Research And Development Of Defect Detection System Of Lifting Motor Based On Audio Signal

With the constant improvement of the mobile phone design level,the major manufacturers have launched their own full screen series mobile phone,because Liu Hai and water droplets screen there is a use of defect,the manufacturers try to make cell phone camera hidden,and use when using lifting motor to drive the way to improve screen ratio,market demand at home and abroad are also rising.At present,the annual production of lifting motors in the world is about 5 million,among which the total production of domestic motor manufacturers accounts for more than 90% of the world.Due to the lifting motor running noise directly affect the user use lifting the experience of cameras,and domestic manufacturers of the factory of the lifting motor audio defect detection method mainly by artificial decibel meter reads the decibel value at runtime,and through the subjective feeling to judge the stand or fall of motor,low real-time detection,artificial detection speed,detection of quality cannot be guaranteed,and how to design a set of intelligent audio defect detection equipment to replace the original manual testing and become the common problem facing each lifting motor manufacturer.First in this paper,the internal structure of lifting motor,and the characteristics of the work,based on the lifting motor end of the work of the audio signal and the environment noise to choose the appropriate low pressure and high precision sensor according to the signal frequency response characteristics to choose the appropriate sampling points and sampling rate,according to the lifting motor internal resistance to choose the appropriate sampling resistance and building hardware platform,choice of America’s national instruments LabVIEW software and form a complete set of NI-9234 build software acquisition card data acquisition system.Then through the preliminary analysis of the audio faults of the motor with the manufacturer,the fault types are mainly divided into two categories: noise and abnormal sound.Collecting the original audio signals of noise,abnormal tone and good quality motors,and using A weighting pretreatment simulating to the human ear hearing characteristics,noise for defects,The original signal was converted to decibel value by octave range decibel spectrum,and statistical threshold was set up for classification,by using wavelet package decomposition to extract the different frequency band energy,analysis of frequency band energy distribution,set A threshold value,establish the motor sound standard library,then artificial classification of fine products and different screening,get good parts and the motor of the original samples,using the short-time Fourier transform,build the time-frequency gray,The CNN convolutional neural network is trained with the time-frequency grayscale map corresponding to the good product and the abnormal sound.The trained CNN convolutional neural network is used to identify the abnormal sound defect of the motor,and compared with SVM,ANN,GBDT network,etc.,the results show that the CNN convolutional neural network based on the spectrum power spectrum has a very good recognition effect on the abnormal sound defect.The deep learning model and the detection of decibel threshold are unified on the LabVIEW software platform.According to the test requirements,the audio signal of the lifting motor when it is working should be collected,and the final determination result is output finally through the detection algorithm deployed in advance.Finally based on the needs of industrial production design code scanning system of the motor to one to one correspondence of testing data,so that the product traceability,convenient engineers to optimize the production process of products and the development of the PC software to soundproof box,motor drive board,and realtime communication code gun and the test procedure,improve the automation level of the testing process and production efficiency.This paper describes in detail the detection principle of motor audio defects and the design process of automatic detection device,and designs a kind of automatic detection system for lifting motor audio defects,which integrates data acquisition,real-time monitoring of data,result display,communication output and data storage,and has the advantages of low cost and good robustness.And through a large number of experiments to verify the effectiveness of the equipment,can fully meet the factory’s automatic production rhythm,can be extended to similar motor production line to use,not only can improve the yield of products,for the labor cost savings also have a great help...