车用三元锂电池CNN-LSTM挤压力学响应预测OA
Prediction of compression mechanical response for automotive ternary lithium-ion batteries using CNN-LSTM
以车用三元锂电池为研究对象,提出一种基于CNN-LSTM 混合模型的电池挤压力学响应预测方法.首先利用 LS-DYNA 有限元仿真软件构建高精度电池单体挤压模型,模拟电池在10、15、20 和25 mm 直径球形与圆柱形压头下的挤压过程;然后基于仿真数据构建包含力-位移曲线的时间序列数据集;最后采用CNN-LSTM 神经网络模型来预测电池在 25 mm 直径压头下的挤压失效情况,并引入 LSTM 模型进行对比.实验结果表明,CNN-LSTM 混合模型能更有效捕捉电池结构非线性变形过程中的时空耦合特征,可为车用三元锂电池机械滥用失效预测提供高精度、强泛化性的分析方法.
Focusing on ternary lithium-ion batteries for vehicles,a prediction method for mechanical response under extrusion is proposed based on CNN-LSTM hybrid model.Firstly,a high-precision battery cell extrusion model is established using LS-DYNA finite element simulation software to simu-late the extrusion process under spherical and cylindrical indenters with diameters of 10,15,20 and 25 mm.Subsequently,a time-series dataset containing force-displacement curves is constructed based on simulation data.Finally,the CNN-LSTM neural network model is employed to predict battery extru-sion failure under indenter with diameter of 25 mm,with comparative analysis against standalone LSTM models.Experimental results demonstrate that the CNN-LSTM hybrid model effectively captures spatiotemporal coupling characteristics during nonlinear structural deformation,establishing a high-pre-cision and strongly generalized analytical method for predicting mechanical abuse failure in automotive ternary lithium-ion batteries.
白圆悦;黎衍;于潇雁;姚立纲
福州大学机械工程及自动化学院,福建 福州 350108福州大学机械工程及自动化学院,福建 福州 350108福州大学机械工程及自动化学院,福建 福州 350108福州大学机械工程及自动化学院,福建 福州 350108
交通工程
三元锂电池力学响应特性失效预测卷积神经网络长短期记忆神经网络
ternary lithium batterymechanical response characteristicsfailure predictionconvolu-tional neural networklong short-term memory neural network
《福州大学学报（自然科学版）》 2026 (3)
276-283,8
国家重点研发计划资助项目(2022YFB4702401)
评论