The Need

Operators of large, complex facilities struggle to balance revenue, maintenance, and regulatory safety constraints under uncertainty. Existing tools typically optimize only a subset of factors without a unified, real‑time view of component health and future degradation. Advanced reactors and other modern plants intensify the challenge given limited operating history and evolving sensor networks. A decision framework that simultaneously ingests live diagnostics, forecasts state transitions, quantifies economic and safety risk, and outputs an optimized, license‑acceptable plan is missing.

The Technology

OSU engineers have developed a novel technology that integrates online monitoring/diagnostics with Markov component behavior models and risk models to drive a Markov Decision Process (and POMDP when states are partially observed). It fuses real‑time component status probabilities, a generation risk assessment (economic value), and a probabilistic risk assessment (safety) to evaluate operational and maintenance strategies and select an asset‑management policy that maximizes value while remaining within licensing/safety boundaries. The output is an optimized, implementable plan for the upcoming operating window.

Commercial Applications

• Nuclear utilities: fleetwide asset‑management and advanced reactor operations optimization.
• Thermal power generation (gas/coal/biomass): value‑optimized maintenance and derate decision support.
• Petrochemical and chemical processing plants: real‑time risk‑informed asset planning with sensor‑rich equipment.
• Heavy industrial facilities with online monitoring: unified economics‑plus‑safety decision support.

Benefits/Advantages

• Unified value + safety optimization: concurrently evaluates GRA (economics) and PRA (safety) to propose license‑acceptable strategies.
• Real‑time, forward‑looking decisions: uses current diagnostics and Markov forecasts of degradation/failure to anticipate future states.
• Scalable to uncertainty: supports POMDP when plant states are inferred rather than observed.
• Operational impact: reduces unnecessary maintenance, avoids unplanned derates/trips, and maximizes revenue within regulatory limits.