Intelligent Diagnosis of Engine Failure in Air Vehicles Using the ALFA Dataset

Abstract

Smart cities enable economic and social development through intelligent solutions to various problems, such as access to essential services, mobility, unnecessary energy consumption, security flaws, etc. Regarding urban mobility problems, smart cities propose the development of Urban Air Mobility (UAM) through a safe, sustainable, and affordable air transport system for passenger mobility, cargo delivery, and emergency services within or between metropolitan areas. However, these vehicles are still incipient and their implementation in cities presents challenges such as failures, security, and safety issues. In this sense and according to the database of the Center for Research and Prevention of Aeronautical Accidents (CENIPA), engine failures are the main causes of problems in air vehicles. Because of that, this study was structured to detect engine failures in electric Vertical Take-Off and Landing aircraft (eVTOLs). We propose a new machine learning algorithm based on Multi-Layer Perceptron, Support Vector Machine, Gradient Boosting, and Random Foresting to the detection of engine failures. The results demonstrate the effectiveness of our technique. Our strategy presents a superior detection, being 21% more effective concerning other recent studies in accuracy, using the same database as the one in this study, and the same engine failure class in aerial vehicles.