Journal of the Air & Waste Management Association

ISSN: 1047-3289 (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/uawm20

On-Road Motor Vehicle Emissions and Fuel

Consumption in Urban Driving Conditions

H.Y. Tong , W.T. Hung & C.S. Cheung

To cite this article: H.Y. Tong , W.T. Hung & C.S. Cheung (2000) On-Road Motor Vehicle
Emissions and Fuel Consumption in Urban Driving Conditions, Journal of the Air & Waste
Management Association, 50:4, 543-554, DOI: 10.1080/10473289.2000.10464041

To link to this article: http://dx.doi.org/10.1080/10473289.2000.10464041

Published online: 27 Dec 2011.

Submit your article to this journal

Article views: 1508

View related articles

Citing articles: 45 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=uawm20

Download by: [112.134.222.141] Date: 07 December 2015, At: 02:51

 TECHNICAL PAPER
On-Road Motor Vehicle Emissions and Fuel Consumption in
Urban Driving Conditions
H.Y. Tong and W.T. Hung
Department of Civil and Structural Engineering, Hong Kong Polytechnic University, Hung Hom, Hong Kong
C.S. Cheung
Department of Mechanical Engineering, Hong Kong Polytechnic University, Hung Hom, Hong Kong
ABSTRACT
This paper reports on the analysis of on-road vehicle speed,
Downloaded by [112.134.222.141] at 02:51 07 December 2015
emission, and fuel consumption data collected by four in-
strumented vehicles. Time-, distance-, and fuel-based aver-
age fuel consumption, as well as CO, HC, NOx, and soot
emission factors, were derived. The influences of instanta-
neous vehicle speed on emissions and fuel consumption
were studied. It was found that the fuel-based emission
factors varied much less than the time- and
distance-based emission factors as instantaneous speed
changed. The trends are similar to the results obtained from
laboratory tests. The low driving speed contributed to a
significant portion of the total emissions over a trip. Fur-
thermore, the on-road data were analyzed using the modal
approach. The four standard driving modes are accelera-
tion, cruising, deceleration, and idling. It was found that
the transient driving modes (i.e., acceleration and decel-
eration) were more polluting than the steady-speed driving
IMPLICATIONS
The emissions and fuel consumption factors used in Hong
Kong were developed in the United States and Europe,
but the driving cycle developed for Hong Kong indicated
substantial differences from those in the United States
and Europe. Therefore, the emissions and fuel consump-
tion factors obtained in this study can be used to esti-
mate vehicular emissions and fuel consumption in Hong
Kong. To achieve more definitive results, more instru-
mented test vehicles could be used to measure on-road
emission rates and vehicle operating parameters. The
current study suggests that the acceleration and decel-
eration modes are more polluting than steady-speed driv-
ing modes. Moreover, low driving speed contributes to a
high percentage of total emissions. Great emphasis should
be placed on minimizing vehicle stops in urban areas to
speed traffic and to smooth acceleration and decelera-
tion. Traffic engineers should devise control measures to
improve traffic progressions and thus the pollutant emis-
sions per vehicle.
Volume 50 April 2000
ISSN 1047-3289 J.Tong, Hung,
Air & Waste and
Manage. Cheung
Assoc. 50:543-554
Copyright 2000 Air & Waste Management Association
modes (i.e., cruising and idling) in terms of g/km and g/
sec. These results indicated that the on-road emission mea-
surement is feasible in deriving vehicle emissions and fuel
consumption factors in urban driving conditions.
INTRODUCTION
Emissions from motor vehicles are a major cause of air
pollution in urban areas of Hong Kong. According to the
Environmental Protection Department of Hong Kong
(HKEPD), CO, NOx, and particulates come mainly from
vehicular emissions. It was indicated that particulates and
NOx emissions were fairly high in urban areas and con-
stantly exceeded the Hong Kong Air Quality Objective
(HKAQO). In 1997, the highest annual particulate level
at street level was almost 68% above the statutory limit
in Hong Kong.1 Therefore, the reduction of motor vehicle
emissions will contribute significantly to the improvement
of air quality in Hong Kong. To achieve this goal, it is
necessary to understand actual vehicle emissions and iden-
tify the important operational factors affecting vehicular
emissions and fuel consumption.
Conventionally, emission testings are conducted by
driving the vehicle through standard driving cycles in a
laboratory on chassis dynamometers. The merit of this
method is simple, but the test conditions are restrictive
in that they may not represent actual on-road driving con-
ditions. A study found that fuel consumption and exhaust
emission depended on whether they were measured at a
steady speed or an on-road driving pattern.2 It was found
that emissions measured on actual speed driving cycles
were generally higher than those measured on the labo-
ratory driving cycles. Even for actual speed driving cycles,
the engine operating conditions differ significantly from
those of the on-road driving test.3,4
In fact, actual on-road driving is more complex than
driving cycle simulations. There are large variations in
operating conditions during real-world driving. Some fac-
tors are difficult to simulate on test beds.5,6 Therefore,
Journal of the Air & Waste Management Association 543
 Tong, Hung, and Cheung
on-road emission, fuel consumption, and engine operating
condition measurements are necessary. However, emis-
sion data directly measured from the tailpipes of on-road
vehicles are limited, and relatively few vehicles have been
