Power Grid Model DS: Simplifying Data Science for Distribution System Analysis
Summary Signal: Presented at the LF Energy Summit Europe 2025:
At the Summit, Peter Salemink and Jaap Schouten from Alliander introduced Power Grid Model DS, a new open source toolkit designed to simplify data science workflows for distribution system analysis. Developed within Alliander and now contributed to the LF Energy ecosystem, the project bridges the Power Grid Model calculation engine with modern data science tools, accelerating collaboration and innovation in grid planning and operation.
Smarter Data Science for the Energy Transition
Across Europe, the energy transition is increasingly constrained by grid congestion.
From the Netherlands to Germany and the UK, overloaded networks are delaying renewable integration and new customer connections.
As Salemink explained, “traditional grid upgrades are too slow.” Building new infrastructure takes years; smarter, data-driven solutions are needed now. By enabling fast and accurate grid calculations, Power Grid Model and its new DS toolkit unlock dynamic congestion management, improve planning, and reduce costs through better investment decisions. Open collaboration is key. Shared tools like Power Grid Model and Power Grid Model DS help DSOs, TSOs, academia, and vendors reduce duplication and accelerate the digital transformation of the grid.
From Library to Toolkit: Expanding the Power Grid Model
The Power Grid Model library is an open source, high-performance calculation engine for distribution systems, providing methods for:
- Power flow (balanced and unbalanced)
- State estimation using statistical methods
- Short-circuit calculations
It’s built for cross-platform performance (Windows, Linux, macOS), native parallelization, and a well-defined API, making it suitable for both research and operational environments. Over the past year, the community—featuring Alliander, Enexis, Stedin, TU Delft, Eindhoven University of Technology, RSZ VR, Ponder Power, and Subtimize—has contributed several key enhancements:
- Columnar data support for improved database integration
- Generic branch modeling for flexible asset representation
- Asymmetric line modeling (TU/e contribution)
- Current sensor and phasor (PMU) measurement support for state estimation
- Full observability checks for radial networks, with meshed grid checks in development
Building on this foundation, Power Grid Model DS adds a powerful new layer—tailored specifically for data scientists and Python developers.
Architecture: Bridging Engineering and Data Science
Jaap Schouten described Power Grid Model DS as a lightweight Python wrapper around the Power Grid Model engine. It combines:
- NumPy for high-speed numerical computation
- NetworkX for graph-based network modeling
- An integrated Grid class that synchronizes analytical and topological views of the network
This architecture allows users to:
- Represent network elements (nodes, transformers, lines) as both arrays and graphs
- Query structural properties such as cycles, paths, and connectivity
- Seamlessly switch between analytical (data-driven) and structural (topological) perspectives
This dual representation supports rapid prototyping while remaining robust and production-ready for operational deployments.
Integrating Data Science Workflows
Power Grid Model DS aligns with the full data science workflow for power system analysis:
- Raw Data Ingestion – importing and cleaning real-world grid data
- Modeling – creating structured, analyzable grid representations
- Simulation – running fast load flow or scenario studies
- Impact Assessment – evaluating voltage, congestion, or capacity outcomes
- Visualization – exploring network behavior for debugging and validation
Visualization remains one of the most requested features, giving developers a clearer understanding of network states and improving developer experience.
Real-World Use Cases
Within Alliander, Power Grid Model DS is used across multiple analytical and operational contexts:
- Short-term system operations – forecasting and preventing congestion 24–48 hours in advance
- Outage and maintenance planning – simulating switching and restoration strategies
- Customer connection studies – assessing connection feasibility and local network impact
- Long-term investment planning – simulating decades of grid evolution to guide investment decisions
Each application leverages the same shared data science foundation, ensuring consistent, scalable, and maintainable workflows.
Watch the full presentation on youtube here: https://youtu.be/8kA7IzLPdDQ?si=8fyExG8ehtR5ztZk
Community and Collaboration
The Power Grid Model ecosystem continues to grow as an open collaboration among DSOs, universities, and software developers. In 2025, Alliander contributed Power Grid Model DS to LF Energy, formalizing its governance and enabling wider participation from across the energy sector. The team invites contributions via the Power Grid Model GitHub repository and encourages developers to join annual community meetups to share use cases, align roadmaps, and expand the shared open source foundation for smart grids.
FAQ
What is Power Grid Model DS?
A Python toolkit that simplifies data science workflows for distribution grid analysis, built on the Power Grid Model calculation engine.
Who develops it?
Developed by Alliander and contributed to LF Energy, with participation from Enexis, Stedin, TU Delft, Eindhoven University of Technology, and other partners.
Which technologies does it use?
Python, NumPy, and NetworkX.
What types of analyses are supported?
Power flow, state estimation (including PMU-based measurements), short-circuit calculations, and visualization for both balanced and unbalanced grids.
Where can I learn more or contribute?
Visit github.com/PowerGridModel/power-grid-model.
Keywords
Power Grid Model DS, Power Grid Model, LF Energy, Alliander, distribution systems, data science, NumPy, NetworkX, open source, grid analysis
About Power Grid Model
Power Grid Model is an open source library for distribution system analysis, providing high-performance calculations for power flow, state estimation, and short-circuit studies. Developed under LF Energy, it enables collaboration among DSOs, TSOs, academia, and technology vendors to accelerate smart grid innovation.
About LF Energy
LF Energy, an initiative of the Linux Foundation, brings together utilities, technology vendors, researchers, and regulators to accelerate the energy sector’s digital and decarbonized transformation through open source software, shared data models, and collaborative governance.
Learn more at https://www.lfenergy.org.
