Machine Learning-based Fault Detection and Diagnosis in Aircraft Systems

Abstract

This study offers a novel method for improving fault detection in aircraft systems by incorporating machine learning (ML). With an interpretivism philosophy, the study develops a technical framework by using a descriptive design and a deductive approach. Through the utilization of secondary data, the study assesses different machine learning algorithms, tackles technical difficulties, and alongside verifies the practical use of the framework. The effectiveness of supervised and unsupervised machine learning models is revealed through performance evaluation, which emphasizes accuracy and flexibility. Innovative solutions are employed to tackle technical implementation challenges, which include data compatibility as well as real-time processing. The usefulness of the developed framework is demonstrated in a variety of fault scenarios through testing in both simulated and real-world aviation scenarios. Industry standard alignment is ensured by expert validation. The study represents an important development in proactive fault detection for aircraft systems by providing a strong technical methodology.