tested for each study.5 St. Denis et al.4 tested a petrol van
in California. They found that the data obtained from
the on-road test differ significantly from those obtained
from the Federal Test Procedure (FTP) driving cycle. Cicero-
Fernandez et al.7 examined the grade effect by a single
test vehicle driving on a hill. Both of these studies have
used limited numbers of vehicles for their on-road test-
ing, which limited the scope of these studies. Several other
variables of interest for vehicle operating parameters could
not be assessed.
In this study, emissions were measured directly from
the tailpipe of four test vehicles under actual on-road driv-
ing conditions. The test vehicles were chosen to repre-
Downloaded by [112.134.222.141] at 02:51 07 December 2015
sent the typical traffic mix in the urban areas of Hong
Kong, which enabled the investigation of more variables
of interest. Although only one vehicle from each vehicle
category was tested, it is believed that the study would
produce indicative results for actual on-road vehicle emis-
sion behavior of contemporary driving patterns in urban
areas, and the ways in which such behavior differs from
mandated certification procedures. The study can also
produce supplementary results to laboratory tests.
This paper describes a series of on-road test runs that
were conducted in the urban areas of the Hong Kong
Special Administrative Region. Four test vehicles were
employed to collect on-road speed, emission, and fuel
consumption data. Average and modal emissions, as well
as fuel consumption factors, were then calculated in
terms of g/sec, g/km, and g/kg fuel. Finally, the behav-
iors of the exhaust emissions and fuel consumption were
examined with reference to operational variables such
as speed and acceleration.
DATA COLLECTION
In order to capture different on-road driving patterns in
urban areas, test vehicles were driven in two typical ur-
ban districtsMongKok and Centrallocated in the
Kowloon Peninsula (KLN) and Hong Kong Island (HKI),
respectively. The roads in KLN are more spread out on
flat land, while those in the HKI are bounded on the
northern shore of the island, one side by the hill and
the other side by Victoria Harbor. MongKok and Central
were chosen to represent two distinct driving environ-
ments in Hong Kong that have a great impact on air
pollution. Air quality in these areas is of great concern
and the government has installed several roadside air
quality monitoring stations to monitor traffic emissions.
Central is the central business district. MongKok is
known to be the most crowded and densest area in KLN.
544 Journal of the Air & Waste Management Association
The vehicle speed seldom exceeds 50 km/hr. The driving
cycle for Hong Kong developed from these two districts
has an average speed of 15.5 km/hr, average running
speed of 21.7 km/hr, and idling and cruising proportions
of 31.4 and 9.1%, respectively.8 Driving in Hong Kong is
thus relatively slow compared to cities in the United
States and Europe, due to the large number of signal junc-
tions in Hong Kong.
Four vehicles of different models, sizes, and weights
were employed in the collection of instantaneous speed,
emissions, and fuel consumption data. The specifications
of these test vehicles are shown in Table 1. These four
test vehicles were chosen to represent the typical traffic
mix in Hong Kong. The Hong Kong Annual Traffic Cen-
sus 19979 stated that among all vehicles traveling in the
urban areas of HKI during peak hours, 50.6% were pas-
senger cars, 31.4% were light-duty vans and diesel taxis,
and 7.7% were buses. In the urban areas of KLN, the
peak hour traffic mix was composed of 41.0% passenger
cars, 27.4 % light-duty vans and diesel taxis, and 6.5%
buses. Although buses occupy only a small proportion
of the traffic mix, they are a major source of vehicular
particulate emissions.
The test vehicles were driven by professional drivers.
The passenger car, petrol van, and diesel van were driven
by the same driver, and the double-decker bus was driven
by a professional bus driver. This reduced the effect of
different driver behaviors. The driving environments for
the test vehicles were basically the same. The drivers were
asked to drive randomly through the above-mentioned
districts. A single test run was performed for the double-
decker bus, while two test runs were performed for the
other three test vehicles. The reason for not choosing a
fixed route is that it may not necessarily reflect the whole
picture of urban driving. On the other hand, changing
routes for each test run allows the testing of different driv-
ing conditions in the prescribed urban districts.
The arrangement of the on-board measurement sys-
tem is outlined in Figure 1. The gas filter connected with
the gas analyzer is used to filter out water vapor and par-
ticulates in diesel vehicles. Otherwise, the gas analyzer
would easily be blocked by the particulate emissions. En-
gine speed and the transmission shaft rotational speed
were measured by infrared photoelectric sensors. Incor-
porated with a PICO data acquisition system, the speed
data were collected into a computer. Vehicle speed could
then be calculated from the shaft speed by a simple equa-
tion. The error and response time of the speed measure-
ment were less than 2% and 0.01 sec, respectively. The
resolution of the speed measurement was lower than 0.03
km/hr for the four test vehicles.
An Econotest fuel flow meter was connected to the
fuel supply pipe to measure the instantaneous fuel
Volume 50 April 2000
 Tong, Hung, and Cheung
Downloaded by [112.134.222.141] at 02:51 07 December 2015

