基于改进深度置信网络的采煤机摇臂传动系统故障诊断研究

采煤机是一个集机械、电气和液压为一体的大型复杂系统,工作环境恶劣,如果出现故障将会导致整个采煤工作的中断,造成巨大的经济损失。采煤机摇臂传动故障作为整机的主要故障,是故障监测研究的重点。提出一种基于改进深度置信网络的采煤机摇臂传动系统故障诊断方法,对摇臂传动信号进行频段分解,通过不同的频段阈值进行降噪处理,提取故障特征信息,完成采煤机摇臂传动故障分类。将深度置信网络引入故障诊断,通过对采集的故障状态信号进行迭代训练深度学习,得出与故障模型的对应关系,并采用粒子群优化算法对故障模型进行迭代优化,应用于摇臂传动的故障诊断识别。结果表明,故障特征提取准确,故障诊断精度高。

Research on Fault Diagnosis of Shearer Ranging Arm Drive System Based on Improved Deep Belief Network

Shearer is a large-scale complex system which integrates mechanical,electrical and hydraulic systems.The working environment is bad.If there is a fault,it will lead to the interruption of the whole mining work,resulting in huge economic losses.As the main fault of the whole machine,the fault of shearer ranging arm drive is the focus of fault monitoring research.A fault diagnosis method based on the improved deep belief network was proposed,which decomposes the frequency band of the transmission signal of the ranging arm,reduces the noise through different frequency band thresholds,extracts the fault feature information,and completes the fault classification of the transmission of the ranging arm of the shearer.The deep belief network was introduced into the fault diagnosis,and the corresponding relationship with the fault model was obtained through the deep learning of the iterative training of the collected fault state signal.The particle swarm optimization algorithm was used to carry out the iterative optimization of the fault model,which is applied to the fault diagnosis and identification of the ranging arm transmission.The results show that the fault feature extraction is accurate and the fault diagnosis accuracy is high.