Health Assessment Of Lithium-ion Battery Based On Optimized Long Short-term Memory Network

With the improvement of people’s awareness of environmental protection and energy storage technology,lithium batteries are widely used in new energy electric vehicles,aerospace satellites,electronic equipment and other fields due to their high power,high storage energy density,and long life.However,with the use of lithium batteries,their performance will continue to degrade,and accidents such as explosions may even occur.Therefore,more and more attention is paid to the safety and reliability of lithium batteries,which is of great significance to the study of the health status of lithium batteries.Based on data-driven thinking,this article applies deep learning methods to the health assessment of lithium batteries.The specific work is as follows:First of all,in view of the influence of noise and other factors in the data acquisition process,the adaptive noise complete empirical mode decomposition method is adopted to smooth the data and extract the trend characteristics.In addition,the phase space reconstruction method is used to determine the optimal embedding dimension to solve the problem that the sliding window in the neural network is difficult to determine.In order to avoid the disappearance of the gradient,a long and short-term memory neural network is used as a predictive model.In the prediction process,due to the large number of model parameters,the previous experience tuning makes the model stability poor and the accuracy is low.Therefore,the genetic algorithm is used to optimize the key parameters of the network model,thereby improving the accuracy of the model for predicting the future health status of lithium batteries.Secondly,the problem of long and short-term memory neural networks being unable to learn back-to-front information and information overload.This paper adopts an improved method based on optimizing the two-way long and short-term memory network and the attention mechanism.This method superimposes two long-term and short-term memory networks with different learning directions for information extraction,which can learn both historical data and future data..The addition of the attention mechanism can focus on hidden states with high attention weights and improve prediction efficiency.Finally,the particle swarm optimization algorithm is used to optimize the network model structure to realize the assessment of the health status of the lithium battery.Finally,the method proposed in this paper is simulated on the NASA data set and compared with other common algorithms.The experimental results show that the method proposed in this paper greatly improves the performance of the model prediction,the convergence speed is faster,and the current health status of the lithium battery can be better evaluated,and it provides an effective strategy for the maintenance of the lithium battery.