摘要
研究动力电池剩余电量问题,电池剩余电量的准确预测一直以来都是一个难点。由于剩余电量与电流、内阻和温度有关,电池内部复杂的电化学反应导致了电池电压不能线性反映电池的剩余电量,传统的方法往往会有较大的误差。BP神经网络的特点是可以逼近任意的非线性函数,而BP神经网络并非完美的神经网络,采用用遗传算法优化BP神经网络可以克服其缺点,更好的预测电池的剩余电量。实验结果证明,所用的GA-BP神经网络方法具有反应快,误差小的特点,达到了预先设计的目的。
Prediction of remaining battery power has always been a difficulty.The complex electrochemical reaction inside the battery leads to fact that battery voltage does not linearly reflect the remaining battery power,and the traditional methods often have large errors.BP neural network can approach any nonlinear function,however,BP neural network is not a perfect neural network.Using genetic algorithm BP neural network can overcome its shortcomings and better predict the remaining battery capacity.Experiments and data simulation results show that the designed GA-BP neural network has fast response characteristics with small error.
出处
《计算机仿真》
CSCD
北大核心
2011年第12期323-326,334,共5页
Computer Simulation
关键词
遗传算法
神经网络
电动汽车
电池剩余电量
预测
Genetic algorithms
Neural network
Electric cars
Residual capacity for Battery
Prediction