【摘要】 矿山大型固定机械，如提升机和主通风机等是煤矿安全生产的至关重要设备。由于矿山机械的工作环境非常恶劣，机械故障点比较隐蔽，且故障原因具有多样性特征，导致诊断难度较高。提出基于深度学习的矿山机械集群故障智能诊断方法。利用深度学习中的卡尔曼滤波对矿山机械集群振动信号完成降噪处理。利用深度神经网络中的卷积核提取故障信号特征，将提取的特征输入到深度神经网络模型中计算出准则函数。根据准则函数实现故障类型聚类处理，输出故障分类结果，完成矿山机械集群故障的智能诊断。实验结果表明，所研究方法的智能诊断准确率可稳定在95%以上，故障诊断的耗时平均为45.3ms,机械设备集群故障诊断的振动信号频率与实际信号完全一致。更多还原

【Abstract】 Large fixed machines in mines are very important equipment for safety production. Because the working environment of mining machinery is very bad, the fault points are relatively hidden, resulting in high diagnostic difficulty. In this paper, an intelligent method of fault diagnosis of mining machinery cluster was presented based on deep learning. Firstly, the Kalman filter in deep learning was used to reduce the noise from the vibration signal of the mining machinery cluster. Secondly, the convolution kernel in the deep neural network was used to extract the fault signal features, and then these features were input into the deep neural network model for calculating the criterion function. Based on the criterion function, the fault types were clustered. Finally, the fault classification results were out. Thus, the intelligent fault diagnosis of the mining machinery cluster was achieved. Experimental results show that the intelligent diagnosis accuracy of the proposed method can be stabilized at 95%, and the average time required in fault diagnosis is 45.3ms. Meanwhile, the frequency of the vibration signal of fault diagnosis of mechanical devices is completely consistent with the actual signal.更多还原