【摘要】 为解决煤矿通风机轴承故障特征难以有效提取及传统诊断方法识别精度不足的问题，提出了一种基于小波包能量谱与粒子群优化-支持向量机（PSO-SVM）融合策略的故障诊断方法。首先，采用小波阈值降噪预处理原始振动信号，优化参数包括小波基、分解层数、rigrsure阈值及硬阈值函数，实现噪声抑制与故障特征保留；然后，通过对比时频域及变分模态分解（VMD）-样本熵与小波包能量谱，小波包能量谱在表征故障敏感特征方面具有最优性能；最后，基于提取的特征向量，采用PSO算法对SVM参数进行优化，从而构建高精度故障识别模型。实验结果表明，该模型可有效区分6类轴承状态，训练集与测试集的准确率分别为96.43%和97.22%，较传统SVM提升47.22个百分点。基于LabVIEW与MATLAB联合编程开发诊断系统，实现振动信号实时采集、降噪、特征提取及在线识别。实测结果表明，该方法具有良好的诊断性能与工程实用价值。更多还原

【Abstract】 To solve the problems of difficult extraction of fault features and insufficient recognition accuracy of traditional diagnostic methods for coal mine ventilator bearing, proposed a fault diagnosis method based on the fusion of wavelet packet energy spectrum and Particle Swarm OptimizationSupport Vector Machine(PSO-SVM). First, the wavelet threshold denoising is used to preprocess the original vibration signals, with optimized parameters including wavelet basis, decomposition level,rigrsure threshold, and hard threshold function, realizing noise suppression while retaining fault features. Then, by comparing time-frequency domain analysis, Variational Mode Decomposition(VMD)-sample entropy with wave let packet energy spectrum, it is verified that the wavelet packet energy spectrum exhibits the optimal performance in characterizing fault-sensitive features. Finally, based on the extracted feature vectors, the PSO algorithm is adopted to optimize parameters of SVM, thereby constructing a high-precision fault recognition model. Experimental results show that this model can effectively distinguish 6 types of bearing states, with training set and test set accuracies reaching96.43% and 97.22% respectively, which is 47.22 percentage points higher than that of the traditional SVM. A diagnostic system was developed based on LabVIEW-MATLAB joint programming, realizing real-time acquisition, denoising, feature extraction, and online recognition of vibration signals. Field test results demonstrate that this method has excellent diagnostic performance and engineering practical value.更多还原