Definition
Percentage of PRA model input parameters whose values reflect plant equipment condition data less than N days old, measuring how current the risk model is relative to the decisions it informs.
The Measurement Gap
Risk-informed decision-making in nuclear plants depends on the PRA model reflecting the plant's current equipment condition. But PRA inputs are typically updated on batch schedules (quarterly, during outages, or when a specific event triggers a revision), while the plant's condition changes continuously. Between updates, the plant makes risk-informed decisions (Maintenance Rule (a)(4) assessments, online maintenance configurations, outage work sequencing) using risk numbers that may not represent the current state. No legacy metric measures PRA input staleness. Schedule compliance tracks whether work gets done on time. PM completion rate tracks maintenance execution. Maintenance Rule (a)(2) criteria track system performance against goals. None of them measure whether the PRA model's inputs reflect the equipment condition data the plant is actually generating. The invisible failure mode is that PRA-informed decisions appear quantitatively justified but are based on a model whose inputs have drifted from the plant's actual condition, and no metric makes this gap visible.
Decision Supported
Are risk-informed decisions being made with risk numbers that represent the plant's current state, or with stale inputs from the last model update?
Formula
Plain language: Inventory every component reliability and availability parameter that feeds the PRA model. For each, check when it was last updated with actual plant performance data. Count the fraction that are current (updated within the currency window). That's the currency index.
Notation:
PRICI = Σ (I(age_i ≤ N)) / P
Where:
I(condition) = 1 if condition is true, 0 otherwise
Variables:
- age_i = calendar days since PRA input parameter i was last updated with actual plant data (days)
- N = currency window threshold (recommended: 90 days for components in active service, 30 days during outage periods when configurations change rapidly)
- P = total number of PRA input parameters (count, typically 500–2,000 for a full-scope PRA)
- I = indicator function (1 if current, 0 if stale)
Sign convention: PRICI = 1.0 means all inputs are current. Lower is worse. 0.0 means no inputs have been updated within the currency window.
Aggregation period: Snapshot at each PRA model invocation (report the currency at the time the risk-informed decision was made, not an average).
Edge Cases
- Parameters for components that haven't operated since last update: if the component has been in standby and the PRA input reflects standby reliability, the parameter is current by definition. Don't penalize for non-operation
- PRA model updates are batch processes: many plants update PRA inputs quarterly or during outages. PRICI will drop between updates. That's the point: the metric makes visible the risk of inter-update staleness.
- Some parameters change slowly (large motor reliability) vs. quickly (valve stroke time trending): weight by rate of expected change if a refined version is needed, but the unweighted version is the right starting point
Proposed formula, pending validation pilots.