A lithium-ion battery RUL prediction method based on improved variational modal decomposition and integrated depth model

In the context of energy shortage, lithium-ion batteries have become the highest consumer demand and most widely used class of batteries in the global battery market with their excellent performance of high-energy density, small size, low rate of self-discharge, good cycling performance and long lifetime, and the research into predicting the remaining useful life (RUL) of lithium-ion batteries has also grown in importance. In this essay, a new method to predict battery RUE, based on improved variational modal decomposition (VMD) with integrated depth model is proposed, which improves the prediction accuracy of a single model under a single scale signal. Firstly, the general deterioration trend and local random fluctuation components of signal are obtained by performing a multiscale decomposition on battery capacity data. Secondly, the time convolutional network (TCN) and the convolutional neural network and long and short-term memory neural network combined (CNN-LSTM) are used to model each fluctuation component and global degradation trend respectively. Thirdly, the predict result of each sub-model is combined to get the ultimate battery RUE. predict result. Finally, lithium -ion battery datascts from NASA arc utilized to test the accuracy of the prediction model, and the result demonstrates the excellent stability and prediction accuracy of the suggested approach.