US Power-Grid Outage Stress →

A critical-infrastructure monitoring agent polls ADW-251 hourly and pages the operations center whenever outage_stress_score spikes above 50 (representing roughly 100,000+ customers out, given the log10 scaling) with a rising trend, appending counties_affected from the IOM payload to pre-position field crews. The backtest shows the score swung from 60.2 to 5.0 across its two observations — a 91.7% drop — illustrating how the log-scaled score captures the difference between a major storm event and a normal grid day; source_lineage (DOE EAGLE-I/ORNL ODIN) confirms the data is the authoritative federal outage feed rather than a commercial aggregator. This eliminates the 1–2 hour lag that operations teams currently experience when piecing together state utility reports.

A catastrophe claims manager at a property insurer uses the total_customers_out and counties_affected fields from ADW-251 to trigger surge-staffing of claims adjusters in the affected regions before the phone queues spike, rather than waiting for policyholders to call in. Deploying adjusters 4–8 hours earlier in a multi-county outage event measurably compresses claim-cycle time and reduces the cost of temporary housing authorizations — value that the free PowerOutage.us alternative does not deliver in a structured, API-ready format.