Abstract
{ "background": "Manufacturing systems in developing economies face complex, multi-level risks that are challenging to quantify using traditional risk assessment frameworks. There is a recognised need for robust, data-driven methodologies that can integrate sparse and heterogeneous operational data to inform targeted engineering interventions.", "purpose and objectives": "This study presents and evaluates a novel Bayesian hierarchical modelling framework designed to quantify and reduce systemic risk in manufacturing plants. The primary objective is to assess the methodological efficacy of this approach in a real-world industrial context.", "methodology": "A Bayesian hierarchical model was formulated and applied to operational performance and failure data collected from multiple manufacturing plants. The core model structure is $y{ij} \\sim \\text{Normal}(\\alphaj + \\beta X{ij}, \\sigmay)$, with $\\alphaj \\sim \\text{Normal}(\\mu{\\alpha}, \\sigma{\\alpha})$, where $i$ indexes observations and $j$ indexes plants. Inference was performed using Markov chain Monte Carlo simulation, with posterior credible intervals used to quantify uncertainty in risk estimates.", "findings": "The model successfully identified significant variation in baseline risk levels between plants, with a central 95% credible interval for the standard deviation of plant-level intercepts ($\\sigma{\\alpha}$) ranging from 1.8 to 3.2 on a standardised risk scale. A key theme was the predominant influence of maintenance protocol adherence over equipment age in reducing systemic failure risk.", "conclusion": "The Bayesian hierarchical model provides a statistically robust and operationally informative framework for risk assessment in manufacturing systems, effectively handling data limitations common in industrialising contexts.", "recommendations": "Manufacturing engineers should adopt hierarchical modelling techniques to prioritise risk mitigation investments. Further research should integrate real-time sensor data into the model's observational layer for dynamic risk forecasting.", "key words": "Bayesian inference, hierarchical modelling, risk assessment, manufacturing systems, industrial engineering, probabilistic methods", "contribution statement": "This paper introduces a novel application