基于粒子滤波算法的车载储能元件SOH预测方法研究

Research on SOH prediction method of vehicle energy storage element based on particle filter algorithm

储能元件作为无接触网供电城市轨道交通车辆的重要组成元件,其健康状态(SOH)直接决定着列车是否能够安全可靠的运行。将由大量数据拟合得到的双指数经验退化模型和适用于解决非线性复杂系统的粒子滤波算法相结合,对车载储能元件蓄电池SOH预测。结果表明,该算法虽能较好地跟踪蓄电池的容量退化过程,但其精度有待提高,因此,引入遗传算法,利用遗传算法产生新粒子的优势提高预测结果的精度,根据实验结果,遗传粒子滤波算法能更精确地预测出电池健康状态。

As an important component of power supply without overhead contact system for urban rail transit vehicles,energy storage components’state of health(SOH)directly determines whether the train can run safely and reliably.In this paper,the SOH prediction of vehicle-mounted storage energy component storage battery was made by combining the double-exponential empirical degradation model with a large amount of data matching and the particle filter algorithm suitable for solving nonlinear complex systems.The results show that although the algorithm can better track the degradation process of battery capacity,its accuracy needs to be improved.Therefore,the genetic algorithm is introduced to improve the accuracy of prediction results by utilizing the advantage of the genetic algorithm to generate new particles.According to the experimental results,the genetic particle filter algorithm can predict the battery health more accurately.