Figure 1. Arrangement of the on-board measurement system.

consumption. Then, instantaneous fuel consumption rate
could be obtained by simple calculations. The error and
response times of the fuel flow meter were less than 1
mL/sec and 0.1 sec, respectively.
Instantaneous concentrations of CO, CO2, NOx, HC,
and O2 were measured by a Flux-2000 five-gas analyzer.
In the gas analyzer, the CO, CO2, and HC measurements
were recorded using the Non-Dispersive Infrared (NDIR)
method, while the O2 and NOx measurements were taped
using electrochemical transducers. The exhaust gas ana-
lyzer was calibrated with standard gases before each of
the test runs. The error and response timed of the gas
analyzer were less than 2% and 0.5 sec, respectively, for
all of the gas pollutants.
According to the HKEPD, diesel vehicles are respon-
sible for an absolute majority of vehicular particulate
emissions, so particulate emission measurements in
terms of smoke intensity were performed in particular
for diesel vehicles. Instantaneous smoke intensity from
diesel vehicles was collected by a LUCAS Hartridge smoke
meter. The analogue smoke signal in DC voltage from
the smoke meter was transmitted to the computer
through the PICO data acquisition system. The smoke
meter was calibrated before performing the test runs.

Table 1. Specifications of the test vehicles.

Passenger Car Petrol Van Diesel Van Double-Decker Bus

Gear mode automatic transmission manual transmission manual transmission automatic transmission
Vehicle type Sunny, 4 seats Urvan, 9 seats Hiace, 6 seats 6LXC, 100 seats
Manufacturer Nissan Nissan Toyota Gardner
Model year 1997 1990 1990 1988
Weight (kg) 2040 2750 2800 10,000
Max power (kw) 81 64 60 150
Engine type carburetor injection injection injection
Bore stroke (mm2) 73.6 88.0 85 86 96 96 120.65 152.4
Engine capacity (L) 1.498 1.952 2.779 10.45
Cylinder number 4-cyl., OHC, in-line 4-cyl., in-line 4-cyl., DOHC, in-line 6-cyl., in-line
Fuel type petrol petrol diesel diesel
Catalytic converter with without without without

Volume 50 April 2000 Journal of the Air & Waste Management Association 545
 Tong, Hung, and Cheung
The error and response time of the smoke meter were
less than 2% and 0.1 sec. respectively.
The fuel consumption rate and gas emission concen-
trations were related to the speed measurement. However,
a time lag exists between the fuel meter, gas analyzer, and
speed measurement. The experiment was conducted to
determine the time lag under various engine speeds by
the method of sudden acceleration and one-cylinder mis-
fire. The time lag of the instrument was calibrated and
was taken to be 5 sec in the data analysis.
The data were collected at the morning peak traffic
period, from 8:00 to 11:00 a.m., and the off-peak period,
from 2:00 to 5:00 p.m., from March to June 1998 during
calm and dry weather conditions.
METHODOLOGY
Data obtained for each test vehicle were grouped together.
Downloaded by [112.134.222.141] at 02:51 07 December 2015
Thus, there are four sets of instantaneous speed, emissions,
and fuel consumption data by vehicle types: the light-duty
diesel van, petrol passenger car, light-duty petrol van, and
double-decker public bus. The equations for instantaneous
mass emission rate (g/sec) are shown in the appendix. The
instantaneous emission rates of CO, HC, and NOx of the
petrol vehicles, in g/sec, were calculated by standard meth-
ods described in the SAE Handbook.10 For the diesel vehicles,
the mass concentrations of soot were obtained from the
smoke intensity based on a conversion chart in the SAE
Handbook.11 The mass emission rate of soot can be calcu-
lated accordingly when the instantaneous engine speed and
its swept volume are known. These data form a basis for
emission and fuel consumption analysis of this study.
Emissions and fuel consumption are usually calcu-
lated as average value over several test runs and then
Instantaneous vehicle speed (km/hr)
Figure 2. Speed distributions of test vehicles.
546 Journal of the Air & Waste Management Association
parameterized in terms of the corresponding average
speed. In fact, on-road driving is random combinations
of the four standard driving modes: acceleration, cruis-
ing, deceleration, and idling. It is of great interest to char-
acterize emissions and fuel consumption behaviors during
different driving modes. In the following sections, both
global and modal analysis will be used to analyze the emis-
sions and fuel consumption data.
Vehicular emissions can be expressed in terms of
grams of pollutants emitted per unit time, per unit dis-
tance traveled, or per unit fuel consumed. Accordingly,
three terms are defined to describe vehicular emissions
and fuel consumption: emissions, fuel consumption rate
(g/sec); emissions, fuel consumption factor (g/km); and
emissions index (g/kg fuel).
It is a common practice to describe emissions and
fuel consumption from motor vehicles by an average value
over a trip. In general, average vehicular emissions and
fuel consumption over a trip were expressed in distance-
based (g/km) unit or fuel-based (g/kg fuel) unit. Average
emissions, fuel consumption factors, and indices of the
vehicles were calculated by eqs 13. All of the factors and
indices were calculated over the whole data sample for
each vehicle and were compared to the fleet average emis-
sion factors used by the HKEPD.
Average emission factor [g/km]= (1)
Average Fuel Consumption Factor [g/km]= (2)
v
e
Average Emission Index [g/(kg fuel)]= (3)
Volume 50 April 2000

where e is the instantaneous emission rates of the subject
gas pollutant, f is the instantaneous fuel consumption rate,
and v is the instantaneous vehicle speed.
With the acquired instantaneous speed data, accel-
eration rates were calculated using the method of central
difference. The corresponding emissions and fuel con-
sumption data were then classified according to various
driving modes. The four standard driving modes were
defined as follows:
idling mode: zero speed and acceleration.
acceleration mode: positive incremental speed
changes of more than 0.1 m/sec/sec during the
1-sec interval.
cruising mode: absolute incremental speed
changes of less than or equal to 0.1 m/sec/sec
during the 1-sec interval.
deceleration mode: negative incremental speed
Downloaded by [112.134.222.141] at 02:51 07 December 2015
changes of more than 0.1 m/sec/sec during the
1-sec interval.
The value of 0.1 m/sec/sec has been used extensively
by other researchers in defining driving modes.8,12-16 For
each vehicle, the instantaneous emission and fuel consump-
tion rates (g/sec) for each driving mode were averaged to
give the modal emission and fuel consumption rates.
RESULTS AND DISCUSSION
Description of the Data Sample
Characteristics of the four sets of data for each test ve-
hicle are shown in Table 2. It can be seen that the average
speeds are generally higher than those of the Hong Kong
driving cycle. The speed distributions for each vehicle
during the test runs are shown in Figure 2. The bars are
the proportions of each speed class, while the lines repre-
sent the cumulative proportions. It can be seen that for
more than 90% of the time, test vehicles are driving at
speeds lower than 50 km/hr, a typical driving phenom-
enon in urban areas. About 30% of the time, the petrol
van idled, which is the largest percentage among the four
test vehicles. However, very few data were recorded at the
speed range of 010 km/hr, indicating that the petrol van
accelerated and decelerated hard during the test runs. The
passenger car and the double-decker bus spent most of
Table 2. Characteristics of test runs.
Duration Average Speed Total Distance
(sec) (km/hr)
Passenger Car 2613 19.47
Petrol Van 6500 23.85
Diesel Van 3133 24.91
Diesel Bus 1690 15.32
Volume 50 April 2000
Tong, Hung, and Cheung
the time driving at lower speeds. For over 90% of the time,
the double-decker bus was driven at speeds lower than 30
km/hr. For the diesel van, the driving time spread through
the speed range below 50 km/hr. The largest proportion
was in the speed range of 1015 km/hr.
Global Analysis
Average Emission and Fuel Consumption. Average emission,
fuel consumption factors, and indices of the test vehicles
are shown in Table 3. The double-decker bus was shown to
have the largest emission factors and indices probably be-
cause it was the oldest and had the largest engine capacity
as well as the highest maximum power output. Apart from
the bus, the highest CO emissions were found in the petrol
van, followed by the passenger car. The smaller size and
engine capacity of the passenger car may contribute to the
relatively low emission level, but the major factor was the
catalytic converter in the passenger car, which can oxidize
a majority of the HCs to CO2 and H2O. However, the emis-
sions and fuel consumption of the catalyzed passenger car
were still higher than expected, probably due to the low
speed and irregularity of driving in Hong Kong,8 which
restricted the optimal performance of the catalytic con-
verter. Moreover, the maximum power output of the cata-
lyzed passenger car was higher than that of the petrol van.
CO and HC emissions from the diesel van were lower
than those from the petrol vehicles. NOx emissions from
the petrol van, which did not have a catalytic converter,
were higher than those of the diesel van. However, the
passenger car emitted less NOx than the diesel van. In
general, diesel engines have a fuel economy advantage
over petrol engines, owing to efficient combustion, and
as a result, diesel engines emit less CO, HC, and NOx.17
However, if the engines were installed in vehicles and used
practically, emissions and fuel consumption would be
governed by vehicle operating variables and the driving
conditions, such as engine capacity, vehicle speed, pay-
load, road grades (i.e., uphill or downhill), and so on. It
can be observed that the fuel consumption factor of the
diesel van was larger than those of the petrol vehicles,
probably because of its larger engine capacity.
Table 4 shows the measured average emission factors
of the four test vehicles and the fleet average emission
factors for the corresponding type of vehicle in Hong
Kong. The fleet average emission factors are provided by
the HKEPD. However, the vehicle fleet used with HKEPD
(km) classifications is not the same as that shown in Table 4.
The diesel and petrol vans were categorized by the HKEPD
14.13 as a public light bus and a passenger van, respectively.
43.06 And it is understood that a significant proportion of pas-
21.68 senger vans are in fact diesel vehicles.
7.19
The HKEPD fleet average emission factors for the pas-
senger van were clearly not derived solely for petrol vans.
Journal of the Air & Waste Management Association 547
 Tong, Hung, and Cheung

