Verslag: Workshop ‘Data-driven Innovations for the Energy Transitions’

TU Delft, Mondai | House of AI, September 21 2023, 14:30 - 18:00

On Thursday, September 21st, the Commit2Data community focused on energy came together with the NL AIC working group on Energy & Sustainability at Mondai House of AI. With the Commit2Data program approaching its end, goal of this workshop was to connect to the NLAIC working group Energy & Sustainability in order to integrate its insights and foster new collaborations, thereby extending the program's research efforts.

The Data-driven Innovations for the Energy Transitions workshop provided valuable insights into the application of AI and data-driven approaches in addressing energy challenges. It highlighted the need for continued research, collaboration, and the training of both current and future workforces to harness the potential of AI in the energy sector.

Mathijs de Weerdt, chair of the NL AIC's working group on Energy & Sustainability, opened the workshop with a warm welcome to the participants, setting the stage for a day of insightful discussions.

Marjan Popov talked about the ReSident project; which main focus is using big data to safeguard future electricity grids by employing AI for real-time simulation, detection, and disturbance modelling, particularly for event classification tasks in energy grids.

Simon van Mourik talked about how technology can be used in combination with biology for energy management in greenhouses, where growers need to provide settings and setpoints to the controller. One of the key points of the discussion was whether growers can utilize the internal energy balance of crops to store energy in starch and sugar.

Eric Pauwels introduced the Energy Intranets (NEAT) project, which aims to better predict and utilize flexibility of supply and demand in energy networks. Key points of the presentation included collaboration with Sympower for optimizing electric vehicle charging, modelling network flows considering weather and demand, and the significance of precise price predictions and one-day-ahead simulations.

Machteld van den Broek discussed climate and power system simulations, focusing on ensuring reliability during adverse conditions, using climate simulations and machine learning to predict renewable resource fluctuations, and developing tools for complex energy problems.

George Rouwhorst discussed the challenges of growing energy demand and congestion in distribution networks, emphasizing the importance of data analysis for network visibility, optimizing load management, and addressing privacy issues.

This led to Vincent Campfens' presentation on synthetic data developed by BlueGen.ai to tackle privacy concerns and advance research, with a focus on application in the electricity sector.

Then, a panel discussion moderated by Neil Yorke-Smith explored various aspects of AI adoption in the energy sector. Key themes included the need for more AI training programs; the challenges of implementing AI in traditional industries; the difficulty of explaining complex AI models to stakeholders; ethical, societal, and legal considerations in AI adoption.

The participants concluded the workshop discussions over drinks, fostering further collaboration and knowledge exchange.