure trace and
the pressure trace calculated by the model. The results have been very satisfactory. The
relative errors in matching the pressure traces have been of the order of OS%, with a
maximum error of 1% over all engine speeds, equivalence ratios and manifold pressures. The
ranges for and n are: 15's < 30' and 1.15 < n < 1.45.
'Ihe engine model uses Woschni's correlation (Woschni, 1967) to estimate engine heat
transfer, This correlation is given as:
-
q = h A (T T
,)
with
where q is the overall heat transfer rate, A is the cylinder area, h is the heat transfer
coefficient in W/mzK, B is the cylinder bore in m,p is the pressure in Wa,T is the massaveraged temperature in K, and w is a measure of the gas velocity inside the cylinder, given
as:
Where C, = 2.28 during the compression and expansion periods; Sp is the mean piston speed;
C2= 0.00324; Vd is the displaced vohme; Trypr and V, are temperature, pressure and volume
at a reference state; and ppmis the difference between the cylinder pressure and the motored
pressure. The variables are a function of time, and are calculated for each crank angle degree
of engine rotation. It was found during the analysis that the heat transfer correlation
underpredicts heat transfer losses. Therefore, the original values of the constants C1and C,
given above were multiplied by 1.8, resulting in a better match with the experimental data.
The engine model includes a friction model and a supercharger/turbochargermodel to
predict brake thermal efficiency. The friction model uses a detailed correlation developed by
Patton et al., 1989. Supercharger and turbocharger performance is calculated by using a
thermodynamic model and assuming a constant (0.7) isentropic efficiency for both the
turbine and the compressor. Selection of a supercharger to match the energy control demands
is outside the current experience of the authors. However, a detailed supercharger map could
be incorporated into the model if fixher refinement is desired. A water-cooled intercooler is
assumed with a thermal effectiveness of 0.7. A model for NO, emissions prediction is
incorporated in the engine model. Emissions of NO, are calculated as a function of engine
equivalence ratio (Figure 2). A correction is used for supercharged and turbocharged
operation, to take into account the higher intake temperature resulting from the
compression process and the less than perfect effectiveness of the intercooler. Based on
reported data for typical engines (Heywood, 1988), volumetric efficiency is assumed to vary
from 85% at low engine speeds, to a maximum of 95% at 4000 rpm, down to 90% at 5000
rpm.
?he engine model is validated by comparing the calculated and experimental indicated
efficiencies. Figure 3 shows experimental (from Figure 1) and model results as a function of
equivalence ratio. The figure indicates that the model predicts absolute values as well as trends
with good accuracy for engine indicated efficiency, over the whole range of operating
conditions, with the maximum error of the order of 1%. NO, emissions are also predicted to
within a good approximation with the correlation of NO, as a function of equivalence ratio.
No validation is done for brake thermal efficiency, because the Onan engine used in the
experiment has substantially more friction per cylinder than a current automotive engine, for
which Patton’s correlation applies.
The engine model is then applied to predicting the engine and vehicle performance that
result if a 4cylinder (1.97 liter) hydrogen engine is built with the same cylinder
characteristics of the Onan engine. The geometry of the engine cylinders is not changed in
the analysis, because small changes in geometry may result in significant changes in
efficiency. It is expected, however, that larger engine cylinders will improve engine
efficiency.
Using the engine model for predicting vehicle performance requires extrapolating from the
engine speeds used in the experiment (1200-1 800 rpm) to engine speeds that are required for
vehicle operation. A maximum engine speed of 5000 rpm is assumed. It is recognized that
this extrapolation may result in errors. Errors are due to turbulence variations with engine
speed which influence heat release rate, thus changing both timing losses and heat trausfer
losses. However, Figure 3 shows that the model predicts the efficiency trends with good
accuracy for the range in which experimental data exist, and it is considered that the model
can do a reasonable job at predicting efficiency for high engine speeds. In addition to this,
engines in conventional and series hybrid vehicles are most often operated at low to
moderate speeds, which is the range for which the model has been validated. For the
conventional vehicle analyzed in a later section of this paper the mean engine speed is 1500
rpm for the urban cycle and 2700 rpm for the highway cycle. Maximum engine speed for the
driving cycles is 3300 rpm. The series hybrid vehicle is set to operate at a constant 2400 rpm
during the driving cycles.
