Reasons behind this project

The power system and energy systems at large experience a paradigm shift with many novel, active components connected. Appraising the behaviour and reactions of such new items is critical for network operators to simulate and anticipate system operation, and for this, available, reliable up to date and possibly specifically targeted component models are required.

At present, on one hand, operators manage to gather and build pragmatically a model database of the components directly connected to their network. They remain however blind or use rough proxies to cope with:

• Components indirectly connected (e.g. to neighboring, coupled networks, such as electricity or gas grids; and/or quite likely in a soon future: across sectors)
• Future connected ones (components at design stage, latest technology and proprietary manufacturer models).

On the other hand, research also suffers from the lack of all the necessary component models, and is slowed down by the effort required to circumvent it.

Having “a complete library of reliable energy system components models easily available for all” sets however many challenges:

• “easily available for all” calls for an open source principle, implying to have solved Intellectual Property Rights (IPR) and cyber-security issues with manufacturers.
• many diverse kind of simulations/studies (e.g. steady state / slow / quick / transient dynamic), call for as many fit for purpose models per component
• The large and fast increasing number of components, calls for distributed efforts, hence enabling crowdsourcing, and pointing to (advertising), already existing resources on the web
• Reliability for end-users implies that every model is auditable enough, documented, tested and transparently testable, and possibly verified (while preserving cyber-security).
• The diversity of the used formats and possible languages makes the comparison difficult even for the same component.

As a result, test benches, performance target standards, and reference cases shall also be available in the same framework, and very naturally as well test or real datasets.

Such a structured standard framework shall also ease the integration of novel technologies and ease interfacing with and interoperability of simulation tools.

Project Objectives

COLib (Collaborative Opensource Library) shall organise and nudge this data & models collection effort to build the “Wikipedia” equivalent for network models. The effort is gradual:

• Develop the technical framework, to make the models and their documentation and assessments available; enabling global overview of the models and the test cases, models & datasets upload, modification and assessment; interconnecting with other facilities (see LFE, EPRI, G-PST/pillar, IEEE…)
• Map & solve IPR, other legal & cyber-security issues, especially with component manufacturers and software vendors (e.g. accreditation levels to access the resources);
• Start to populate the database with existing components and test cases for the different time simulation cases (steady state / slow / quick / transient dynamic), coming from existing portals and initiatives (e.g. G-PST/pillar 5, IEEE, EPRI, etc.). The use of practical feedback will lead to the adaptation of the framework.
• Add new components. After the first populating of the library, a list of missing components is identified, and a selection of to-be-added network component is made. Prior directions can be given to:
  • 1/ numerous, wide-spreading power-electronic components (e.g. COLib-gdcc);
  • 2/ novel technologies (likely to challenge the initial design);
  • 3/ existing network component that need improvement, standard models that fits with the behaviour of specific models
  • 4/ the completeness of network components for a given dynamic type
• Adding new test cases. It is very likely that after the first populating of the library, most of existing test cases are either unit test or small standard existing cases. Prior directions can be given to:
  • 1/ large scale power system test cases, with realistic public data or private real networks
  • 2/ extreme cases to illustrate specific events or phenomena (loss of synchronism, voltage collapse, converter-driven stability )
• Foster a living open source project and community, so that it is widely used, while ensuring a continuous populate of the library and improving its structure over time

Project Members:

COLib is a common project. Its architecture has been set up and coordinated by CRESYM’s staff in 2023.

Standards and new models and test cases are being prepared in 2024, esp. by FG-IEE (e.g. standard machine controls, Kundur two-area test case, Simplified Electrolyser model, EPRI grid forming inverter model, IEC generic type 3 wind turbine model…)

Credit: the very nice picture is by Max Langelott on Unsplash.