Dynamic Multi-Objective Optimization for Gas Turbine Operation

This paper presents a systematic method for the optimal settings of gas turbine operation under changing environment. After building the high-fidelity model of the gas turbine, four decision variables including fuel flow, high and low pressure compressor variable inlet guide vane and cooling coefficient were selected, while the pollutant emissions, cost of gas turbine and system output power are chosen as response variables to be optimized. A dynamic multi-objective optimization algorithm based on segmented prediction (DMOEA-BSP) has been developed to find the optimal operating conditions under the varying ambient pressure and temperature. Comparing to the classical DNSGA-II algorithm, the simulation showed that proposed DMOEA-BSP algorithm achieves a better operating performance.