International Journal of Hydrogen Energy 32 (2007) 1597 – 1604

www.elsevier.com/locate/ijhydene

Assessing current vehicle performance and simulating the performance of

hydrogen and hybrid cars
Bent SZrensen ∗
Roskilde University, Institute for Nature, Systems and Models; Energy, Environment and Climate group, Universitetsvej 1, DK-4000 Roskilde, Denmark

Available online 27 November 2006

Abstract
A measure of the efficiency in transforming energy input into transport work is defined and applied to road vehicles as well as to sea,
air and rail vehicles for passenger or freight transportation. The insight obtained with this measure is compared with the results of applying
the conventional measure of kilometres per unit of energy for current fleets of vehicles. Then, simulation methods are used to assess the
performance of fuel cell vehicles, electric vehicles and hybrids between the two. The latter are found to provide an optimum performance for
a small, efficient passenger car.
䉷 2006 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved.

Keywords: Vehicle efficiency; Transport work; Hydrogen vehicles; Fuel cell hybrid cars

1. Introduction
The on-the-road performance of passenger cars are often dis-
cussed in terms of kilometres driven per MJ (or other unit of
energy). This does not do justice to the different payloads that
different cars are certified to carry, perhaps considering the
poor correlation between required, actually used and available
passenger and cargo space characterising the vehicle choice
of many passenger car owners. Campaigns have occasionally
been directed at increasing customer usage of the available pay-
load (better trip planning, pool arrangements), but automobile
sales material often mentions only km/MJ overall performance,
if anything at all related to energy efficiency (mentioning en-
ergy performance is mandatory in some European countries,
but mostly not done in North America). Below, I present per-
formance data in terms of transport work carried out, such as
kilometres multiplied by maximum payload that may be ob-
tained per unit of energy input. In addition to being a prerequi-
site for encouraging better payload utilisation, this also allows
transport modes other than passenger cars to be compared on a
equal footing. Such comparisons are made for passenger cars,
motorcycles, buses, trucks, rail-based trains, ships and aircraft
carriers, with payloads that can comprise different ratios be-
tween passenger and cargo transportation, as the payload is in
all cases expressed in kilograms. This transport work perfor-
mance measure may be expressed in terms of the maximum
certified payload, or it may be expressed in terms of the av-
erage actual payload, in order to discuss the role of the pay-
load utilisation fraction (such as seat occupancy for passenger
modes).
With the assessment tool described above at hand, a discus-
sion of hydrogen hybrid road vehicles is carried out using sim-
ulation techniques. The issue is to determine the optimum bat-
tery and fuel cell ratings, with endpoints corresponding to pure
electric or pure fuel cell vehicles. The behaviour is non-linear,
because the total power requirements for a given driving cycle
change as function of vehicle mass, which again—depending
on battery type used—is not constant as the ratio between bat-
tery and fuel cell ratings are altered. Details of the assumptions
will be given in Section 4.

2. Performance of passenger cars

∗ Fax: +45 4674 3020.
E-mail address: boson@ruc.dk Fig. 1 gives the conventional performance measure for a
URL: http://mmf.ruc.dk/energy. number of current passenger cars, in km/MJ, but with the
0360-3199/$ - see front matter 䉷 2006 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijhydene.2006.10.037
1598 B. Srensen / International Journal of Hydrogen Energy 32 (2007) 1597 – 1604

1.00
Audi A2 3L:VW Lupo 3L diesel
0.90

Smart fortwo diesel

0.80
Toyota Prius
performance (km/MJ)
0.70 Citroën C2
diesel Renault Clio diesel

0.60 Skoda Octavia diesel

0.50

0.40
Hyunday
0.30 Santa Fe
Chevrolet Toyota
Corvette Cherokee Landcruiser
0.20 jeep
VW Touareg V8
Landrover Range Rover

0.10
300 400 500 600 700 800 900
payload (kg)

Fig. 1. Conventional passenger car performance measure (km/MJ) for cars on the 2004 Danish market, based on Refs. [1,2] and presented as function of
maximum payload.

