Research on LPV-based model of a turbofan engine

Two system identification approaches are discussed to approximate the nonlinear dynamics of a turbofan engine by constructing linear parameter varying (LPV) models in this paper. The state variables in several steady points from the idle to maximum condition are determined based on the thermodynamic characteristics of engine chambers. The small perturbation method is given out and utilized to obtain the coefficient matrices of state variable model. In order to acquire the better representation of the engine rotating components dynamics, rotor acceleration speeds are added to state variables. The matrix coefficients with acceleration speeds is obtained by the Genetic Algorithm (GA) initially, and then calculated with linear least square fitting method. The LPV-based model is built up based on the state variable model in various conditions, and the fuel flow is recognized as the index. The simulation experiments on a turbofan engine are carried out, and the comparisons of different state variable model results are also represented. It shows that both methods are effective to represent dynamic performance of the engine.