Integrated Approach to Processing Diagnostic Data Based on Heterogeneous Cognitive Models

Abstract

The paper presents the reasonable feasibility of using measuring and expert information in diagnosing industrial equipment. The problems encountered in the diagnosis of industrial equipment (IE), theoretical provisions for the construction of models for diagnosing equipment and requirements for models built are considered. The authors proposed a comprehensive approach based on the building heterogeneous cognitive models (HCM) and mixed production rules for processing diagnostic information (DI) obtained from measuring instruments and experts. HCM here is a directed weighted graph with marked vertices and edges with various types of factors and relationships. A model in the form of HCM for diagnosing electrical equipment (EE) in the oil industry is proposed. As an example, a script of possible risk situation developments related to the equipment developed using pulse modeling, and its analysis are presented. The proposed approach allows one not only to calculate the predicted values of the controlled parameters of the equipment, but also to increase the accuracy and completeness of the IE diagnostics. It takes into account not only the measurement data, but also the knowledge from the duty staff, the chief engineer, which leads to the IE timely repair and maintenance.