Prediction of compressor nominal characteristics of a turboprop engine using artificial neural networks for build standard assessment

Abstract

Abstract Compressor characteristics of a single spool turboprop engine have been studied in this paper. It has been brought outhow constant power lines in the compressor characteristics of these compressors make them different from others. Constant speed lines and constant power lines have also been highlighted. A novel method of modeling of compressorof a single spool turboprop engine has also been studied in this paper. Application of neural networks in prediction of compressor characteristics has been investigated. Multilayer Perceptron feed forward neural network has been considered with different transfer functions to assess the potential capability of network in extrapolation and interpolation. Effectiveness of prediction with and without engine bleed valve open and anti-ice valve open situations have been assessed. Network Predictionshas been compared with engine test data to assess the accuracy of prediction and to quantify the build variation in the manufacture of engines. Capability of network with limited test data to predict the complete performance has also been assessed and presented in this paper.