"background": "Industrial machinery fleets in developing economies face significant operational risks, yet structured, data-driven methodologies for quantifying and mitigating these risks are scarce. Existing frameworks often lack empirical validation in real-world settings, particularly in West African industrial contexts. ", "purpose and objectives": "This data descriptor presents a novel methodological framework for evaluating machinery fleet systems and provides a corresponding dataset from a randomised field trial. The primary objective was to empirically measure the efficacy of a structured risk-reduction protocol on machinery availability and incident frequency. ", "methodology": "A randomised controlled trial was conducted with a fleet of heavy industrial equipment. Participants were randomly assigned to an intervention group, which implemented a predictive maintenance and operator training protocol, or a control group maintaining standard practice. The primary analysis used a generalised linear mixed model: \ (E (Y{it) ) = \0 + \1 Ti + \2 Xit + ui +, where Yit is the monthly incident count for machine i at time t, Ti is the treatment indicator, and uᵢ is a random machine effect. Robust standard errors were clustered at the operator level. ", "findings": "The intervention yielded a statistically significant reduction in unplanned downtime. Specifically, machinery in the treatment group exhibited a 34. 7% lower incidence rate of major mechanical failures (95% CI: 22. 1% to 45. 2%) compared to the control group over the trial period. ", "conclusion": "The methodological framework proved viable for field deployment and generated high-quality, structured data. The trial demonstrates that a randomised design can be successfully implemented to rigorously assess engineering risk interventions in an industrial setting. ", "recommendations": "Future research should apply this methodology to other machinery classes and geographic regions to test its generalisability. Industrial operators are encouraged to adopt similar randomised evaluations for new safety or maintenance protocols. ", "key words": "Predictive maintenance, randomised controlled
Boateng et al. (2026) studied this question.