DevelopersDevOpsFeaturedOpen SourceState of EnergyVideo

LF Energy’s Power Grid Model Project Tackles Complex Grid Calculations Efficiently


Guest: Tony Xiang (LinkedIn)
Organization/Company: LF Energy (Twitter) | Alliander (Twitter)
Show: State of Energy

The traditional energy sector is undergoing significant change as distribution system operators (DSOs) now have to factor in electrical vehicles and solar panels making it more challenging to predict peak and off-peak times. To tackle this problem, Alliander developed its own solution called Power Grid Model to help with grid calculations.

In this episode of State of Energy, Tony Xiang, Principal Scientist at Alliander while at the 2023 LF Energy Summit in Paris, talks about the Power Grid Model project, its purpose, and the community around the project.

  • Xiang explains that Alliander was looking for good power grid calculation software five years ago to do grid calculations based on future needs. However, they had struggled to find appropriate tools, either commercially or open source, which led to them building their own in-house solution.
  • The way we are consuming energy is changing, from the traditional distribution system, which used to be predominantly predictable to now having a more unpredictable pattern due to more electrical vehicles and solar panels. Xiang talks about why this is leading to an increased need for scenario analysis and what this involves.
  • Xiang discusses the reasons why Alliander decided to open source the Power Grid Model saying that it was initially used internally but they believed that the solution would be useful to colleagues at other DSOs, before deciding to open source it.
  • The project is now hosted by LF Energy and has two major DSOs who are trying to implement the library into their production environment to run several scenario calculations and long-term planning. Xiang talks about the others that make up the community.
  • Unlike other projects that are hosted by LF Energy, this project is already in production. Many people at Alliander trust and rely on the performance and validity of the library and it is used in many critical processes internally.
  • Xiang discusses how the project has a feature request where the community can suggest new features or functions. Although these requests do not benefit Alliander directly, they can benefit the community as a whole. He talks about how this helps them see what the community wants as a whole and how this is their guiding compass.
  • The Power Grid Model project has a community developer-driven development cycle which has a bi-annual meetup where they gather together all the internal and external stakeholders to look at the community requests and map out a roadmap for the next period from there.
  • The characteristics of distribution grids vary significantly between countries and regions, and as such, Xiang feels that although it would be theoretically possible that this project could also benefit Asia or American DSOs, that it would require major revisions and function additions in order to make this possible.
  • Since Alliander is the Dutch DSO, it makes sense that the parties most interested in this project would be other European DSOs, as they would share some of the common problems they are facing. Xiang discusses how interested parties can get started with the project.
  • Xiang talks about the two key ways AI is being used in the project: when you apply the AI machine learning algorithms the back-end calculation produces the test or training side for the AI to further proceed and also, in the algorithms themselves.

This summary was written by Emily Nicholls.