Development of battery management system for hybrid electric propulsion system.

Because of the high overall efficiency and low emissions, Hybrid Electric Propulsion System (HEPS) have become an attractive research area. In this research, a parallel HEPS architecture is adopted and a Hardware test platform is constructed. As a relative new power source in powertrains, battery system plays an important role in HEPS. Hence, a Battery Management System (BMS) is investigated in this research. Battery pack State of Charge (SOC) is a key feedback value in HEPS control. In order to estimate SOC, firstly, an operation-classification adaptive battery model is proposed for Li-Po batteries. Considering the fact that model parameter accuracy is of importance in model-based system state estimation method, an event triggered Adaptive Genetic Algorithm (AGA) is applied for online parameter identification. Secondly, the Extended Kalman Filter (EKF) is applied for single battery cell SOC estimation. Finally, a fuzzy estimator is proposed for battery pack SOC estimation based on maximum/minimum cell voltages and SOC values. Experimental results show that the proposed AGA can effectively track battery parameter variation and SOC estimation error for single cell as well as for the battery pack are both less than 1%. Moreover, considering the Li-Po battery characteristics, a converter based battery cell balancing method is proposed. Simulation result shows that proposed balancing method can be effective in balancing battery cells. In addition, in relation to safety and reliability concerns, a Discrete Wavelet Transform (DWT) based battery circuit detection method is proposed and simulation results showing its feasibility are presented.