Markov chain-based model for IoT-driven maintenance planning with human error and spare part considerations

Abstract

This study introduces a Markov-based optimization framework for maintenance and spare parts inventory management, enhancing cost efficiency and operational reliability in cement production. By leveraging steady-state probabilities, the model integrates real-time equipment monitoring via the Industrial Internet of Things (IoT), reducing manual inspections and mitigating human errors. A comprehensive analysis demonstrates that level-2 preventive maintenance (PM) achieves the highest steady-state probability, effectively balancing cost minimization and system reliability over a 36-period planning horizon. Key optimization variables include the IoT adoption rate (γ = 0.72), human error probability (HEP) ($p^{\left[Total\right]}$ = 0.137), and total cost objective ($z^{\left[Total\right]}$= 3151,385 currency units). The model dynamically adjusts inventory replenishment policies to minimize stockouts and reliance on costly emergency orders. Results indicate that the proposed framework significantly improves maintenance scheduling, optimizes resource allocation, and reduces operational downtime. Furthermore, the study underscores the model's adaptability and its potential for integration with predictive analytics, paving the way for intelligent, data-driven maintenance strategies. These findings provide a strong foundation for advancing industrial maintenance optimization and operational efficiency.