基于PSO-SVM的电动汽车电池SOC估算方法

SOC estimation method of electric vehicle battery based on PSO-SVM

针对常用的SOC估算方法依赖于所建立电池组模型的精确性而没有考虑环境温度对SOC值的影响,基于SVM模型建立相应的电池组SOC估算模型,并用PSO算法优化SVM模型的参数。仿真实验表明,SVM模型的估算效果要优于BP神经网络模型;在对SVM的参数整定中,PSO算法优于网格搜索法。

The commonly used methods for SOC estimation, which were depended on the accuracy of the model of battery pack, and without considering the effects of temperature, so a model based on SVM to estimate the SOC of the battery pack of Electric Vehicle was established. And the parameters of SVM model was optimized by using PSO algorithm. The results of simulation shows that the estimate accuracy of SVM model is superior to the BP neural network model. PSO was better than the commonly used grid search method in terms of tuning parameters of SVM.