Engine brake thermal efficiencies are required for applying the engine code to vehicle
calculations, and brake thermal efficiencies are calculated with a friction and a
superchargerhrbocharger model that have not been validated for this particular application.
While the model cannot replace experimental runs, it is considered that the brake thermal
efficiencies calculated with the model give a good idea of the performance that can be
obtained with such an engine. Engine emissions are very insensitive to engine operating
conditions other than maximum temperature within the cylinder (Figure 2), and it is
therefore expected that the engine NO, model can provide accurate predictions for emissions
levels throughout the operating range.
The engine model is applied to generate engine emissions and performance maps, necessary
for predicting vehicle performance. A conventional engine has only one degree of freedom
for controlling the output tdrque at any given speed: the inlet manifold pressure. This is due
to the use of three-way catalysts that require near-stoichiometric operation for high
conversion efficiency. A hydrogen engine has two degrees of freedom, because equivalence
ratio can also be varied. Generating an engine map therefore requires determining a control
strategy that specifies how to adjust these two parameters to obtain the desired torque for any
given engine speed. In this analysis, an optimizer (Haney et al., 1992) is used to determine
the combination of equivalence ratio and inlet manifold pressure that satisfies the torque
requirement while providing the maximum engine brake thermal efficiency. Equivalence ratio
in the optimization is restricted to values less than 0.5, and inlet manifold pressure is kept
under 2 bar. Another constraint is used in the optimization: engine NO, emissions are less
than 10 ppm under all operating conditions. An engine generating 10 ppm of NO, is below
the EZEV standards, provided that it is installed in a vehicle with a fuel economy of 40 mpg
or higher.
Lean operation results in low power output, and therefore turbocharged or supercharged
operation is required for providing a reasonably high power output. Both supercharged and
turbocharged operation have been considered for generating the engine performance map.
The performance maps for both cases are very similar, with turbocharged operation having a
slight efficiency advantage over supercharged operation. Only the results for supercharged
operation are shown in this paper. Supercharged operation is preferred to turbocharged
operation due to the lag time that may exist in turbocharged operation.
Results
Figures 4-7 show the predictions for engine efficiency and emissions maps, as well as the
optimum control strategy for inlet equivalence ratio and pressure, for the 4cylinder
supercharged engine.
Figure 4 shows lines of constant brake thermal efficiency (in percent) as a finction of engine
speed and engine torque. The figure also shows a dotted line corresponding to the conditions
at which the engine generates 10 ppm of NO,, and a dashed line that indicates the maximum
torque that can be obtained within the upper bounds of equivalence ratio (0.5) and inlet
pressure (2 bar) used in the analysis. The 10 ppm NO, curve is the lower of the two, and
therefore sets the limit on the maximum torque and power that can be obtained fiom the
engine. ?he maximum power that can be obtained while satisfying the NO, restriction
approaches 60 kW at 5000 rpm. The contour lines in this and the following figures spread
beyond the 10 pprn line, to show the potential power gains obtained by relaxing this
restriction. A square in the figure indicates the approximate range of experimental conditions
covered.
Figure 4 shows that the engine is predicted to have a broad area of high efficiency, for
intermediate speeds and high torques. The efficiency drops for lower speeds due to increased
heat transfer losses, and for higher speeds due to increased friction. As expected, the
efficiency drops to zero as the load is reduced. However, the drop occurs more slowly than in
conventional engines, because the equivalence ratio can be reduced as the load is reduced,
resulting in lower throttling losses.
Figure 5 shows contours of NO, emissions in'ppm as a function of engine speed and torque.
Engine levels are near zero (a
ppm) over the low load range, which is the range at which the
engine is operated most of the time during city and highway driving in conventional cars.
Emissions increase slowly as the torque increases, until the restriction of 10 ppm is
approached. When this restriction is approached, the operating conditions in the engine are
adjusted so that the 10 ppm line is pushed as high as possible by increasing the manifold
pressure without further increases in equivalence ratio, at the cost of some losses in
efficiency. This explains the great distance between the 8 ppm and the 10 ppm lines shown
in the figure. Emission levels shown in the figure are expected to be valid over the lifetime of
the engine, since no catalytic converter is used to control emissions. Emission levels are only
a function of equivalence ratio and manifold pressure, and are therefore not expected to
increase with use, as occurs in gasoline engines due to catalytic converter deterioration.
