A Genetic Algorithm Approach to Estimate Performance Status of Gas Turbines
Accurate estimation of performance status of a gas turbine engine at certain ambient and operating condition based on measured gas path parameters is very important for both engine designers and users alike. It could be a very challenging task for engine performance engineers to estimate the value of component design parameters in order to match measured gas path parameters when the number of design point component parameters and the number of measurable performance parameters become large. Such status estimation can be used to distinguish the performance difference among fleet engines and build accurate engine models at an artificial design point for individual engines, which is also crucially important for gas path diagnostic analysis. In this paper, a gas turbine design point performance adaptation approach based on the integration of gas turbine thermodynamic performance modelling and a Genetic Algorithm has been developed in order to estimate the design point component parameters and match the available gas path measurements of real engines. In the approach, the initially unknown component parameters may be compressor pressure ratios and efficiencies, turbine entry temperature, turbine efficiencies, air mass flow rate, cooling flows, by-pass ratio, etc. The engine measurable performance parameters may be thrust and specific fuel consumption for aero engines, shaft power and thermal efficiency for industrial engines, gas path pressures and temperatures, etc. The developed adaptation approach has been applied to a design point performance status estimation of an industrial gas turbine engine GE LM2500+ operating in Manx Electricity Authority (MEA), UK. The application shows that the adaptation approach is very effective and robust in producing a model engine that matches the actual engine performance with acceptable computation speed. Theoretically the developed techniques can be applied to different gas turbine engines.