Skip to main content

Demonstration & Validation

Four demonstration nodes will address key energy challenges,
with datasets and synthetic data supporting the testing of use cases.

Multi energy and sector coupling [Node: DTU, Denmark]

This use case will demonstrate how AI technology can enable, improve, and optimize the synergies between electric power grid operation and district heating systems in the triangle area in Jutland, Denmark (CDK) and on the Danish Island Bornholm in the Baltic Sea (BHV).

Objectives

  1. Test AI Technology on Virtual Power Systems: Demonstrate AI-EFFECT functionality for testing AI technology on virtual power systems.
  2. Optimize Real Single- & Multi-Energy Systems: Demonstrate AI-EFFECT functionality for testing AI optimization of real single- and multi-energy systems.
  3. Ensure AI Interpretability & Certification: Demonstrate AI-EFFECT functionality for AI interpretability, validation, verification, and end-to-end certification from lab-testing to real-world operation.

Expected Impact

This initiative aims to show how to test, validate, and deploy AI tools to improve the operation and efficiency of single- and multi-energy systems. The node specifically focuses on demonstrating the validation of various AI technologies for efficient district heating network operations. Through targeted use cases, it will validate AI tools for forecasting, plant operations, and sector coupling. The node will showcase how AI tools can achieve near-zero-emission district heating by coordinating multiple energy vectors to lower overall system costs and minimize CO2 emissions.

Photo Credits: Willem de Kam & TU Delft.

Transmission Congestion Management Laboratory [Node: TU Delft Netherlands]

This AI-EFFECT node facility extends the Control Room of the Future (CRoF) at TU Delft with AI-testing abilities focusing on grid data synthetising and AI-algorithm verifications. This node would serve as a bedrock for safe experimentation and validation of AI algorithms.

The central ambition of this facility is to hone AI-based algorithms for congestion management, a critical function of control rooms. By providing a digital-physical and controlled environment replete with processes for synthetic grid data generation, the facility would allow for rigorous testing and refinement of these algorithms, ensuring they are robust and effective in managing the grid’s congestion, e.g., through sequential topological reconfigurations.

Objectives

  1. Controlled Environment for AI Validation: Develop an operational controlled environment to validate AI solutions, allowing technology developers to rigorously test and enhance AI algorithms for congestion management, ensuring interoperability and reliability.
  2. Hyper-Realistic Testing with Digital Tools: Define a hyper-realistic testing environment using synthesized grid data, Digital Twins, and Real-Time Digital Simulators (RTDS) to safely explore AI solutions across diverse operational scenarios.
  3. AI System Evaluation Processes: Develop comprehensive evaluation processes to ensure AI-driven grid security, compliance, and readiness for future challenges.

Expected Impact

This node will feature (1) a new data synthesizer for grid data tailored to testing objectives for AI algorithms, and (2) a verification package with guidelines for synthesized grids. These advancements aim to support the development of innovative AI-based congestion management tools and facilitate testing human-AI interactions within power grids. By combining advanced AI algorithms with human oversight, the CRoF seeks to synthesize grid data and verify AI solutions for effective congestion management on synthesized networks.

Energy efficiency, management and sharing for local energy communities [Node: INESC TEC)

Effective energy data management and sharing is crucial for advancing energy efficiency and fostering sustainable practices. Traditionally, data management platforms, notably for smart meters, have been advocated by DSOs and/or TSOs. DSOs are progressively incorporating data on innovative services, extending beyond smart meter data to include customer data management and collaborative data sharing with TSOs. However, the prerequisite for explicit consent from consumers (data owners) to share their data with third-party service providers presents challenges, potentially disrupting various use cases and business models.

This issue becomes particularly pronounced when dealing with behind-the-meter data, as barriers related to connectivity, data privacy, and security hinder seamless access. The challenges persist even when access to data is granted, engaging directly with end-users—consumers— is an ongoing challenge. This lack of direct engagement hampers essential activities like service design, co-creation, and the testing of AI-based solutions tailored to the specific needs and preferences of consumers. By addressing these limitations, the energy sector can unlock the full potential of AI for the benefit of both consumers and the broader energy community.

Objectives

  1. Enable a Local Energy Data Space for Collaboration: Facilitate access to a local Energy Data Space where AI-based energy service providers (e.g., Watt-IS) can co-create with consumers and prosumers. This environment supports users in managing, sharing, and protecting their personal data rights, adhering to the EU Data Governance Act principles.
  2. Develop an Integrated Data Ecosystem: Establish a comprehensive data ecosystem that includes both real-time and historical data while actively engaging citizens. This ecosystem aims to encourage the development of data-centric services that inspire consumer behavior change, promote DER adoption, and support net-zero goals for the local energy system.

Expected Impact

The initiative will create value for AI-based energy service providers and other stakeholders, including DSOs, by enhancing flexibility in grid constraint management. It will foster collaborations with service providers across sectors like electric mobility and smart appliances, promoting cross-sector innovation. Additionally, the Energy Data Space will set the standard for implementing strong security and privacy measures to protect personal energy data, ensuring responsible and trustworthy AI testing and certification.

© Photo: Fraunhofer FIT

Distribution Network Congestion Management for Renewable Integration [Node: Fraunhofer]

The business motivation for using AI in power distribution systems is to enhance operational efficiency, facilitate the connection of renewable resources, optimize resource allocation, improve reliability, and enable more effective decision-making. AI algorithms can potentially enhance the accuracy, robustness, scalability, real-time performance, security of distribution networks, but there is a need for seamless integration of AI algorithms and models within the distribution system IT and OT infrastructure as part of business as usual and to augment classical engineering methods.

Given the safety and security requirements, it is crucial to validate the performance and functionality of AI systems to ensure they meet the specific requirements of power distribution applications and deliver reliable and efficient results.

Objectives

  1. Create a Realistic Distribution System Model: Develop a close-to-reality model of a distribution system environment that bypasses the reliability limitations of real systems, enabling robust testing of AI algorithms.
  2. Define AI Testing Requirements for Congestion Management: Focus on specific network segments to outline the testing needs of AI systems for managing congestion in distribution networks and facilitating distributed energy resources.
  3. Evaluate AI Performance: Test AI algorithms and models within the modeled system, assessing their performance in comparison to classical methods, such as optimization techniques.

Expected Impact

The flexible, reconfigurable laboratory setup offers emulation of real cyber-physical distribution systems tailored to the needs of various AI tools and services. The environment can be scaled by integrating simulators for customized scenario configurations and data generation, serving as an ideal testbed for refining AI solutions before their deployment in real-world operations. This use case aims to develop a comprehensive, testable framework for AI in distribution network applications, with a particular emphasis on AI for congestion management.