基于混合ISSA-LSTM的锂离子电池剩余使用寿命预测

准确预测锂离子电池剩余使用寿命(remaining useful life, RUL)对降低电池使用风险和维护设备稳定性方面具有重要意义。为了提高锂离子电池RUL预测的稳定性和结果的准确性,提出一种基于混合改进麻雀搜索算法(improved sparrow search algorithm, ISSA)与长短期记忆(long short-term memory, LSTM)神经网络的锂电池RUL预测模型。首先,用均值化方法对原始数据中的异常值进行处理。然后,结合Tent混沌映射、自适应权重以及反向学习策略和柯西变异扰动策略优化麻雀搜索算法,再利用改进麻雀搜索算法对LSTM模型的参数进行优化。最后,采用改进的混合ISSA-LSTM模型并完成RUL预测。采用NASA公开数据集对本模型进行验证。结果表明,该模型的平均绝对误差、均方根误差和平均相对百分比误差控制在0.016 47、0.022 84和1.2048%以内,能够有效地提高锂离子电池RUL的预测精度。

Remaining useful life prediction of lithium-ion batteries based on hybrid ISSA-LSTM

Accurate prediction of the remaining useful life (RUL) of lithium-ion batteries is of great significance to reduce the risk of battery use and maintain the stability of equipment. To improve the stability and accuracy of RUL prediction of the lithium-ion battery, this paper proposes an RUL prediction model based on the hybrid improved sparrow search algorithm (ISSA) and long short-term memory (LSTM) neural network. First, the outliers in the original data are decomposed using averaging. Second, a Tent chaotic map, adaptive weight, an opposition-based learning strategy, and the Cauchy variation perturbation strategy are combined to optimize the sparrow search algorithm. The parameters of the LSTM model are optimized by the improved sparrow search algorithm. Finally, the improved hybrid ISSA-LSTM model is used to complete the RUL prediction. The NASA public data set is used to verify the model. The experimental results show that the mean absolute, root mean square, and average relative percentage errors of the model are controlled within 0.016 47, 0.022 84, and 1.2048%, which can effectively improve the prediction accuracy of the RUL of lithium- ion batteries.