A JOINT PROBABILISTIC FRAMEWORK TO INTEGRATE RAM SIMULATION AND ECONOMIC FORECASTING WITHOUT DOUBLE‑COUNTING DOWNTIME

Abstract

Reliability, Availability, and Maintainability (RAM) simulation and economic lifecycle forecasting are routinely performed as sequential or parallel analyses in asset-intensive industries. This separation creates a systemic error: downtime events are either counted twice (once in RAM metrics, once in financial models) or counted inconsistently, leading to distorted net present value (NPV) estimates and suboptimal investment decisions. This paper proposes a joint probabilistic framework (JPF) that integrates RAM simulation and economic forecasting within a unified Monte Carlo engine, enforcing a single source of truth for downtime events across reliability and financial domains. Using a case study of a natural gas compression station with three compressor trains, we compare three methods: (1) sequential RAM-then-economic analysis (baseline), (2) parallel analysis with manual reconciliation (industry practice), and (3) the proposed JPF. Over 10,000 simulated operating years, the baseline method double-counts downtime in 34% of scenarios, overstating lifecycle costs by 18.7% on average. The proposed JPF eliminates double-counting entirely, reduces NPV estimate variance by 52%, and correctly identifies the optimal maintenance strategy (predictive vs. preventive) that sequential analysis misclassifies. The framework offers a rigorous, computationally tractable approach for integrated asset investment planning.