CAREER: Enhancing Temperature Visualization in Boiling Fluid over Finned Surfaces using Deep Learning-Enhanced Laser-Induced Fluorescence

Arizona State University, Scottsdale AZ

Investigators

Abstract

This project utilizes deep learning-assisted experimental techniques to visually investigate temperature changes in boiling fluids. Boiling heat transfer plays a pivotal role in various industries such as aviation, space exploration, electric vehicles, and industrial heat. Despite its significance, fundamental questions persist regarding boiling heat transfer, including an understanding of unique flow patterns, temperature distributions, bubble sizes, and trajectories. These challenges arise from difficulties in modeling and visualizing temperatures. Understanding temperature distributions and driving processes is crucial for the development of next-generation thermal management systems. The outcomes of this project are expected to be pertinent to industrial applications by establishing knowledge and metrology for complex heat transfer systems. One specific area of interest is enhancing the energy efficiency of heat exchangers. Cooling for data centers (that use heat exchangers) accounts for approximately 1% of all electricity produced in the US, resulting in a cost of $34 billion and 137 million metric tons of carbon dioxide annually. Therefore, exploring new opportunities in advanced temperature-metrology and analysis for heat exchanger performance improvements will have a tremendous impact on energy resources. Additionally, the project contributes to education through three tasks: (1) creating educational videos and exercises for machine learning modeling; (2) developing structured learning activities for K-12 and undergraduate students; and (3) collaborating with Intel Corporation to inspire students in non-academic research settings. Understanding temperature changes in boiling fluids over finned surfaces is currently limited. There is a lack of understanding regarding the spatiotemporal variation of temperature field in boiling fluids over finned surface, which represent a complex fundamental mode of heat transfer. The research proposes a novel temperature visualization method that integrates laser-based diagnostic tools and advanced deep learning methods to enable the measurement in boiling fluids in complex geometries. The project hypothesizes that advances in deep learning can reconstruct temperature fields in fluids from sparse measurements and correct visualization artifacts, enabling the visualization of spatiotemporal temperature variations in boiling fluids. If successful, the proposed research can significantly advance the fundamental understanding of the following thermal transport phenomena: (1) The impact of flow-structure interaction on the temperature field, thermal boundary layer, and superheated liquid layer development, (2) The mixing behavior of the thermal boundary layer during vapor evaporation, bubble departure from the surface, and rewetting of the dry spot after microlayer evaporation, and (3) The distribution of local heat transfer coefficients on finned surfaces at different phases of the boiling process. The improved understanding will contribute to the design of more effective finned heat exchangers. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.