Identification of twin shaft gas turbine based on hybrid decoupled state multiple model approach

Abstract The work presented in this paper focuses on presenting an hybrid identification method for a nonlinear dynamic gas turbine, from a real time input and outputs data exploitation, with the fuel flow as the input and the rotational speed of high pressure and low pressure turbine as outputs. The multi model, which are in the form of a weighted combination of local linear state space models, offer an interesting alternative of the nonlinear models because it takes into account a several operating modes. The models are identified with the help of decoupled models using a hybrid approach between parametric estimation using artificial intelligence algorithms.