【摘要】 针对涡旋压缩机故障特征难以提取以及在小样本下故障类别难以区分的问题，提出了一种基于小波包能量谱与独立成分分析相结合的故障特征提取方法，结合支持向量机建立故障分类模型，实现非平稳信号的准确识别。首先对信号进行小波包分解与重构，通过重构系数获取分解后不同频带所对应的小波包能量作为故障特征值；再利用独立成分分析法对故障特征值进行优化得到特征向量值；最后构造支持向量机，输入故障特征向量进行训练和测试，可得出涡旋压缩机故障诊断的准确率。实验和仿真分析结果验证了该方法的有效性，其诊断准确率可达94.5%。更多还原

【Abstract】 Aiming at difficult of extracting scroll compressor fault features and distinguishing fault categories under small samples, it proposes a wavelet packet energy spectrum and Independent component analysis based on wavelet packet fusion fault feature extraction method. Combined with Support Vector Machine, a fault classification model is established to realize the accurate recognition of non-stationary signals. Firstly, the wavelet packet decomposition and reconstruction of signals are carried out, and the corresponding wavelet packet energy in different frequency bands is obtained by the reconstruction coefficient as the fault eigenvalue, and the eigenvector value is optimized by the independent component analysis method. Finally, the accuracy of scroll compressor fault diagnosis can be obtained by constructing support vector machine and inputting fault feature vector for training and testing. The effectiveness of the method is verified by experimental simulation analysis and result verification, and the diagnostic accuracy can reach 94.5%.更多还原