Real-time energy management based on aging- and temperature-conscious MPC for hybrid electric trains

Abstract

To fully exploit the potential of the Hybrid Energy Storage System (HESS) in urban hybrid electric trains, a real-time energy management strategy (EMS) based on Aging- and Temperature-Conscious Model Predictive Control (ATC-MPC) is proposed to tackle challenges associated with lithium-ion battery (Li-B) aging and thermal issues. First, an aging model based on Dynamic Equivalent Ampere-hour Throughput (DEAT) is proposed for Li-B. Furthermore, a control-oriented electro-thermal-aging coupled (ETAC) model is developed to characterize the real-time dynamic interactions among the electrical, thermal, and aging states of Li-B. Subsequently, a mechanistic-LSTM enhanced power demand prediction method is proposed to improve the prediction accuracy under dynamic operating conditions. Moreover, the power distribution of the HESS is structured as a rolling optimization problem within the MPC prediction horizon, solved using the Sequential Quadratic Programming (SQP) method. Within the MPC framework, a comprehensive objective function incorporating aging and temperature consciousness is proposed, with penalty factors defined for Li-B temperature rise and supercapacitor (SC) state of charge (SOC). Finally, simulation tests conducted on the hardware-in-the-loop (HIL) platform validate the efficacy of the proposed strategy.