562 Asset reliability was characterized by analyzing four years of historical maintenance data from SAP records (2020–2024). This data enabled statistical fitting for over 50 critical asset groups using TwoParameter Weibull distributions. This method accurately represents variable failure rates, capturing both early-life failures (β < 1) and the wear-out phases of rotating equipment, a technique successfully applied in brownfield expansion studies by (A. Noiseux & J. P. Côté, 2010). 3.3 Baseline Validation Protocol The reliability analysis focused on identified constraint areas, particularly the classification, pumping, and dewatering-related constraint areas. The resulting MTBF, mean time to repair (MTTR), and Weibull parameters used for the simulation baseline are summarized in Table 1. The model executed 50 Monte Carlo replicates per scenario, ensuring a confidence interval greater than 95% for the final throughput results. Table 1 – Reliability Parameters for Representative Critical Assets Included in the Baseline Area Asset Description MTBF (h) MTTR (h) Parameter β Parameter η 2030 FI-2030-01 Horizontal Belt Filter 149.9 1.38 0.6093 149.9 2020 BP-2020-03 Centrifugal Pump 283.2 2.87 0.4268 283.2 2020 MI-2020-01 Scrubber Drum 203.1 1.8 0.4001 203.1 2020 PN-2020-03 Schenck Screen 95.68 0.91 0.4826 95.68 2030 TR-2030-01 Conveyor Belt 299.7 0.96 0.4956 299.7 The model's validity was established through a calibration phase using the 2024 plant configuration and operational data. The system was stress-tested under high-grade conditions (19.44 % P2O5), with a success criterion defined as an absolute deviation in annual throughput of less than ±1% relative to historical production. This ensures that all expansion projections are derived from a robust technical baseline that accurately mirrors the plant's physical and operational constraints. 3.4 Simulation Execution and Performance Metrics To ensure statistical stability, each simulation scenario was executed using a warm-up period of 120 simulated hours, allowing internal buffers and recycle loops to reach operational equilibrium—a practice recommended for infrastructure expansion planning by (J. T. McGinty et al., 2010). The model executed 50 Monte Carlo replicates per scenario, with each replicate representing a full calendar year (8,760 hours). The primary output metrics included Effective Utilization (Eᵤ) and dynamic capacity constraints. By tracking these KPIs, the simulation captured the volumetric constraints triggered by the 76% increase in mass pull, consistent with the flowsheet simulation approaches for specialized technologies documented by (C. Stewart, 2005). 3.5 Model Calibration and Validation The calibration of the stochastic discrete event simulation (DES) model was a fundamental phase in establishing the study's technical authority, transforming the simulation into a verified analytical tool. Utilizing high-fidelity historical data from the 2024 operational cycle, the model’s internal logic and throughput parameters were iteratively refined to replicate the plant's physical response. This rigorous calibration achieved a global throughput variance of less than 1% relative to historical production records, confirming the accuracy of the modeled interactions between asset reliability and process constraints. As a result, the validated model provides a robust baseline for stress-testing expansion scenarios, ensuring that all
RkJQdWJsaXNoZXIy MTM0Mzk2