针对电梯故障诊断中特征提取困难和故障样本数量少问题,提出了应用小波包分解和最小支持向量机( LS-SVM) 相结合进行电梯急停智能故障诊断的方法。借助小波包分解,该方法首先提取电梯轿厢振动信号作为特征向量,然后利用 LS-SVM 分类模型对故障进行辨识。实验证明,小波包与 LS-SVM 相融合的故障诊断与识别技术可发挥两者的优势,该方法对电梯急停故障的诊断具有较好的诊断效果。 Aiming at the difficulty of feature extraction and the number of fault samples in elevator fault diagnosis, a method of intelligent fault diagnosis of elevator emergency stop using wavelet packet decomposition and least support vector machine (LS-SVM) is proposed. With wavelet packet decomposition, this method first extracts the elevator car vibration signal as the eigenvector, and then uses the LS-SVM classification model to identify the fault. Experiments show that the fault diagnosis and identification technology combined wavelet packet with LS-SVM can give play to the advantages of both. This method has a good diagnostic effect for the diagnosis of elevator emergency stop.