 |
Zhenyi YuanAssistant Professor |
Upcoming Events & Recent News
[Dec 06-13, 2026]: NeurIPS AI Foundations for Power Grids Workshop at Sydney, Australia.
Short Biography
I am a tenure-track assistant professor at the IoT Thrust of HKUST Guangzhou, and a cross-campus affiliated assistant professor at HKUST. I did my Ph.D. in Engineering Sciences (Mechanical Engineering) with Jorge Cortés at UC San Diego, and B.S. + M.S. in Control Science and Engineering with Ligang Wu at HIT Honors School. Prior to HKUST, I held short visiting and research positions at CUHK Information Engineering Department, hosted by Changhong Zhao. I was also involved in several industrial projects on DER management, collaborating with NREL and TotalEnergies.
Research Interests
I am broadly interested in Control, AI, and Power Systems. Recently I work on:
Data-driven control of dynamical systems using Koopman operators
Control-theoretical perspectives on LLM alignment and interpretability
Machine learning for power flow analysis, control, and optimization
But if you work on any aspect of my broaden research interests, feel free to reach out, I am sure we can find something interesting to discuss.
Want to Read Something?
Here are two entry points into my work.
Learning + Power Systems
Provably Stable Voltage Control
How can learned voltage controllers improve network operation without giving up closed-loop stability? This line of work develops stability-aware local controllers and extends them to richer information and communication structures.
- Start here Learning provably stable local Volt/Var controllers for efficient network operation IEEE Transactions on Power Systems, 2024
- Design constraints Constraints on OPF surrogates for learning stable local Volt/Var controllers IEEE Control Systems Letters, 2023
- Broader networks Stability constrained voltage control in distribution grids with arbitrary communication infrastructures IEEE Transactions on Smart Grid, 2026
- Learning perspective Structured Equilibrium Learning for OPF Imitation NeurIPS Workshop on AI Foundations for Power Grids, 2026
Data + Dynamical Systems
Data-Driven Control of Complex Systems
When first-principles models are difficult to obtain, structured representations learned from trajectories can still support prediction and control. This framework connects bilinear and Koopman models with optimal and predictive control.
- Foundations Formulas for data-driven control: Stabilization, optimality, and robustness IEEE Transactions on Automatic Control, 2019
- Bilinear systems Data-driven optimal control of bilinear systems IEEE Control Systems Letters, 2022
- Nonlinear systems Data-enabled predictive control for nonlinear systems based on a Koopman bilinear realization IEEE Conference on Decision and Control, 2025