Adaptive energy management strategy for hybrid electric vehicle based on power prediction

Abstract

The advantages of hybrid electric vehicle (HEV) in energy-saving strongly rely on the energy management strategy. Equivalent consumption minimization strategy (ECMS) is one kind of well-known instantaneous strategies with the capability of online application. Due to the complicated driving conditions, there are some limitations for the application of ECMS with fixed equivalence factor (EF). Therefore, the adaptive ECMS (A-ECMS) based on the average predicted power is proposed and studied. The average power is predicted by a polynomial function with the average velocity and average acceleration as prediction inputs. The coefficients of the prediction model are updated by the recursive least squares algorithm. By investigating the impacts of different power requirements and EF on the battery charging/discharging behaviors, the explicit expression model of adaptive EF based on the average predicted power and battery SOC is established. The particle swarm optimization (PSO) method is further adopted to derive the parameters of the explicit expression model. Simulation results demonstrate that when the average power is higher than 0.5 kW, the prediction accuracy of the proposed power prediction model is higher than 91%. Compared with the constant ECMS which uses the corresponding optimal EF under different test cycles, although the equivalent fuel economy with the proposed A-ECMS slightly deteriorates, most of the deterioration is smaller than 5%, while the battery charging sustainability for all test cycles are well maintained. The proposed strategy ensures good cycle adaptability in achieving good fuel economy and battery charging sustainability.