Aircraft Engine Fault Detection Algorithm Based on Multivariate Time Series Sensor Data

In the rapidly evolving landscape of civil air transportation, ensuring the safety and reliability of aircraft engines remains a critical concern. This study presents a novel aircraft engine fault detection algorithm based on multivariate time series sensor data, with the aim of enhancing the reliability and safety of air transportation. The proposed algorithm utilizes advanced data analysis techniques and intelligent algorithms to accurately predict potential engine failures, thereby facilitating proactive maintenance and minimizing aircraft delays. The study reviews state-of-the-art methods in aircraft engine fault detection, highlighting advances in dynamic radius support vector data description, modified moving window kernel principal component analysis, hybrid extended Kalman filter, and cross-domain guided learning methods, among others. In addition, the study proposes a comprehensive methodology for aircraft engine fault detection that integrates sensor data acquisition, pre-processing, fault detection algorithms, and performance optimization using multivariate time series analysis. Experimental results demonstrate the effectiveness of the proposed algorithm in achieving high detection rates and specificity for known fault types, while addressing challenges related to computational complexity and generalization.