01Scoped source
Network model
Topology & geography
Lines, substations, transformers, switches, feeder extent, and the identities that hold the network together.
ArcGIS · model exports · asset records
For utilities & electric co-ops
Your grid, ready to answer back.
Pilot a scoped distribution model on one operating surface: geographic map, topology-derived one-line, Radar, replay, and the grid agent beside the network. Keep the model, timestamp, and evidence visible at every step.
Your modelTopology + assets
Your stateOperations + work
Outside contextWeather + market
One investigationAgent + evidence
One network · four ways to read it
The agent belongs beside the grid.
Map, Network, and Radar are representations of the same modeled feeder. The active surface, selection, and time context should move into the investigation with you.
Illustrative utility workspace
NREL SMART-DS · synthetic model · private
Geographic networkSynthetic demonstration
Illustrative answers over a public synthetic distribution model. Private data access, connectors, and workloads are scoped and governed per deployment.
Start with the systems you have.
A Private Grid engagement scopes the connection boundary with your team. These are not presented as one-click integrations: identity, freshness, permissions, and ownership are validated source by source.
02Scoped source
Operating state
Telemetry & control
Measurements, device state, availability, and the timestamped operating frame your team is reviewing.
SCADA · EMS/ADMS · historian
03Scoped source
Customer & outage
Service context
Outage and customer context can join the network when those sources are included in the deployment boundary.
OMS · AMI · service territory
04Scoped source
Work state
Planned change
The planned work, model case, and assumptions that explain why the observed network is not always the base model.
Switching plans · work orders · studies
05Scoped source
External context
Outside the fence
Wholesale and public conditions remain separately sourced and timestamped beside private operating state.
ISO · weather · forecasts · public record
06Scoped source
Agent context
The active investigation
The selected network object and modeled time cursor travel with the question instead of being rebuilt in another tool.
Selection · topology · cursor · evidence
Shared time
One clock—with the seams still visible.
Replay the private model without losing the selected feeder or network state. When public context is live or uses a different timestamp, the product labels that boundary instead of pretending every feed is aligned.
LayerBeforeCursorNext
Topology state
Before: Base modelCursor: Switch stateNext: Selected branch
Operating frame
Before: Modeled replayCursor: 3:00 PMNext: Current frame
Public context
Before: Own timestampCursor: ERCOT · liveNext: Clearly labeled
Agent context
Before: SelectionCursor: Cursor carriedNext: Evidence attached
Illustrative modeled replay. Public and private timestamps stay independently inspectable.
Questions to scope
Start where the decision gets hard.
“What changed before this feeder's modeled demand increased?”
Align the operating frame, selected feeder, and preceding model state.
“Which branches sit downstream of this open point?”
Trace the topology from source to switch without redrawing the one-line.
“Where do the model and connected operating state disagree?”
Keep the case, timestamp, source, and discrepancy attached to the review.
“What should the morning operating review inspect first?”
Read private network context beside weather and the public bulk-system picture.
Engineering models
Bring the model you trust. Keep its assumptions visible.
The public map is observational, not a power-flow solver. Network studies and private workloads are scoped around your existing engineering process and deployment requirements.
Bring the case
Start from the network model and engineering outputs your team already trusts. Ask the Grid does not quietly substitute a new system of record.
Scope the workload
Model runs, contingency checks, or power-flow outputs are defined per deployment—inside a private boundary or connected from an existing tool.
Keep assumptions visible
Case name, topology state, inputs, timestamp, and result stay attached so an agent answer never becomes an unexplained number.
Deployment pattern · defined with the utility
- 01boundary
Scope
- one feeder
- one model
- one decision
- 02sources
Connect
- model export
- operating state
- public context
- 03review
Validate
- identity
- timestamp
- assumptions
- 04agent
Investigate
- map
- network
- Radar
- evidence
A bounded first deployment
Bring us one feeder and one hard question.
Start with a synthetic or exported model, a narrow operating-data boundary, and a decision your team can validate. Prove the shape before expanding the network or source footprint.