"background": "The reliability of power-distribution networks is a critical concern for infrastructure development and economic stability. Existing reliability assessments often lack methodological rigour in accounting for the hierarchical structure of network components and regional operational variances. ", "purpose and objectives": "This study aims to methodologically evaluate power-distribution equipment systems and develop a robust statistical model to analyse the determinants of system reliability. The objective is to quantify the influence of equipment age, maintenance regimes, and environmental factors on failure rates. ", "methodology": "A multilevel regression modelling approach was employed, treating individual substations and transformers as nested within municipal regions. The core model is specified as \ () = \0 + \1 X{ij + uj + eij, where is the failure rate for asset i in region j, Xij are asset-level covariates, and uⱼ is a region-level random effect. Analysis utilised a bespoke national dataset of asset performance and maintenance records. ", "findings": "Transformer age exhibited a statistically significant positive relationship with failure rates, with each additional decade in service associated with a 34% increase (95% CI: 28% to 40%). The intra-class correlation coefficient indicated that 22% of the variance in reliability outcomes was attributable to differences between municipal regions. ", "conclusion": "The methodological evaluation confirms that a multilevel framework is essential for accurate reliability analysis in this context, as it captures significant regional heterogeneity. Asset ageing is a predominant, quantifiable driver of declining network performance. ", "recommendations": "Asset replacement programmes should be prioritised based on the quantified age-reliability relationship. Utilities should adopt hierarchical modelling in their planning tools to better allocate resources across diverse regional networks. ", "key words": "power distribution, reliability engineering, multilevel modelling, infrastructure asset management, regression analysis", "contribution statement": "This paper provides a novel methodological framework for reliability analysis
Lebo van der Merwe (Tue,) studied this question.