Fault Diagnosis Of Rotor System And Shearer Cutting Condition Recognition Based On Current Signal

With the improvement of automation and intelligence,mechanical equipment as the core equipment in industrial production what plays an important role.In order to ensure the safe,reliable,efficient operation of the equipment and to avoid the occurrence of accidents and economic losses,it is very important to research on the fault diagnosis and operation state recognition of mechanical equipment.In recent years,the motor current signal analysis as a new detection technology gradually favored by the majority of scholars,it had became a research hotspots to use motor current signal analysis to fault diagnosis and operation state recognition of mechanical equipment.In this paper,the author explores the method of motor current feature extraction,fault diagnosis of rotor system and shearer cutting condition recognition.The main work is as follows:1,The author analyzed the influence of load torque variation on motor current signal.And found that the load torque fluctuation is reflected in the frequency spectrum of the motor current signal.The frequency components on the two sides of the current fundamental frequencyieff ?0.The correctness of the method is verified by the simulation of the motor model in Maltab/Simulink.2,Aiming at the difficulty of extracting the characteristics of the motor current signal and the characteristic frequency is easily annihilated by the power frequency,the author proposes a method of combining wavelet threshold denoising,EEMD and cross correlation analysis.By applying a sinusoidal torqueexcitation on the rotor test bench to simulate the torque change and collecting the motor current signal for processing,the results show that the method of cross correlation analysis to filter the IMF component can select the IMF component quickly and effectively suppress the interference of 50 Hz frequency.The frequency of torque ripple is extracted.It proves that the method is feasible.3,Aiming at the unbalance and misalignment fault of rotor system,the author proposes a method of EEMD-PCA to extract the characteristic parameters of motor current signal.By collecting the motor current signal in rotor system fault simulation test bench,and the fault was identified by BP neural network and support vector machine method.The results show that the feature extraction using EEMD-PCA can effectively improve the recognition effect and the EEMD-PCA-SVM method of the fault recognition rate reached 93.4%,above the EEMD-PCA-BP method.4,Aiming to the identification of coal and rock condition during the operation of shearer cutting,the author used the wavelet packet energy method to extract the characteristic of the motor current signal and used SVM for identification.Finally,the recognition algorithm is optimized from two aspects of feature vector and support vector machine.The results show that the accuracy of the PSO-SVM algorithm is more than 90%,the method of PSO-SVM is useful.