Archakova A. Detection of properties change of controlled plants based on set-membership identification approach

The dissertation is dedicated to the solution of the problem of early diagnostics of controlled plants in real time and development of methods for objects properties change detection under conditions of apriori uncertainty about the characteristics of the object, signals and noise using set-membership identification approach. Fault detection methods are improved. Variations of set-membership identification algorithms are developed supplemented by elective modification policy. A new algorithm for set-membership identification of dynamic objects is developed. Algorithms of set-membership identification for non-stationary objects are proposed. The algorithms proposed are designed for functioning in real time and under uncertainty about parameters and operational disturbances.