Comparison of linear models for gas turbine performance

Abstract

Linearization of turbofan engines is an effective method for performance control and fault diagnosis. In this study, three different linearization techniques, including the partial derivative method, optimized fitting method and equilibrium manifold, are discussed and compared. First, an optimized fitting method is developed based on the least-square method and an optimization algorithm. To avoid trapping in local optimization solution, the initial values used in the optimization approach are obtained through the partial derivative method. Second, to verify the accuracy and effectiveness of the linear model in the flight envelope, the result of linear modeling method based on equilibrium manifold is analyzed in detail. Finally, an overall assessment of the merits or weaknesses of linearization models is provided based on the obtained results.