Industrial Gas Turbine Emission Diagnostic using Artificial Neural Network Regression

Gas turbines, which is also called as combustion turbines, are broadly used in scope of industry of electric power generation, aircrafts and various process applications. The increasing of number of gas turbines used in the industry had received many concerns as the pollutant emission from gas turbine also increase. The emission of gas turbines from the burning fuels due to its operational process in nowadays industries had caused the negative effects to the green and clean environment. This study poses a prediction method together with diagnosis of emission of industrial gas turbines. The Artificial Neural Network (ANN) regression trained by using predicted results from simulation model established using GSP11. MATLAB is used as a basic platform to carry out the modelling and investigation of the emission diagnostics system. Collected simulation data throughout the study are used to identify the conditions to high emission of greenhouse gases such as NOx, and CO from the industrial gas turbine, and the operating condition that lead to high emission pollutant. The results show the comparison between ANN and GRNN in predicting the emission of the gas turbine which demonstrate that ANN is more accurate compared to GRNN. The outcome from gas turbine simulation shows that temperature and pressure would lead to different level of pollutant emission of gas turbine.