Bayesian Hierarchical Model Replication for Measuring System Reliability in Industrial Machinery Fleets in Kenya

The reliability of industrial machinery fleets is crucial for optimising maintenance schedules, reducing downtime, and ensuring operational efficiency in Kenya's manufacturing sector. A replication study using data from three randomly selected industrial sites in Nairobi, Mombasa, and Eldoret. The analysis employs a Bayesian hierarchical linear regression model with robust standard errors to account for site-specific variations. The model accurately predicted system failure rates with an average absolute error of ±3% across all sites, indicating high reliability estimates within the specified confidence intervals. This study confirms the utility and accuracy of the Bayesian hierarchical model in assessing industrial machinery reliability in Kenya's diverse geographical settings. The findings suggest that policy makers should consider implementing this method for fleet maintenance planning to enhance overall system performance. The maintenance outcome was modelled as $Y_{it}=\beta_0+\beta_1X_{it}+u_i+\varepsilon_{it}$, with robustness checked using heteroskedasticity-consistent errors.