Method for Selecting Diagnostic Parameters Using Information Technologies Under a Set of Various Types of Information

The rapid growth of electrical systems (ES) is a result of the scientific and technological progress development. ES are characterized by a large number of elements of different types and interacting with each other by the nature and degree of element loads. Diagnosis of such systems is accompanied by a large amount of initial information with a significant degree of uncertainty due to the random factor on the one hand, and fuzzy information on the other hand. To apply methods of diagnostics and to accurately assess the ES state, the authors propose a methodology for selecting significant diagnostic parameters that characterize the ES state on the basis of a set of statistical methods and artificial intelligence methods under various types of factors. A generalized scheme for selecting parameters taking into account quantitative information and qualitative (fuzzy) information is provided. Methodology of selecting diagnostic parameters is described in detail, therefore allows one to select the most significant parameters, while screening out insignificant parameters (reducing the dimension of factor space), which affect the ES state. The results of experiments to solve practical problems on the basis of the developed methodology are presented and analyzed.