Design And Research Of Fault Diagnosis System For Gear-Rack Of Three Gorges Ship Lift

The Three Gorges ship lift is a typical vertical ship lift with gear-rack.It can be lifted and lowered by the meshing of gear-rack and driving system.In order to ensure the safety of the ship lift,it is of theoretical significance and engineering value to design the on-line monitoring and fault diagnosis system of the gear-rack of the ship lift.Based on the actual situation of the gear-rack of Three Gorges ship lift,this paper designs the fault diagnosis system of the gear-rack of the Three Gorges ship lift,which mainly includes the following contents:(1)According to the actual situation of gear-rack and gear shaft bearing of Three Gorges ship lift,the fault diagnosis system of gear-rack of Three Gorges ship lift is designed.The structure and hardware selection of the gear-rack fault diagnosis system are described in detail.The software for fault signal analysis and diagnosis is designed by MATLAB.The system realizes the detection,acquisition and wireless transmission of vibration signal,and the software realizes the real-time monitoring,diagnosis and recording of signal.(2)In the design of software gear-rack fault diagnosis for Three Gorges ship lift,a comprehensive fault diagnosis method based on the variational mode decomposition,one-dimensional convolution neural network and Low-Rank decomposition is proposed.The method of variable mode decomposition is adopted to reduce the noise of the signal and the separation of the bearing vibration signal from the rack and gear vibration signal.The fault diagnosis of the rack and gear vibration signal is realized by using the method of one-dimensional convolution neural network through training and learning.The accuracy rate of identifying the six states of the gear-rack is 93.5%,and the accuracy rate of identifying the fault signal segment is 91.0%,which proves that1D-CNN has high precision.Because the vibration signal of bearing fault has low rank in time and frequency domain and the noise signal has sparsity in time and frequency domain,the method of Low-Rank decomposition is used to reconstruct the transient signal of bearing fault and fault diagnosis.The fault characteristic frequency of the reconstructed signal of three kinds of bearing fault has little difference with the theoretical value.The difference between the fault characteristic frequency of LowRank decomposition of outer ring fault and the theoretical value is 1.4%,the difference between the fault characteristic frequency of low rank decomposition of inner ring fault and the theoretical value is 0.2%,and the difference between the fault characteristic frequency of low rank decomposition of rolling element fault and the theoretical value is 0.5%.(3)In order to verify the comprehensive fault diagnosis method based on the variational mode decomposition,one-dimensional convolution neural network and Low-Rank decomposition,the vibration signal of gear-rack and bearing of Three Gorges ship lift needs to be simulated because of the lack of prototype observation data.According to the actual situation of Three Gorges ship lift,the types and characteristics of the vibration signals of gear-rack fault vibration and bearing fault vibration are analyzed.The vibration signal of gear-rack is simulated by using the impact series model.Based on the theoretical value of fault characteristic frequency of various kinds of bearing faults,the vibration signal of gear shaft bearing is simulated by pulse signal.Considering the disturbance factors such as water body shaking,the vibration of the ship bearing chamber,the internal vibration of the main reducer,the vibration of the synchronous shaft,the vibration of the gear bracket mechanism,the pump,the water conveyance system,the lubrication pump station,the hydraulic system and the noise,etc.,the vibration signal database of the gear-rack of the Three Gorges ship lift was established by MATLAB.