Integrating an autonomous agent-based power flow control system into control center software

Conference: ETG-Kongress 2021 - ETG-Fachtagung
03/18/2021 - 03/19/2021 at Online

Proceedings: ETG-Fb. 163: ETG-Kongress 2021

Pages: 6Language: englishTyp: PDF

Authors:
Pohl, Oliver; Hito, Louay; Ibrahim, Hadi; Al Samman, Omar; Haeger, Ulf (TU Dortmund – ie3, Dortmund, Germany)
Kentchim Tamgue, Ruben; Kubis, Andreas; Heine, Michael (PSI Software AG, Aschaffenburg, Germany)

Abstract:
While some control actions within electrical (sub-)transmission networks have been automated, Power Flow Control (PFC) actions still remain mostly under manual control by a control center operator. In the recent past this has become troublesome since power flows are not as predictable and unidirectional anymore as they used to be. At the same time the rising numbers of installed PFC devices complicate the task of finding optimal set points for them. A Multi-Agent System (MAS) can potentially use these devices for automated curative PFC – but it has to be integrated into control center software so an operator can anticipate agent behavior and include it in system analyses. In this paper a PFC MAS implemented on commercial off-the-shelf control hardware is tested in a Hardware-in-the-Loop (HIL) simulation along with supervision and control software implemented within a control center demonstrator. The agents receive measurements from a Real-Time Simulation (RTS), communicate with each other, activate flexible power to alleviate overloads and forward all their data to the control center for monitoring purposes. Furthermore, an MAS behavior prognosis tool for the control center is developed to include MAS activity within their overall system analyses. The developed systems are tested within a network use case in which a Contingency Analysis (CA) shows that one power line would approach its thermal line rating in case of a specific contingency. The prognosis made by the control center suggests that the MAS will be able to take care of this overload, so no preventive actions are necessary. The results show that the MAS can indeed alleviate the overload in due time while the control center is continuously informed about its actions.