Building a city-scale water digital twin

Across the Sultanate's interconnected grid, operators face an emerging challenge: the pace of demand variability now exceeds what classical SCADA loops were designed to absorb. This piece distills what our field teams learned from working alongside distribution control rooms in Muscat, Sohar and the Dhofar interconnect.

Why it matters now

Rooftop solar, EV depots and seasonal cooling loads have collapsed the gap between forecast horizons. The control room needs decisions that are both faster and more explainable, and that's where the architecture below comes in.

The reference architecture

We layer telemetry, a domain feature store, and a constrained policy engine. Each agent ships with guardrails reviewed by the engineering desk, and every action is logged into the audit ledger so the regulator can replay decisions on demand.

What teams measured

• 18% reduction in SAIDI on the pilot feeders within two quarters.
• 27% faster fault isolation through agentic switching recommendations.
• Zero RLS / data-governance incidents thanks to scoped service accounts.

Where to start

Pick one feeder, one substation and one outcome. Wire telemetry into a sandbox, shadow the operators for four weeks, and only then promote the copilot into the live console. The discipline pays back in trust.