PowerGridModel webinar Q/A

[figueroa1395]

On January 20th, 2025, the webinar titled The Power Grid Model Story: Advancing Grid Calculations Through Collaborative Open Source Innovation was carried out. In this discussion, you will find all the questions that were asked during the webinar, and their respective answers.

Q: Is the Power Grid Model suitable to be used by TSOs as well?
A: Even though the mathematics involving TSOs and DSOs aren't very different, the Power Grid Model has been specifically developed for DSOs; for instance, we don't support all the components of a TSO (high voltage grid). However, excluding such components, it should be possible to use the Power Grid Model for TSOs as the general load flow is the same.

In addition, the Power Grid Model is currently focused on "steady state" load flow calculations. So, currently it is not fit to do calculations on a very short timescale (say shorter than 100 milliseconds).

Q: Is the Power Grid Model suitable for both balanced and unbalanced networks?
A: Yes, the Power Grid Model offers the possibility to perform calculations for both balanced (symmetric) and unbalanced (asymmetric) calculations.

Q: How was the internal development of the Power Grid Model balanced with commercially available tools and how did the vendors respond?
A: In the beginning, we had a strong incentive to make a generic tool which could be used in different scenarios and projects. Additionally, one of the problems that we identified at the time, was that the market was missing strong software packages, as the existing solutions were mainly focused on providing a good desktop application with interesting user and visualization tools, aimed at the desktop user. Therefore, we took a completely different approach. The main target of the Power Grid Model is always a programmer, data scientist, software engineer, among other professionals, who want to incorporate the Power Grid Model into a bigger application.

Moreover, we also try to interact with commercial parties, as we offer them very good open source packages, free for use and with a very commercial friendly license. The commercial vendors are free to think about the possibility of increasing their commercial value by incorporating the Power Grid Model as a calculation engine to be used in a bigger application that can be profitable in a commercial environment.

Q: Is it possible to get involved in helping to further improve the Power Grid Model project as someone external to Alliander?
A: Absolutely! Anyone can contribute to the project on many different levels: requesting new features, reporting bugs, developing a feature within the calculation core (C++) or at the user interface level (Python), providing additional tests or validation cases, etc. If you are interested in contributing, a good place to begin are the good first issue issues in our issue board, together with the CODE_OF_CONDUCT.md, CONTRIBUTING.md and GOVERNANCE.md files. Furthermore, feel free to contact the maintainers of the project, or visit the LF Energy website with lots of handy links on the Power Grid Model.

Q: What is the scope of the Power Grid Model?
A: We scoped the Power Grid Model to the bear minimum of the physical modeling. This means that we try to model only the network and its components, while we stay away from business rules, which are usually very specific to every use case and application. In this way we try to ensure the broader applicability of the model, as well as to inspire more collaboration with other projects, both commercial and open source, because there is plenty of room to build an application based the Power Grid Model as its backend.

Q: What solutions do you offer for model validation and convergency issues?
A: Model validation is a very important task for the whole project, because we are performing physical calculations; if we don't deliver reliable results, nobody will use the Power Grid Model. For this reason, we have more than 150 validation cases in our project, which derive from many existing projects, both commercial software and other open source projects, as well as many hand-made cases which we corroborate by doing the calculations by hand first, all of which must always yield the correct results through automatic testing. Thus if anyone in the world wishes to contribute to the project, all of the validation cases will run against the proposed changes and the changes are only accepted after all the cases pass successfully.

On the other hand, we have a Data Validator module which checks for certain data errors. Non convergence is very case specific and hence we do not offer more solutions. Alliander internally has a set of rules based on our distribution grid which weeds out many data errors causing non convergence.

Q: Which vendor tools was the Power Grid Model benchmarked against?
A: Among others: Pandapower and Vision network design tool. More details can be found in one of our published papers, Power grid model: a high-performance distribution grid calculation library

Q: Do you have any tips on persuading a company to embrace open-source tools?
A: One of the main problems we notice with other stakeholders, is that many still see open source as something scary, or something which is not ready to be used in a production environment. However, it is important to point out the role that open source already plays in our current landscape. For instance, many organizations don't know that open source is a critical enabler for innovation, which is already used in many places not only within internal tools, but also in software bought to external vendors. In short, demonstrate the already critical importance and presence of open source software.

Secondly, further elaborate the open source strategy such that the benefits of open source not only come from a user perspective, but also from a collaboration for R&D and innovation. Make the conversation very strategic, as open source is a very good way to bring more external knowledge and resources into your ecosystem.

Q: Does the Power Grid Model include some visualization-methods of the input and the output? For example, is there a visualization of the input-net topology (lines, nodes, sources, etc.)?
A: We are currently working on this within the newly open sourced package Power Grid Model Data Science, and we hope to develop this feature in spite of its challenges. For now we have open sourced tools to interact with and modify the network ,within this data science toolkit.

Q: How did you interface with GIS and SCADA?
A: We use the Common Information Model (CIM) to model our network data, and have established near real time data streams from our SCADA. If you are interested in a specific interface, feel free to post a more specific question here or in our discussion board, as well as to contact any of the maintainers of the project.

Q: Does the Power Grid Model Data Science package use the CIM model for MV/LV grids?
A: Internally, the Power Grid Model Data Science package is very similar to and is based on the Power Grid Model package data structures (they match), so it is possible to interact with the CIM model; nevertheless, there are other specific tools designed for the CIMconversion itself. Moreover, we have our own CIM implementation within Alliander, so we don't provide a dedicated open source converter yet, but we hope that it can be done in the future. If you would like to collaborate on this topic, please contact us.

Q: How long have you been running these grid planning tools for?
A: The Power Grid Model project started more than four years ago. All the packages within the project have been running for quite some time now and all of them are still being developed further.

Q: Do you use the CIM model for the LV grid? Is it the same for the HV and MV grids?
A: The CIM model is applicable, and in fact the same, for all types of grid: HV, MV and LV.