Temperature prediction of lithium-ion battery based on artificial neural network model

Accurate temperature prediction is one of the most critical problems to improve battery performance, and prevent thermal runaway. However, the heat generation and heat dissipation of lithium-ion batteries have complex nonlinear characteristics and are easily affected by external factors, therefore it is difficult to accurately predict the battery temperature. In recent years, artificial neural network (ANN) has been widely used in many fields of lithium ion batteries due to its unique advantages in dealing with highly non-linear problems, such as battery modeling and SOC estimation, residual life (RUL) prediction and battery temperature prediction. However, there are few studies on temperature prediction of lithium ion batteries in foam metal thermal management system, and the current research has not reached an accurate conclusion to explain which neural network is better for temperature prediction. Therefore, an artificial neural network approach was used to estimate the temperature change of lithium-ion batteries in the metal foam thermal management system. Back propagation neural network (BP-NN), radial basis functions neural network (RBF-NN) and Elman neural networks (Elman-NN) were respectively applied to establish the temperature prediction model, and the temperature prediction performance of different neural network modeling techniques were compared. In order to verify the accuracy and validity of the neural network thermal model, the performance tests under the sample condition and the new condition were carried out respectively. The predicted result data and temperature contrast diagram of sample and test conditions are obtained. Elman neural network model has better adaptability and generalization ability, and the training time of Elman neural network model is shorter. It is more suitable for the temperature prediction of LIBs under metal foam and forced air cooling system.

Automated summary

Accurate temperature prediction is crucial for improving battery performance and preventing thermal runaway in lithium-ion batteries. Artificial neural networks (ANNs) have been widely used for battery temperature prediction due to their ability to handle highly non-linear problems. In this study, the performance of different ANN modeling techniques, including back propagation neural network (BP-NN), radial basis functions neural network (RBF-NN), and Elman neural networks (Elman-NN), were compared for temperature prediction in a metal foam thermal management system. The results showed that the Elman neural network model had better adaptability and generalization ability, with shorter training time, making it more suitable for temperature prediction in lithium-ion batteries under a metal foam and forced air cooling system.