Adaptive energy management strategy based on a model predictive control with real-time tuning weight for hybrid energy storage system

Abstract

The effective utilization of ultra-capacitor (UC) in the energy allocation process is crucial for improving the efficiency of the energy management strategy (EMS) for hybrid energy storage system (HESS). In this paper, we present an adaptive energy management strategy framework based on a model predictive control (MPC) with real-time tuning weight to optimize UC utilization and extend battery lifetime. Firstly, the adaptive autoregressive integrated moving average model (AARIMA) is developed to accurately predict vehicle velocity and road gradient for online power demand computation, based on local historical information. The differencing order and lags of the model in AARIMA are updated using the augmented Dickey-Fuller (ADF) test and Bayesian Information Criterion (BIC). Secondly, the MPC is developed to optimize power allocation using quadratic programming. Finally, the weight in the cost function is real-time tuned via the fuzzy logic rule according to the predicted power. Furthermore, the proposed strategy is validated using a scale-down hardware-in-the-loop verification platform. Simulation results demonstrate that the UC utilization ratio has a 9.91% improvement, and the battery lifetime has a prolonged by 24.38% across two tested actual and two typical driving cycles. Consequently, the proposed EMS achieves optimization of both energy management efficiency and battery lifetime.