The method of express-diagnostics of the roughness of the surface layer of machine parts on basis of a probabilistic model with hidden states

Abstract

The article suggests a method of express-diagnostics of the roughness of the surface layer of machine parts and shows its efficiency. The authors developed a graphical probabilistic model, providing high quality of diagnosis with small number of statistical data. Express diagnostics performs determination of the quality of surface layer on basis of profilogram classification. The diagnostic process is running as the comparison of profilograms with prepared models of different classes of roughness using statistical sampling. Program realization of the model is developed, specialists were trained and experimentally investigated the accuracy of the diagnostics.