基于超声波技术和随机森林模型的电池浸润程度分析

Analysis of Battery Wetting Degree Based on Ultrasonic Technology and Random Forest Model

摘要: 近年来，锂离子电池的需求量随着新能源汽车的发展急速上升。然而，锂电池在生产过程中存在一些一致性方面的挑战。其中，电池浸润一致性是影响电池寿命和性能的关键因素之一，不均匀的电解液分布可能引发电池起火、爆炸等危险。超声波检测技术用于评估电池电解液的浸润一致性是一种高效且无损的方法。传统的超声波检测电池浸润一致性技术通常采用超声C扫图像与经验法相结合的方式，但依靠人为经验看图分析，会存在较大的误差。为了克服传统技术的局限性，本文提出通过超声波A扫波形进行波形分析，以磷酸铁锂软包电池为研究对象，分析不同注液量电池的浸润区域波形差异。介绍了基于随机森林模型进行不同浸润程度自动识别的具体流程。结果表明，该模型具有较高的识别精度，对未浸润和浸润不良电池的分类准确率达到了100%，而浸润良好电池的分类准确率为97.8%，过注液电池的准确率也达到了95%。本文为识别电池的不同浸润程度提供一定的理论指导。

Abstract: In recent years, the demand for lithium-ion batteries has surged rapidly with the development of new energy vehicles. However, there are some consistency challenges in the production process of lithium batteries. Among them, the consistency of battery wetting is one of the key factors affecting battery life and performance. Uneven distribution of electrolyte may cause dangerous incidents like battery fires and explosions. Ultrasonic detection technology is an efficient and non- destructive method used to evaluate the consistency of battery electrolyte wetting. Traditional ultrasonic tech-niques for detecting battery wetting consistency usually combine ultrasonic C-scan imaging with empirical analysis. However, relying on manual image analysis can lead to significant errors. To overcome the limitations of traditional techniques, this paper proposes waveform analysis through ultrasonic A-scan waveforms. Using lithium iron phosphate soft pack batteries as the research sub-ject, the paper analyzes the waveform differences in the wetting areas of batteries with different electrolyte volumes. The paper introduces the specific process of automatic identification of differ-ent wetting degrees based on the random forest model. The results show that this model has high recognition accuracy, achieving a classification accuracy of 100% for unwetted and poorly wetted batteries, 97.8% for well-wetted batteries, and 95% for over-wetted batteries. This paper provides some theoretical guidance for identifying the different degrees of battery wetting.