Dissertation Introduces AI-Powered Hybrid Model for HVAC Systems

Congratulations to Po-Ching Hsu on successfully defending his Ph.D. dissertation, “Experimental Investigation and Data-Driven Modeling for Variable Refrigerant Systems.”

Hsu’s research focused on variable refrigerant flow (VRF) systems – complex HVAC configurations that can deliver high energy efficiency by using multiple indoor units for zone-specific temperature control. He investigated modeling methods to improve VRF efficiency and enhance real-time control.

For his dissertation, Hsu developed a hybrid model that combines physics-based modeling with machine learning. The hybrid approach delivered higher predictive accuracy than either method alone, helping a system operate more efficiently by better matching heating and cooling supply with demand.

Hsu conducted his research through CEEE’s Consortium for Energy Efficiency and Heat Pumps (EEHP), under the guidance of EEHP Director and Research Professor Yunho Hwang. Next, Hsu will join Solstice Advanced Materials in Buffalo, New York, as an R&D engineer.

“My time at CEEE provided me with the opportunity to work on impactful research in HVAC and energy systems, while building strong skills in both experimentation and modeling,” Hsu says. “I am especially grateful for the supportive and collaborative environment, which encouraged me to explore new ideas and grow as an independent researcher. This experience has played a key role in preparing me for my next step in industry.”

A searchable listing of CEEE theses and dissertations published 2003–2026 is available online.

Published May 18, 2026