VALUE DECENTRALIZATIONPowered by Frontier ProtocolRestoring session…VALUE DECENTRALIZATIONPowered by Frontier ProtocolRestoring session…Set a battery policy, follow changing demand and a grid outage, and compare modeled cost, unmet demand, and carbon. Explore the separate classic 100 MWh mix too.
THE IDEA IN ONE PICTURE
Save battery power for an outage, use it now, or buy electricity from the grid. Every choice changes the day ahead.
Choose when to charge, discharge, and buy power.
Follow six turns of weather, demand, and grid outages.
Cost, unmet demand, and carbon stay separate.
Concept illustration. The day is a deterministic simulation, not a live electricity network.
START HERE · A SHORT GUIDED EXPERIMENT
Run the current day, tighten the grid import cap, and see the consequences. Change only one setting so you can explain what happened.
The starter holds a battery reserve. Record its cost, unmet demand, and carbon for this day.
Try changing Grid import cap from 12 to 2 MWh per period. It limits all grid purchases, including battery charging. Leave the other rules unchanged.
Change grid import cap ↓Compare all three outcomes with your previous run. Then inspect the 16:00 outage on Storm day to see where energy was missing.
Learning progress only. No score, reward, AI fee, or payment. A tie is a valid result.
A day in the community · Simulated Practice
Write a dispatch policy, then follow six turns as sun, demand and grid prices change. Your battery carries each decision into the next turn.
Day simulator v1 · Separate from Classic mix. Compare modeled cost, unserved energy and operational carbon only within the same day context.
A four-hour grid outage arrives as solar falls and demand peaks. Changing the day starts a separate comparison.
01 · Set the policy
Renewables serve demand first. Surplus charges storage. Battery supplies a shortage when the grid fails, the price threshold is reached, or the import cap is insufficient. The online reserve still applies until an outage.
The grid then serves remaining demand and may charge storage below your price threshold. Battery starts at 6 MWh; capacity is 20 MWh. Every four-hour turn allows up to 8 MWh charging input and 8 MWh discharge output (2 MW average). Each conversion is 90% efficient. No simultaneous charging and discharging, and no exports.
These published numbers are an educational aggregate model. They do not represent voltage, power flow, battery aging or real grid safety.
Start from a preset or change the four rules.
02 · Follow the energy
Started at 6 MWh
Conversion loss 0 MWh · Curtailed renewables 0 MWh. Arrows show energy totals for this four-hour period.
03 · Compare the full day
End-of-day storage: 8.811 MWh; no resale or terminal credit. Solar and wind generation costs include curtailed energy. Initial 6 MWh carries $300 and 2,100 kgCO₂ for every policy. Construction emissions and battery degradation are outside this model.
Run a policy to compare with all three public baselines. A second run also compares your previous revision.
Comparisons use only the published references and your previous run. No global winner or frontier expansion is claimed. Revisions remain in memory until you leave this mode or reload.
04 · Reproduce the decision
The JSON records the version, exact policy, public day, every energy flow and result hashes. Re-running it checks the same deterministic model.
0xd893c2d83f5ec6194a3d6fce0ef49572d96cc1fc27c5047d0e15b0a557a4cdcb0x65721103ba7f54ce2bca7aecd235aade3d17311d5ea8f3ec7be2d165c37a0cbe0xe26063e04bd61e130fcbcc368cbc5dd769af8c624da3d7da144d921083b77dbf{
"version": "microgrid-policy-v1",
"reservePercent": 50,
"dischargePrice": 100,
"chargeBelowPrice": 0,
"maxGridMwh": 12
}Offline entry: @frontier/microgrid-dispatch/practice → verifyMicrogridReplay(json). Context binds the complete public day, constraints, battery rules and independent metrics.
Local Practice only. No account submission, paid pool, live grid operation, AI execution or Hidden Final is connected to this simulator.
These terms describe the original reference challenge. The editable Practice lab above uses its own versioned context and evidence; its results are not entries in this challenge and do not earn rewards.