01
Modeled eligibility
A workload fits the authored model conditions, requires a revision, fails a stated requirement, or cannot be assessed. A result applies only to its scope and window.
Proof before autonomy
Evidence informs a decision. Authority stays explicit.
A modeled result, a useful business outcome, acceptance of a report, and permission to operate are different things. A capacity assessment must keep those boundaries visible.
What each result means
Model screening uses allow / repair / reject / no-proof for a stated profile and evidence set. The labels do not authorize equipment changes or establish safe sellable MW.
02
Commercial usefulness
The decision owner determines whether a profile meets the service and business need. A smaller modeled increment is not automatically a useful offer.
03
Report acceptance
The customer reviews contracted deliverables against agreed criteria. Accepting a report is not acceptance of an operational action.
04
Operating permission
Facilities and operational authorities control site changes. Engineering, safety, contractual, and regulatory review are separate from this public demonstration.
A conditional path to more authority
Shadow Mode would observe without control. Advisory Mode would put proposals before operators. Bounded Autonomy would require a verified dispatch envelope, explicit permissions, rollback, and audit evidence. Expanded Autonomy would require further evidence. These are development stages, not services proven by the synthetic sample.
What evidence must travel with a number
A capacity report needs the requested and revised profiles, reference load, meter boundary, time interval, binding conditions, assumptions, unknowns, dataset/model versions, and review status. A safety report requires its own supported scope; a software check cannot stand in for it.
Related operator resources
Continue the proof path
- A concise brief, with a traceable supporting record.Review the contents of a capacity decision package: profiles, conditions, model records, unknowns, review responsibilities, and publication versions.
- See the decision, the conditions, and the unanswered question.Inspect a synthetic capacity decision brief. Compare requested and revised profiles, model screening, unknowns, and the unresolved business decision.
- Evidence for safe, usable, auditable capacityPublication-gated methods, synthetic traces, ledgers, specifications, and proof artifacts for evaluating virtual capacity claims.
- Connect a capacity question to its evidence.Understand the proposed GridNinja assessment role alongside existing monitoring, models, controls, and commercial planning, with explicit evidence limits.
Proof before autonomy
Start with one capacity question.
Agree the decision, evidence boundary, responsibilities, and paid scope before work begins.