Driving cycle construction and combined driving cycle prediction for fuzzy energy management of electric vehicles

Energy management strategies (EMSs) play an important role in battery electric vehicles (BEVs). However, the efficiency of an EMS is significantly affected by the driving cycle (DC). On the one hand, because of the differences in driving type, torque characteristics, and speed response of BEVs differ from those of internal combustion-powered vehicles. Meanwhile, on the other hand, typical DCs that are widely used as evaluation indexes cannot reflect changes in the road slope and the driving characteristics of BEVs in a specific city. To solve this problem, a novel EMS based on combined DC prediction (CDCP) is proposed, and three efforts are made. First, a large volume of driving data is collected based on a BEV. The DC for a specific city is constructed using the principal component analysis andK-means cluster algorithm combined with 15 characteristic parameters of speed, acceleration, slope, and running state rate. Second, CDCP is adopted to overcome the disadvantage that the speed cannot be predicted initially using standard rolling prediction. The state transmission matrix based on the constructed DC is employed to predict the speed when the vehicle starts to run. Third, a fuzzy EMS that considers the CDCP is proposed in order to adapt to the real-time changes of DCs. Compared with the other predictions, the simulation results show that the proposed EMS has a longer driving range and lower energy consumption rate.

Automated summary

A novel EMS based on combined DC prediction (CDCP) was proposed to adapt to real-time changes of the driving cycle (DC), utilizing driving data collection, construction of city-specific DC, and adoption of CDCP and fuzzy EMS. Simulation results show that the proposed EMS outperforms other predictions with longer driving range and lower energy consumption rate.