Lithium battery RUL prediction method and system based on composite neural network

The invention relates to a lithium battery RUL prediction method and system based on a composite neural network, and belongs to the technical field of battery monitoring. The method comprises the following steps: data preprocessing and grey correlation analysis; and a variant Transform neural network is designed. According to the invention, a variant Transform neural network added with a ProbSparse self-attention mechanism is designed for related data of a lithium battery. Firstly, a multi-head self-attention mechanism is changed into a ProbSparse multi-head self-attention mechanism to avoid clues of sequence information leakage of a Transform encoder, the problems of position insensitivity and high data complexity of the model when the model processes such tasks are solved, and then LayerNorm in a feedforward network sub-layer is changed into BatchNorm to improve calculation efficiency and enable the model to be more adaptive to prediction of RUL. 本发明涉及一种基于复合神经网络的锂电池RUL预测方法及系统，属于电池监测技术领域。该方法包括以下步骤：数据预处理及灰色关联