This content is not included in your SAE MOBILUS subscription, or you are not logged in.
Analyzing Fuel Savings of an Aerodynamic Drag Reduction Device with the Aid of a Robust Linear Least Squares Method
ISSN: 1946-391X, e-ISSN: 1946-3928
Published September 30, 2014 by SAE International in United States
Citation: Van der Krieke, J. and Van Raemdonck, G., "Analyzing Fuel Savings of an Aerodynamic Drag Reduction Device with the Aid of a Robust Linear Least Squares Method," SAE Int. J. Commer. Veh. 7(2):675-684, 2014, https://doi.org/10.4271/2014-01-2450.
Improving the aerodynamic drag level of semi-trailers will contribute largely to reduce the fuel consumption and the emissions of harmful gases of heavy duty vehicles. The final step in product validation of aerodynamic drag reduction devices is often conducting fuel savings test during operational activities.
During an operational test, data is gathered for a period when the vehicle is not equipped with an aerodynamic device and consequently for a period with the device equipped. A simple fuel consumption comparison between the periods does not give the desired accurate result as the operating conditions are different for the control and test period. In an attempt to take these varying conditions into account, the average fuel consumption per ride is modeled as a linear function of several independent variables: the wind conditions, the outside temperature, the humidity, the payload, the road inclination and the presence of the drag reduction device. The coefficients in this linear relationship are determined using a robust linear least squares algorithm.
Applying this approach for a single vehicle test resulted in the finding of a fuel savings of 1.51 l/100 km. In contrast, a simple comparison of the fuel consumption during the control and test periods showed a fuel saving of 1.1 l/100 km when highway speeds are considered.
Residuals show little dependence on the independent variables, demonstrating the model fit is fairly good. Still the analysis could be improved if a more sophisticated model for fuel consumption is used, incorporating more relevant independent variables and more accurate (perhaps non-linear) relationships. Such a model could also contribute to more reliable results in one-day track tests.