Figure 6 shows contour lines of equivalence ratio as a function of engine speed and torque.
Equivalence ratios shown in the figure are optimum values, that result in the maximum
possible brake thermal efficiency, while meeting the 10 ppm NO, limit. The figure includes a
dotted line for the 10 ppm NO, limit, and a dashed line for maximum engine torque for the
maximum equivalence ratio (0.5) and pressure (2 bar) considered in the analysis. Figure 6
shows that engine equivalence ratio is reduced down to 0.23 at the low load conditions, to
reduce throttling losses. Equivalence ratio is then increased as the torque is increased, until
the 10 ppm NO, limit is approached. At this point, equivalence ratio cannot be increased any
further, and additional power is obtained by supercharging (Figure 7).Equivalence ratio has to
be reduced as the inlet pressure increases, to compensate for the higher temperature of the
intake gases. The reduction of equivalence ratio with increasing torque appears in the figure
as sharp corners in the equivalence ratio lines at about 80 Nm.
Figure 7 shows contour lines of optimum inlet pressure (in bars) as a function of engine speed
and torque. Pressures in the figure are selected to provide the maximum possible brake
thermal efficiency, while meeting the 10 pprn NO, limit. The figure includes a dotted line for
the 10 ppm NO, limit, and a dashed line for the maximum engine torque. Inlet pressure is
kept relatively high (0.5 bar) at the very low load conditions to reduce throttling losses. The
engine operates without supercharging over most of the low-load conditions that are required
for typical urban and highway driving. When the 10 ppm NO, limit is approached, pressure is
increased rapidly to provide the required power without increasing equivalence ratio (which
would increase combustion temperature and therefore NOJ.
Application to Conventional and Series Hybrid Vehicles
The engine efficiency and NO, maps presented in the previous section are now used in
predicting vehicle fuel economy, performance and emissions for a conventional and a series
hybrid vehicle. This is accomplished by incorporating the engine maps into an existing
vehicle evaluation code (Aceves and Smith, 1995). The main characteristics of the two
vehicles are listed in Table 2. Both vehicles have a low weight, with the series hybrid vehicle
weighing 100 kg more than the conventional car, due to the additional components required
in the series hybrid power train.The engine and the hydrogen storage tank for the series
hybrid vehicle can be downsized to reduce the weight differential between the two cars. 'This
possibility, however, is not considered in this analysis. It is assumed that liquid hydrogen
cryogenic storage is used, since hydride storage would result in a substantially increased
vehicle weight. Liquid hydrogen storage also has a reasonable volume (about 110 liters, 30
gallons, for 5 kg of hydrogen), compared with the volume required for compressed hydrogen
(about 220 liters, 60 gallons, at 34 MPa, 5000 psi). The conventional engine has a 5-speed
transmission, and the series hybrid a single-speed transmission. Both transmissions have been
optimized by finding the reduction ratios (and shift points for the conventional car) that
result in maximum vehicle efficiency. The series hybrid car uses a high efficiency flywheel
(Post et al., 1993), a permanent magnet generator, and an induction motor.
Conversion between ppm of NO,., obtained fiom the engine maps, and grams per km (mile),
specified in the emissions regulabons, requires a knowledge of the composition of NO,
(fraction of NO and NOz in the mixture). It was found in the experiment that NO2 emissions
are a significant part of the total NO, emissions, due to the very low equivalence ratios used.
The calculation of g r a m s h of NO, done in this analysis assumes that half of the NO,
produced per unit volume is NOz. This is a conservative assumption, since NO2 emission
levels are lower over most operating conditions.
Table 3 shows the results of the analysis. Fuel economy is given as gasolineequivalent energy
consumption. The conventional vehicle has a reasonably high fuel economy (17.4 M i t e r ,
41 mpg). The vehicle range (330 km,205 miles) is lower than obtained in gasoline vehicles,
but is still high enough to not represent a limitation on the applicability of the vehicle to
most circumstances. The conventional vehicle has reasonable acceleration and hill climbing
performance. The most desirable feature of the conventional car is the emission levels.
Emissions are projected to be a factor of 75 lower than the CARB ULEV requirements, and
therefore a factor of 7.5 lower than EZEV. 'The reason for the emissions to be so low is that
the engine is operated most of the time at low torque, generating much less than the 10 ppm
maximum allowable NO, (Figure 5). Using the 10 ppm NO, restriction guarantees that any
driver obtains EZEV emissions, regardless of how aggressively they may drive.