Diesel-fueled passenger vans could lower the fleet emis-
sion factors of CO, HC, and NOx derived for all passenger
vans. Thus, the measured average emission factors for the
test petrol van are significantly smaller than those of the
HKEPD fleet average emission factors for passenger vans.
However, the deviations between the measured emission
factors of the test diesel van and the HKEPD fleet emis-
sion factors for the public light bus are relatively smaller
than those of the test petrol van, because the public light
buses in Hong Kong are basically diesel-fueled. Thus the
derived fleet average emission factors are close to those
derived for our diesel van.
The differences between the measured and fleet emis-
sion factors for the passenger car were smaller than those
for the bus. For the passenger car, the measured emission
factors were quite close to the HKEPD fleet emission fac-
tors, especially for HC emissions, which only differed by
Downloaded by [112.134.222.141] at 02:51 07 December 2015
However, the measured bus emission factors deviated sig-
nificantly from the HKEPD fleet emission factors. The mea-
sured emission factors overestimated CO, HC, and NOx but
underestimated soot emissions. The large differences may
be due to the age of the test bus. The fleet average emission
factors from the HKEPD were derived to represent the av-
erage emission pattern of the buses in Hong Kong, which
were composed of a small proportion of vehicles 10 years
old or more. Therefore, the test bus, which is 10 years old,
should have had higher emission factors than the fleet av-
erage emission factor from the HKEPD.
Influence of Instantaneous Speed. To investigate the influ-
ence of instantaneous speed on emissions and fuel con-
sumption, records of instantaneous emissions and fuel
consumption were grouped according to the
instantaneous speed. Usually, instantaneous emission and

6%. This result indicated that the emission data of the pas- fuel consumption are expressed in terms of g/sec rather
senger car could represent, to a certain extent, the average than g/km and g/kg fuel. However, it is of great interest to
emission pattern of passenger cars in Hong Kong. examine the trends of distance-based and fuel-based emis-
sions as well as fuel consumption as ve-
Table 3. Average emission factors and indices of each vehicle. hicle speed changes. For each speed range,
the average emission and fuel consump-
Passenger Car Petrol Van Diesel Van Double-Decker Bus tion rates, factors, and indices were calcu-
lated by eqs 13 and plotted against the
CO g/km 14.83 21.02 3.72 24.33 corresponding speed range. Figures 35
g/kg fuel 210.10 294.84 47.67 show the plots of the mean emissions and
HC g/km 0.99 2.79 0.84 4.32 fuel consumption in different units aver-
g/kg fuel 14.10 39.10 10.75 aged for specific speed range. The curves
NOx g/km 1.03 2.64 1.20 25.63 were constructed in 5-km/hr resolution of
g/kg fuel 14.62 37.00 15.39 instantaneous speed.
Soot g/km 0.10 0.83 Figure 3 shows the emissions and fuel con-
g/kg fuel 1.30 sumption factors (g/km) of each test vehicle.
Fuel g/km 70.59 71.30 78.03 There is a very clear and distinct trend for
the four test vehicles. It is shown that the
emission factors of CO, HC, NOx, and soot,
as well as the fuel consumption factors of
Table 4. Average emission factors compared with fleet average emission factors from HKEPD.
the test vehicles, drop as the instantaneous
speed increases. For the petrol van, when the
Passenger Car Petrol Van Diesel Van Double-Decker Bus
instantaneous speed increased from 510 to
520 km/hr, the emissions factors decreased
CO
dramatically while the fuel consumption fac-
Measured factors (g/km) 14.83 21.02 3.72 24.33
HKEPD factors (g/km) 11.71 1.26 1.25 8.89 tor increased a little and then dropped gradu-
HC ally after the speed exceeded 20 km/hr. For
Measured factors (g/km) 0.99 2.79 0.84 4.32 the other three test vehicles, all of the emis-
HKEPD factors (g/km) 0.93 0.71 0.69 2.30 sion and fuel consumption factors decreased
NOx as speed increased, especially for the double-
Measured factors (g/km) 1.03 2.64 1.20 25.63 decker bus. For an instantaneous speed in-
HKEPD factors (g/km) 1.63 2.31 2.21 11.71 crease from 510 to 1015 km/hr, the CO,
Soot HC, NOx, and soot emission factors of the
Measured factors (g/km) 0.10 0.83 double-decker bus generally decreased more
HKEPD factors (g/km) 0.59 1.38 than 70%. The decreasing trend as instanta-
neous speed increased can be explained by

