"background": "The reliability of industrial machinery fleets is a critical determinant of productivity and economic output in developing industrial economies. However, rigorous, field-based methodologies for assessing the operational reliability of such systems in these contexts are notably absent from the literature, leading to reliance on theoretical models or aggregated failure data. ", "purpose and objectives": "This paper presents a methodological evaluation of a randomised field trial (RFT) designed to quantify the in-service reliability of heavy machinery fleets within the Ethiopian industrial sector. The primary objective was to develop and test a protocol for collecting high-fidelity, time-to-failure data under real operating conditions. ", "methodology": "A stratified randomised sampling approach was employed across four major industrial zones. A total of 127 units from three machinery classes (excavators, haul trucks, loaders) were enrolled. Each unit was instrumented for continuous condition monitoring. Reliability was modelled using a Weibull proportional hazards model: h (t|X) = \\{\ \ (\ \) ^\-1 \ (\T X), where X represents covariates including duty cycle and maintenance regime. Inference was based on robust standard errors clustered by fleet operator. ", "findings": "The RFT protocol proved feasible, achieving a 94% data completeness rate. The analysis revealed a significant positive association between scheduled, predictive maintenance and time-to-failure, with a hazard ratio of 0. 62 (95% CI: 0. 51 to 0. 75). Operational intensity was the dominant covariate affecting failure rates across all machinery classes. ", "conclusion": "The randomised field trial constitutes a viable and robust methodological framework for generating empirical reliability data in an industrialising context. It moves reliability engineering practice beyond assumptions based on data from mature economies. ", "recommendations": "Industry regulators and large fleet operators should adopt similar field-based reliability assessment protocols to inform maintenance policy and capital planning. Further trials should incorporate
Mekonnen et al. (Tue,) studied this question.