ENERGY RESILIENCE / EXPLAINABLE OPTIMIZATION

POWER THAT
PLANS FOR PEOPLE.

GridKind is a constraint-based microgrid digital twin that lets neighborhoods rehearse an outage and prove who remains protected.

RUN THE TWIN ↓

REAL WEATHER. AUDITABLE DECISIONS.
Hardware-ready without pretending simulated sensors are physical.

WORKING OPTIMIZER / 01

NEIGHBORHOOD
RESILIENCE TWIN

Change the emergency, reserve policy, data source, and public-interest objective. GridKind searches feasible allocations and exposes its decision trace.

FALLBACK SNAPSHOTMELBOURNE 27°CCLOUD 24%3,135 FEASIBLE PLANS TESTED
FORECAST TELEMETRYSOLAR INPUT 66%
☀SOLAR CANOPY205 KWH AVAILABLE
▥COMMUNITY BATTERY128 KWH DISPATCHABLE
✚CARE CLINIC100% SERVED
⌂COOLING SHELTER100% SERVED
▦72 HOMES100% FLEXIBLE LOAD
NOW
+4H
+8H
+12H
+16H
+20H
+24H
VERIFIABLE METHOD / 02

NOT A BLACK
BOX.

GridKind enumerates feasible load allocations under a hard energy budget. It scores each plan using explicit community priorities, then publishes the winning constraints and trade-offs.

01

GROUND INPUTS

Live Open-Meteo conditions drive solar availability, with a transparent offline fallback.

02

SEARCH FEASIBLE PLANS

Clinic, shelter, and household allocations must fit the calculated energy budget.

03

EXPLAIN THE WINNER

Weights, coverage, curtailed demand, and reserve effects remain visible to the operator.

HARDWARE-READY PATH / 03
01OPEN DATAWeather · load · storage
→
02SENSOR BRIDGEESP32-shaped payload lab
→
03CONSTRAINT SOLVER3,135 feasible plans
→
04AUDITABLE ACTIONDecision + reason + impact

Prototype boundary: live weather is real; load and storage values are documented scenario assumptions. Sensor Lab emulates the JSON payload expected from an ESP32 and is clearly labeled as simulated.