The four cylinder engine has also been optimized considering a 100 ppm maximum NO, to
observe how this change &e& vehicle performance and emissions. It is found that changing
the maximum NO, limit causes very little change in driving cycle emissions and fbel
economy, because the urban and highway cycles are driven at low torque conditions, for
which the NO, limit does not have an effect. The major difference is an increase in the
maximum engine power, fiom 60 kW to 75 kW, which results in a significant gain in
performance (the time for 0 to 97 km/hr acceleration drops to 7.9 s, and the maximum
continuous slope at 97 k m h increases to 16%). However, an aggressive driver may operate
the vehicle at conditions that result in NO, emissions that do not meet the EZEV limits.
'Ihe series hybrid vehicle has a very high fuel economy (26.9 kmA, 63.2 mpg), a range
similar to conventional gasoline vehicles (508 km, 316 miles) and reasonable performance.
Emissions are higher than for conventional cars, but still within EZEV limits. Emissions are
higher than for conventional cars because the engine in a series hybrid vehicle operates at or
near maximum efficiency, to maximize fuel economy. In this case, maximum engine
efficiency (36%) is obtained near the 10 ppm emission limit (Figures 4 and 5).
Fuel economy projected for the series hybrid engine vehicle is low compared to the 33.5 kmll
(79 mpg) predicted for a hydrogen series hybrid vehicle in a previous work by the authors
(Aceves and Smith, 1995). The results are different because the engine model used in this
analysis is based on experimental data for a particular engine (Onan), while the previous work
indicates improvements that are likely to be obtained in a future optimized hydrogen engine.
Conclusions
This paper presents the development and validation of a simplified (zerodimensional) engine
model. The model is applied to a hydrogen engine which has been experimentally tested. The
model predicts accurately engine efficiency and NO, emissions over the 111 range of
operating conditions. The validated model is then used to generate engine performance and
emission maps for supercharged engine operation. The performance maps are then
incorporated into a vehicle evaluation code to obtain performance and emissions for
hydrogen-fbeled conventional and series hybrid vehicles. Analysis of these two vehicles yields
the following results:
1. The conventional vehicle has a high fuel economy, reasonable acceleration and hill
climbing performance and a short but acceptable range. Emissions out of the
conventional car are projected to be a fixtor of 75 lower than the CAW ULEX
requirements, and therefore a kctor of 7.5 lower than EZEV. The engine control
strategy presented in this paper guarantees that the conventional vehicle achieves EZEV
emissions levels regardless of how the car is driven.
2. The series hybrid vehicle has a very high fuel economy, a range similar to conventional
gasoline vehicles, and reasonable performance. Emissions are higher than for
conventional cars, but still within EZEV limits. Emissions in a series hybrid vehicle are
intrinsically independent of driver's input.
These results indicate that lean-burn hydrogen spark-ignited engines provide basically the
same benefits as a fuel cell, with a technology that is well-known and can be applied
immediately. Emission control is achieved without a catalytic converter, and emission levels
are therefore not expected to deteriorate with time.
Plans for Future Work
We plan to continue with the analysis of the experimental data generated by the SNLC
researchers for the next generation of optimized hydrogen engine (a modified Perkins engine).
Analysis of a different engine will allow us to compare the combustion and turbulence
characteristics existing in the two engines, and to evaluate how these conditions affect engine
efficiency and emissions. These comparisons are very important in achieving the goal of
establishing a design guide for optimized hydrogen engine design. It is expected that the larger
combustion chamber of the Perkins engine will show lower heat transfer losses and hence
higher efficiency.
We also plan to use the results of the engine performance and emissions analysis with our
vehicle simulation code and optimizer to determine vehicle and engine characteristics that
result in an optimum vehicle design and point to desired engine and vehicle characteristics that
result in better performance.
A third activity is to work jointly with the Los Alamos researchers in correlating the engine
ignition characteristics with the turbulence characteristics obtained with KIVA. ’Ihis activity is
expected to improve the level of understanding of the flame propagation process in the engine,
which will result in improved efficiency of future engine designs.
References
1. Aceves, S.M., and Smith, J.R., 1995, “A Hybrid Vehicle Evaluation Code and Its
Application to Vehicle Design,” SAE paper 950491.
