Research On Aero-Engine Control Based On Multiple Adaptive Method

Aero-engine system is a strong nonlinear system and there is no accurate model to describe.It is not easy to control and its control problem has been a hot spot in control field.A nonlinear model of engine is developed according to the working principle of the engine and the aerodynamic characteristics between the components.The modeling process is very complex,and the accurate algebraic expression of the nonlinear model is not available.This thesis taking the aero-engine as the research object,based on the different operating modes of the aero-engine,and the parameters transition,established models of aero-engine and designed a multiple model adaptive method.The main contents of this thesis are described as follows:First,summarize the model building method and the control method of aero-engine.Aero-engine works in a wide range of external conditions and internal parameters,and there are strict requirements for strength and aerodynamic constraints.So we must use a certain control strategy to meet the requirements of the engine.In this thesis,we take the turbofan engine as an example.Aimed at the different operating modes of the aero-engine set up models,which is LPV normalized decomposition.It is of great significance to the design of control method with theory guarantee and practical engineering application.Then,aimed at the changing of the operating modes of aero-engine,parameters usually change very quickly.It is easy to think of using multiple model adaptive method.Traditional multiple model adaptive method switches from a controller to another one.The process of switching may cause mutation,turbulence,instability and so on.In this thesis,the multiple model adaptive method we use can switch smoothly.It can not only avoid mutation,turbulence and instability,the number of models reduces greatly.Each of the selected model produces the estimate of system vector,and the estimate depends on all the models output.It forms the convex hull of estimate which changes over time.The convergence speed accelerates greatly.We use this method to control the aero-engine system.Finally,we analyze the stability of multiple model adaptive method and perform MATLAB simulation in order to verify the accuracy of this method.Moreover,the effectiveness of the proposed method is verified by comparing the controller based on the traditional adaptive method.