Gerald Gebhardt | Searcharxiv

arXiv2026

Evaluation of Various Objective Functions for Optimal Reactive Power Flow Including Transformer Tap Changer Optimisation

Modern distribution grids with high penetration of renewable generation provide substantial flexibility through distributed reactive power sources and transformer tap changers. This high degree of freedom can be exploited for optimisation. However, choosing an objective function for optimisation is not trivial, e.~g. minimising grid losses may lead to overvoltages and minimising voltage deviations may lead to higher reactive power flows to neighbouring system operators. Thus, this paper deals with the design of an objective function for the centralised optimisation of distributed reactive power sources and transformer tap changers. Different objectives for characteristic network quantities are investigated for the optimisation and optimised in a combined manner and separately. The consequences of optimising conflicting target values are then analysed. For the optimisation, various grid usage cases of a 110 kV benchmark power grid from SimBench are examined. The investigated power grid is characterised by a high proportion of renewable energy plants. The optimisation is carried out in a data-driven, object-oriented manner using the interior point method with open source software. At the end of the paper, a meaningful optimisation function with combined weighted objectives is derived and the results are analysed.

math.OC↗

arXiv2022

Event-Triggered l2-Optimal Formation Control with State-Estimation for Agents Modeled as LPV Systems

This paper proposes a distributed scheme with different estimators for the event-triggered formation control of polytopic homogeneously scheduled linear parameter-varying (LPV) multi-agent systems (MAS). Each agent consists of a time-triggered inner feedback loop and a larger event-triggered outer feedback loop to track a formation reference signal and reject input and output noise. If a local event-trigger condition is violated, the event-triggered outer feedback loop is closed through the communication network. The event-trigger condition is only based on locally available information. To design the controller, a synthesis problem is formulated as a linear matrix inequality of the size of a single agent under the assumption, that local estimators trigger intercommunication events with neighboring agents if the event-trigger condition is violated. The design procedure guarantees stability and bounded l2-performance. Furthermore, the estimators are interchangeable for a given controller. We compare in simulation zero-order hold, open-loop estimation, and closed-loop estimation strategies. Simulation trials are carried out with non-holonomic dynamic unicycles modeled as polytopic LPV systems.

eess.SY↗