论文部分内容阅读
针对采煤机滚动轴承常见的突发问题诊断准确性不高和速度慢,以小波包和RBF神经网络为基础,提出了由小波包分解提取各个节点特征能量谱与自适应步长萤火虫算法优化的RBF神经网络进行分类辨识的采煤机滚动轴承故障诊断方法.对振动传感器输出的信号进行小波包分解,运用基于代价函数的局域判别基(LDB)算法对小波包分解进行裁剪,获取最优的特征能量谱,经处理后作为特征向量训练ASGSO-RBF神经网络,建立诊断模型.实验结果表明:所建模型的故障辨识正确率达到95.8%以上,相较于其他算法模型具有更低的误报率和漏报率,诊断精度及诊断效率更高.
Aimed at the low diagnostic accuracy and slowness of common sudden problems of shearer rolling bearings, wavelet packet decomposition and RBF neural network are proposed to optimize the energy spectrum of each node extracted by wavelet packet decomposition and the adaptive step firefly algorithm RBF neural network for fault diagnosis of shearer rolling bearing.Based on the wavelet packet decomposition of the signal output by the vibration sensor and the wavelet packet decomposition using the local discriminant base (LDB) algorithm based on the cost function, the optimal The energy spectrum of the ASGSO-RBF neural network was trained as a feature vector to establish a diagnostic model.The experimental results show that the accuracy of the model is above 95.8%, which has lower false positives than other algorithms Rate and false negative rate, diagnostic accuracy and diagnostic efficiency higher.