Great Britain’s energy landscape is changing at an increasing rate, not only in terms of how energy is generated and distributed but also in how it is used by consumers. This increases the complexity for grid operators in forecasting power system conditions and in developing dispatch control strategies that can utilise more distributed intermittent renewable resources whilst retaining the reliability of supply required. Traditional dispatch tools and processes are becoming increasingly insufficient and cannot take advantage of the advancement of digital technologies such as Artificial Intelligence and Machine Learning. This project builds upon work undertaken in NIA2_NGESO0013 Advance Dispatch Optimisation and will design a blueprint for transformation of control room tools and processes to meet the System Operator needs of the future.
Benefits
Multiple benefits can be realised from the delivery of this project:
Short Term Benefits
- Identification of the detailed requirements for the system components and adaptive input models
- Development of the sequencing required to deliver a control room dispatch optimiser tool capable of assessing the multiple operational scenarios inclusion of distributed intermittent renewable resources present.
Longer Term Benefits
- Optimised forecasting and dispatch by leveraging flexible demand and storage technologies reducing Balancing Mechanism (BM) costs
- Management of increasingly complex grid operations through utilising adaptive input models and machine learning.
- Improved performance monitoring, evaluation and feedback of the system through machine learning and operator education.
- Improved operator situational awareness through the use of multi-dimensional visualisation tools utilising highly accurate input data, probabilistic trajectories, input scenario information, optimisation results and scenarios.
Learnings
Outcomes
The project has delivered all its original objectives as well as evolving the solution End Vision.
Capability Framework – The capability framework that was developed as part of the DOT project aimed to break down the vision defined by Google Tapestry into distinct (to-be) capabilities. These direct and value-adding capabilities (L0) were then broken down into enabling capabilities (L1), which in sum provide the means to the L0 capability.
Gap Analysis Report - The comprehensive gap analysis report provides an in-depth explanation of the capabilities and the associated value that each of these capabilities may offer.
Google Tapestry created a strategic vision to design an efficient dispatch process that is fit for purpose for the energy system of the future. Tapestry introduced several concepts to achieve the desired outcome, including:
- Automated insights through adaptive machine-learning input data models.
- Probabilistic trajectories of various system states.
- A series of look-ahead time-coupled security-constrained economic dispatch optimisation engines create a system operating plan (SOP), instructions and reserves.
- Enhanced or automated operator decision support.
- Automated performance monitoring.
The DOT Project, following the Google Tapestry study, aimed to meticulously explore the effective and efficient implementation of the recommendations outlined in the Tapestry report. The project involved a comprehensive analysis of the current dispatch process and the identification of future capabilities necessary for realising the envisioned objectives. After the identification of prospective or future capabilities, a thorough gap analysis was undertaken to assess the disparity between the current "as-is" state and the envisioned "to-be" state to devise a path for their realization.
Work Package Report – Closing the gaps identified in the gap analysis section will necessitate the initiation of a series of strategic initiatives. The total scope of work has been systematically broken down into a series of workstreams and modular work packages.
The work package report provides a comprehensive overview of the work packages, including their detailed descriptions, the duration of each initiative, the dependencies involved, and the corresponding effort needed for successful completion.
The work packages have been categorized as follows:
- Regulatory Framework Agreement and/or Stakeholder engagement
- Value and/or feasibility analysis
- Design
- Agile Development
- Research
Roadmap Report – The Roadmap has been designed to orchestrate the essential initiatives (work packages) that are required to be delivered as part of a comprehensive strategy to address the gaps identified in the gap analysis. The roadmap presents a visual representation of the proposed timelines, detailing the sequence of work packages, their associated dependencies, and scheduled milestones.
Input Data Model Report – Within the Data model report deliverable, the focus is on the Adaptive Input Data Models – exploring current and planned capabilities, the required final capability, as well as the associated gaps and next steps. The model groups discussed are:
- Adaptive Generation Models2
-Thermal
-Renewable
-Grid scale duration limited assets, such as batteries and pumped storage
[added following Google X clarification and not explicitly referenced within the Tapestry report]
- Adaptive Transmission Model
- Adaptive Interconnector Models
- Adaptive Distributed Energy Resources (DER) Models
- Adaptive Net Demand Models
-Demand Forecast and Consumer Behaviour
-Embedded DER
Architectural View – To ensure a technically sound roadmap the project created a high-level view of a potential logical architecture for the Advanced Dispatch Optimization (ADO) system. The project created four key architectural artefacts:
A System Context Diagram: This provides an overview of the ADO system and its relationship to other systems and components.
Architectural Principles: They outline the guiding principles and constraints that have shaped the architecture of the ADO system.
Architectural Decisions: To capture the key decisions made during the design process, including the rationale behind each decision.
A Component Model: This diagram provides a detailed view of the components that make up the ADO system, including their relationships and interactions.
Lessons Learnt
This was an ambitious project that sought to understand the multitude of Control Room systems and the data inputs and outputs involved. Utilising an agile approach and understanding that not necessarily all the data that was requested would be available, or available in a format that was easily ingestible within the project’s timescales, was beneficial in ensuring the project were pragmatic in what level of granularity could be achieved when producing the Data Model report especially with regard to legacy system data.
Having a technical delivery partner who is experienced in Control Room operations, systems and processes meant the project could call on their knowledge and experience as well as the ESO SMEs and deliver more efficiently than a partner with little or no Control Room knowledge.
The project team gained valuable insight by spending time in the Control Room as well as holding in person meetings with SMEs at key points in the project to elicit requirements or review outputs.
The process for ensuring the IBM team were set up with all the relevant access to ESO systems and software took far longer than anyone anticipated.
Dissemination
The project featured in a presentation on the Virtual Energy System at the Energy Innovation Summit in November 2023 and can be found via the following link:
EIS presentation
ADO was also mentioned in an article in Utility Week ‘We’re only just getting started with AI’ by Dan Clarke, Head of Innovation, Energy Networks Association – Published 20th May 2024
