Prediction of turbojet performance by using artificial neural network

Abstract

Artificial intelligent represents a sophisticated tool for solving a complicated problem in the world. In this work, the artificial neural network (ANN) was utilized to predict turbojet performance at different operational conditions. The computational approach was used to solve turbojet governing equations in Matlab/Simulink. The obtained data were processed to build-up an artificial neural network model. The network organized into three layers, including input, hidden and output respectively. While a backpropagation was used as a training algorithm. The developed model was examined in different values of operational and design parameters such as flight Mach No., fight altitude, compression ratio of compressor and turbine inlet temperature. It showed a good capability for disclosing turbojet behaviour and attained a satisfactory agreement with the analytical model. Therefore, the proposed model will be helpful in pre-specify of turbojet design parameters moreover prediction at various flight speed and atmospheric conditions.