Abstract
{ "background": "The reliability assessment of power-distribution equipment in South Africa is critical for infrastructure planning and maintenance. Current reliability models often rely on laboratory data or historical failure rates, which may not accurately capture performance under diverse, real-world operational and environmental stresses.", "purpose and objectives": "This working paper presents a methodological evaluation of a randomised field trial (RFT) designed to measure the in-service reliability of key distribution equipment, including transformers and circuit breakers. The primary objective is to assess the RFT's methodological rigour and feasibility for generating robust reliability data.", "methodology": "A stratified randomised design was implemented, allocating units of equipment from multiple manufacturers across different climatic and load zones. Reliability was modelled using a Weibull proportional hazards model: $\\lambda(t|\\mathbf{x}) = \\frac{\\beta}{\\eta} \\left( \\frac{t}{\\eta} \\right)^{\\beta-1} \\exp(\\mathbf{x}^T \\boldsymbol{\\gamma})$. The analysis considers right-censored data and reports 95% confidence intervals for hazard ratios.", "findings": "The methodological evaluation indicates that the RFT design successfully generated time-to-failure data under representative conditions. A key theme was the significant variation in reliability linked to environmental stressors, with equipment in coastal regions showing a hazard ratio of 1.8 (95% CI: 1.3, 2.5) compared to inland temperate zones. Practical challenges in maintaining randomisation protocols in a live network were identified.", "conclusion": "The randomised field trial is a methodologically sound approach for collecting high-fidelity reliability data, though it requires substantial logistical coordination. It provides a superior evidence base compared to retrospective analyses.", "recommendations": "Future trials should incorporate more granular environmental monitoring and extend the asset sample to include newer composite materials. Network operators should consider adopting RFT principles for targeted asset cohorts.", "key words": "reliability engineering, power distribution, field trial, randomised experiment