Gear Fault Diagnosis Based On Synchronous Averging Kinematic Error In Spatial Domain

Due to its advantages of accurate, reliable transmission, high efficiency, high power and speed range of gear transmission, the gear transmission has been widely used in modern machinery. It is necessary to diagnose and predict the gear fault since it will seriously affect the transmission reliability. Diagnosis based on motor driver is a new research hotspot, of which the time and frequency domain method is suitable for steady speed situa tion and the frequent speed up and down in practical conditions leads to lim itations of applying the method. To solve this, this paper proposes the Synchronous Averaging Kinematic Error method.Firstly, this paper discusses the meaning and mechanism of kinematic error, compares the performance of methods in T-F domain and spatial domain in steady and variable speed situation, and analyzes failure causes T-F methods when speed changes. Designs a cascade state feedback observer of shaft torque and kinematic error. The appropriate motor drive system control algorithm is the only necessary part, without need to install sensors on the load side, which can greatly reduce the cost of diagnosis. The effect of the state feedback gain coefficient on the performance of the observer is studied, and the conclusion is drawn that the two gain coefficients are related to the stability of the observer and the bandwidth.Later in this paper, the resampling technique in spatial domain is proposed. Combing with the kinematic error and position signal in time domain, the reconstruction of kinematic error is implemented, and the cubic spline interpolaion algorithm improves the sampling rate. Synchronous averaging method is applied to improve SNR and benefits gear fault dianosis.Single stage gear transmission experiment platform is set up. Observer performance is tested under different load. The results show that the designed observer can track on the shaft torque change well. The position and kinematic error signal are collected in steady and variable speed state, and then the signal resampling process is completed in off-line mode. It turns out that the synchronous averaging method can effectively reduce noise and improve SNR. Through the experiment results illustrate the kinematic error method in spatial domain can be used to solve both the steady state speed and variable speed gear fault diagnosis. The reconstrcuted KE signal can be used to establish the space position distribution map, which will be very beneficial to the fault d iagnosis and maintenance of the equipment.