548 Journal of the Air & Waste Management Association Volume 50 April 2000
 Tong, Hung, and Cheung

Fuel CO HC NOX Fuel CO HC NOX

HC, NOX Emission Factor (g/km)

HC, NOX Emission Factor (g/km)

CO Emission Factor (g/km)

CO Emission Factor (g/km)

Fuel Consumption Factor

Fuel Consumption Factor

Instantaneous Vehicle Speed (km/hr) Instantaneous Vehicle Speed (km/hr)

(a) Petrol Passenger Car (b) Petrol Van

Fuel CO HC NOX Soot Fuel NOX HC Soot

Fuel Consumption Factor (g/km)

Fuel Consumption Factor (g/km)

HC, Soot Emission Factor (g/km)

CO, HC, NOX Soot Emission
Factor (g/km)
Downloaded by [112.134.222.141] at 02:51 07 December 2015

Instantaneous Vehicle Speed (km/hr) Instantaneous Vehicle Speed (km/hr)

(c) Diesel Van (d) Double-Decker Bus

Figures 3a3d. Emissions and fuel consumption factors (g/km) against instantaneous vehicle speed.

Fuel CO HC NOX Fuel CO HC NOX

HC, NOX Emission Rate (g/sec)

HC, NOX Emission Rate (g/sec)

CO Emission Rate (g/sec)
CO Emission Rate (g/sec)

Fuel Consumption Rate

Fuel Consumption Rate

Instantaneous Vehicle Speed (km/hr) Instantaneous Vehicle Speed (km/hr)

(a) Petrol Passenger Car (b) Petrol Van

Fuel CO HC NOX Soot Fuel NOX HC Soot

HC, Soot Emission Rates (g/sec)
CO, HC, NOX Soot Emission

CO, NOX Emission Rates (g/sec)

Fuel Consumption Rate (g/sec)

Rates (g/sec)

Instantaneous Vehicle Speed (km/hr) Instantaneous Vehicle Speed (km/hr)

(c) Diesel Van (d) Double-Decker Bus

Figures 4a4d. Emissions and fuel consumption rates (g/sec) against instantaneous vehicle speed.

Volume 50 April 2000 Journal of the Air & Waste Management Association 549
 Tong, Hung, and Cheung

the calculation method of the distance-based emission fac- emission rates fluctuated in a small margin. For the double-
tors (g/km). It can be observed from eqs 1 and 2 that emis- decker bus, the emission and fuel consumption rates
sion and fuel consumption rates (g/sec) were divided by the tended to increase with an increase in vehicle speed. How-
vehicle speed to obtain the emission and fuel consumption ever, larger local fluctuations were observed at vehicle
factors (g/km). Therefore, the emission and fuel consump- speeds higher than 40 km/hr.
tion factors (g/km) would be inversely proportional to the Unlike those of the emission rates and factors, emis-
vehicle speed. sion indices shown in Figure 5, expressed in terms of g/kg
Figure 4 shows the relationship between instanta- fuel, were less sensitive to the instantaneous speed, which
neous speed of the test vehicles and the emissions and agrees with the results from other researchers.18,19 Actu-
fuel consumption rates (g/sec). When the emission and ally, the fuel-based emission index (g/kg fuel) to some ex-
fuel consumption are expressed in terms of g/sec, the tent reflects the fuel combustion efficiency, which is
trends are reversed when compared with those expressed governed by many factors such as vehicle characteristics,
in terms of g/km (i.e., emission factors). In general, higher operating conditions, and fuel type. Therefore, it is diffi-
emission and fuel consumption rates were observed at cult to characterize fuel-based emission indices (g/kg fuel)
higher vehicle speeds, because vehicles have to consume with vehicle speed alone. However, some important points
more fuel to generate enough power and maintain en- can still be extracted from Figure 5. It can be observed
gine operation at higher speeds. that diesel fuel is more efficient than petrol in terms of
Downloaded by [112.134.222.141] at 02:51 07 December 2015

For the petrol-fueled vehicles, the trend of CO and emitting less CO, HC, and NOx. Except for idling, the com-
NOx emission rates as well as the fuel consumption rates bustion efficiency of the diesel van was relatively stable
was generally upward, while the HC emission rates showed at speeds lower than 45 km/hr. The petrol van showed an
an upward convex shape with a maximum in the 3040 increasing efficiency as its speed increased.
km/hr region. The CO, HC, and NOx emission rates, as In brief, instantaneous vehicle speed had the greatest
well as the fuel consumption rates of the diesel van, in- impact on emission and fuel consumption factors, followed
creased as vehicle speed increases, while the soot by emission and fuel consumption rates. The fuel-based

CO HC NOX CO HC NOX

HC, NOX Emission Indices [g/(kg fuel)]

HC, NOX Emission Indices [g/(kg fuel)]
CO Emission Indices [g/(kg fuel)]

CO Emission Indices [g/(kg fuel)]

Instantaneous Vehicle Speed (km/hr) Instantaneous Vehicle Speed (km/hr)

CO HC NOX Soot
CO Emission Indices [g/(kg fuel)]

HC, NOX, Soot Emission

Indices [g/(kg fuel)]

Instantaneous Vehicle Speed (km/hr)

Figures 5a5c. Emissions and fuel consumption indices (g/kg fuel) against instantaneous vehicle speed.

550 Journal of the Air & Waste Management Association Volume 50 April 2000
 Tong, Hung, and Cheung

Table 5. Modal emission and fuel consumption rates.

