Alliander, the Netherlands’ electricity and gas distribution system operator (DSO), initiated a Short-Term-Forecasting (STEF) project to anticipate the load on the grid within the next few hours to days. Forecasting future load in the short term is essential to avoid local congestion, analyze grid safety, and optimize the use of existing assets. The stack has since been open sourced as LF Energy’s OpenSTEF Project.

In this episode of State of Energy recorded at the 2023 LF Energy Summit in Paris, Frederik Stoel, OpenSTEF Technical Steering Committee Member as well as Data Science Engineer at Alliander, talks about the status of the project and what lies ahead.

A lot of renewable production and changes in the consumption of generation are causing the loads on the grid to fluctuate more, making forecasts more difficult to accurately predict. OpenSTEF, one of LF Energy’s latest open source projects, helps deliver self-correcting and accurate forecasts of the load on the grid and either energy consumption or generation from renewable sources.

In this episode of TFiR State of Energy, Swapnil Bhartiya sits down with Jonas van den Bogaard, Solution Architect at Alliander, and Frank Kreuwel, Product Owner and Data Scientist at Alliander, to discuss their OpenSTEF project and how it aims to help energy companies.

Kreuwel says, “The grid that's currently in place actually was installed decades ago. And decades ago, electric vehicles, or putting solar panels on your house or wind turbines somewhere in the field really wasn't a thing. It didn't exist yet. And so the grid that we put into the ground 30 years ago wasn't designed to have this both inflow and also outflow of energy.”

Key highlights of this video are:
Changes in the consumption of the generation are making it more difficult to forecast the load on the grid. Having accurate predictions of the load on the grid, especially for the short-term, is discussed.
The grid that was implemented 30 years ago was not designed for today’s energy needs. Kreuwel discusses some of the challenges this creates.
Alliander already uses OpenSTEF internally for three main goals for use cases. Kreuwel discusses these goals and how it is being used in Alliander in further detail in the video.
Bogaard discusses why Alliander felt it was important to open source the OpenSTEF project and what benefits it believes this will bring to the community as well as to Alliander.
OpenSTEF is one of the key systems used by Alliander to make forecasts for the national grid operator, to manage and balance the grid and to deploy new smart solutions. Kreuwel discusses how mature the project is and how Alliander are making it usable beyond the scope of Alliander use cases.
Organizations in the energy sector who need accurate forecasts would be interested in this project, but Kreuwel discusses how the project could be useful from an academic standpoint as well.
Those wanting to become a part of the community and collaborate can go to https://www.lfenergy.org/projects/openstef/ for further information or check out the Github community at GitHub.com/OpenSTEF.
Kreuwel discusses the process of how new features are assessed and introduced. He explains how new versions of the OpenSTEF package are released and how the roadmap is determined by each contributor from their own point of view.
Alliander uses as many open source packages as makes sense, such as the machine learning pipelines and machine learning algorithms. Kreuwel explains some of the open source components used in the OpenSTEF project.
It is becoming increasingly important for organizations to reduce their carbon emissions and carbon footprint. Kreuwel explains how the OpenSTEF project is helping organizations achieve this.
Alliander wants to enable the transition to more renewables, getting the most out of the grid while making sure we use all the capacity in a grid. Kreuwel explains how the tooling from OpenSTEF is helping create a new level of distribution system operator with accurate assessment on capacity and forecasting.
Kreuwel explains how short-term forecasting helps organizations manage things in real-time but while long-term analysis goes hand in hand with this, they are two separate applications. He explains why this is the case.

This video was originally posted to https://fosdem.org/2024/schedule/event/fosdem-2024-2211-openstef-opensource-short-term-energy-forecasting/. Thank you to FOSDEM for sharing under CC-BY-2.0-BE-DEED (http://creativecommons.org/licenses/by/2.0/be/deed.en).

"OpenSTEF: Opensource Short Term Energy Forecasting" was presented in the FOSDEM 2024 Energy Devroom by Jonita Ruiter of Alliander.

