Using statistical and artificial intelligence approach to predict the exhaust gas temperature of a micro gas turbine engine

To study the relation between the amount of Exhaust Gas Temperature (EGT) as the output quantity of an experimental gas turbine engine and the parameter of engine rotational speed (RPM), as its input quantity, two different data mining approaches were employed in the present work. Artificial Neural Network (ANN) and Multiple Polynomial Regression (MPR) techniques were used to predict the nonlinear relation between the input and output of the engine. The related experiments were already performed by using an experimental micro gas turbine engine with an engine rotational speed in the range of 0 ~ 108000 RPM. The results show that, in general, both the ANN and MPR approaches have good predicting capability for estimating exhaust gas temperature values. Also the results of using the ANN and MPR approaches show that the degree of agreement between the predicted and measured values of exhaust gas temperature is higher for the case of employing the ANN approach. In other words, the ANN has better predicting capability for estimation of exhaust gas temperature than the MPR method.