Abstract:
Vehicular fuel consumption is a growing energy, environmental and economic concern for
countries around the globe. It is predominantly affected by factors related to road and traffic
conditions. The fuel consumption issue is more magnified in urban areas like Addis Ababa due
to the existing congested and interrupted traffic operating conditions. However, the problem
is less studied at the local level due to the absence of models well confirmed with the local
operating condition. Therefore, the purpose of this study is to calibrate and develop fuel
consumption models for passenger cars under urban interrupted traffic operating conditions
in Addis Ababa city. The study was entirely based on vehicle performance, driving cycle, and
trajectory data collected using an OBD-II scanner mounted on a test vehicle. The study also
utilized road condition and specification data obtained from a field test and secondary sources.
As a part of analyses, the study comprised characterization of the local driving cycle,
calibration of HDM-4 fuel consumption model, and development of regression models. The
local driving cycle was characterized by average driving speed, positive acceleration, and
frequency of complete stop of 20.07 km/hr, 0.28 m/sec2,
and 3.54 stops/km, respectively. It was
also described by a proportion of idle, cruise, and acceleration time of 33.20, 11.26, and
29.12%, respectively. From the calibration analysis, calibration factors of Kcr2=0.320,
Ki=0.551, and Kpea=0.678 were found to adjust the over-prediction of the model. The
calibrated model yielded an overall prediction accuracy of RMSE=29.55 mL/km, MAE=19.77
mL/km, and a model bias of PBIAS=1.32%. The sensitivity analysis based on the calibrated
model also showed that a unit change on road grade (1.0%) and IRI (1.0m/km) at 40 km/hr
increase fuel consumption by 24.24 and 1.5%, respectively. At last, the impact of traffic on fuel
consumption was studied using a linear regression model of idle rate, stop rate, driving speed,
and PKE of having an adjusted R
2
value of 94.0%. The study also unveiled that vehicular fuel
consumption in urban roads is 104.72 and 44.75% sensitive for 5.0 km/hr speed reduction and
0.1 m/s2
increments of PKE, respectively. Finally, it can be concluded that road grade,
pavement roughness, and traffic condition are the most influential factors affecting vehicular
fuel consumption. Therefore, calibration factors and models resulted from this study can be
reliably approaches to investigate these factors and implement energy-efficient measures.
