Application of Helium-Filled Soap Bubbles Towards Double-exposure Multi-pulse 3D/3C Velocimetry in a Turbulence Chamber

The temporal resolution of time-resolved particle tracking velocimetry is typically limited by the maximum frame rate of high-speed cameras, while the achievable pulse rate of illumination sources, such as lasers and LEDs, can be one or more orders of magnitude higher. This work presents a technique to extend the conventional limits of time-resolved particle tracking velocimetry by encoding temporal information within LED pulses of long-exposure particle images. The tracers are helium-filled soap bubbles in air, generated using an in-house built seeder. A machine-learning framework leveraging recent advancements in computer vision is used to separate particle images at high densities into their "single exposure" equivalents. The framework is tested both synthetically and using experimental 3D/3C velocity field reconstructions of flow inside a novel zero-mean homogeneous turbulence chamber. The framework can be used to effectively collect time-resolved data with low-speed cameras, reducing the heavy data storage requirements of high-speed cameras and improving statistical convergence in turbulent flows.