Abstract: For analog circuit Prognostics and Health Management （PHM） applications, a fault diagnosis method combining Principal Component Analysis （PCA） and Extreme Learning Machine （ELM） is discussed. The fault samples are obtained based on Sallen-Key bandpass filter design failure mode, and the fault feature extraction is performed by PCA. It trains the ELM by fault samples to obtain fault diagnosis model. Simulation results show that the method has a high recognition rate and robustness, with the value of research and application in engineering practice.

Key words: analog circuit, Principal Component Analysis（PCA）, Extreme Learning Machine（ELM）, fault diagnosis

摘要： 针对模拟电路的故障诊断和健康管理（PHM）的应用，提出了结合主成分分析（PCA）和极限学习机（ELM）的故障诊断方法。该方法用Sallen-Key带通滤波器来获取故障样本，并通过PCA进行故障特征提取。根据故障样本对ELM进行训练来获得故障诊断模型。实验结果表明，该实现方法识别率高、鲁棒性好，在工程实际中具有研究和应用价值。

关键词: 模拟电路, 主成分分析, 极限学习机器, 故障诊断