基于布谷鸟搜索优化神经网络的锂电池荷电状态预测

Lithium battery charge status prediction based on cuckoo search optimization neural networks

摘要:

锂电池荷电状态(SOC)的预测是电动汽车锂电池管理系统中最为关键的技术之一。为实现对SOC的高精度的预测，提出了一种基于布谷鸟搜索算法（CS）优化的误差反向传播（BP）神经网络的锂电池SOC预测方法，该方法的核心也是难点之一，在于优化BP神经网络的初始权值和阈值，从而可以改善易陷入局部最优的情况，减小算法对初始值的依赖。MATLAB仿真结果表明，CS-BP神经网络算法的均方根误差值比BP算法的均方根误差值平均降低了0.0106，CS-BP算法具有更高的预测精度和极强的泛化性能。

Abstract:

The prediction of lithium battery charge status (SOC) is one of the most critical technologies in lithium battery management system of electric vehicles. In order to realize the high-precision prediction of SOC, a lithium battery SOC prediction method based on Cuckoo search algorithm (CS) optimization is proposed, the core of which is to optimize the initial weight and threshold of BP neural network, so as to overcome the disadvantages of local optimality and reduce the algorithm's dependence on the initial value. The MATLAB simulation results show that the average square root error value of CS-BP neural network algorithm is 0.0106 lower than that of BP algorithm, and the CS-BP algorithm has better prediction accuracy and generalization performance.