基于卷积神经网络与双向长短时融合的锂离子电池剩余使用寿命预测
Remaining Useful Life Prediction of Lithium-Ion Battery Based on CNN and BiLSTM Fusion
-
摘要: 针对锂离子电池剩余使用寿命(remaining useful life, RUL)传统预测方法的精确度与稳定性较低等问题, 融合卷积神经网络(convolutional neural network, CNN)和双向长短期记忆(bidirectional long short-term memory, BiLSTM)神经网络的特点, 设计一种锂离子电池剩余使用寿命预测方法。为了充分使用电池数据的时间序列特性, 使用一维卷积神经网络(one-dimensional convolutional neural network, 1D CNN)提取锂离子电池容量数据深层特征, 利用BiLSTM神经网络的记忆功能保留数据中的重要信息, 预测电池RUL变化趋势。通过采用NASA(National Aeronautics and Space Administration)的锂离子电池数据, 与1D CNN模型、LSTM模型、BiLSTM模型、1D CNN-LSTM模型进行预测对比。经实验结果表明, 1D CNN-BiLSTM具有更高的预测稳定性和精度。Abstract: Aiming at the low accuracy and stability of the traditional prediction method of the remaining useful life(RUL) of lithium-ion batteries, in this study, the convolutional neural network(CNN)and bidirectional long short-term memory(BiL STM) neural network are integrated, and a method for predicting the RUL of lithium-ion batteries is designed. To make full use of the time-series characteristics of lithium-ion battery data, one-dimensional CNN(1D CNN) is used to extract the deep characteristics of battery capacity data, and the memory function of the BiL STM neural network is selected to retain important information in the data and predict the trend of the battery RUL change. Through the use of lithium-ion battery data from the National Aeronautics and Space Administration, the predictions are compared with the 1D CNN, LSTM, BiL STM, and 1D CNN-LSTM models. The experimental results show that 1D CNN-BiLSTM can more accurately predict the RUL of lithium-ion batteries and improve the stability of predicting the RUL of lithium-ion batteries.
下载:
