Sizing Optimization and Energy Management Strategy for Hybrid Energy Storage System Using Multiobjective Optimization and Random Forests

Abstract

Sizing optimization and energy management strategy (EMS) are two key points for the application of the hybrid energy storage system (HESS) in electric vehicles. This article aims to conduct the sizing optimization of HESS and apply an adaptive real-time EMS for practice. First, considering the system cost and battery lifespan, the multiobjective grey wolf optimizer is used to obtain the Pareto front. Second, with optimal parameters, the offline optimal power splitting results by dynamic programming (DP) under different driving patterns are analyzed. Then, the random forests (RF) method is used to learn control rules from the DP results. Driving pattern recognition (DPR) is implemented by the support vector machine (SVM). The intelligent EMS is composed of RF to guide power distribution and SVM to realize DPR. Finally, a combined load cycle involving different driving patterns is used for verification. Results illustrate that the proposed adaptive RF-based EMS can demonstrate a notable superiority in terms of battery protection, ultra-capacitor utilization, and system efficiency. Compared with the ordinary RF-based EMS without DPR, the proposed method can reduce total energy loss by 0.74%–9.49%, and reduce the battery Ah-throughput by 0.5%–19.83% under unknown driving cycles.