400
Skoda Octavia diesel
Audi A2 3L Renault Clio diesel
VW Lupo 3L
350 Toyota Yaris diesel Hyunday Getz diesel
index (km/MJ times kg-payload)

Ranault Megane diesel

300

250 Smart fortwo

diesel
Hyunday
Santa Fe
200 Toyota
Landcruiser

150
Cherokee
jeep

100
VW Touareg V8

50
300 400 500 600 700 800 900
payload (kg)

Fig. 2. Transport work performance measure (km/MJ times maximum payload in kg) for cars on the 2004 Danish market, based on Refs. [1,2] and presented
as function of maximum payload.

payload mentioned along the abscissa. Payloads are the offi- cars have emerged over the last 2–4 years, with equally effi-
cially permitted maximum loads provided in the certification cient diesel engines (and now with particle filters for greatly re-
documents of each vehicle [1], and the fuel efficiencies are the ducing environmental impacts) but without the automated gear-
official test results for the New European Union Driving Cycle, box (although with efficiency-optimised gear exchange ratios).
as used for taxation purposes in Europe [2]. One notes the wide This trend started in France (Citroën C2, C3 and their Peugeot
spread in performance, indicating that properties other than fuel counterparts, as well as Renault Clio) but has now been taken
efficiency are important in consumer choices. The top scor- up by most automakers outside the USA. Other cars standing
ing vehicles are the 1999 Audi A2 and VW Lupo common-rail out in Fig. 1 are the Toyota Prius gasoline-battery hybrid and
diesel cars with a proprietary high-efficiency automatic trans- the Smart two-seater diesel. At the other end one finds cer-
mission (a computer-controlled, driver-independent gear chang- tain sports cars and utility vehicles (4-wheel driven jeeps or
ing device). These cars are no longer in production, but other tanks constructed for difficult off-road terrain, but often used as
 B. Srensen / International Journal of Hydrogen Energy 32 (2007) 1597 – 1604 1599

150.00

Index (km/MJ times kg of average payload)

124.00

98.00

72.00

46.00

20.00
0 60

Fig. 3. Same as Fig. 2 but with all payloads reduced from maximum permitted to actual average.

Average performance
10000
Index (km/MJ times av. payload)

1000

100

10

1
n
in
)

io
ar

s
5t

tra
bu

at
in

ft
(<
rc

s
ck

ra
tra

vi
e

bu

in

ne ip
er

er
ge
cl

ck

rc
la
tru

ra

ge sh
cy

ut

ut
ity

ity

ai
ru
en

ra
tt
m

m
vy
or

o
rc

rc
tt

e
gh
ss

rg
m

ut
ot

te

te
h
pa

he

i
co

co

ca
lig

fre

ro
m

in

in

vehicle

Fig. 4. Performance of different passenger and freight transport modes in terms of average transport work, based on US 2002 data (averages based on [3])
with indication of ranges based on other sources [1,4,5].

standard cars on paved highways with one person and little C3 and C5 relative to their positions in Fig. 1). However, the
luggage in the car). advantage of a high maximum payload disappears if it is not
Now the same data are translated into a measure of trans- used. Replacing the actual maximum payload for each vehicle
port work efficiency, by multiplying the km/MJ figure by the by the average of 157 kg, the ranking reverts to the conven-
maximum payload in kg. The results are shown in Fig. 2, while tional km/MJ level, as illustrated in Fig. 3. Most publications
Fig. 3 shows the corresponding quantities for the current aver- of vehicle km per unit of energy, e.g. by European Departments
age load factor of the vehicles (taken as 1.57 persons per vehicle of Transportation, assume such an average occupancy of the
[3], with an average weight of 100 kg per person, comprising vehicle.
luggage carried). The data shown in Fig. 2 shows the perhaps
surprising result, that the Skoda Octavia diesel car is the most 3. Other transport modes
efficient one, because its high permitted payload outweighs the
lower km/MJ-value compared e.g. to the VW Lupo 3L. It is The payload-weighed measure of performance allows a com-
also remarkable, that the order no longer favours the lowest parison of passenger cars with other road and non-road vehi-
mass car (as seen e.g. by the reverse ordering of Citroën C2, cles in a reasonably unbiased way. Fig. 4 shows the vehicle
1600 B. Srensen / International Journal of Hydrogen Energy 32 (2007) 1597 – 1604

150
Driving cycle and achieved speed cyc_kpr_r
100 kpha

km/h 50

0
0x104 1000 2000 3000 4000 5000 6000 7000 8000

2
fc_pwr_out_a
fc_pwr_out_a

1.5
Power form
1 fuel cell
(x10 kW)
0.5

0
0x107 1000 2000 3000 4000 5000 6000 7000 8000
sec

Fig. 5. Driving cycle used in simulations (top, identical to realised speed curve) and fuel cell power output (bottom) as function of time, for a 30 kW pure
fuel cell vehicle. The 30 kW peaks are not shown, in order to better view the behaviour below 20 kW. Calculated energy use corresponds to 2.3 litres of
gasoline equivalent per 100 km. The average performance efficiency is 1.17 km/MJ and the transport work efficiency is 398 kg km/MJ. The vehicle range on
4 kg hydrogen is 675 km.

