基于物理信息机器学习的锂电池温度估计OA

Under the background of carbon peaking and carbon neutrality,lithium batteries are widely used in the transportation sector,and the performance and safety of batteries have received extensive attention.Real-time estimation of the actual operating temperature of the battery is the key to ensuring battery safety.In view of the temperature estimation problem of lithium batteries,a lithium battery temperature estimation method based on physi-cal information machine learning is proposed.The effectiveness of the proposed method was verified within a wide temperature range of 5 to 25 ℃.Based on the aggregated parameter thermal model of lithium batteries,the least square method and genetic algorithm are used to identify the parameters of the lithium-ion battery thermal model,providing prior knowledge of battery temperature for machine learning.Features related to temperature,such as heat generation,are input into the machine learning framework as supplementary content to improve the estimation accuracy.The machine learning model combining convolutional neural networks,long short-term memory neural networks and attention mechanisms is successively integrated with the thermal model to improve the estimation accu-racy of temperature.The experimental results show that,compared with the thermal model and the pure data-driven method,the estimation accuracy of the proposed method has been increased by 68.0 percentage points and 49.3 percentage points respectively.