Quantitative reliability and airworthiness analysis of propulsion systems of multi-rotor UAVs for certification in civil airspace

• Main Author: Govind, Siddesh
• Other Authors: Mir Feroskhan
• Format: Thesis-Master by Research
• Language: English
• Published: Nanyang Technological University 2025
• Subjects: Engineering; UAV reliability; Airworthiness certification; Multi-rotor UAVs; Propulsion systems; Failure mode and effects analysis (FMEA); Markov chains; Degradation modelling; Weibull analysis; Predictive maintenance; Unmanned aerial vehicles (UAVs; Urban air mobility (UAM); Bayesian modelling; Deep learning in UAVs; LSTM networks; Reliability engineering; Drone performance evaluation; Experimental testbeds; Component-level analysis; System-level reliability; Predictive modelling
• Online Access: https://hdl.handle.net/10356/182045
• Institution: Nanyang Technological University

Unmanned Aerial Vehicles (UAVs) have become increasingly integral to a diverse array of industrial applications, including military, civilian, and commercial uses. Ensuring their airworthiness is critical for the safe and effective operation of UAVs. This necessitates advancements in sophisticated urban air traffic management strategies and the implementation of robust evaluation tools. Assessing UAV airworthiness requires a detailed examination of their reliability or health state, utilizing available data and model-based evaluation criteria. Evaluating the reliability of UAVs involves analysing their capacity to perform designated functions reliably, safely, and efficiently over a specified period. This critical evaluation requires continuous monitoring and analysis of data pertaining to the UAV’s operational performance, including indicators such as reliability, maintainability, and availability. In this context, this thesis explores and establishes diverse methodologies adaptable to various scenarios, considering the format and quantity of available source data. By employing both statistical and data-driven approaches, the research developed models for assessing the reliability and performance degradation of UAVs at both the component and system levels. Through FMEA, critical failure modes in UAV propulsion systems were identified, revealing that motor failures account for approximately 45% of total failures. Weibull analysis predicted motor failures with a 90% confidence interval, highlighting wear-out mechanisms as predominant. Additionally, LSTM-based models for predicting UAV performance degradation achieved an accuracy of 85% in forecasting remaining useful life (RUL). These models demonstrated significant potential in enhancing predictive maintenance strategies, reducing downtime by up to 30%, and improving the overall reliability and safety of UAV operations. This work aims to refine UAV airworthiness evaluation methodologies, ensuring their safe integration and operation in complex airspace systems.