This project aims to lay out the design principles for a whole system balancing optimisation engine, that include the interaction with the underlying power system model, this security constrained economic dispatch (SCED) approach will assist NESO in modernising their process and systems. This project will include looking at architecture and road-mapping of development and integration with the wider systems, including the Open Balancing Platform.
The project will consider the strategic direction of control room processes, impact of approach on people and systems, by understanding balancing from a wider whole energy system perspective and how it can be included into a more comprehensive and systematic balancing process while creating a robust design that is future proof.
Benefits
- Strategic Clarity: Will provide a clear vision and strategic direction required for control room optimisation that will reduce uncertainty for following development phases
- Design Robustness: Will deliver a robust design that will be capable of dealing with the growing levels of uncertainty and need for whole system optimisation that will reduce the efforts required to deliver that phase of future projects
- Accelerate Benefit Realisation: The output of this phase will enable the NESO to start to put in place the building blocks to start the successfully delivery of this vision, this will reduce the time required to deliver benefits to the consumer
Learnings
Outcomes
The proposed Grand Optimiser design would transform NESO’s control room operations, moving from fragmented, deterministic tools to a unified, AI-powered, probabilistic optimisation approach that combines energy balance, reserves, and transmission netwrok model into a single security-constrained optimisation process.
This design is based on an AI-assisted, human in control philosophy that presents suggestions and information to operators, using enhanced visualisation and AI-powered insights, so that operators can make better informed and well-justified decisions. A key innovation in this approach is to unify and automate the security-constrained optimisation engines such that they produce consistent and comprehensive solutions over a range of timeframes and outcomes. These probabilistic trajectory results may then be processed by AI-agents and other tools to provide insights on the impact of uncertainty on dispatch outcomes to enhance operator decisions and equip operators with proactive tools to manage uncertainty and operational risk in a systematic manner.
The system enables operators to visualise probabilistic dispatch trajectories across various scenarios which will enhance their situational awareness of potential dispatch actions to manage uncertainty in a proactive, systematic process. These probabilistic scenarios will provide a dispatch trajectory of operation, at five-minute intervals, extending out through the current dispatch window (60 to 90 minutes) and a scheduling trajectory, at 30-minute intervals, extending out 36 hours. Trajectories are provided for multiple scenarios, so operators will be armed with significantly more information regarding the impact of various pending decisions.
Additionally, the trajectory information will be leveraged in agentic AI tooling in the operator interface to provide suggested strategies for operators to consider and to provide insights to operators on trends and relationships between the scenarios to aid in decision making. The AI-enhanced decision support tools are enabled by intelligent processing of potential scenarios that can suggest the most likely outcomes and provide justification for dispatch actions through automated logging mechanisms.
Some key potential benefits of the design include:
> Maximised Economic Efficiency. Co-optimised reserve, energy, and transmission analysis increases the efficiency of scheduling and dispatch.
> Enhanced Grid Reliability. The unified Forward Reliability Scheduling and Forward Advisory Dispatch optimiser engines integrate a representation of the full AC Transmission Model directly into the decision logic. This ensures that every recommendation is physically feasible and respects thermal, voltage, and stability constraints by design.
> Proactive Uncertainty Management. Instead of relying on static safety margins, the system leverages adaptive models to generate probabilistic risk envelopes for wind, solar, and demand. This allows operators to visualise specific risks (e.g., "20% chance of a wind drop-off") and take targeted, data-based actions.
> Operational Clarity & Trust. The system is built on a principle of "Explainability," providing clear, auditable justifications for every automated recommendation (e.g., specific constraint costs or price components). This reduces the cognitive load on operators and provides a robust digital audit trail for regulatory compliance.
> Future-Proof Scalability. Built on a modular, microservices-based architecture, the Grand Optimiser is designed to evolve alongside the energy transition. It allows for the seamless integration of new solvers, rapid adaptation to market rule changes , and the scaling of compute power.
The proposed Grand Optimiser is based on a modular design to provide flexibility in future integration of new tools and to scale the functionality over time.
Lessons Learnt
The design documents in scope for this project outlined the proposed architecture of the Forward Reliability Scheduling (FRS) and Forward Advisory Dispatch (FAD) optimisers along with the transmission model, adaptive model forecasts, a data ingestion layer and a “what if” capability to test data combinations.The report documents helped to assess the plan for the implementation of the solution and provided visibility on its potential strengths and on other aspects that will need to be validated during the implementation.
To take the design into production and implementation, it is necessary to conduct a validation and verification stage, where the design is tested in a modular approach before end-to-end build. This is part of the roadmap delivered in WP4.