Driving Mode Modal Emission Rate (mg/sec)

CO HC NOx Soot Fuel

Passenger Car Acceleration 9.54 0.69 0.62 62.62

Cruising 9.15 0.49 0.77 39.10
Deceleration 9.96 0.58 0.69 28.11
Idling 2.99 0.36 0.14 18.11
Petrol Van Acceleration 15.14 1.85 1.96 67.29
Cruising 14.52 1.70 1.81 52.14
Deceleration 17.30 1.91 2.33 52.16
Idling 8.39 1.88 0.81 12.71
Diesel Van Acceleration 2.71 0.65 0.91 0.08 62.02
Cruising 2.64 0.54 0.79 0.07 52.47
Deceleration 2.67 0.65 0.89 0.07 56.01
Idling 1.33 0.22 0.44 0.01 18.52
Downloaded by [112.134.222.141] at 02:51 07 December 2015

Double-Decker Bus Acceleration 12.63 2.36 13.94 0.66

Cruising 8.16 1.36 10.62 0.17
Deceleration 10.92 1.96 11.65 0.31
Idling 7.50 1.15 4.04 0.05

emission indices reflected the fuel combustion efficiency
and varied much less than distance- and time-based emis-
sions as instantaneous speed changed.
Modal Analysis
The modal emission and fuel consumption rates (g/sec), fac-
tors (g/km), and indices (g/kg fuel) are shown in Tables 57.
Table 5 shows the modal emission and fuel consumption
rates. It can be seen that fuel consumption and CO, HC,
NOx, and soot emission rates of the four test vehicles dur-
ing the acceleration mode were comparatively higher than
other driving modes. In the acceleration process, the engine
needs more fuel to generate enough power to accelerate.
The higher the acceleration rate, the more fuel is needed.
Therefore, fuel consumption and emissions increase. Dur-
ing deceleration, the engine does not necessarily gener-
ate power. However, the fuel flow rate cannot be stopped
immediately when acceleration or cruising mode suddenly
changes to deceleration mode. Excess fuel thus continues
flowing at the early phase of deceleration, especially for
hard acceleration/deceleration changes.20 For the passen-
ger car, the fuel consumption rate during deceleration was
lower than the cruising mode. The reason is that the av-
erage deceleration rate during all deceleration phases of
the passenger car is 0.436 m/sec/sec, which is smaller
than those of the other three test vehicles.
In cruising mode, fuel is used to maintain the vehicle
at a certain speed so that emissions and fuel consump-
tion are generally lower than in acceleration and decel-
eration modes. However, the difference in emissions
during cruising and acceleration could be higher than
observed. The cruising NOx emission rate of the passenger
car was even higher than that of the acceleration. In the
urban areas of Hong Kong, signal-controlled intersections
are very close to one another so that each individual cruis-
ing period is short, less than 5 sec, just like low accelera-
tion or deceleration mode. As a result, emission factors
during cruising would be relatively close to those of the
acceleration and deceleration modes.
It is clear that idling emissions and fuel consumption
rates are the lowest. A small amount of fuel is needed to
maintain engine operation; thus, the idling emissions and
fuel consumption rates are significantly lower than those of
the other driving modes. However, differences in HC emis-
sion rates between idling mode and the other driving modes
are comparatively smaller than those in fuel consumption,
as well as CO, NOx, and soot emissions. This may be caused
by evaporation of HCs in the unburned fuel.
Table 6 shows the distance-based modal emission fac-
tors. As they are distance-based, idling emission factors
could not be calculated. The comparisons of modal emis-
sion and fuel consumption factors are almost similar to
the modal emission and fuel consumption rates. However,
the double-decker bus had significantly larger cruising emis-
sion factors than those of the acceleration and decelera-
tion modes, because the average speed during cruising mode
was 11 km/hr while those of the acceleration and decelera-
tion modes were 21 and 20 km/hr, respectively. It has been
demonstrated that average speed can explain up to 70% of
distance-based emissions.5,21-23 The lower the average speed,
the higher the emissions. Therefore, combined with the
effect of the short cruising periods described earlier, the

Volume 50 April 2000 Journal of the Air & Waste Management Association 551
 Tong, Hung, and Cheung

Table 6. Modal emission and fuel consumption factors.

Driving Mode Modal Emission Factors (g/km)

CO HC NOx Soot Fuel

Passenger Car Acceleration 14.41 1.04 0.94 94.58

Cruising 11.68 0.63 0.98 49.92
Deceleration 14.40 0.84 0.99 40.66
Petrol Van Acceleration 17.38 2.13 2.25 77.28
Cruising 17.58 2.05 2.19 63.15
Deceleration 18.68 2.07 2.52 56.32
Diesel Van Acceleration 3.48 0.83 1.17 0.11 79.77
Cruising 3.70 0.75 1.11 0.09 73.53
Deceleration 3.55 0.86 1.18 0.10 74.38
Double-Decker Bus Acceleration 21.01 3.93 23.20 1.10
Cruising 26.91 4.50 35.06 0.56
Deceleration 20.17 3.62 21.52 0.58
Downloaded by [112.134.222.141] at 02:51 07 December 2015

cruising emission factors of the double-decker bus are larger the distance- and time-based emissions, meaning that the
than those of the acceleration and deceleration modes. highest fuel combustion efficiency occurred during the
Table 7 shows the fuel-based modal emission indices. acceleration mode. The reason might be that the higher
The emission indices of the diesel van were significantly engine temperature during acceleration facilitates the fuel
smaller than those of the petrol vehicles. As discussed combustion process.
earlier, the diesel van was more fuel-efficient than the
petrol vehicles. Therefore, the fuel-based emission indi- CONCLUSIONS
ces should have been smaller than those of the petrol ve- Four instrumented vehicles were used to conduct on-road
hicles. It can be observed that idling emission indices were emission and fuel consumption tests. Instantaneous emission
basically higher than those of the acceleration and decel- concentrations were measured directly from the tailpipe of
eration modes for the petrol van and diesel van. the on-road driving test vehicles. Emission and fuel consump-
However, low idling emission indices were observed for tion rates (g/sec) and factors (g/km) as well as indices (g/kg
the passenger car, possibly because of the catalytic con- fuel) were then derived and analyzed at microscopic (i.e.,
verter. During idling, a passenger car needs more fuel to emission and fuel consumption against instantaneous vehicle
maintain engine and catalytic converter operation and speed plots) and macroscopic (i.e., global and modal analy-
thus decreases the emission indices. sis) levels. Average factors and indices calculated over the whole
Meanwhile, the smallest emission indices were ob- data sample for specific vehicles were obtained. It was found
served for the acceleration mode, which was different from that the double-decker bus had the largest emission factors.
Modal emission and fuel consump-
Table 7. Modal emission indices. tion rates and factors as well as indices
for each standard driving mode (i.e., ac-
Driving Mode Modal Emission Index (g/kg fuel) celeration, cruising, deceleration, and
CO HC NOx Soot idling) were also derived. It was found
that the transient driving modes (i.e., ac-
Passenger Car Acceleration 152.37 10.97 9.94 celeration and deceleration) were signifi-
Cruising 234.01 12.55 19.68 cantly more polluting than steady-speed
Deceleration 354.28 20.59 24.46 driving modes (i.e., cruising and idling)
Idling 164.89 19.70 7.72 in terms of g/km and g/sec, except for
Petrol Van Acceleration 224.91 27.51 29.15 the bus. But the difference in emissions
Cruising 278.43 32.52 34.70 for cruising and acceleration modes was
Deceleration 331.62 36.69 44.67 smaller than those established in labora-
Idling 660.31 147.77 64.09
tory tests, due to the particular driving
Diesel Van Acceleration 43.68 10.46 14.61 1.36
conditions in the urban areas of Hong
Cruising 50.36 10.23 15.12 1.28
Kong, where congestion time is long and
Deceleration 47.67 11.58 15.92 1.30
thus acceleration and deceleration peri-
Idling 71.59 11.79 23.89 0.80
ods are comparatively longer than in U.S.
and European cities.

