Context
Transmission system operators in high-voltage and extra-high-voltage networks must integrate renewable energy sources efficiently. The operational environment is defined by static thermal constraints, manual topology decisions, and power flow optimisation within fixed parameters.
Challenge
Operators face a three-way optimisation problem: static line ratings underestimate real capacity, topology decisions are made in isolation due to the large combinatorial space, and power flow optimisation is performed within suboptimal constraints. This leads to congestion, renewable curtailment, elevated balancing costs, and a gap between theoretical and actual renewable integration capacity.
AI Solution
The solution is an AI orchestration layer that coordinates Dynamic Line Rating (DLR), Topology Control, and Optimal Power Flow (OPF) tools using a live, physics-calibrated Digital Twin. DLR provides real-time thermal capacity per line segment without hardware sensors. Topology Control uses a deep reinforcement learning agent to recommend optimal switching configurations. OPF optimises generation dispatch using dynamic constraints and confirmed topology. Outputs are injected into the Energy Management System as dynamic updates and recommendations.
Impact & evidence
The solution has delivered increased grid stability, asset reliability improvements, reduced outages, faster operations, better forecasting, and enhanced integration of distributed energy resources (DER). Quantitative evidence includes a 20–38% increase in transmission capacity, 18–28% reduction in renewable curtailment, and 22–35% reduction in congestion costs. Impact is further supported by a 71% operator acceptance rate for AI recommendations and validation against historical SCADA events.
Implementation / Adoption
The solution is commercially deployed across multiple organisations. Implementation involves EMS integration via open APIs, validation of asset registers, SCADA historian access, Digital Twin build, DLR engine deployment, EMS integration and testing, topology control training, and OPF configuration. The process spans approximately 11 months, with defined phases from contract alignment to full deployment.
Data
Required data includes SCADA/EMS telemetry (branch flows, bus voltages, generator output, breaker states), a transmission asset register (topology, conductor parameters), numerical weather prediction (NWP) service data, and real-time/historical renewable generation data. Data cleansing and augmentation were necessary.