TSO-DSO Coordinated Distributed Optimizations in Grid Operations Considering System Operator Sovereignty

The replacement of fossil power plants by renewable distributed energy sources, as one part of the current energy system transition, shifts electric generation capacity and corresponding ancillary services from transmission system operators (TSOs) to distribution system operators (DSOs). In addition, parts of the German and European power systems are increasingly operated near their capacity limits. Consequently, appropriate coordination between TSOs and DSOs is necessary to ensure the security of supply and to improve system operation behavior. To exploit all potential, the coordinated grid operation needs to consider reactive power and voltage management in addition to active power management. This thesis reviews existing and develops new methods to optimize operational degrees of freedom in voltage and reactive power management by coordinating independently acting system operators (SOs) whose knowledge of other SOs’ grids is limited. It focuses on distributed optimization methods that promise to achieve near-optimal solutions while dividing the overall optimization problem into subproblems corresponding to the areas of the SOs. Three classes of distributed optimization methods are identified. At least one method per class is described and simulated in detail. For simulations, a sample of the SimBench dataset is applied including 261 buses, two TSOs, and two DSOs. While iterative methods repeatedly show undesired, divergent behavior, the sequential method with consecutive optimizations of DSOs, TSOs, and again DSOs proves to be most practicable. However, the need for improvement also exists with the sequential method due to inaccuracies within the methodic procedure, suboptimal simulation results, and an inadequate option to impartially balance the interests of SOs. This thesis proposes a simple and a more advanced method of a sequential approach to coordinate the voltage magnitudes and reactive power exchanges at SOs’ interfaces. Both methods prove to be practicable, satisfy operational limits, and achieve better results than investigated methods and the model of state-of-the-art grid operation. The more advanced method provides a fair-minded balancing of the interests of the SOs and achieves near-optimal results in two of the four combinations of objective functions analyzed. Finally, the thesis advises on how to address potential challenges of the new, more advanced method and proposes the next steps for the implementation in SOs’ operational planning.