"background": "Industrial machinery fleets are critical to economic productivity, yet systematic reliability data for such assets in sub-Saharan Africa is scarce. Current maintenance strategies are often reactive or based on manufacturer schedules not calibrated to local operating conditions, leading to costly downtime. ", "purpose and objectives": "This study aimed to develop and apply a novel randomised field trial (RFT) methodology to empirically assess the reliability of industrial machinery fleets under typical Kenyan operational environments. The objective was to generate locally validated failure rate data to inform maintenance optimisation. ", "methodology": "A randomised controlled trial was designed, assigning 120 heavy-duty vehicles from a national fleet to either a condition-based maintenance (CBM) group or a standard schedule-based maintenance (SBM) control group. Operational data were collected over a defined period. Reliability was modelled using a Weibull proportional hazards model: h (t|X) = \\{\ \ (\ \) ^\-1 \ (\ X), where X indicates the treatment group. Inference used robust standard errors clustered by vehicle type. ", "findings": "Machines under the CBM protocol exhibited a 32% lower hazard of catastrophic failure (95% CI: 18% to 44%) compared to the SBM group. The shape parameter (\) estimate of 2. 1 indicated an increasing failure rate with time for both groups, underscoring wear-out dominance. ", "conclusion": The RFT proved a viable method for generating rigorous reliability data in this context. Condition-based approaches, informed by local operational data, significantly enhance reliability compared to rigid calendar-based schedules. ", "recommendations": "Fleet managers should adopt condition-based maintenance strategies underpinned by localised reliability monitoring. Further RFTs should be conducted across different machinery classes and sectors to build a comprehensive reliability database. ", "key words": "reliability engineering, randomised field trial, condition-based maintenance, industrial machinery, asset
Wanjiku Mwangi (Sun,) studied this question.