To improve the accuracy and stability of power battery state of charge (SOC) estimation, this book proposes a SOC estimation method for power lithium batteries based on the fusion of deep learning and filtering algorithms. More specifically, the book proposes a SOC estimation method for Li-ion batteries using bi-directional long and short-term memory neural networks (BiLSTM), which overcomes the problem that long and short-term memory neural networks (LSTM) pose, because they can only learn in one direction, resulting in poor feature extraction and memory effect.The book provides some technical references for the design, matching, and application of power lithium-ion battery management systems, and contributes to the development of new energy technology applications.
