Lithium battery remaining life prediction method

Abstract

The invention discloses a lithium battery residual life prediction method, which comprises the following steps: acquiring the capacities of a plurality of charge-discharge cycles of a lithium battery,and normalizing the capacities; performing window division on the normalized capacitances to obtain a training data set; inputting the training data set into a degradation state model comprising a convolutional neural network and a long and short memory recurrent neural network for training; inputting the last window data of the training data set into a trained degradation state model for slidingprediction until the predicted capacitance reaches a capacity degradation threshold point; and predicting the remaining life of the to-be-tested lithium battery according to the sliding cycle numbercorresponding to the predicted capacity value. According to the method, the feature extraction capability of the convolutional neural network and the time sequence prediction capability of the long and short memory recurrent neural network are fused, the degradation features of the lithium battery are effectively extracted and predicted, and the prediction precision is improved. And the window size of the degradation index is automatically determined by using the false nearest neighbor method, so that the calculation efficiency is improved.

Description

technical field [0001] The invention belongs to the technical field of lithium battery degradation state monitoring, and more specifically relates to a method for predicting the remaining life of a lithium battery. Background technique [0002] Lithium batteries have the characteristics of high energy density and low self-discharge rate, and are the main power source for electric vehicles, mobile phones, and aviation systems. The failure of lithium batteries will cause equipment performance degradation, increase maintenance losses, and even cause sudden disasters. At the same time, replacing lithium batteries too early will cause greater economic losses. The remaining life of lithium batteries can represent the real-time state of machinery, therefore, accurate prediction of the remaining life of lithium batteries is of critical significance to the reliability and safety of electric equipment. [0003] Lithium battery life prediction methods can be divided into model-based ...

AI Technical Summary

Problems solved by technology

However, the selection of the size of the sliding window in the above methods is based on manual experience debugging, which consumes a lot of training resources and is inefficient.

Since the separate cyclic neural network is not as strong as the feature extraction ability of the convolutional neural network, and the separate convolutional neural network does not have the ability to predict the time series of the cyclic neural network; the training process of the correlation vector machine and the support vector regression is obtained from all training data Part of the data is selected for training, resulting in a certain training error, so the predicted life is not accurate enough, and it is difficult to meet the actual work requirements

[0005] In general, the existing data-driven lithium battery remaining life prediction methods have the problem of low prediction accuracy

Images