基于音频特征识别的锂电池热失控监测方法OA
A Monitoring Method for Lithium-Ion Battery Thermal Runaway Based on Audio Feature Recognition
锂电池在新能源汽车、电网储能电站等领域已实现规模化应用,但是锂电体系的热稳定性脆弱仍是制约其安全应用的瓶颈问题.锂电池安全阀爆开时产生的音频,可对热失控过程实现提前预警.提出了一种基于音频特征识别的锂电池热失控监测方法,首先使用变分模态分解与小波阈值协同去噪;然后融合传统梅尔频率倒谱系数与时域统计特征进行提取;最后采用贝叶斯优化支持向量机模型进行模式识别.实验结果表明,所提出的方法对三元锂电池和磷酸铁锂电池的安全阀音频信号的识别效果较好,识别准确率可达94.32%,可为锂电池热失控预警提供参考.
Lithium-ion batteries have been widely used in new energy vehicles and grid energy storage stations.However,the poor thermal stability of lithium battery systems remains a major challenge limiting their safe application.The acoustic signal generated when the safety valve of a lithium-ion battery bursts can provide an early warning of the thermal runaway process.In this study,a monitoring method for lithium-ion battery thermal runaway audio feature recognition is proposed.First,variational mode decomposition(VMD)combined with wavelet thresholding is employed for collaborative denoising.Then,traditional Mel-frequency cepstral coefficients(MFCCs)are fused with time-domain statistical features for feature extraction.Finally,a Bayesian-optimized support vector machine(SVM)model is adopted for pattern recognition.Experimental results demonstrate that the proposed method achieves effective recognition of safety valve acoustic signals from ternary lithium batteries and lithium iron phosphate batteries,with a recognition accuracy of up to 94.32%.The proposed method provides a promising approach for the early warning and monitoring of lithium-ion battery thermal runaway.
闫浩杰;谢家乐;史庆武
佳木斯大学 信息电子技术学院,黑龙江 佳木斯 154007华北电力大学 燕赵电力实验室,河北 保定 071003佳木斯大学 信息电子技术学院,黑龙江 佳木斯 154007
信息技术与安全科学
锂电池热失控音频特征识别声信号模式识别
lithium-ion batterythermal runawayaudio feature recognitionacoustic signalpattern recognition
《江汉大学学报(自然科学版)》 2026 (4)
26-36,11
国家自然科学基金项目(52207235)
评论