Energy Management Strategy for PHEV Based on Condition Recognition

As for the disadvantage of poor adaptability of equivalent fuel consumption minimum strategy (ECMS) under different working conditions, the energy management strategy of variable equivalent factor ECMS based on condition recognition algorithm is designed to improve fuel economy of parallel hybrid electric vehicles (PHEV). The vehicle equivalent fuel consumption is optimized as the optimization objective. Three typical working conditions were selected to establish the SVM classification model, the sample features were selected by recursive feature elimination method, and the whale algorithm was used to optimize the parameters of the SVM. The simulated annealing algorithm was used to solve the ECMS equivalent factors of the three types of working conditions for offline global optimal solution, and were stored in the equivalent factor library respectively. The target optimized working conditions were identified by the trained support vector machine classifier. Different types of working conditions were treated with different equivalent factors for torque distribution. Compared with the logic threshold energy management strategy, the variable equivalent factor ECMS energy management strategy based on the condition recognition algorithm reduced the variation of State of charge (SOC) by 8.67%, and the fuel saving rate by 13.11%. Compared with the ECMS strategy before optimization, the SOC variation of battery was reduced by 3.47%, and the fuel saving rate was about 6.63%. The variable equivalent factor ECMS energy management strategy based on the driving condition recognition algorithm can effectively reduce the fuel consumption and improve the economy of the PHEV.