Research On Bearing State Detection And Positioning Based On Sound

Bearing has the advantages of easy installation,high efficiency,low resistance and convenient lubrication,which supports main shaft and transmits torque is extensively used in rotating machinery and equipment.Carrying out research on detection and positioning of bearing operation status has a profound impact on understanding its operation status and timely discovering abnormalities,ensuring the operation of rotating machinery and has extremely importance for the development of modern industry.Sound is a natural signal and the most convenient and widely used signal interaction method for human beings.Sound diagnosis technology can achieve non-contact measurement and practical signal acquisition,it has good application potential.Using a microphone to collect sound signals during the operation of sliding bearings then analyzing and processing this sound signal can achieve status detection of sliding bearings,and using a microphone array to collect sound signals for localization can provide suggestions for subsequent bearing troubleshooting and maintenance.In this paper,the research on bearing state detection and positioning based on sound signal is carried out,the main work of this paper as follow:First of all,in complex working environment there are some difficulties to install contact sensors such as acceleration sensors,which cause state detection of the sliding bearing based on vibration signal has limitation.Therefore,this paper proposes to apply sound signals to the state detection of sliding bearing.An improved wavelet packet denoising and resonance sparse decomposition method for sliding bearing sound diagnosis is proposed in order to address the issue that the sound signal fault s severely disturbed by the noise emitted by other components and feature of sliding bearing is not obvious,and abundant of impact components will be generated in the sound signal when it fails.Secondly,conducting simulation and experiment verification research on the sliding bearing acoustic diagnosis method.Establish a simulation signal according to the characteristics of the sound signal of sliding bearing failure.The wavelet basis function,decomposition level,threshold function,and threshold are the four aspects that will affect noise reduction result.This paper uses the grid screening to find the optimal parameter combinations.At the same time,proposing a threshold that is suitable for the decomposition scale,which improves the traditional threshold calculation form does not take the wavelet packet decomposition scale into consideration.Collecting sound signal of sliding bearing from the rotor test bench,and use the method proposed in this paper to denoise and extract fault features of the sliding bearing fault sound signal can successfully extract fault frequency that is 96 Hz,which realizes the state detection of the sound signal of sliding bearings,providing a new idea to solve the problem of sliding bearing state detection.Then,carrying out in-depth research on microphone array sound source localization.This paper skips the improvement of traditional microphone array sound source localization methods and combines the microphone array structure with localization algorithms.The cube microphone array is designed,and the theoretical model of sound source location is established by deriving the acoustic wave equation and combining the spherical near-field acoustic holography,and then transform the theoretical model into the real model of sound source location based on the cube microphone array.Some parameters in the sound source localization model are determined according to the number of microphones and the Spherical t-designs integration method.Numerical simulation is conducted on the sound source localization model of cube microphone array using spherical wave sound field.The result shows that when the reconstruction surface is the outer envelope sphere of the cube microphone array and the wave number is 1,the reconstruction error is the smallest.Finally,using the cube microphone array to collect sound source at frequencies of100 Hz,1000Hz,and 2000 Hz respectively to do simulations and experiments of sound source localization.Both simulation and experiment results indicate that the frequencies at 100 Hz and 1000 Hz,using the sound source location model of cube microphone array can accurately determine the location of the sound source with minimal error.The error of the sound source location is very high at 2000 Hz,resulting in the misjudgment of the sound.Therefore,using the cube microphone array to collect sound signal can effectively reconstruct the sound field near the sound source if the frequency of the sound is low,then determine the position of the sound source according to the sound pressure distribution of the reconstructed sound field near the sound source.Combining the range of fault characteristic frequencies of sliding bearing,applying this model to sliding bearings is expected to locate the position of sliding bearing faults.