Optical Correction for High-Speed Visualization of Two-phase Flow Boiling: Ray

Understanding heat transfer mechanisms in two-phase flow boiling requires highly accurate, high-speed optical visualization. However, capturing high-quality images in cylindrical systems remains technically challenging due to optical aberrations inherent to the geometry. Depending on the fluid phase, the tube acts either as a convergent optical system under liquid-dominated conditions or as a divergent system under vapor-dominated conditions, leading to non-uniform illumination and loss of crucial structural data near the tube walls. This work presents a systematic approach to correct these distortions by developing an analytical optical model based on a thick-lens ray-tracing matrix method implemented in MATLAB® R2024b. To counteract the convergent effect in liquid-dominated flows, a biconcave divergent lens with a radius of 12.4 mm was designed and positioned between the light source and the tube. Conversely, to address the divergent effect in vapor-dominated flows, a positive meniscus convergent lens was designed and placed between the tube and the camera. Due to geometric constraints, the outer surface of this convergent lens was optimized using regression analysis, yielding an eighth-order polynomial profile. Both corrective elements were fabricated using a cost-effective Creality 3D resin printer and subjected to post-printing curing and polishing. Experimental validation with a high-speed camera at 2000 fps demonstrated significant improvements in image quality across both flow regimes. The divergent lens enabled detailed observation of vapor plugs, slugs, and small bubbles by eliminating central light concentration. The convergent lens effectively mitigated peripheral dark zones under high vapor quality conditions, rendering interfacial flow structures highly visible. Despite minor experimental challenges, such as mechanical alignment and optical magnification of the convergent lens, the results highlight the effectiveness of integrating optical analytical modeling with rapid-prototyping 3D-printed lenses. This approach offers a versatile, low-cost framework for improving visualization accuracy in flow boiling research.