A diagnostics tool for aero-engines health monitoring using machine learning technique

In this work an integrated heath monitoring platform is proposed and developed for performance analysis and degradation diagnostics of gas turbine engines. The aim is to link engine measurable data to its health status. A numerical tool has been implemented in order to calculate engine performance in design condition and to create a database of expected vales. Then different degradation levels have been introduced in the two main components, compressor and turbine of a single spool turbojet and the diagnostics instruments have been trained to detect the component fault. In order to evaluate the performance prediction two different machine learning based techniques, namely, artificial neural network (ANN) and support vector machine (SVM) have been compared. Synthetic data generation has been carried out to show how the degradation effects can affect the engine performance. The two main degradation causes considered are the compressor fouling and turbine erosion. The machine learning techniques were applied with two aims: aero-engine performance prediction and health diagnostics. The study was carried out based on three samples flights, whose data were used for the training and testing process of the prediction and diagnostics tools. The knowledge and the continuous monitoring of the engine health status can be crucial for maintenance and fleet management operations.