Flow boiling can remove large heat loads efficiently, but pressure and temperature oscillations can make a cooling system difficult to operate and control. This work studies active flow control as a way to mitigate those instabilities rather than treating them as an unavoidable operating limit.
The research connects two-phase-flow experiments with machine-learning-based control for thermal management. The long-term goal is a controller that can keep a boiling system in a useful operating window as heat load and flow conditions change. I presented this work at ITherm 2026, where it received the Best On-Site Poster Award.