基于LSTM和注意力机制的锂电池荷电状况预测

Prediction of the State of Charge of Lithium Battery Based on LSTM and Attention Mechanism

为提升锂电池的荷电状况预测精度,文章提出一种改进的注意力LSTM算法。先将美国航空航天局提供的电池数据分成测试集与训练集;再使用一维卷积神经网络算法提取训练集的特征信息,然后通过LSTM网络保留特征数据的历史信息;为减少LSTM算法的计算量引入特殊注意力机制,再使用Softmax函数输出预测的荷电状况;最后使用测试集验证训练好的算法。实验成果表明,提升的注意力LSTM方法与无损卡尔曼滤波、LSTM方法相比预测效果更佳。

An improved attentional LSTM algorithm was proposed to improve the prediction accuracy of lithium battery.Firstly,the battery data provided by NASA were divided into test set and training set.Then one-dimensional convolutional neural network algorithm was used to extract the feature information of the training set,and LSTM network was used to preserve the historical information of the feature data.In order to reduce the computation amount of LSTM algorithm,a special attention mechanism was introduced,and then Softmax function was used to output the predicted charge state.Finally,a test set was used to verify the trained algorithm.Experimental results showed that the improved attentional LSTM was better than the lossless Kalman filter and LSTM.