Digital twin for Electronic Centralized Aircraft Monitoring by machine learning algorithms

Abstract

Electronic Centralized Aircraft Monitoring (ECAM) parameters play a vital role in the operation of an aircraft to reduce the workload of the cockpit crew. A wide-body commercial aircraft with a triple-spool turbofan engine is examined within the scope of the study. This study is focused on the estimation of the ECAM primary engine parameters: Engine Pressure Ratio, Exhaust Gas Temperature, Fuel Flow, and Shaft Speeds without any additional measurement for data continuity. The recorded flight data obtained from a commercial aircraft is processed with machine learning methods, and the most suitable estimation method is tried to be determined. Correlation analysis is carried out for each data in the study to show strong predictor candidates. The modeling process is conducted by using MATLAB. Results indicate that the Fine Decision Tree is better at memorizing data, while the Wide Neural Network is better at generalizing data. Computational results show that the developed models are outstandingly precise and accurate to estimate aircraft's ECAM data to ensure flight safety for health and performance monitoring of an engine. Thus, when an unreliable situation occurs while performing the flight in practical conditions, the cockpit crew will be able to overcome this situation by Digital Twin.