1
ess_soc_hist
ess_soc_hist

0.9

0.8

0.7 Battery charge state

0.6
0x104 1000 2000 3000 4000 5000 6000 7000 8000

10
Power from battery (x10 kW) ess_pwr_out_a
ess_pwr_out_a

-5
0x104 1000 2000 3000 4000 5000 6000 7000 8000
sec

Fig. 6. Battery charge state (top) and battery power output (bottom) as function of time, for a 17 kWh pure electric vehicle with 1670 kg NiMH-batteries.
Calculated energy use corresponds to 2.5 litres of gasoline equivalent per 100 km. The vehicle range is only 350 km.

km/MJ times the average payload for current usage patterns,
for land, sea and airborne vehicles. The spread around the av-
erage value indicates an estimate of both variations in available
technology, and also variations that would emerge with differ-
ent load factors, i.e. when charter airlines in the vacation busi-
ness operate completely full aircraft in contrast to the about
70% seat occupancy of many commercial airlines. It is seen
that freight transport by sea or train obtains the highest en-
ergy performance. The average speed or infrastructure needed
to evaluate the cargo-delivery time-use is not included in this
analysis. The lower efficiency of road transport of cargo by lor-
ries is partly due to the decreased energy efficiency associated
with current requirements that lorries follow the (passenger car)
traffic flows on highways, implying the need for larger engine
ratings.
Regarding aircraft efficiency, apart from general aviation, the
regular scheduled air travel is seen on average to be as energy
efficient as automobile travel, when measured in relation to
transport work, i.e. per person-km or per kg-km travelled. It is
only a few decades ago, that air travel was considerable less
efficient than road travel [3]. Train transport of passengers is
seen to have similar energy performance to road and air travel,
inferior only to intercity buses. Decisive for these conclusions
are the load factors for each mode. Although seat occupancy
 B. Srensen / International Journal of Hydrogen Energy 32 (2007) 1597 – 1604 1601

150
driving cycle =
speed
100

km/h
50

0
0 1000 2000 3000 4000 5000 6000 7000 8000

0.9
battery charge state
0.8
fraction

0.7

0.6
0 1000 2000 3000 4000 5000 6000 7000 8000

5
× 10 kW

power from battery

−5
0 1000 2000 3000 4000 5000 6000 7000 8000

2
power from FC
1.5
× 10 kW

0.5

0
0 1000 2000 3000 4000 5000 6000 7000 8000
Time (Sec)

Fig. 7. Performance of PEMFC/Li ion battery hybrid car under the mixed urban/highway driving cycle shown at top. The fuel cell rating is 20 kW and a 5 kWh
Li-ion battery weighing 113 kg is installed. The range is 600 km and the gasoline equivalent fuel use 2.4 litres per 100 km, corresponding to 1.32 km/MJ or a
transport work of 448 kg km/MJ.

may be similar for train and bus travel, the seat arrangement is
usually much more condensed in bus design, which accounts
for the higher efficiency.
For passenger cars, the spread in data are taken from Fig. 2.
The light truck data are from the USA [3] as is the indicated
averages for other modes of transportation. This may be the
reason that the light truck efficiency is closer to that of an
average passenger car than one might have expected. In any
case, the scale used in Fig. 3 is logarithmic, and the differ-
ence between the best light truck performance within the range
given and the best passenger car performance is still a factor
of four.
4. Hybrid car simulation setup
The tool used for this part of the study is the modular
software-package ADVISOR [6], which is well tested and
offer a range of simple, parametrised sub-models or more de-
tailed physical models for the fuel cell stack, the batteries, the
electric motor, the exhaust control, the transmission and entire
power train including controls and control strategies. Each
ADVISOR module can be changed or written anew to accom-
modate particular segments of importance for a given vehicle.
The basic vehicle used for studying hybrid fuel cell/battery
passenger cars is loosely based upon the Volkswagen Lupo,
with a basic mass of 605 kg and total mass ranging from about
1 t (1000 kg) with pure fuel cell propulsion (at a rated power
of 20.30 kW) to nearly 3 t with pure battery operation and
lead-acid batteries. Even at this weight, it is not possible for
the battery version to achieve the average range of 600 km
between recharging/refuelling, which the hydrogen vehicle
achieves. With advanced lithium ion batteries, the mass of the
electric vehicle may be brought down below 2 t, but only the
hybrid cars obtain a manageable total weight. The car is in the
popular a-class and has a 340 kg maximum load of passengers
and/or luggage.
1602 B. Srensen / International Journal of Hydrogen Energy 32 (2007) 1597 – 1604