The energy transition poses new challenges to all parties in the energy sector. For grid operators, the rise in renewable energy and electrification of energy consumption leads to the capacity of the grid to near its physical constraints. Forecasting the load on the grid in the next hours to days is essential for anticipating on local congestion and making the most of existing assets. OpenSTEF provides a complete software stack which forecasts the load on the electricity grid for the next hours to days. Given a timeseries of measured (net) load or generation, a fully automated machine learning pipeline is executed which delivers a probabilistic forecast of future load. During this presentation, the implementation of OpenSTEF to fully automatically forecast the load on the grid at the Dutch Distributed System Operator will be shown. Furthermore, the community built around OpenSTEF will be discussed.

Learn more about LF Energy OpenSTEF at https://lfenergy.org/projects/openstef/.

Welcome to the beginner workshop of OpenSTEF! OpenSTEF provides automated machine learning pipelines to deliver accurate, self-correcting, and explainable forecasts of the load on the grid for the next 48 hours.

In this video, we first give a short presentation explaining the basics around OpenSTEF. Thereafter, we dive into the beginner workshop, which exists of three parts. The workshop documents can be found here:
OpenSTEF/openstef-workshop: Provides Jupyter Notebooks with exercises which enable you to get to know OpenSTEF
https://github.com/OpenSTEF/openstef-workshop

In the workshop, we use Google Colab as a user interface, you can find this here: https://colab.research.google.com/

In this workshop, you will fulfill exercises by answering questions, which will enhance your understanding of OpenSTEF.

Learn more about OpenSTEF at https://lfenergy.org/projects/openstef/. If you have any questions specific to the workshop, feel free to leave a message on the Github page or send an email to korte.termijn.prognoses@alliander.com.

Ready for more of a challenge? Check out our Advanced OpenSTEF Workshop at https://youtu.be/9C2LIcoK7lw

Welcome to the advanced workshop of OpenSTEF! OpenSTEF provides automated machine learning pipelines to deliver accurate, self-correcting, and explainable forecasts of the load on the grid for the next 48 hours.

In this video, we first give a short presentation explaining the basics around OpenSTEF. Thereafter, we dive into the advanced workshop, which exists of three parts. The workshop documents can be found here:
OpenSTEF/openstef-workshop: Provides Jupyter Notebooks with exercises which enable you to get to know OpenSTEF
https://github.com/OpenSTEF/openstef-workshop

In this workshop, you will fulfill exercises by writing code and answering questions, which will enhance your understanding of OpenSTEF.

Learn more about OpenSTEF at https://lfenergy.org/projects/openstef/. If you have any questions specific to the workshop, feel free to leave a message on the Github page or send an email to korte.termijn.prognoses@alliander.com.

Are you new to this field and want similar content geared toward a beginner? Check our our OpenSTEF Workshop for Beginners: https://youtu.be/XNnLvZb5_1o

Net Congestion Forecasting - Daan van Es, Alliander

The energy transition is causing both a rapid growth in net grid loads and fundamental changes in the patterns of power supply and distribution. This leads to congestion on the electricity grid, where some businesses have to wait longer to get their requested connection. One way to mitigate net congestion is by curtailing generation during peak situations. Requests for curtailment have to be sent out by grid operators a day ahead, and thus require accurate load forecasts. In this project demo, I will show how Alliander, a Dutch Distribution System Operator, forecasts grid congestion using openSTEF, an open source Python package for Short Term Energy Forecasting. I will share how we tuned machine learning models to predict peak situations well. If you are also dealing with grid congestion or energy forecasting, please join this talk to exchange experiences and best practices.

Workshop Preview: OpenSTEF Workshop - Bart Pleiter & Daan van Es, Alliander

Join us for a 15-minute workshop preview to learn more about this workshop.

Workshop Description:
OpenSTEF is an opensource project to make short-term energy forecasts. OpenSTEF provides automated machine learning pipelines to deliver accurate and explainable forecasts of the load on the grid for the next 48 hours. Experience OpenSTEF first-hand during this workshop by running OpenSTEF on your device. Dive into the code for an interactive, hands-on learning experience. Master how to leverage OpenSTEF for accurate and efficient energy forecasting. You will also learn how to understand and evaluate the forecasts produced by OpenSTEF. Participants of the workshop are expected to have basic knowledge on Jupyter Notebook and Python.