Excellence lives on exchange and networking among the best researchers worldwide. That is why we want to invite you to join us in a new lecture series, International Excellence Talks, where internationally renowned scientists will give insights into their current research topics and, ideally, inspire a broad audience of scholars. This monthly event will include a 30-minute keynote lecture and a platform for networking and informal exchange among researchers with a wide range of interests. Recordings of previous talks can be found at KIT International Excellence Talks YouTube Channel.

Abstract: The modern low-carbon smart grids face several challenges including the high uncertainty level brought about by the intermittent renewable energy sources, increasing complexity of system modeling due to the integration of converter-interfaced sources (CIGs), and the vulnerability of power grids to cyberattacks because of growing dependence on information technology. In this talk, I will present some recent works of my group addressing the three challenges. To assess the impacts of uncertainty on power system security and operation, we developed polynomial chaos expansion (PCE)-based methods to carry out accurate and efficient probabilistic security assessments including probabilistic power flow, available transfer capability assessments, economic dispatch, microgrid service support, etc. Uncertainty-aware control using energy storage systems is also being designed to reduce the negative impacts of uncertainty and promote the integration of clean energy. To handle the increasing complexity in modeling, analysis and control due to CIGs, we leveraged the high-frequency GPS-synchronized data provided by Wide Area Monitoring System (WAMS) and designed data-driven methods to detect network topology change, conduct wide-area damping control, locate forced oscillations, estimate inertia, maintain voltage and frequency stability for both large grids and microgrids. To investigate the cybersecurity of power grids, we designed false data injection models targeting AC state estimation and automatic generation control (AGC) without the knowledge of system models and parameters. Our current works on the detection and mitigations of the attack models will also be discussed.