Normalized Radial Basis Function Networks Applied to Exergy, NOx, and Creep Life Prediction in an Industrial Gas Turbine

Normalized Radial Basis Function Networks (NRBFN) are known to have good interpolation and extrapolation characteristics. The purpose of the present work is to show the use of NRBFN for capturing exergy, NOx and creep life trends in a gas turbine generator. The models cover the whole operating region. It also addresses the use of Cuckoo search for identifying NRBFN models. The data required for model training and validation are generated using semi-empirical models developed by the authors. As an input, actual data collected over a week are adopted. The tests carried out on a different set of input data showed that the prediction from the NRBFN model closely matched the prediction from the semi-empirical model. The models can be used for performance optimization, multi-state reliability prediction and condition monitoring.