2. Amsden, A.A., 1993, “KIVA-3, A KIVA Program with Block-Structured Mesh for
Complex Geometries,” Los Alamos National Laboratory Report LA-12503-MS.
3. Appleby, A.J., 1994, “Fuel Cells and Hydrogen Fuel,” International Journal of Hydrogen
Energy, Vol. 19, pp. 175-180.
4. California Air Resources Board, Mobile Sources Division, 1995, “Proposed
Ammendments to the Low-Emission Vehicle Regulations to Add an Equivalent ZeroEmission Vehicle (EZEV) Standard and Allow Zero-Emission Vehicle Credit for Hybrid
Electric Vehicles,” Preliminary Draft Staff Report, CAW, El Monte, CA, July 14.
5 . DeLuchi, M.,1992, “Hydrogen Fuel Cell Vehicles,” University of California Davis
Institute of Transportation Studies Report UCD-ITS-RR-92-14.
6. Ferguson, C.R, 1986, “Internal Combustion Engines, Applied Termosciences,” John
Wiley and Sons,New York.
7. Haney, S.W., Barr, W.L., Crotinger, JA., Perkins, L.J., Solomon, C.J., Chaniotakis, E.A.,
Freidberg, J.P., Wei, J., Galambos, J.D., and Mandrekas, J., 1992, “A SUPERCODE for
Systems Analysis of Tokamak Experiments and Reactors,” Fusion Technology, Vol. 21,
p. 1749.
8. Heywood, J.B., 1988, Internal Combustion Engine Fundamentals, McGraw-Hill, New
York.
9. Olsson K., and Johansson, B., 1995, “Combustion Chambers for National Gas SI Engines
Part 2: Combustion and Emissions,” SAE paper 950517.
10. Patton, K.J., Nitschke, RG., and Heywood, J.B., 1989, “Development and Evaluation of
a Friction Model for Spark-Ignition Engines,” SAE Paper 890836.
11. Post, R.F., Fowler, T.K., and Post, S.F., 1993, “A High-Efficiency Electromechanical
Battery,” Proceedings of the IEEE, Vol. 81, pp. 462-474.
12. Smith, J.R, 1994, “Optimized Hydrogen Piston Engines,” Proceedings of the 1994
International Congress on Transportation Electronics, Convergence 1994, SAE, pp. 161166.
13. Smith, J.R, Aceves, S.M., and Van Blarigan, P., 1995, “Series Hybrid Vehicles and
Optimized Hydrogen Engine Design,” SAE Paper 951955.
14. Van Blarigan, P., 1996, “Development of a Hydrogen-Fueled Internal Combustion
Engine Designed for Single SpeedRower Operation,” Submitted to the 1996 S A E Future
Transportation technology Conference and Exposition, Vancouver, BC.
15. Woschni, G., 1967, “Universally Applicable Equation for the Instantaneous Heat
Transfer Coefficient in the Internal Combustion Engine,” SAE Paper 670931.
Table 1. Modified Onan engine characteristics and experimental conditions.
Bore, mm
Stroke, mm
Displacement, cm3
Geometric compression ratio
Experimental range for equivalence ratio
Experimental range for engine speed, rpm
Experimental range for volumetric efficiency, %
82.55
92.08
493.0
14.0
0.2-0.5
1200-1800
90-2 15
Table 2. Main parameters for hydrogen-fueled conventional and series hybrid vehicles.
Vehicle parameter
conventional
test weight, kg (empty wt. + 136 kg)
frontal area, m2
aerodynamic drag coefficient
coefficient of rolling friction
transmission efficiency
transmission gears
accessory load, w
hydrogen storage capacity, kg
engine idling speed, rpm
launch engine RPM,maximum effort acceleration
regenerative braking
generator type
motor type
energy storage device
motor maximum torque, Nm
mobr maximum speed, rpm
flywheel energy storage, kwh
flywheel maximum power, kW
series hybrid
1136
2.04
0.24
0.007
0.94
5
1000
5
600
3 600
no
-
1236
2.04
0.24
0.007
0.95
1
1000
5
Yes
permanent magnet
AC induction
flywheel
95
11000
1
-
100
Table 3. Performance and emissions results for the conventiod and series hybrid hydrogenfueled vehicles.
