Research On Vibration Monitoring And Fault Diagnosis Of Shearer Rocker Arm

The healthy operation of mine equipment transmission system is a prerequisite to ensure safe and efficient production of coal mine.Vibration analysis is an effective technical means for fault diagnosis of mine equipment transmission system,and it is also one of the research hotspots at present.In this thesis,MG620/1660-WD Shearer is taken as the research object,based on the kinematic analysis of its transmission system by using simulation software,a simulation test device of shearer rocker working condition is designed,and suitable sensor measuring points are arranged,in this thesis,time-frequency analysis method is introduced to enhance the dimension of vibration signal,and a deep transfer learning model based on multi-layer neural network model is constructed.Based on the analysis of the structure of the rocker arm transmission system,the three-dimensional modeling of the rocker arm transmission system is carried out by Solid Works.Combined with the working condition of shearer rocker arm,using Adams simulation software,the transmission characteristics of each gear of shearer rocker arm are analyzed,and the angular velocity and angular acceleration curves of each structure are obtained,it provides a theoretical basis for the arrangement of sensor measuring points on the test platform.According to the transmission principle of bearing and gear,the characteristics of its fault vibration signal are analyzed.Based on the analysis of the vibration form of the shearer,the noise reduction methods of the singular value decomposition and EMD signals are discussed,and the order parameters of the two noise reduction methods are adjusted by combining the simulation signals of the running state of the shearer rocker arm,according to the effect of noise reduction,the method of noise reduction suitable for vibration signal of shearer rocker arm is determined.Continuous wavelet transform and Short-time Fourier transform time-frequency analysis method are used to enhance the dimension of the signal.In this thesis,the recognition accuracy of deep transfer learning method in small sample scene is discussed according to the characteristics of less data set for fault recognition of shearer rocker arm.On the basis of summarizing the common fault forms of shearer rocker arm,the data set of fault vibration is collected and established.Based on Resnet50 neural network model,a deep transfer learning model for fault identification of shearer rocker arm is constructed.It divides the training set and the test set,optimizes the model to improve its recognition speed and accuracy,determines the suitable super-parameter by comparative analysis,and establishes the verification set to evaluate the recognition effect of the model.According to the function requirement of vibration monitoring of shearer rocker arm,the working condition simulation and vibration monitoring test-bed of shearer rocker arm are designed,the vibration monitoring system is set up and the sensor measuring points are arranged.The collected vibration data are built into the training set and test set of the fault identification model,and the model is proved to be accurate and reliable by fault identification of the tested rocker arm.The thesis has 76 figures,20 tables,and 91 references.