A K-NN Algorithm-Based ML Model for Predictive Maintenance in Aircraft Engines

Abstract

Engine maintenance is a vital cog in the airworthiness of an airplane that ensures the safety of passengers. Therefore, effective forecasting of the Remaining Useful Life (RUL) of aircraft’s engine components is critical for ensuring the safety and reliability of aircraft operations. Since small defects can lead to catastrophic consequences, precise RUL predictions are indispensable in aviation maintenance practices. This research paper aims to establish the efficacy of Machine Learning (ML) methods in the context of Predictive Maintenance (PdM) of aircraft engines. Several factors influence the RUL of an engine component that can only be considered into account when real-time data analytics is used for the prognosis of a machine's state and PdM. The proposed method employs batch processing and real-time stream data to facilitate health monitoring and predictive analysis of RUL of engine components. It is found to enhance maintenance prediction and optimize the overall service aircraft operators provide. Ultimately, this research has employed various ML algorithms for effective prediction of RUL and it is established that in the field of PdM, the technique of the K-Nearest Neighbours (K-NN) algorithm performs better than other ML algorithms.