Optimization steady state control of turbofan engine in the whole flight envelop based on Support Vector Machines

Abstract

This paper aims to present a method to design the optimization steady state nonlinear controller of turbofan engine which is applicable for the whole flight envelop. First, based on the linear small deviation mathematical models of some representative operation points of turbofan in various flight conditions, the linear quadratic regulators of turbofan are designed respectively and the optimization linear controller parameters are obtained. Then, the SVM method is put forward and applied to synthesize each linear controller to redesign a nonlinear controller via Support Vector Machines (SVM), so the nonlinear controller designed is theoretically applicable in the whole flight envelop. The inputs of SVM are the flight altitude, flight mach number, and the throttle actuator angle. The outputs are the feedback matrix parameters which represent linear controller design results. The SVM is trained and obtained offline. Numerical experiments are carried out to verify the validity of the method. Simulation results show that the relative error of controller design is less than 2% in the whole flight envelop, and the method can meet engine control precision well.