A dual-driven approach for modelling heavy-duty gas turbine based on operational data and intelligent genetic algorithm

Abstract

In order to build an accurate model of heavy-duty gas turbine, a dual-driven approach is proposed based on operational data and intelligent genetic programming considering rotational speed, inlet/outlet temperature, pressure and generation power. The input-output thermodynamic characteristics of the compressor are obtained by genetic programming and net generation power of gas turbine is expressed by polynomial fitting formula equation, whose coefficients are obtained by least square method. Results show that all the models to calculate temperature ratio, pressure ratio and air mass flow ratio of compressor have a good accuracy, which of temperature ratio can reach 0.01. The accuracy of model to calculate generation power value can reach 0.04. This method for holistic modelling can be applied to other kinds of heavy-duty gas turbine.