Desine Of Bearing Fault Detction System Based On Audio Recognition

As one of the key parts of rotating machinery,Bearings play an extremely important role in the country’s heavy industry.In the Industrial production process,the failure of the bearing will directly affect the normal operation of the entire mechanical equipment,and in some serious cases,it will also threaten the personal safety and property.Due to the impact of Sino-US trade frictions,the United States has listed voiceprint technology as one of the 14 banned export technologies.It shows that our country’s audio recognition technology is still restricted,and continuous exploration and progress are needed.Therefore,real-time monitoring of bearing operation based on audio recognition technology,and reduction of losses caused by bearing failures through recognition and alarms have become the research goals of this paper.The system adopts 16-bit DSC digital signal controller.Through the audio collection of bearing fault,signal conditioning,signal sampling and analysis,characteristic parameter extraction and template recognition,and display the type of fault to implement the system.This paper mainly introduces three parts: hardware circuit design,software program design and basic system verification.The hardware circuit introduces the audio signal acquisition circuit,the amplifier circuit,the boost circuit,the low-pass filter circuit,the system power circuit and the communication circuit between the microcontroller and external data storage,touch screen and OLED.The software program design is carried out in the MPLAB-IDE compilation environment using C language and assembly program,the program mainly include A/D sampling program,fast Fourier transform(FFT)program,characteristic parameter extraction program and related hardware module driver.The basic verification of the system includes the testing of the working state of the hardware circuit,the calculation and verification of the sampling accuracy of the system,and the comparison of the results of the FFT between the single-chip microcomputer and MATLAB.The final results show the correctness of the function and the reliability of the system.Finally,the bearing experiment data of Case Western Reserve University in the United States is used to analyze the characteristic parameters of bearing faults,and then the voiceprint matching templates of different fault categories are established,and the deterministic relationship between bearing vibration signals and audio signals is further explored.