150
cyc_kph_r
Driving cycle and achieved speed kpha
100

km/h 50

0
0 1000 2000 3000 4000 5000 6000 7000 8000

0.9
ess_soc_hist
ess_soc_hist

0.8 Battery charge state

0.7

0.6

0.5
0 1000 2000 3000 4000 5000 6000 7000 8000

x 104
2
ess_pwr_out_a

0
power from
-1 battery
ess_pwr_out_a (x10 kW)
-2
0 1000 2000 3000 4000 5000 6000 7000 8000

3
fc_pwr_out_a
fc_pwr_out_a

2
power from
fuel cell
1
(x10 kW)

0
0 1000 2000 3000 4000 5000 6000 7000 8000

Fig. 8. Performance of PEMFC/Li ion battery hybrid car under mixed urban/highway driving cycle shown at top. The fuel cell rating is 30 kW and the car
is otherwise as the one in Fig. 5, and with a minimal 1 kWh Li-ion battery (23 kg). The gasoline equivalent fuel use is 2.3 litres per 100 km and the range
is 675 km. The “dip” in battery charge state around 6000 s may be smoothed by using a more sophisticated control mechanism that connects the fuel cell
directly to the electric motor if the battery charge state is getting poor.

5. Simulation results A fuel cell rating of 30 kW is found to be sufficient for achiev-

Fig. 5 shows the results of a simulation made for a pure fuel
cell vehicle. This car would have a range of 675 km with 4 kg
of hydrogen stored onboard, under average driving conditions
such as the driving cycle shown at the top of Fig. 5 (which is
a composite comprising both the EU and the USA regulatory
standard driving cycles in a mixed city and highway cycle of
total length 89 km). The car load of passengers and freight was
taken as 136 kg for all simulations. For the hybrid car simula-
tions, surplus power from the fuel cell is used to recharge the
batteries, which at the end of the driving cycle are required to
be as well charged as at the cycle start. This requirement fixes
the battery size, once the speed requirements of the driving cy-
cle has been used to fix the fuel cell power rating. The pure
fuel cell vehicle modelled has a small-size high-voltage bat-
tery, through which all power to the electric motor must pass.
ing the performance required by the model driving cycle, with
reasonable latitude for the challenges of alternative cycle pre-
scriptions. The fuel cell characteristics has a maximum effi-
ciency of 59% and is thus rather the goal cell of current R&D,
in terms of efficiency and assumed life of at least 5 years, as
further discussed in [7].
The purely electric vehicle shown in Fig. 6 is based on a
vehicle with 17 kWh of NiMH batteries weighing 1670 kg.
Nickel-metal-hydride batteries are common in current electric
vehicles, but may in the future be replaced by lighter lithium
ion batteries. These batteries have only recently become avail-
able for automotive uses, and are (even more than other types of
batteries) far from economically viable. The battery weight is
680 kg for Li-ion batteries and would be about 2.5 times more
for NiMeH or lead-acid batteries, starting to have a negative
effect on performance due to increased overall car weight, in an
 B. Srensen / International Journal of Hydrogen Energy 32 (2007) 1597 – 1604
Motor/Controller Operation GE 83-kw Ac
induction motor/inverter, tested by VA Tech
200
150
100
50
Torque (Nm)
0 0.5
-50
max cont. motoring torque
-100
max motoring torque
max cont. gen. torque
-150 max gen. torque
actual operating points
-200
0 2000 4000 6000 8000 10000 12000 14000
Speed (rpm)
Fig. 9. Torque-speed diagram for the power delivered by a 65 kWmotor in
a hybrid car with a 30 kW fuel cell and a modest size Li-ion battery also
capable of delivering 30 kW. The calculated limiting curves are indicated by
circles, while the actual operating points during the simulation are marked
with x.
obviously vicious circle. None of the pure battery solutions for
the vehicle in question could fulfil the range requirement (dou-
bling the battery mass in the hope of doubling the range would
entail a severe mass penalty on performance and motor rating).
The hybrid solution shown in Fig. 7 needs at most a fuel cell
rated at 20 kW, plus a modest 5 kWh Li-ion battery weighing
113 kg. At present, the extra cost of 50% more fuel cells or
of a 5 kWh Li ion battery are probably comparable, and both
must come down in order for any of the alternatives to become
economically viable or even interesting. The hybrid solution
has a better performance (higher maximum torque and better
acceleration characteristics) and even pure fuel cell vehicles are
as mentioned above most often equipped with a traction battery
of say 1 kWh.
Fig. 8 shows the simulation results for a similar hybrid with a
larger, 30 kW fuel cell, but with only 1 kWh of Li-ion batteries.
Apart from the larger excursions in battery charging, this vehicle
performs as well as the one illustrated in Fig. 7. As stated above,
“performing well” means achieving the prescribed speed along
