基于长短期记忆(LSTM)神经网络的锂电池极片面密度精细预测方法研究

打开文本图片集

关键词：锂电池极片；面密度测量;X射线测厚仪；大射线斑；小射线斑拟合;LSTM神经网络中图分类号：TM912 文献标志码：A 文章编号：1003-5168（2025）08-0028-06DOI：10.19968/j.cnki.hnkj.1003-5168.2025.08.005

Abstract： [Purposes] The areal density of lithium battery pole piece coating is very important to the capacity，performance and safety of the battery.Especially in the high-speed production line，the measurement accuracy of the coating surface density willdirectly affect the consistency and stability of the battery.However，despite their high resolution，small spot raythickness gauges are expensive and slow to acquire data， making it difficult to meet the needs of large-scale production.Therefore，a fine prediction method of lithium batery pole coating areal density based on Long Short-Term Memory （LSTM） neural network was proposed.[Methods] The LSTM model is trained by the large ray spot measurement data， and the time series characteristics and local variation laws of the areal density data are captured to fit the areal density distribution with small spot resolution，so as to reduce the equipment cost and time consumption under the premise of ensuring the measurement accuracy.[Findings] The experimental results show that the model is applicable to a variety of substrates and samples with defects，especially the fitting effct of the thining area is significantly improved.The model successully achieves the fiting accuracy of the average correlation coeficient of O.999 6，and has a high fitting ability.[Conclusions] This method provides a new idea for the on-line measurement of the coating surface density of lithium battery pole pieces，and has the application potential of achieving both high precision and low cost in highspeed intelligent production lines，and provides effective technical support for production line quality control.

Keywords： lithium battery electrode; areal density measurement; X-ray thickness gauge; large ray spot; small spot fitting;LSTM neural network

0 引言

极片涂布是锂电池制造过程中的重要工序，涂布质量会对锂电池的容量、一致性和安全性产生重要影响。（剩余4557字）