A Novel Concept for Dynamic Network Reduction Dedicated to Real-Time Application

Kuri A, Gäbel S, Burlakin I, Wagner T, Mehlmann G, Luther M (2022)

Publication Language: English

Publication Type: Conference contribution, Original article

Publication year: 2022

Publisher: CIGRE

Conference Proceedings Title: Kyoto Symposium 2022 - CIGRE Japan

Event location: Kyoto

Abstract

An increasing power electronic based generation and distribution causes new requirements for the power system. It pushes the limits of hardware and software resources for modelling and simulation while demanding more accuracy, with modern methods for analysis [1]. Therefore, stakeholders increasingly insist on analyzing the new source’s interactions and impacts on the power grid under offline and real-time environments. In power system engineering, real-time simulators form the interface between the virtual and real-world and assess electrical equipment and control. Thus, they serve as a tool for digital twins. The simulation of dynamic phenomena requires a plausible network model with the device under test to assess those interactions accurately. This is achieved by a suitable dynamic reduction of the grid areas to be investigated, with complex reduction process requirements. Historically the main reason for dynamic network equivalents was the limited CPU capacity enabling steady-state and dynamic calculations on extended networks. The results often had a reasonable degree of approximation. Modern computers are equipped with performant multi-core CPUs and large RAM, yet dynamic network equivalents of extended grids are still used nowadays. One of the main reasons for developing model equivalents is to accurately reproduce and simulate complex inverter-based equipment and depict its interaction with the network in software tools and real-time simulators. Due to the control structure, among other things, integration time steps are often chosen in the microsecond range. Hence, such simulations usually take ample time and require high processing power. This is a usual concern when equipment test matrices with several hundred cases have to be dealt with. Simulating a network and the device under test in a real-time environment has additional challenges. Several approaches and tools to achieve reduced networks are available; however, none of the approaches are customized to achieve reduced models for real-time simulators. The paper highlights a dynamic model reduction methodology that accommodates real-time requirements. The methodology concentrates on achieving a high degree of reduced grid elements focusing on the real-time simulator requirements and developing an interactive user automation process. The dynamic reduction process is based on a hybrid methodology of machine coherency and estimation-based equivalent methods achieved via Multi-Objective Genetic Algorithm NSGA-II presented in [2]. The paper focuses on the next phase of transferring the equivalent developed in an offline environment to the real-time environment considering the hardware requirements and automating the process. The reduced model of the targeted network area consists of a physically realistic grid model rather than a mere mathematical equivalent and is therefore suited for real-time simulations with hardware-in-the-loop (HIL). The complete path and endeavor from developing offline equivalents to emulating the complete system on real-time hardware is the prime focus of the paper.

Authors with CRIS profile

How to cite

APA:

Kuri, A., Gäbel, S., Burlakin, I., Wagner, T., Mehlmann, G., & Luther, M. (2022). A Novel Concept for Dynamic Network Reduction Dedicated to Real-Time Application. In Kyoto Symposium 2022 - CIGRE Japan. Kyoto, JP: CIGRE.

MLA:

Kuri, Ananya, et al. "A Novel Concept for Dynamic Network Reduction Dedicated to Real-Time Application." Proceedings of the CIGRE 2022 Kyoto Symposium, Kyoto CIGRE, 2022.

BibTeX: Download