Intelligent Acoustic Emission Monitoring Of Grinding Surface Roughness Of Engineering Ceramics

Engineering ceramics alumina and zirconia are typical difficult-to-machined materials with high strength and hardness,so grinding is the main machining method.The application of acoustic emission technology in the grinding process of engineering ceramics to realize online real-time monitoring is of great significance to improve the machining efficiency,grinding quality and grinding intelligence of engineering ceramics.The prediction of surface roughness of engineering ceramics by acoustic emission technology is an important step to realize online real-time monitoring.In this paper,the application status of acoustic emission technology in grinding field is described firstly,and the research status of surface roughness prediction by acoustic emission technology is introduced emphatically.The generation mechanism of acoustic emission signal and common analysis methods are summarized.Effective means of noise reduction is an im portant premise of grinding acoustic emission signal analysis,this article adopts the method of the local mean decomposition,the grinding acoustic emission signal is decomposed into a number of product function,using the K-means clustering algorithm sel ected the effective component to preliminary noise of signal,then the ind ependent component analysis(PCA)is adopted to signal further nois e reduction processing,compared with the traditional wavelet threshold denoising method and EMD threshold denoisin g method,the signal-to-noise ratio is improved and the mean square error is smaller.Based on the grinding force and acoustic emission monitoring experiments of alumina and zirconia ceramics,the grinding process was monitored from three aspects of grinding force,grinding acoustic emission and grinding surface roughness,and the changes of grinding force and acoustic emission under different grinding depths and different grinding directions were investigated.The RMS values of grinding force and acoustic emission signal increase with the increase of grinding depth,and the RMS values of grinding force and acoustic emission signal in the direction of grinding are greater than those in the direction of grinding.The effective values of grinding force and aco ustic emission signal of zirconia ceramics are larger than those of alumina,which is related to the two material removal methods.BP model and GA-BP model were established based on the surface roughness prediction experiment of alumina and zirconia cerami cs grinding.The time domain,frequency domain and waveform characteristic parameters of grinding acoustic emission signals were extracted by statistical feature method,and the surface roughness of alumina and zirconia ceramics was predicted by neural net work model.However,as the high frequency acoustic emission signal is weaker than the low frequency acoustic emission signal,the statistical feature method is poor in the characterization of the high frequency information,so the entropy method based on variational mode decomposition is applied.From the perspective of time and frequency,the energy entropy and permutation entropy of each mode component after the variational mode decomposition of AE signals in grinding were extracted.The energy entropy and permutation entropy could reflect the difference of grinding surface roughness in different frequency bands.The entropy method based on variational mode decomposition can predict the surface roughness of alumina and zirconia ceramics better than statis tical feature method.