Journal Design Engineering Masthead
African Civil Engineering Journal | 13 September 2010

Comparative Cost-Effectiveness of Maintenance Methodologies for Industrial Machinery Fleets in Kenya

A Randomised Field Trial
N, j, e, r, i, M, w, a, n, g, i, ,, A, m, i, n, a, J, u, m, a, ,, K, a, m, a, u, O, t, i, e, n, o
Predictive MaintenanceCost-EffectivenessField TrialFleet Management
Predictive maintenance cohort showed 23% lower total costs versus preventive.
Run-to-failure strategies incurred significantly higher unscheduled downtime costs.
Analysis employed a generalised linear model with robust, clustered standard errors.
Findings advocate for investment in condition monitoring and data analytics.

Abstract

{ "background": "Industrial machinery fleets are critical assets for economic development, yet maintenance strategies in many developing economies are often ad hoc, leading to high lifecycle costs and operational downtime. There is a paucity of rigorous field data comparing the long-term cost-effectiveness of structured maintenance methodologies in such contexts.", "purpose and objectives": "This study aimed to empirically compare the cost-effectiveness of three prevalent maintenance methodologies—preventive, predictive, and run-to-failure—for industrial machinery fleets within the local operating environment. The primary objective was to determine which strategy yields the lowest total cost of ownership while maintaining asset availability.", "methodology": "A randomised field trial was conducted using a fleet of 72 similar heavy-duty vehicles from multiple industrial sites. Units were randomly assigned to one of the three maintenance cohorts. Cost-effectiveness was analysed over a full operational cycle using a generalised linear model: $Ci = \\beta0 + \\beta1 Mi + \\beta2 Ui + \\epsiloni$, where $Ci$ is total cost for unit $i$, $Mi$ denotes maintenance cohort, and $Ui$ is utilisation. Inference was based on robust standard errors clustered by site.", "findings": "The predictive maintenance cohort demonstrated superior cost-effectiveness, reducing total maintenance costs by an average of 23% (95% CI: 18% to 28%) compared to the preventive strategy. The run-to-failure approach, while lower in direct maintenance spend, resulted in significantly higher unscheduled downtime costs.", "conclusion": "Predictive maintenance, enabled by condition monitoring, is the most cost-effective methodology for managing industrial machinery fleets in the studied setting, challenging the prevailing reliance on scheduled preventive protocols.", "recommendations": "Fleet managers should invest in condition monitoring technologies and data analytics capabilities to enable a predictive maintenance paradigm. Policymakers should consider initiatives to build local technical capacity for advanced maintenance practices.", "key words": "maintenance strategy, cost-effectiveness, randomised controlled trial