Research On Acoustic Emission Intelligent Monitoring For Grinding Of Ductile Iron And Zirconia Ceramics

Ductile iron QT700-2 is the material of automobile crankshaft.On-line monitoring is needed to ensure the machining quality.AE technology is suitable for nondestructive monitoring.High precision intelligent prediction of grinding surface roughness of ductile iron QT700-2 was studied by combining acoustic emission technology with machine learning and deep learning.At the same time,it exists the phenomenon of ductile grinding to brittle grinding in precision machining of ceramics.Acoustic emission monitoring of brittle-ductile transition in grinding of zirconia ceramics was studied.Firstly,the research status of the above two problems in AE monitoring grinding quality is analyzed.The theories of AE technology,AE signal analysis,processing method,machine learning and deep learning are expounded.On this basis,the acoustic emission intelligent prediction of surface roughness of nodular cast iron grinding and the acoustic emission monitoring of brittle plastic transition of ceramic grinding are studied.Secondly,on the basis of the acoustic emission prediction experiment of grinding surface roughness of ductile iron QT700-2,two methods of artificial experience extraction and convolutional neural network automatic extraction were used to extract the grinding acoustic emission signal characteristics and input it into the BP neural network,support vector regression machine.Use genetic algorithm and particle swarm optimization algorithm to optimize the prediction model respectively,then predict grinding surface roughness of the ductile iron QT700-2.The best prediction model after comparison is CNN-GA-SVR.The relative error of prediction reaches about 1%,and the prediction accuracy is very high.Finally,based on the monitoring AE experiment of brittle-ductile transition in zirconia ceramics(PSZ)micro-cutting and deep grinding,the time-frequency characteristics of AE signals in the phenomena of brittle-ductile transition under short-time fourier transform,empirical mode decomposition and wavelet decomposition are analyzed.The energy change and entropy change of AE signals in the process of brittle-ductile transition are analyzed.The energy of IMF1 and IMF2,the main components of empirical mode decomposition of AE signal,are related to the brittle-ductile transition of zirconia grinding.The global information entropy of grinding AE signal,IMF1 and IMF2,and the wavelet entropy increase at the critical grinding depth of brittle-ductile transition of zirconia grinding.