Fraunhofer Institute for Wind Energy Systems (IWES) kicked off the new year by presenting TenSyGrid research at the 11th Workshop on Matrix Equations and Tensor Techniques (METT XI), held on January 7-9, 2026, in Leuven, Belgium. Around 60 academics mainly from Europe participated in this KU Leuven organised event.
Christoph Kaufmann presented a poster titled “Structure-preserving Tensor Decomposition — Open Problems for Multilinear Models” co-authored by him with Dr. Georg Pangalos from Fraunhofer IWES and Prof. Gerwald Lichtenberg from Hamburg University of Applied Sciences,
The research tackles a critical challenge in modern energy systems: how to create simplified mathematical models of complex energy networks without losing their essential characteristics. The team’s work focuses on a multilinear time-invariant (MTI) models, which are particularly suited for analysing power grids and complex energy networks. These models can be compressed using advanced mathematical techniques, making them more practical for real-world applications.
It is demonstrated by our team that applying standard tensor decomposition algorithms to a power system model increased its complexity, essentially destroying the network structure that makes the model interpretable and physically meaningful.
The poster presents five key challenges that must be solved to create better model reduction methods. These include preserving mathematical consistency, maintaining important system behaviors like equilibrium points, respecting physical laws such as energy conservation and Kirchhoff’s circuit laws, keeping the sparse network structure intact. Also, ensuring that stability properties critical for power system operation are maintained. The research emphasizes that future algorithms must be “structure-aware” to preserve these essential characteristics while still achieving computational efficiency.
During the METT XI workshop, very valuable feedback from the world-leading expert could be received, which will help TenSyGrid team in their research.
This work is particularly relevant for the clean energy transition, where increasingly complex energy networks with distributed generation, storage systems, and flexible loads require sophisticated yet computationally efficient modeling approaches. The MTI Toolbox 2.1 for MATLAB® is available at TenSyGrid Resources page. The METT XI abstract book is available at the workshop website.
