基于幅值平方相干谱的电池故障诊断方法

Battery Fault Diagnosis Method Based on Amplitude Squared Coherence Spectrum

针对锂离子电池故障诊断中存在测量噪声等不确定因素的问题,提出一种基于幅值平方相干谱的电池故障诊断方法。该方法的流程为:首先建立电池的二阶RC等效电路模型,进行参数辨识;然后用正弦电流激励电池模型和实际电池系统,分别获得电池模型和实际系统的频率响应,引入幅值平方相干函数和动态阈值对电池进行评估;判断存在故障后,采用多故障模型的方法生成特定故障的发生概率,确定发生故障的类型,进一步完成故障隔离。以单体锂离子电池为研究对象,设计过充过放实验进行验证,结果表明,与传统方法相比,过充和过放故障误检率分别降低了67%和63%;最后在实际电池系统中验证所提出方法的有效性。

Aiming at the problem of uncertainties such as measurement noise in Li-ion battery fault diagnosis,this paper proposed a battery fault diagnosis method based on amplitude squared coherence spectrum.The process of the method was as follows:firstly established the second-order RC equivalent circuit model of the battery for parameter identification;then used the sinusoidal current to excite the battery model and the actual battery system to obtain the frequency response respectively,and applid the squared amplitude coherence function and the dynamic threshold to evaluate the battery;if there is a fault,the multi-fault model was used to generate the probability of occurrence of the specific fault,then the type of fault was determined to complete fault isolation.Taking the single Li-ion battery as the research object,the overcharge and overdischarge experiments were designed for verification.The results showed that,compared with traditional methods,the false detection rates of overcharge and overdischarge faults were respectively reduced by 67%and 63%.Finally,the effectiveness of the proposed method was verified in a real battery system.