Yenchev S. Synthesis of intelligent control systems for aircraft gas turbine engines

The thesis analyses the synthesis of intelligent control systems of aircraft gas turbine engines principles of construction, functions, tasks, methods and algorithms. The thesis studies the structures and information flow in aviation gas turbine engines electronic control systems and establishes the methods and algorithms for identification of aviation gas turbine engines and information processing in the system of automatic control system of gas turbine engines based on intelligent control methods. The study establishes scientific and methodological bases of designing neural network regulators in the structures of intelligent control systems of aviation gas turbine engines: training, synthesis algorithms and settings; intelligent control systems of gas turbine engines formation and intelligent wavelet filter. The method of synthesis of fuzzy intelligent control systems of gas turbine engines using the method of balanced straight line approximation and algorithmic maintenance of their functioning is developed. The thesis proposes the concept of intelligent control systems of gas turbine engines optimization in terms of system approach. The control part of the systems is a single mathematical operator that converts input information into control effects. The study provides the method and methodology for the intelligent control systems of the gas turbine engine control laws optimization in the course of operation in stable and transient modes according to the criteria: speed, minimum fuel consumption, based on the abovementioned concept. The thesis offers the solution to the problem of failure-resistant control algorithms synthesis of intelligent control systems of gas turbine engines on the basis of fuzzy logic and neural networks. The algorithm and method for the synthesis of neural network regulator as a part of intelligent control systems of gas turbine engines are developed. The study analyses the dependence of intelligent control systems management quality on the choice of neuron network architecture and structure. The failure-resistance algorithm for intelligent control systems of gas turbine engines is proposed. The algorithm is based on the use of FDI method, characterized in that the detection of failures in the system is carried out by comparing elements of intelligent control systems of gas turbine engines with the elements of the reference neural network model of intelligent control systems. The reference model is used in real-time mode. Therefore, the reliability of the failures detection in a wide work and characteristics variation range of intelligent control systems of gas turbine engines.