Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Pioneering Roadmap for ML-Driven Algorithmic Advancements in Electrical Networks

Jochen L. Cremer, Adrian Kelly, Ricardo J. Bessa, Milos Subasic, Panagiotis N. Papadopoulos, Samuel Young, Amar Sagar, Antoine Marot

TL;DR

The paper addresses the fragmentation and safety concerns hindering ML adoption in electrical grids and proposes a goal-oriented, TRL-based innovation roadmap. By analyzing a survey of about 110 experts, it maps the current ecosystem, highlights high-priority use cases like forecasting and situational awareness, and identifies governance and data challenges. It then details a structured journey from proofs of principals to testing and production, including switchback strategies to adapt TRLs in response to data, performance, and end-user feedback, all within a framework that leverages TEF facilities and open-source collaborations. The resulting roadmap aims to align operators, academics, and labs to accelerate safer, more scalable AI-enabled grid tools and decision-support systems, with practical guidance on governance, testing, and interdisciplinary collaboration.

Abstract

Advanced control, operation, and planning tools of electrical networks with ML are not straightforward. 110 experts were surveyed to show where and how ML algorithms could advance. This paper assesses this survey and research environment. Then, it develops an innovation roadmap that helps align our research community with a goal-oriented realisation of the opportunities that AI upholds. This paper finds that the R&D environment of system operators (and the surrounding research ecosystem) needs adaptation to enable faster developments with AI while maintaining high testing quality and safety. This roadmap serves system operators, academics, and labs advancing next-generation electrical network tools.

A Pioneering Roadmap for ML-Driven Algorithmic Advancements in Electrical Networks

TL;DR

The paper addresses the fragmentation and safety concerns hindering ML adoption in electrical grids and proposes a goal-oriented, TRL-based innovation roadmap. By analyzing a survey of about 110 experts, it maps the current ecosystem, highlights high-priority use cases like forecasting and situational awareness, and identifies governance and data challenges. It then details a structured journey from proofs of principals to testing and production, including switchback strategies to adapt TRLs in response to data, performance, and end-user feedback, all within a framework that leverages TEF facilities and open-source collaborations. The resulting roadmap aims to align operators, academics, and labs to accelerate safer, more scalable AI-enabled grid tools and decision-support systems, with practical guidance on governance, testing, and interdisciplinary collaboration.

Abstract

Advanced control, operation, and planning tools of electrical networks with ML are not straightforward. 110 experts were surveyed to show where and how ML algorithms could advance. This paper assesses this survey and research environment. Then, it develops an innovation roadmap that helps align our research community with a goal-oriented realisation of the opportunities that AI upholds. This paper finds that the R&D environment of system operators (and the surrounding research ecosystem) needs adaptation to enable faster developments with AI while maintaining high testing quality and safety. This roadmap serves system operators, academics, and labs advancing next-generation electrical network tools.
Paper Structure (16 sections, 2 figures, 1 table)

This paper contains 16 sections, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Innovating with ML
  3. Survey findings
  4. Innovation environment
  5. Interdisciplinary approach
  6. A Use Cases Implementation Roadmap
  7. Developing Proof of Principals (TRL 1-2)
  8. Developing Proof of Concepts (TRL 3-5)
  9. Testing and Experimentation (TRL 6-9)
  10. Project Management
  11. Switchbacks
  12. Circumstantial Discovery Switchback
  13. Predefined, Embedded Switchback
  14. Switchback When Requirements Not Met
  15. Direct, Planned Switchbacks
  16. ...and 1 more sections

Figures (2)

  • Figure 1: Innovation with ML for the electrical network (modified from MTRL23).
  • Figure 2: Circumstantial discovery switchback (a), predefined, embedded switchback (b). AI starts often at TRL $4$ making it challenging to switchback to lower TRLs. (c) Switchbacks to ML toolboxes if use-case-specific tests and requirements were not met, (d) direct switchback and rapid prototyping between modifying ML toolboxes and final products. Figure from MTRL23.