26778eba-0e1f-4376-8d56-0b7aacbae636
Private Server Persistent Server Active Server Uptime: 0s
Auto-refresh: 30s
Instance
Privacy
Server Access
Private Server
Persistent Server (Active · Saved permanently)
Server Uptime:
0s
Demand
Schedule
Time to next demand shift
0s
Live dynamic countdown based on heartbeat timestamp
Next Demand
Info
Total Next Demand
0 MW
U1: 0 MW
U2: 0 MW
⏸
Sync Paused (Tab Inactive)
Click anywhere or switch back to this tab to resume live metric updates.
Reactor Metrics History
Drag horizontally to pan through history · Scroll wheel / pinch to zoom
Time Window:
Score Calculation Breakdown
Job ID:
Loading scoring calculation breakdown...
Failed to load score breakdown
Server history or snapshots may be unavailable
Top Score
--
/ 100
Raw: --
PPS Wt: 100.0%
Pen: --
Evaluation Window
Half-life: 10m
-- min
1.50x curr • 1.25x prev priority
Evaluated: --
Total Stored: --
Unit Synergy
-- dual active
+25% bonus when both units run
Single: --
Inactive: --
Events & Graces
Tolerance
Demand Graces (60s):
--
LOOP Outages Excused:
--
SCRAMs: --
Trips: --
Point Generation Sustainability
Current Snapshot: --/s (1.50x wt) •
Previous: --/s (1.25x wt) •
Weighted Avg: --/s •
Max Target: 2.9/s (U1: 1.2, U2: 1.7)
Points Efficiency Weight
--
Unit 1 Performance
-- pts avg
Generation Active
--
Avg MW Output
--
Demand Downtime
--
60s Grace Periods
--
LOOP Outages Excused
--
Avg Turbine Health
--
Avg Points/Sec
--
Demand Band (±40)
--
Unit 2 Performance
-- pts avg
Generation Active
--
Avg MW Output
--
Demand Downtime
--
60s Grace Periods
--
LOOP Outages Excused
--
Avg Turbine Health
--
Avg Points/Sec
--
Demand Band (±40)
--
Algorithm Reference & Debug Rules
In-Code Toggle: DEBUG_MODE
• 60s Grace Period: First 60s after demand change awards 75 pts to allow rod/flow adjustments.
• Demand Downtime: Active demand with APRM < 5% or Out <= 0 MW drops score to 5.0 pts.
• ±40 MW Demand Deadband: Output within ±40 MW of demand awards 100% target fulfillment; only extreme deviations past ±40 MW reduce score.
• LOOP Outage (Partial Deduction): Loss of Offsite Power (Offsite False or Demand -2) awards 50 pts; deducts some points, but not as heavily as unexcused active demand downtime.
• Refueling Outage Prediction: Unit 2 refuels (clean shutdown during event with no preceding demand downtime, depressurized, rods inserted) are evaluated by duration up to 5 hours. The fastest 50m refuel maintains high efficiency (70.0 pts, ~1,000 pts pred loss vs running). Extended refuels decay to 20.0 pts at 5h (~30,000 pts pred loss). Outages with preceding demand downtime or exceeding 5h decay to 5.0 pts (unexcused downtime).
• Multi-Unit Synergy: 1.25x (+25%) when both units run, 1.0x for one, 0.7x when both idle.
• Point Generation Weight & Recency: Point generation acts as an overall sustainability weight. Maximum plant point generation is 2.9 pts/s (1.2 pts/s Unit 1 + 1.7 pts/s Unit 2 with demand bonus). Evaluated snapshots are recency-weighted with a 10-minute half-life (600s), with the current snapshot boosted by 1.50x and previous snapshot by 1.25x. If a server does not sustain the maximum 2.9 pts/s across its history, its score is directly scaled down by its point efficiency factor (efficiency = min(1.0, pps / 2.9)), causing it to lose score.
• SCRAM/Trip Deductions: -60% snapshot multiplier and -5 pts deduction per incident.
Evaluated Snapshots Log
| Time / Age | Weight | Unit 1 Status | U1 Score | Unit 2 Status | U2 Score | Synergy | Snap Score |
|---|
Edit scoring weights in server/scoring.py → calculate_server_top_score()