Development of a Predictive Tool for the Parametric Analysis of a Turbofan Engine

Abstract

Parametric cycle analysis, an on-design engine study, specifies the required design characteristics that optimize engine performance. This study aimed to conduct a parametric analysis of a low-bypass turbofan engine with an afterburner, F100-PW229, and develop a technique for estimating its performance based on data using machine learning and deep learning. Commercially available gas turbine simulation software, GasTurb 14, was used to create a dataset of engine performance response variables and input design parameters. The effects of the Mach number, fan pressure ratio, altitude, turbine entry temperature, and bypass ratio on the specific thrust, propulsive efficiency, specific fuel consumption, and total fuel flow were investigated. Regression learning models and deep neural networks were then programmed on this dataset to predict responses for new input data. In MATLAB, a total of 24 regression models were trained with cross-validation, and the model with the least root mean square error was selected as the final model. The machine learning regression models produced reliable output parameter predictions, with the least root mean square error of 9.076 × 10−5. Among the numerous regression models tested, Gaussian process regression, the quadratic support vector machine, and the wide neural network emerged to be the most successful in predicting turbofan engine performance metrics. A multilayer perceptron model was coded in Python with two hidden layers that accurately predicted the performance parameters. The mean square error value on test data was found to be as low as 0.0046. In comparison to intensive computational simulations, machine learning and deep learning models offer an efficient method for conducting parametric analysis of turbofan engines.