A New Methodology For Engine Installation Effect Prediction Using Machine Learning

This paper presents a machine learning approach to predict the effects of installing new air intake system, thus reducing the need for extensive flight tests. The following method is used: firstly, a machine learning model is used to predict the characteristics of the flow entering the engine; secondly, an engine thermodynamic model is used to predict the engine's response to disturbance. Several models were evaluated, and the gradient boosting model proved to be the most promising, with very good R² scores and low MAEs to introduce an error as low as possible in the evaluation of installation effects. Validation by comparison with real flight tests shows that the model is able to accurately predict the aerodynamic characteristics entering the engine. Finally, this model is applied to a real case study and the installation effects computed with the method explained in this paper are used and show a very good prediction of the installation effects with less than 1\% absolute error on the prediction. This methodology offers a new way of reducing air intake development costs and improving overall performance by reliably estimating installation effects.