the driving cycle, and leaving the battery as well charged at the
end of the cycle as at start. It should be noted that although the
30 kW rating seems to be required only a few times during the
driving cycle, this does not mean that the power level could
be set lower: the standard driving cycles test specific types
of performance required in the driving of most users, and the
fact that a given power level occurs only once or a few times
in the prescribed driving cycle does not mean that it can be
disregarded in actual driving.
Fig. 9 shows the torque delivered by the motor of a hybrid
vehicle, as function of rotational speed. The behaviour shows
that high efficiencies (operating points marked by x) have been
obtained without stressing the motor by taking it to limiting
values. The presence of a high-voltage battery supports this
1603
Motor/controller Efficiency (driving only, not regen)
1
0.9
0.8
0.7
0.6
efficiency
0.4
0.3
0.2
0.1
0
0 1000 2000 3000 4000 5000 6000 7000 8000
time (s)
Fig. 10. Instantaneous efficiency of the electric motor and its controller for
the hybrid vehicle as function of time during the traversing of the prescribed
driving cycle with its stop/go patterns (see previous figures).
torque behaviour. Finally, Fig. 10 shows the instantaneous mo-
tor/controller efficiency as function of simulation time. It is seen
that an efficiency around 90% is obtained except for the stop/go
situations, notably during the urban parts of the driving cycle.
6. Discussion
The key advantage of the car used in the present simula-
tion study is that it has high efficiency in the conventional
sense (mass, air and rolling resistance), before adding a fuel
cell/electric motor also of high conversion efficiency. Many
current fuel cell prototypes put 60.100 kW of fuel cells into a
basic car of poor efficiency, which makes little sense consid-
ering that the fuel cell cost is the most difficult obstacle. The
assessment technique presented here offers a way of appraising
vehicle efficiencies suitable for application to many different
types of vehicles, without bias against variations in the relative
passenger and freight-carrying capacity.
In general, one would like to be able to compare advantages,
such as those of high efficiency or low pollution propulsion
systems, with the cost associated with each feature. In the case
of efficiency, there is hardly any cost penalty: the common rail,
computer controlled compression ignition diesel cars coming
on the market these years cost exactly the same as the less than
half as efficient average car. Some people are willing to spend
more money on an inefficient car than on the most efficient
ones offered in the marketplace, for reasons of biased adver-
tising and the archaic notion that less efficient cars rank higher
as status symbols. These market imperfections are difficult to
meaningfully include in a scientific context.
As regards the cost of future cars such as electric vehicles,
fuel cell or fuel cell hybrid vehicles, there is a strong motivation
for using the most efficient basic car before contemplating new
propulsion systems: it is that the most costly items such as fuel
1604 B. Srensen / International Journal of Hydrogen Energy 32 (2007) 1597 – 1604
cells and batteries can be rated much lower if a basic car with
high efficiency is used. Many observers doubt if fuel cells with
ever become so cheap that a rated power of about say 100 kW
for a passenger car can be justified. One should remember that
current goals of fuel cell development include a 5000 km or
5 year lifetime, implying that the fuel cell stacks will have
to become replaced 4–5 times during the lifetime of current
vehicle generations. It is thus the cost of such 4–5 fuel cell units
that should be compared with alternatives (e.g. bio-diesel) for
viability [8].
References
[1] Car Dealer Information. Technical data for new cars, 2005; http://www.
biltorvet.dk/nyebiler/ fabrikat.asp.
[2] Danish Traffic Agency. New passenger car energy classification, 2005;
http://www.hvorlangtpaaliteren.dk.
[3] US Department of Energy. Transportation energy data book 24th ed.,
ORNL 2005.
[4] Åkerman J. Sustainable air transport—on track in 2050. Transportation
Res Part D 2005;10:111–26.
[5] Léonardi J, Baumgartner M. CO2 efficiency in road freight transportation:
status quo, measures and potential. Transportation Res Part D 2004;9:
451–64.
[6] Markel T. et al. ADVISOR. J. Power Sources 2002;110:255–66.
[7] SZrensen, B. Hydrogen and fuel cells. Elsevier Academic Press, Boston
2005, 450 p.
[8] SZrensen, B. Comparison between hydrogen fuel cell vehicles and bio-
diesel vehicles. In Proc. 16th World Hydrogen Energy Conf., Lyon 2006,
Paper 111, CDROM published by IHEA, AFH2 and EHA, 2006.