Research On Coupling Characteristics Of Current And Torsional Excitation Of Rotor System

Now rotating machinery is developing toward large-scale,complex and automation,study on the core component of the rotor system is becoming more and more important.In the electric motor drives the rotor systems,The electromechanical coupling effect has a non negligible impact on the performance and service life of rotor systems.The rotor system with torque excitation,the motor current and electromagnetic torque are closely related.Because of the change of electromagnetic torque of the rotor system,shafting torsional vibration of rotor system became complex.Considering the elastic damping axis and the rigid disk,build electromechanical coupling model of rotor system using the torque and current as interface.In the MATLAB/simulink environment,built dynamic imitation model of rotor system and solve analytical solution of model using four order Runge Kutta method,get curve of torsional vibration,rotor speed,electric current and electromagnetic torque with different torque excitation.Using classical mechanics and electromagnetic analysis the simulation results.Using FFT,time frequency analysis and wavelet analysis technology,study on variation of the motor current of the rotor system under different torque excitation.In order to realize the recognition of torque simply and rapidly,put forward a kind of torque identification of rotor system base on singular value decomposition,wavelet packet energy spectrum and BP neural network.Design rotor test rig,verify the correctness of the model and research method through the test.Through the comparison of rotor system of motor current test data and the simulation data of different types of torque excitation,the study found the spectrum of current signal decide by electromagnetic torque.Types of current spectrum and the corresponding energy are different with different torque excitation.Rotor torque excitation a role in the modulation of motor current,the torque excitation curve be similar envelope curve of current signal.The current signal after computing are decomposed and reconstructed by wavelet analysis,get the current envelope curves of different torque excitation.Using neural network can identify the the rotor torque excitation type accurately and quickly.Provides a reliable and effective methods for fault diagnosis of rotor system.