DIGITAL TWIN–RAM CO‑SIMULATION FOR MAINTENANCE AND SPARING DECISIONS IN PROCESS FACILITIES

Abstract

Maintenance scheduling and spare-parts inventory planning in process facilities depend critically on accurate, dynamically updated reliability, availability, and maintainability (RAM) models. Static RAM analyses — computed once from historical failure data and applied unchanged across multi-year planning horizons — fail to reflect the evolving condition of assets, the consequences of actual maintenance decisions, and the stochastic interaction between equipment failure events and spare-parts stockout risks. This paper presents a Digital Twin (DT)–RAM co-simulation framework that couples a physics-informed Bayesian RAM model with a high-fidelity Monte Carlo co-simulation engine to provide continuously updated, probabilistically rigorous inputs to maintenance scheduling and spare-parts optimisation. The DT layer ingests real-time condition-monitoring (CM) sensor streams, CMMS work-order records, and inventory transaction data to update Weibull failure-rate posteriors on both a quarterly calendar cycle and an event-triggered basis, while the Monte Carlo co-simulation propagates these posteriors into system-level availability forecasts, spare-parts demand distributions, and total maintenance cost trajectories. A stochastic spare-parts optimisation module derives dynamic reorder points and stock levels that minimise the joint probability of stockout and excess carrying cost under the updated posteriors. The framework is validated at a heavy petrochemical olefins complex in India across a 36-month monitoring and 24-month forecasting window covering 47 critical rotating and static equipment items. Results demonstrate a 78.7% reduction in availability forecast RMSE relative to static RAM, a 15.7% reduction in total maintenance and sparing cost, and a 67% mean reduction in critical-spare stockout events. The framework is directly compatible with SAP PM and OSIsoft PI data architectures and meets ISO 55000-aligned asset management requirements.