W WattsMIND Grid Agent

Theme 03

Transmission intelligence across screening, dispatch, markets, and verification.

For transmission operators and ISOs/RTOs, WattsMIND Grid Agent turns grid cases, operating state, topology changes, objectives, and reliability constraints into fast voltage prediction, contingency screening, dispatch feasibility, market sensitivity, and verification targets.

High-level framework

One transmission engine, multiple operating questions.

Operating coverage
  1. Voltage prediction
  2. Contingency screening
  3. OPF dispatch
  4. Market sensitivity
  5. Verification targets

WattsMIND Grid Agent is short for Manifold-Informed Neural Dual-encoder. For transmission operations, it maps grid cases, injections, topology changes, reliability constraints, dispatch context, and market conditions into decision products: voltage prediction, contingency screening, constraint-risk ranking, dispatch feasibility, market sensitivity, and verification targets.

WattsMIND Grid Agent high-level framework for transmission prediction, contingency screening, OPF dispatch, and verification
High-level product flow for transmission: operating context enters WattsMIND Grid Agent; prediction, screening, OPF dispatch, feasibility, sensitivity, and verification products go back to grid workflows.

Application 01

Voltage prediction under contingencies.

The core transmission use case is fast contingency screening. WattsMIND Grid Agent predicts both voltage magnitude and phase angle while the network topology changes under single- and double-contingency regimes, giving operators an early view of where the grid is becoming stressed.

Experiment: ACTIVSg500 transmission benchmark under Base-N, N-1, and N-2 operating regimes. The evidence includes benchmark-level error, inference speed, and same-bus true/predicted voltage traces over sampled operating points.

ACTIVSg500 transmission benchmarkBase-N, N-1, and N-2 operating regimesFast screening followed by simulator-backed verification
0.002854average complex voltage RMSE
0.000537average voltage magnitude MAE, p.u.
1.65 msaverage forward-pass time per sample
N-1 / N-2contingency regimes included
ACTIVSg500 contingency benchmark summary for voltage prediction error and forward-pass time
Quantitative benchmark summary: complex-voltage RMSE, voltage-magnitude MAE, and forward-pass time across Base-N, N-1, and N-2 transmission regimes.
True and predicted voltage magnitude and phase angle under N-1 and N-2 contingencies
Voltage prediction examples under N-1 and N-2 contingency regimes.

Application 02

OPF and uncertain OPF for fast feasible dispatch.

Beyond screening contingencies, WattsMIND Grid Agent can support operating decisions that must balance economics, voltage recovery, thermal security, generator limits, and ramping behavior. The product view is not a replacement for the EMS solver; it is a fast feasibility-aware decision layer that proposes high-quality dispatch targets before slower optimization becomes the bottleneck.

Experiment: IEEE 30-bus AC OPF over a 300-step dispatch rollout, evaluated in both nominal conditions and branch-admittance uncertainty. The uncertain OPF setting perturbs the network model while preserving the same AC-flow, thermal, voltage, generator, and ramp-limit requirements.

IEEE 30-bus AC OPFNominal and branch-admittance-uncertain regimes300-step dispatch rolloutThermal, voltage, generator, and ramp limits
100.0%feasible dispatch under uncertain OPF
0.74%mean gap to nonlinear oracle under uncertainty
12.6 msuncertain OPF decision latency per step
97.5xlower latency than MPC in uncertain OPF
WattsMIND Grid Agent OPF and uncertain OPF summary with feasibility, objective gap, and decision latency
Application 02 result summary: nominal OPF and branch-admittance-uncertain OPF on the IEEE 30-bus benchmark, including feasibility, objective quality, and decision latency.
OPF dispatch behavior showing objective tracking, priority shift, feasibility, objective value, and latency
OPF dispatch behavior over a rollout: objective tracking, priority reweighting across voltage, thermal, and economic terms, and benchmark feasibility with millisecond-scale decisions.

Workflow

From transmission context to screening, dispatch, and verification.

01

Read the transmission context

Start from the grid model, injections, topology, limits, operating objectives, uncertainty, and market context.

02

Predict and screen

Estimate voltage magnitude, phase angle, stressed constraints, and contingency response across operating regimes.

03

Optimize dispatch signals

Support OPF-style setpoints, congestion analytics, and market-sensitive feasibility checks under nominal or uncertain models.

04

Verify operating actions

Send selected cases, dispatch candidates, and risk flags to simulator-backed verification and operator procedures.

Applications

  • N-k contingency and voltage prediction
  • Fast screening for stressed buses and operating constraints
  • Dispatch-feasibility and congestion analytics
  • Network-aware market sensitivity
  • Operator-facing verification targets

Why it matters

  • Separates high-risk operating cases from routine cases earlier in the study loop.
  • Tracks voltage magnitude, phase angle, constraints, and feasible response across changing grid regimes.
  • Keeps fast prediction tied to simulator-backed verification, not blind automation.

All research themes

Theme 01 Utilities & Co-ops Theme 02 Embedded Engines Theme 03 Transmission & ISOs/RTOs Theme 04 Energy Markets Theme 05 Flexible Loads