Numerical Methods for Stability and Controller Synthesis in Power Systems with Periodic Dynamics

• Applicant: Professor Dr.-Ing. Johannes Schiffer
• Subject Area: Automation, Mechatronics, Control Systems, Intelligent Technical Systems, Robotics
• Term: since 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 545632618

The global transition to renewable energy sources presents significant challenges for the operation of future climate-neutral power systems, particularly in alternating current (AC) systems. A major structural change is the replacement of conventional synchronous generators with inverter-interfaced devices at various voltage levels. This leads to much faster frequency dynamics in the grid. Such inverter-dominated power systems are therefore termed low-inertia power systems. To ensure the affordable, efficient and sustainable operation of future low-inertia systems, there is a critical need for novel and flexible analysis and control solutions. Given their highly disruptive nature, it is impossible to successfully handle these changes exclusively via numerical simulations, nor purely local analysis and synthesis approaches based on linearization of the system dynamics. Instead, new advanced formal methods and techniques suitable for analyzing and designing global properties in power systems are needed. Since AC power systems inherently exhibit periodic behavior, resulting in the existence of multiple equilibria, their global analysis and synthesis are very challenging. To meet this challenge, our research approach is guided by the observation that by exploiting the system’s periodicity, it is possible to relax the usual definiteness requirements of Lyapunov theory while still rigorously guaranteeing stability and robustness properties. By using this property, the partners have jointly developed the Leonov function framework, which specifically addresses the periodic nature of power system dynamics. Inspired by recent advances in deep learning and physics-inspired neural networks (PINNs), the main goals of SyNNuM are geared towards the next development steps in the Leonov framework, namely the derivation of numerical methods to efficiently construct Leonov and Control Leonov functions. In addition, we aim to further relax the requirements of stability analysis and control design via the Leonov method. The outcomes will serve as a link between cutting-edge theoretical advances for control synthesis and a significant application area pertaining to climate-neutral future energy systems, which are at the center of numerous national and European research projects. To this end, we will also experimentally demonstrate the potential of the derived new methods on a Power-Hardware-in-the-Loop setup for low-inertia power systems. The consortium, comprising the Chair of Control Systems and Network Control Technology at Brandenburg University of Technology Cottbus-Senftenberg (BTU), Germany, and the Valse team of Inria, France, combines complementary expertise and aims to realize the high transfer potential of the project outcomes while achieving scientific excellence. This has already been demonstrated by a long-standing, successful collaboration on the project’s subjects.

• DFG Programme: Research Grants
• International Connection: France
• Partner Organisation: Agence Nationale de la Recherche / The French National Research Agency
• Cooperation Partner: Dr. Denis Efimov