Linear Transfer Functions Extraction for a Power Generation Three-Shaft Gas Turbine Based on Actual Specifications

Abstract

The design and testing of various types of controllers require accurate and reliable transfer functions that are compatible with the performance condition of the system. In this paper, the linear and the nonlinear methods have been utilized for extracting the transfer functions for a three-shaft industrial gas turbine. The main variables are the fuel input and environment temperature, while the main outputs are the LPT exhaust gas temperature and the generated power. According to the nonlinear structure of the system and the input variation intervals, the quasi-amplitude-modulated pseudo random binary sequence signals have been utilized for generating input parameters. Finally, the results extracted from the selected methods were compared to the outputs of real performance data under loading and unloading conditions. Based on the comparison between the accuracies obtained by the results, the auto-regressive moving average with eXogenous, try and error transfer function and linearized Hammerstein model methods are proposed, respectively. The outcome of using the controller indicated higher compatibility from transfer functions as compared to the reference dynamic model.