Využití Weibullova rozdělení při diagnostice technologických procesů

Abstract

The doctoral thesis “UTILIZATION OF WEIBULL DISTRIBUTION IN DIAGNOSTICS OF TECHNOLOGICAL PROCESSES” is focused on development and testing of methodology for creation of model based on Weibull probability distribution. Primary applicable are of these models are dependability, technical diagnostics and maintenance. Algorithm of developed methodology for models creation with Weibull distribution principle is based on statistics-evolutionary approach. Proposed methodology is tested on three mechanical fault mechanisms – atmospheric corrosion, creep and abrasive wear and two electrical fault mechanisms – overload and feature change. Tested fault examples were from operational environment of metallurgical and energetic industry, from automotive area and from laboratory experiments. The results confirm the presumption, that for creation of such models are suitable artificial intelligence methods. The thesis has a character of operational-basic research, when the results will be able to applied not only in operation in dependability area, technical diagnostics and maintenance, but also in other areas like materials and constructions. Also the results could create a base knowledge for various research projects concerning the fault mechanisms, when at sufficient amount of data from various areas would be able to create a knowledge base for fault mechanisms models. This base then could be used in machine construction area and maintenance.