552 Journal of the Air & Waste Management Association Volume 50 April 2000

Microscopic analysis indicated that instantaneous ve-
hicle speed had the greatest impact on the distance-based
emission and fuel consumption factors, followed by the
time-based emission and fuel consumption rates. The fuel-
based emission and fuel consumption indices to some ex-
tent reflected the fuel combustion efficiency and varied
much less than distance- and time-based emissions and
fuel consumption as instantaneous speed changed. The
trends of the emission and fuel consumption rates, fac-
tors, and indices as instantaneous speed varied are basi-
cally the same as those established in the laboratory test.
The speed distribution showed that idling and low
speed ranges occupied a large proportion of total driving
time in urban areas. However, microscopic analysis dem-
onstrated that low speed ranges have high emission fac-
tors. Moreover, modal analysis indicated that idle emission
indices were higher than those of the other driving modes
Downloaded by [112.134.222.141] at 02:51 07 December 2015
as well as the average emission indices for several orders
of magnitude. In other words, idling and low speed driv-
ing contributed to a high percentage of total emissions
over a vehicle trip, again due to the heavy traffic and driv-
ing conditions in the urban areas of Hong Kong.
A limitation of this study was the use of a single test
vehicle for each category, but the results highlighted gen-
eral emission behaviors under actual on-road driving con-
ditions in urban areas. The results indicated that on-road
driving emission measurement is feasible. To achieve more
definitive results, more instrumented vehicles could be
used to measure on-road emission rates and vehicle oper-
ating parameters.
ACKNOWLEDGMENTS
The work described in this paper was supported by the
research fellowship of the Department of Civil and Struc-
tural Engineering, The Hong Kong Polytechnic Univer-
sity, and a grant from the Research Grants Council of the
Hong Kong Special Administrative Region (Project No.
PolyU141/96E).
REFERENCES
1. Chang, A. Air Quality in Hong Kong 1997; EPD/TR1/98; Air Services
Group, Hong Kong Environmental Protection Department: Hong
Kong, 1997.
2. Joumard, R.; Jost, P.; Hickman, J. Influence of Instantaneous Speed
and Acceleration on Hot Passenger Car Emissions and Fuel Consump-
tion; SAE Technical Paper 950928; Society of Automotive Engineers:
Warrendale, PA, 1995.
3. Rapone M.; Ragione, L.D.; DAniello, F.; Luzar, V. Experimental Evalu-
ation of Fuel Consumption and Emissions in Congested Urban Traf-
fic; SAE Technical Paper 952401; Society of Automotive Engineers:
Warrendale, PA, 1995.
4. St. Denis, M.J.; Cicero-Fernandez, P.; Winer, A.M.; Butler, J.W.; Jesion,
G. Effects of In-Use Driving Conditions and Vehicle/Engine Operat-
ing Parameters on Off-Cycle Events: Comparison with Federal Test
Procedure Conditions; J. Air & Waste Manage. Assoc. 1994, 44, 31-38.
5. Cheung, C.S.; Zhao, H.; Hung, W.T. Determination of Emission Fac-
tors of Motor Vehicles from On-Road Emissions Measurement; Trans.
CSICE 1999, 17, 47-52.
Volume 50 April 2000
Tong, Hung, and Cheung
6. Foss, P.W. Development of an On-Road Vehicle Emissions Labora-
tory. In Proceedings of FISITA 92, The Vehicle and the Environment, Vol.
2; Institute of Mechanical Engineers: London, 1992; pp 85-87.
7. Cicero-Fernandez, P.; Long, J.R; Winer, A.M. Effects of Grades and
Other Loads on On-Road Emissions of Hydrocarbons and Carbon
Monoxide; J. Air & Waste Manage. Assoc. 1997, 47, 898-904.
8. Tong, H.Y.; Hung, W.T.; Cheung, C.S. Development of a Driving Cycle
for Hong Kong; Atmos. Environ. 1999, 33, 2323-2335.
9. Hong Kong Transport Department Annual Traffic Census 1997; Hong
Kong Government: Hong Kong, 1997.
10. Test Procedure for the Measurement of Gaseous Exhaust Emissions from
Small Utility Engines; SAE J1088 FEB93, 1996 SAE Handbook, Vol. 1:
Materials, Fuels, Emissions and Noise; Society of Automotive Engi-
neers: Warrendale, PA, 1996.
11. Diesel Engine Smoke Measurement; SAE J255 FEB93, 1996 SAE Hand-
book, Vol.1: Materials, Fuels, Emissions and Noise; Society of Auto-
motive Engineers: Warrendale, PA, 1996.
12. Biggs, D.C.; Akcelik, R. An Energy-Related Model of Instantaneous
Fuel Consumption; Traffic Eng. Control 1986, 27, 320-325.
13. Biggs, D.C.; Akcelik, R. Models for Estimation of Car Fuel Consump-
tion in Urban Traffic; ITE J. 1986, 56(7), 29-32.
14. Bowyer, D.P.; Akcelik, R.; Biggs, D.C. Guide to Fuel Consumption Analy-
sis; Special Report No. 32; Australian Road Research Board:
Nunawading, Victoria, Australia, 1985.
15. Bowyer, D.P.; Akcelik, R.; Biggs, D.C. Fuel Consumption Analyses for
Urban Traffic Management; ITE J. 1986, 56(12), 31-34.
16. Cernuschi, S.; Guigliano, M.; Cemin, A.; Giovannini, I. Modal Analy-
sis of Vehicle Emission Factors; Sci. Total Environ. 1995, 169, 175-183.
17. Saleh, W.; Nelson, J.D. Modelling Vehicle Emissions Using the TEMIS
Program, Part 2: Case Studies. In Proceedings of the Institute of Me-
chanical Engineers, Part A, J. Power & Energy 1998, 212, 205-212.
18. Singer, B.C.; Harley, R.A. A Fuel-Based Motor Vehicle Emission In-
ventory; J. Air & Waste Manage. Assoc. 1996, 46, 581-593.
19. Singer, B.C.; Kirchstetter, T.W.; Harley, R.A.; Kendall, G.R.; Hesson,
J.M. A Fuel-Based Approach to Estimating Motor Vehicle Cold-Start
Emissions; J. Air & Waste Manage. Assoc. 1999, 49, 125-135.
20. Carlock, M.A. Laboratory Tests of Modal Emission and Off Cycle Cor-
rection to FTP-75. In Proceedings of the National Conference on Trans-
portation Planning and Air Quality, American Society of Civil Engineers,
New York, 1992; pp 211-218.
21. Alimoradian, B. Transient Simulation of Vehicle Fuel Consumption
and Emissions, Ph.D. Thesis, University of Melbourne, Melbourne,
Australia, 1986.
22. Andre, M.; Pronello, C. Relative Influence of Acceleration and Speed
on Emissions Under Actual Driving Conditions; Int. J. Vehicle Design
1997, 18, 340-353.
23. Zhao, H.; Cheung, C.S.; Hung, W.T. Investigation of Motor Vehicle
On-Road Exhaust Emissions Behavior; J. Combust. Sci. Technol. 1999,
5, 100-107.
APPENDIX
Calculation of Mass Emissions in g/sec
Petrol Vehicles. The mass emission of CO, HC, and NOx
for the petrol vehicles in g/sec were calculated by the fol-
lowing equations:
where MCO is the molecular weight of CO, 28.01; MHC is
the molecular weight of the exhaust HCs (based on 1 C
atom), 12.011 + 1.008y + 15.999z; y is the HC atomic ra-
tio in the fuel (assuming the exhaust HC = fuel); z is the
Journal of the Air & Waste Management Association 553
 Tong, Hung, and Cheung

