Robust Multi-Objective Optimization for Gas Turbine Operation Based on Kriging Surrogate Model

• DOI: 10.23919/ccc52363.2021.9550112
• Publication Date: 2021-10-07T04:24:31Z
• Authors: Hao Xia; Peilin Jia; Liang Ma
• Keywords: Surrogate model; Latin Hypercube Sampling; Robustness

This paper presents a systematic method for the optimal settings of gas turbine operation. After building high-fidelity model turbine, fuel flow, variable inlet guide-vane(VIGV) high and low pressure compressors, ambient air temperature are selected as decision variables, while pollutant emissions output power whole system chosen response variables to be optimized. In order guarantee accurate prediction, performance three different types surrogate models, namely polynomial surface (PRS), Kriging radial basis function (RBF) models built with sampling points generated by Latin hypercube design (LHD) method. With best fitting data, it is prediction during optimization process. It also shown that NSGA-II algorithm suitable multi-objective turbine. The robustness Pareto solutions checked under varying temperature. By considering solution robustness, nature change perturbation, operating conditions identified.