Extreme vibration environmental poses significant challenges to the fatigue life of aircraft cables. This study aims to propose a novel method to investigate the fretting wear life of aircraft cable under random vibration.
In the presented method, the cable degradation data and machine learning (ML) are used to capture the complex, non-linear relationships between the wear rate (Δs/Δt) and the initial worn thickness. Three ML algorithms – k-nearest neighbor (KNN), support vector regression (SVR) and random forest regression (RFR) – are compared for their performance in surrogate modeling. A physical boundary condition is incorporated to expand the sample set, and its influence on model training is systematically evaluated. The proposed methodology uses a time-efficient alternative for wear lifetime assessment by integrating surrogate modeling with Monte Carlo simulation.
The SVR model demonstrates greater potential in predicting the wear rate compared to KNN and RF. The SVR model, combined with the Monte Carlo algorithm, predicts a wear life of 14,700 h for the cable under random vibration conditions.
In open literature, it is rare to find the fretting wear life of cable insulation, particularly under extreme vibration environmental. The study significantly enhances predictive maintenance strategies and improves the reliability of aircraft cables.
The peer review history for this article is available at: https://publons.com/publon/10.1108/ILT-02-2025-0070/
