Analysis of performance deterioration of a micro gas turbine and the use of neural network for predicting deteriorated component characteristics

Deteriorated performance data of a micro gas turbine were generated and the artificial neural network was applied to predict the deteriorated component characteristics. A program to simulate operation of a micro gas turbine was set up and deterioration of each component (compressor, turbine and recuperator) was modeled by changes in the component characteristic parameters such as compressor and turbine efficiency, their flow capacities and recuperator effectiveness and pressure drop. Single and double faults (degradation of single and two parameters) were simulated. The neural network was trained with a majority of the generated deterioration data. Then, the remaining data were used to check the predictability of the neural network. Given measurable performance parameters as inputs to the neural network, characteristic parameters of each component were predicted and compared with original data. The neural network produced sufficiently accurate prediction. Using a smaller number of input parameters decreased prediction accuracy. However, an acceptable accuracy was observed even without information on several input parameters. This paper was recommended for publication in revised form by Associate Editor Ohchae Kwon Mr. J. E. Yoon received his MS degree from Dept. of Mechanical Engineering, Inha University in 2008. His thesis topic was test and simulation of micro gas turbines. He has been working at LG Digital Appliance Company. Mr. J.J. Lee received his MS degree from Dept. of Mechanical Engineering, Inha University in 2006, and is now Doctoral student at the same department. His research topics include simulation and diagnosis of gas turbines. Prof. T.S. Kim received his PhD degree from Dept. of Mechanical Engineering, Seoul Na-tional University in 1995. He has been with Dept of Mechanical Engineering, Inha Univeristy since 2000, and is Associate Professor as of Oct. 2008. His research area is aero-thermodynamc simulation and test of gas turbine systems including microturbine and their components. His recent research concern also includes analysis on fuel cells and fuel cell/gas turbine hybrid systems. Prof. J.L. Shon received his PhD degree from Dept. of Mechanical Engineering, The University of Alabama in Huntsville in 1986. He has been with School of Mechanical & Aerospace Engineering, Seoul National University since 2000, and is BK Associate Professor as of Oct. 2008. His research area is design, simulation and test of gas turbine system and components. He is also interested in researches on fuel cells and fuel cell/gas turbine hybrid systems.