Study Of Ceramic Materials Grindability System

Grindability reflects the easy degree of certain material to be grind, also is the combination between grinding processes and material properties, structure and components, which could connect the design of ceramic materials with grinding processes. Through the analysis of ceramic grindability, rational materials and grinding parameters can be selected. Also, ceramic materials can be ground efficiency. Based on the theory of ceramic grindability, basic ceramic grindability system was studied considering ceramic mechanical properties and technical parameters with the application of system analysis and modeling method.Based on the idea of performances determined by structures, ceramic mechanical properties parameters, the main element in ceramic grindability system, were studied using nonparametric statistics method firstly. The distribution characteristics of parameters were studied using goodness of fit test. Meanwhile, the linear and nonlinear relativity of various parameters were studied using rank relational coefficient and grey correlative coefficient. The results, which could provide the basis to design and select ceramic materials, were compatible with the current theories and became primary foundation for ceramic grindability study.According to the theory of ceramic grindablity, ceramic mechanical and technical parameters were selected rationally for system modeling. And then an interpretative structural modeling(ISM) of ceramic grindablity system was established, which could intuitionisticly show the hierarchical structure and could be applied to control the grinding process. Through the calculations of ultimate super matrix using analytic network process, scale-weight of effect parameters of grindability was determined, which could provid a new method for the evaluation of ceramic grindability.Applying modern multivariate statistics, the shortcomings of multiple linear regression used in ceramic grindability system modeling was analyzed and existing problems of ceramic grindability multivariate data fitting of small sample was proposed. A PLSR model of grindability system was established and higher fitting effects were obtained. Regression equation was consistent with the existing theories and experimental conclusions, which could be applied to predict the grinding force and surface roughness and provide guidance to the machining of ceramic materials and the selection of technical parameters.The ceramic grindability system model was established by applying BPNN and GPNN, which could better solve the linear assumption problem of linear regression model. GRNN could fit the multivariate data small sample, and is better for the study on ceramic grinding process.