Strengthening occupational safety, health, and environment risk assessments in cement manufacturing under stochastic production pressure using hybrid predictive models

• Published in: Discover applied sciences Vol. 7; no. 9; p. 991
• Main Authors: Ssemuddu, Jeffy Briton, Olupot, Peter Wilberforce, Kirabira, John Baptist, Okure, Mackay, Kayenga, Tendo Joshua
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 26.08.2025; Springer Nature B.V
• Subjects: Accuracy; Algorithms; Applied and Technical Physics; Carcinogens; Cement; Cement industry; Cement plants; Chemistry/Food Science; Compliance; Confined spaces; Earth Sciences; Engineering; Environment; Failure; Fractal analysis; Hazard assessment; Hazard identification; Health risk assessment; Manufacturing; Materials Science; Mechanical properties; Occupational health; Occupational safety; Optimization; Prediction models; Productivity; Protective equipment; Quarrying; Questionnaires; Ratings; Raw materials; Risk assessment; Search algorithms; Sustainable development; Swarm intelligence; Triangulation; Working hours; Workplace diversity; Sub-Saharan Africa; Predictive modelling; Cement industry; Risk assessment; Risk ratings; Occupational Safety; Production pressure
• DOI: 10.1007/s42452-025-07594-x
• ISSN: 3004-9261; 2523-3963

Cement manufacturing is considered a high-risk industry due to hazardous operations such as quarrying, maintenance, and blending, which frequently lead to occupational injuries and elevated incident rates. The increasing demand for cement in Sub-Saharan Africa intensifies stochastic production pressure, further escalating these risks. This study identifies and assesses common hazards associated with routine cement manufacturing activities and predicts risk ratings under dynamic and hypothetical production pressure scenarios in two Ugandan cement plants. The study surveyed over 400 workers to assess hazard awareness. It validated the findings through direct field observations and structured checklist-based site inspections, enabling robust triangulation between perceived and observed risks. Using a 6 × 6 risk assessment matrix, baseline ratings were determined, while predictive modeling using the different Artificial Neural Fuzzy Interface System (ANFIS) algorithms like Anti Colony Optimization, Bee Colony Optimization, and Swarm Fractal Search analyzed risk fluctuations under varying production pressures and working hours. Results indicated that confined spaces, unsafe behaviors, mechanical, and electrical hazards posed the most significant risks, with risk ratings exceeding 30 on a 0–36 scale under normal conditions and escalating to 35–36 under increased production pressure and non-standardized working shifts. Implementing adaptive and predictive risk assessment models, providing advanced Personal Protective Equipment, and enhancing the quality of task supervision are essential to mitigating these risks. These findings contribute directly to Sustainable Development Goals 3, 8, and 9 by promoting improved occupational health, safer work environments, and industrial innovation in high-risk sectors.