【摘要】 针对单一的BP神经网络在进行锂离子电池的荷电状态(state of charge,SOC)估计时存在估计精度不高的问题，提出了一种基于改进灰狼优化算法(improved grey wolf optimization algorithm,IGWO)的BP神经网络来估计锂离子电池SOC的方法。通过采用改进的灰狼优化算法来优化BP神经网络的权值和阈值，来克服单一的BP神经网络容易陷入局部最优的缺陷，并且加快了收敛速度。经仿真实验表明，BP神经网络估计锂电池SOC的平均绝对误差为6.39%，而基于IGWO-BP神经网络估计锂电池SOC的平均绝对误差为1.45%，相比于单一的BP神经网络，平均绝对误差减少了4.94%，证明提出的算法提高了SOC的估计精度。更多还原

【Abstract】 Aiming at the problem of low estimation accuracy when estimating the SOC(State of charge, SOC) of lithium ion batteries with a single BP neural network, a new method based on the Improved Grey Wolf Optimization Algorithm(IGWO) was proposed. A method for estimating the SOC of lithium ion batteries using BP neural network.By adopting the improved gray wolf optimization algorithm to optimize the weights and thresholds of the BP neural network, the defect that a single BP neural network is easy to fall into the local optimum is overcome, and the convergence speed is accelerated. The simulation experiments show that the average absolute error of BP neural network estimation of lithium battery SOC is 6.39%, and the average absolute error of lithium battery SOC estimation based on IGWO-BP neural network is 1.45%, compared with the average absolute error of single BP neural network. It is reduced by 4.94%, indicating that the algorithm proposed effectively improves the estimation accuracy of SOC.更多还原