Application Research Of Motor Running State Monitoring And Fault Diagnosis Technology Based On Time-Frequency Heat Map

The wide application of high-end intelligent equipment has made the importance of motors increasingly prominent.As the main power output equipment for life and production,motors play a vital role in residents’ lives and social development.Once the motor fails,it will bring economic losses and affect production efficiency,or cause catastrophic accidents and threaten personal safety.At present,most of the motor faults are based on known models or experimental conditions,and less consideration is given to the impact of the work site load on the motor.The personalized automatic online monitoring system for the motor is a difficult problem in modern fault diagnosis.Aiming at the problems of diverse motor operating environments,complex demand conditions and poor adaptability of conventional diagnostic algorithms,this paper proposes an on-line motor monitoring and fault diagnosis method based on time-frequency heat map.By introducing manual intervention,this method improves the environmental adaptability of the diagnosis system,and improves the interpretability of the diagnosis results through the visualization of the diagnosis process.Taking vibration as the starting point,this thesis analyzes and studies the formation mechanism of motor fault,and summarizes the typical fault vibration model and characteristic frequency.This thesis analyzes and processes the motor fault signal from the perspective of signal,analyzes several typical time-frequency processing methods,and analyzes the fault signal from time domain,frequency domain and time-frequency domain.The best wavelet base selection based on Tamura texture feature is studied to obtain the best time-frequency feature image.Feature enhancement of time-frequency feature image from the perspective of image;Local binary feature extraction is used to obtain time-frequency image features;According to the local binary feature vector of time-frequency image,support vector machine is selected to diagnose motor fault.Dynamic fault diagnosis is proposed and manual intervention is introduced to make the motor fault diagnosis algorithm have the ability of real-time verification and improvement,have better environmental adaptability,visualize the diagnosis process and make the diagnosis results interpretable.The vibration acquisition and analysis system is designed and developed,and the vibration sensor and data acquisition card are selected and arranged;The upper computer software is designed,developed and verified.The system has been applied to a large motor maintenance factory in Suzhou,and has achieved good results in troubleshooting at the factory end.The off-line analysis function is verified to ensure the practicability and reliability of the system.