"background": "Reliability engineering in manufacturing contexts of developing economies is under-researched, with a paucity of longitudinal studies that account for operational and environmental variability. Existing assessments often rely on cross-sectional data, which fails to capture the dynamic nature of system performance and maintenance effects. ", "purpose and objectives": "This study aims to methodologically evaluate system reliability within manufacturing plants and to develop a robust panel-data estimation framework for quantifying reliability metrics. The objective is to identify key operational determinants of system failure and to provide a validated model for predictive maintenance planning. ", "methodology": "A panel dataset was constructed from repeated technical audits and maintenance logs across multiple plants. System reliability was modelled using a generalised linear mixed model (GLMM) with a logit link for failure events. The core specification was \ (p{it) = \0 + \1 Xit + \ +, where pit is the probability of system failure for plant i at time t, Xit is a vector of time-varying covariates, and \ᵢ captures plant-specific random effects. Inference was based on robust standard errors clustered at the plant level. ", "findings": "The panel model revealed that preventive maintenance adherence had a significant negative effect on system failure probability (coefficient: -0. 87, 95% CI: -1. 24 to -0. 50). A one standard deviation increase in scheduled maintenance compliance was associated with a 32% reduction in the odds of a major failure event. Ambient humidity and power fluctuation severity were also statistically significant covariates. ", "conclusion": "The methodological approach demonstrates that panel-data techniques are essential for accurately modelling manufacturing system reliability in this context, as they control for unobserved heterogeneity. The findings confirm the critical role of disciplined maintenance regimes in sustaining operational integrity. ", "recommendations": "Manufacturing plant managers should implement structured data collection systems to enable panel analysis for reliability forecasting.
Sarr et al. (Sat,) studied this question.