Vehicle parameter
conventional
series hybrid
fuel economy', urban cycle, M i t e r (mpg)
15.0 (35.3)
24.7 (58.2)
30.0 (70.6)
fuel economy', highway cycle, M i t e r (mpg) 21.7 (51.1)
fuel economy', combined cycle, kmfliter (mpg)
17.4 (41.0)
26.9 (63.2)
0.012 (0.019)
NO, emissions, urban cycle, g h (glmile)
0.0020 (0.0032)
NO, emissions, highway cycle, g/km @/mile)
0.0013 (0.0020)
0.0097 (0.015)
0.011
(0.018)
NO, emissions, combined cycle, g/km (glmile) 0.0016 (0.0027)
10.0
10.0
time for 0-97 km/h (0-60 mph), s
13.1
maximum climbing slope at 97 k m h (60 mph), %
6.0
vehicIe range, combined cycle, km (miles)
330 (205)
508 (316)
solid symbols naturally
open symbols
A
43
421-
A
H A
A
0
41 L
A
A
0
40
39 38
-
@a
a
37
36 35
34 33
0
0
0
:.
0
-
0.2
A
0.3
1200rpm
1500 rpm
1800rpm
0.4
I
0.5
0.6
equivalence ratio
Figure 1. Indicated efficiency for the Onan engine as a function of equivalence ratio, for all
the experimental points at MBT timing obtained in the analysis. Engine speeds and
supercharged operation are indicated with different symbols.
140
'
solid symbols: naturally aspirated
open symbols: supercharged
A
0
I
.-%5
I
60
40
E
8 20
0-
e
0.2
0.3
0.4
0.5
0.6
equivalence ratio
Figure 2. Emission of NO, in parts per million for the Onan engine as a function of
equivalence ratio, for all the experimental points at MBT timing obtained in the analysis.
Engine speeds and supercharged operation are indicated with different symbols.
solid symbols
open symbols
A
A
B
A
A
a
0
0
U
0
0
0
1200rprn
I 1500rprn
A
1800 rprn
“*E
33
~
0.2
0.3
0.4
0.5
0.6
equivalence ratio
Figure 3. Indicated efficiency for the Onan engine as a function of equivalence ratio. The
figure includes both the experimental results and the model predictions.
E
z
4e
0
c.
1000
2000
3000
4000
5000
engine speed, rpm
Figure 4. Contour lines of constant brake thermal efficiency (in percent) as a function of
engine speed and engine torque. The dotted line corresponds to the conditions at which the
engine generates 10 ppm of NO,; the dashed line indicates the maximum engine torque that
can be obtained within the constraints of maximum equivalence ratio (0.5) and inlet pressure
(2 bar); and the square indicates the approximate area in which the experimental data were
taken.
180
-
maximum
,r,.+
100
10
40
20
0
--
-.-.-...-.
-#.*--.--B
.
e
.
-
-
- -.a
.
w
-
-
I
1000
1
I
1
2000
3000
4000
so00
engine speed, rpm
Figure 5. Contours of NO, emissions in ppm as a function of engine speed and torque, for the
optimized hydrogen engine. ?he dashed line indicates the maximum engine torque that can be
obtained within the constraints of maximum equivalence ratio (0.5) and inlet pressure (2 bar).
E
2
Q
c
'P
Q
engine speed, rpm
Figure 6. Contour lines of equivalence ratio as a function of engine speed and torque.
Equivalence ratios shown in the figure are optimum values, that result in the maximum
possible brake thermal efficiency, while meeting the 10 ppm NO, limit. The figure includes a
dotted line for the 10 ppm NO, limit, and a dashed line for the maximum engine torque that
can be obtained within the constraints of maximum equivalence ratio (0.5) and inlet pressure
(2 bar).
180
160
E
z
140
120
100
80
60
40
20
0
1000
2000
3000
4000
5000
engine speed, rpm
Figure 7. Contour lines of optimum inlet pressure (in bars) as a h c t i o n of engine speed and
torque, Pressures in the figure are selected to provide the maximum possible brake thermal
efficiency, while meeting the 10 ppm NO, limit. The figure includes a dotted line for the 10
ppm NO, limit, and a dashed line for the maximum engine torque that can be obtained within
the constraints of maximum equivalence ratio (0.5) and inlet pressure (2 bar).