Syrotkina O. Methods of SCADA Diagnostics Based on M-tuples and K-valued logic

This dissertation is devoted to solving a relevant scientific task of quality assurance for the distributed SCADA system while the system is in operation. This is achieved by performing self-diagnostics through detecting, localizing and identifying failures in real-time. We further developed an object and classification model of SCADA self-diagnostics using the methodology of expert systems. We suggested a method of forming a search space to establish the diagnosis based on this methodology. The method processes and analyzes a full flow of low-level diagnostic information generated by SCADA. It uses mathematical methods to work with the data structure m-tuples based on ordered sets we developed for the first time. We improved a structural and logical model to diagnose SCADA operability. For the first time the author obtained the following analytical dependencies: an admissible state change of the controlled parameter while it is passing through the SCADA hierarchy levels; detection and localization of a failure in the system; separation between independent and dependent failures. The author proposed the method to analyze the change of the information flow state while it is passing through SCADA structural elements and hierarchy levels on the basis of analytical dependencies obtained. The method allows human operators to control the system operability in real-time in case of incomplete / unreliable / absent data in the system's structural elements. The integrated use of models, methods and software application developed allows the maintenance personnel of any enterprise to carry out timely diagnostics of emergency situations and reasons for failures; make quick decisions for the organization of SCADA repair and recovery work; repeatedly reduce the recovery time of SCADA performance after reversible failures in the event the system has mechanisms to automatically self-recover.