电动汽车用MH/Ni电池剩余容量智能预测研究  

Study of Intelligent Prediction of the SOC of MH/Ni Battery for Electric Vehicle

作　　者：周红丽[1] 何莉萍[1] 钟志华[1] 丁舟波[1] 高学锋 廖晓军[1] 吴振军[1] 陈宗璋[1] 

机构地区：[1]湖南大学,湖南长沙410082 [2]深圳市力可兴电池有限公司,广东深圳518040

出　　处：《机械与电子》2006年第10期7-10,共4页Machinery & Electronics

基　　金：二一一"电动汽车研发项目(900215180)

摘　　要：为有效地对电动车电池剩余容量进行预测,在分析了自适应模糊神经推理系统(ANFIS)的网络结构后,利用湖南大学自主研发的EV3电动汽车充放电实验数据,建立了MH/Ni电池的AN-FIS电压降模型,并进一步应用改进后的尺度化共轭梯度算法的BP神经网络,得到开路电压与剩余容量的对应关系.实验和仿真结果表明,智能方法在预测电池剩余容量方面,能满足精度要求,模糊控制与神经网络的结合,可以建立一个精确而有效的SOC智能预测系统.In order to predict the SOC of batteries for electric vehicles accurately,a voltage descent model has been built up based on ANFIS analysis and the charge/discharge experimental data of EV -3 developed by Hunan University. Then a BP neural network has also been proposed with the ＂trainscg＂ mathematical method. The BP neural network indicates the correlation between open circuit voltage and SOC. Tests and simulation results show that the intelligent methods can accurately predict the SOC of MH/Ni batteries. The combination of fuzzy control and neural network can achieve an effective way of predicting the SOC of batteries.

关 键 词：电动汽车 剩余电量预测 电压 自适应模糊神经推理系统 BP网络 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]