CABLEGNOSIS Publication in IEEE –Power Cable Ageing Classification Research
As part of the project’s dissemination activities, CABLEGNOSIS researchers (Mohsen Abdolahi, Wenjuan Song and Mohammad Yazdani-Asrami), from the “Propulsion, Electrification and Superconductivity” group, Jamse Watt school of engineering, University of Glasgow, United Kingdom, published a conference paper titled as “ML-Assisted Ageing Classification of XLPE Power Cables using an Adaptive Neuro-Fuzzy Inference System” in IEEE Xplore on 11 November 2025.
The ageing of XLPE cable insulation reduces transmission and distribution reliability, while traditional statistical and empirical models fail to capture its nonlinear, multi-stress ageing behaviour. Our study develops an machine learning-based (namely ANFIS) model to classify the health index of aged XLPE cables, with parameters optimised through sensitivity analysis and performance assessed using commonly-used evaluation metrics. The proposed model in this conference paper, learned ageing patterns and accurately predicted health indices for unlabelled and historical data, demonstrating strong generalisation capability. Moreover, ANFIS provides interpretable rule-based outputs that support expert decision-making. ANFIS effectively handled nonlinear input relationships and achieved classification accuracy well-above 98%. The advances offered by the research in this paper is well-aligned with the objectives of the CABLEGNOSIS project.
The findings of this paper offers several benefits for real-world cable asset management, as they enable data-driven prioritisation of maintenance and replacement, thereby improving system reliability and cost efficiency.
CABLEGNOSIS IEEE Xplore paper on ML-Assisted Ageing Classification of XLPE Power Cables using an Adaptive Neuro-Fuzzy Inference System
