The Experimental Study Of Cermet And Cemented Carbides Based On Ultra High-speed Grinding

With the development of society’s productivity and the progress of science andtechnology, the difficult-to-grinding and hard brittle materials, such as cermet,cemented carbides and so on, has been used widely in various industry fields.Becauseof its physical character of high strength，high hardness and good wearability,thedifficult-to-grinding and hard brittle materials are difficult to process.At present, ultrahigh-speed grinding is commonly used in the process of the difficult-to-grinding andhard brittle materials,and is the trend of the grinding process of thedifficult-to-grinding and hard brittle materials. On the basis of the previous researchresults, the grinding force, grinding simulation temperature and acoustic emissionsignal of GN20and PA30were researched with ultra high-speed grinding in this paper.In this study, the application of acoustic emission technology of ultra high-speedgrinding wheel passivation detection was also researched.This paper mainly content as follows:(1) The material characteristics of cermet GN20and cemented carbides PA30andthe ultra high-speed flat grinder were introduced; collecting the grinding force ofcermet GN20and cemented carbides PA30in the process of ultra high-speed grinding;researching the ultra high-speed grinding temperature of GN20through simulationanalysis; analyzing the influence of grinding parameter to the grinding force andgrinding simulation temperature of ultra high-speed grinding.(2) On the basis of AE test results of GN20and PA30, the influence of grindingparameters (the feeding speed of work table，grinding wheel linear velocity andgrinding depth) and workpiece material to the RMS of AE signal was studied.Thefrequency distribution of AE signal of GN20and PA30were researched through thefrequency analysis.In this paper, the selection principles of wavelet basis of AE signalanalysis were related and the neural network model of grinding wheel passivationdetection was built based on acoustic emission technique and wavelet theory withwavelet energy coefficient method.The study of experimentals indicates that there is a good relationship betweengrinding parameters of experimentals and the grinding force, grinding simulationtemperature and acoustic emission RMS of GN20and PA30.In the study of AEexperimentals of cermet GN20and cemented carbides PA30, the feeding speed ofwork table was trated as the main operation parameters of researching wheel dull.We took the BP neural network as the recognizer of wheel dull and confirmed the relevantparameters of neural network through the simulating optimization.The trained modelcould detect the passivation degree of wheel accurately. This provides a kind of viablemethod which has very high practical value for confirming wheel dressing c ycle,reducing processing cost and improving surface quality.