Lithium battery capacity prediction method based on sparrow algorithm to optimize convolutional neural network and bidirectional long short-term memory network model

To address the challenge of inevitable battery capacity degradation over extended usage, we have explored and applied a neural network that combines convolutional neural network (CNN) and Bidirectional Long Short-Term Memory (BiLSTM) for battery capacity prediction. In this model, CNN helps mitigate the fluctuations in prediction results by extracting data features, while BiLSTM predicts the battery capacity's aging trend. We optimized the parameters of the CNN-BiLSTM model using the Sparrow Search Algorithm (SSA) and evaluated the performance of the SSA-CNN-BiLSTM network using the leave-one-out method (LOOCV). The experiments demonstrate that the model can provide accurate predictions of battery capacity, particularly in scenarios with a large sample size and an extended time frame.