| Summary: | The purpose of this research is to develop a representative driving cycle for fuel cell logistics vehicles running on the roads of Guangdong Province for subsequent energy management research and control system optimization. Firstly, we collected and preliminarily screened the 42-day driving data of a logistics vehicle through the remote monitoring platform, and determined the vehicle characteristic signal vector for analysis. Secondly, the principal component analysis method is used to reduce the dimensionality of these characteristic parameters, avoiding the linear correlation between them and increase the comprehensiveness of the upcoming clustering. Next, the dimensionality-reduced data are fed to a clustering machine. K-means clustering method is used to gather the segmented road sections into highway, urban road, national highway and others. Finally, several segments are chosen in accordance to the occurrence possibility of the four types of road conditions, minimizing the deviation with the original data. By joining the segments and using a moving average filtering window, a typical driving cycle for this fuel cell logistics vehicle on a fixed route is constructed. Some statistical methods are done to validate the driving cycle.The effectiveness analysis shows the driving cycle we constructed has a high degree of overlap with the original data. This positive result provides a solid foundation for our follow-up research, and we can also apply this method to develop other urban driving cycles of fuel cell logistics vehicle.
|
|---|