基于物理信息神经网络的电池组SoC估计的研究OA

State of Charge(SoC)estimation for battery packs is a critical issue in Battery Management Systems(BMS),especially in multi-cell battery packs where traditional SoC estimation methods exhibit significant errors due to differences among individual cells.To improve the accuracy of SoC estimation,this paper proposes a method based on Physics-Informed Neural Networks(PINNs)for estimating the SoC of battery packs.By integrating classical physical models of battery SoC with deep learning,this method utilizes information such as the voltage and temperature of each cell to estimate the battery pack's SoC.Experimental verification on a real vehicle datasets shows that the average absolute error of the proposed method on the test set is 0.89%,and the root mean square error is 1.12%,which are reduced by 72.0%and 53.3%respectively compared with traditional voltage-based methods and ablation models.The experimental results demonstrate that a strategy combining physical constraints with data-driven approaches can significantly enhance SoC estimation accuracy,providing a more precise and reliable SoC estimation method for BMS.