O/C atomic ratio in the fuel (assuming the exhaust HC = [(14.007 + 15.999) + (14.0078 + 15.999 x 2) 0.1] / 1.1 =
fuel); MNOx is the molecular weight of NOx, 31.46; Pambient is the ambient air pressure; Tambient is the am-
[(14.007 + 15.999) + (14.0078 + 15.999 2) 0.1] / 1.1 = 31.46; bient air temperature, 300 K; and Ru is the universal gas
Mexhaust is the molecular weight of the exhaust constant, 8314.2 J/mol-K.
(13.88 HCppm 10-6) + (28.01 CO% 10-2) + (44.01
CO2% 10-2) + (46.01 x NOxppm 10-6) + (32.00 O2% 10-2) +
(2.016 H2% x 10-2) + 18.01 (1K) +
HCppm NOx ppm 28.01 ;
100 10 4 CO% CO2% 10 4 O2% ( H2% ) 100 (1 K ) 100

K is [1 + 0.005 (CO%+ CO2%) y 0.01 H2%]-1 ; and H2%
0.5 y CO% (CO% + CO2% )
is
CO% + 3 CO2%

Diesel Vehicles. The mass emissions of CO, HC, and NOx

for the diesel vehicles in g/sec were calculated by the fol-
lowing equations:

(m )
Downloaded by [112.134.222.141] at 02:51 07 December 2015

+ m fuel Pambient MCO CO% 10 2

COg / sec =
air
About the Authors
Tembient Ru Hing-Yan Tong is a doctoral candidate in Civil Engineering
in the Department of Civil and Structural Engineering De-

(m )
partment at Hong Kong Polytechnic University. He can be
+ m fuel Pambient MHC HCppm 10 6 contacted by regular mail at Department of Civil and Struc-
=
air
HCg / sec tural Engineering, Hong Kong Polytechnic University, Hung
Tambient Ru Hom, Hong Kong SAR; or via e-mail at 97980212r@
polyu.edu.hk. Dr. Wing-Tat Hung is currently an assistant
professor in the Department of Civil and Structural Engi-

NOx g / sec =
(m air )
+ m fuel Pambient MNO
NOxx
NOx
x ppm 10
6 neering, Hong Kong Polytechnic University. He can be con-
tacted by regular mail at Department of Civil and Structural
Tambient Ru Engineering, Hong Kong Polytechnic University, Hung Hom,
Hong Kong SAR; or via e-mail at cewthung@polyu.edu.hk.
Dr. Cheung Chun-shun is currently an assistant professor
where MCO is the molecular weight of CO, 28.01; MHC is
in the Department of Mechanical Engineering, Hong Kong
the molecular weight of the exhaust HCs (based on 1 C Polytechnic University. He can be contacted by regular mail
atom), 12.011 + 1.008y + 15.999z; y is the HC atomic ra- at Department of Mechanical Engineering, Hong Kong Poly-
tio in the fuel (assuming the exhaust HC = fuel); z is the technic University, Hung Hom, Hong Kong SAR; or via e-
O/C atomic ratio in the fuel (assuming the exhaust HC = mail at mmcsc@polyu.edu.hk.
fuel); MNOx is the molecular weight of NOx,

554 Journal of the Air & Waste Management Association Volume 50 April 2000