Koopman operator techniques for power grid dynamic data

By Professor Yoshihiko Susuki (Kyoto University, Japan)

The use of massive quantities of data in power system engineering has attracted much interest in the era of data science and AI. For real-world infrastructures that are complex and safety-critical, such as power systems, a mathematical foundation for data utilization that guarantees reliability and interpretability has been required. Since 2010, I have been working on applying a mathematical concept for nonlinear dynamical systems, called the Koopman operator, to power systems engineering. The Koopman operator enables the development of analysis and control techniques of power systems that entirely use data, while priori physical models are used. In this presentation, I will talk about my recent developments in applying the Koopman operator technique to power transmission systems, which include modal analysis and inertia estimation from dynamic data, and linear prediction of large-signal dynamics and its application to emergency control

Website
https://y-susuki.github.io/