More Like this

1. A Novel Vision-Based Framework for Detecting Changes in Heat Flux in Pool Boiling Using Morphological Similarity

   https://doi.org/10.1115/IPACK2025-163830  

   Nirapure, Pranay; Singh, Ayushman; Rangarajan, Srikanth; Sammakia, Bahgat (October 2025, American Society of Mechanical Engineers)

   Abstract This work presents a robust visual diagnostic framework for analyzing pool boiling phenomena, aimed at quantifying relative changes in heat flux and providing a foundation for future predictive models with limited temperature data. Instead of directly estimating heat flux from visual features, the proposed approach focuses on identifying relative changes in boiling visual patterns to detect a shift in heat flux and infer proximity to the critical heat flux (CHF), which is a key threshold beyond which thermal performance rapidly deteriorates. At the core of this framework is a novel metric, morphological similarity, which captures variations in bubble shape and spatial organization using Scale Invariant Feature Transform (SIFT)-based feature matching between image pairs. Experimental validation on two different heated surfaces demonstrates that morphological similarity reliably tracks changes in heat flux and strongly correlates with surface superheat, a key thermal parameter. Requiring only minimal data, this method generalizes across two surfaces and offers a scalable, real-time alternative for thermal diagnostics in electronics cooling systems. 

   more » « less
2. Visual-data sampling strategies for pool boiling heat flux prediction: Enhancing pathways to data-driven digital twins

   https://doi.org/10.1016/j.applthermaleng.2026.129997  

   Nirapure, Pranay; Singh, Ayushman; Rangarajan, Srikanth; Sammakia, Bahgat (April 2026, Applied Thermal Engineering)

   This work aims to develop a data acquisition methodology for visual data-based deep learning models to predict boiling heat flux. Conventional deep learning models often rely on a fixed architecture and uniform training data sampling across the nucleate boiling regime, which can lead to suboptimal performance because visual patterns are sensitive to changes in heat flux. To address this challenge, we developed an approach that uses the Index of Visual Similarity (IVS) to guide training data sampling (𝛥𝑄). To investigate the problem, we developed an interpolation framework that uses high-speed pool-boiling images and leverages global visual features via a VGG-16-based convolutional neural network to estimate heat flux. This model has been used to test the developed IVS-based methodology. The methodology uses IVS as a visual indicator to assess the degree of change in identifiable physical features with respect to heat flux. Hence, changes in IVS can be used to guide the selection of an optimal data sampling distribution across the heat flux range under consideration. The study further examines how the heat flux sampling used for training influences both model accuracy and computation overhead. A Genetic Algorithm (GA) is developed and employed to optimize the training interval, balancing predictive accuracy against computational cost. The optimization results reveal variations in boiling visual patterns across the explored heat flux range of 50-90 W/cm2 . These findings from the genetic algorithm were used to validate the IVS-guided heat flux sampling. Finally, it was observed that an IVS-guided training reduced training time by 39.1% compared with training at the lowest uniform sampling rate (𝛥𝑄 = 1 W/𝑐𝑚2 ), while maintaining high accuracy, with a mean absolute percentage error of 1.59%. Overall, this work demonstrates that efficient data sampling strategies can significantly reduce computational overhead without compromising accuracy, providing new opportunities for AI-driven thermal diagnostics. 

   more » « less
3. A Convolution Neural Network Design for Combined Image and Sensor Data Analysis to Determine Droplet Vaporization Regime and Heat Transfer Performance

   https://doi.org/10.1115/1.4069330  

   Njike, Ursan Tchouteng; Silva, Anisa; Carey, Van P (November 2025, ASME Journal of Heat and Mass Transfer)

   Abstract Combining high-speed video cameras and optical measurement techniques with digital sensors controlled by a data acquisition system can provide an effective means of exploring boiling process thermophysics and heat transfer mechanisms. Imaging can provide qualitative and quantitative information that complements data provided by temperature, pressure, and other sensors. This paper summarizes the results of an exploration of machine learning strategies to optimally combine and analyze boiling process images and digital sensor information from experiments. We specifically sought a convolution neural network (CNN) to analyze the vaporization of deposited water droplets on superheated surfaces that may have varying degrees of nucleate boiling effects. Two specialized CNN models were developed in this study that can simultaneously analyze both image and digital data. One of our CNN model designs (case B) was trained to take an image of the vaporization process and nonthermal digital data as input and predict thermal heat transfer performance. This model predicts performance remarkably well given its nonthermal inputs, matching independent heat flux test data to a root-mean-square percent error (RMSPE) of 10.3%. This model appears to learn how the variations of nucleate boiling, vapor recoil activity, and local dryout over the surface vary with surface temperature and/or heat flux from changes in boiling system images. We also describe a CNN model (case C) that takes digital nonthermal data, digital thermal data, and image information and provides a high-fidelity prediction of vaporization heat transfer performance. This model predicted performance very well—better than our conventional fit to data (case A) and on par with best fits to quality nucleate boiling heat transfer data in the literature. This type of trained model fit independent heat flux test data to an RMSPE of 5.8%. Our results indicate that training this type of model which predicts performance from input image information and digital operating condition thermal data makes the resulting predictive model more accurate and robust. The successful use of the hybrid CNN models described here suggests that there is a strong correlation between two-phase morphology variations and changes in heat transfer performance. The hybrid CNN modeling approach developed in this research appears to be a promising strategy for analyzing experimental data for physical systems that are best investigated experimentally with combined use of imaging and digital sensor instrumentation. Possible use of this type of modeling in other systems is also discussed. 

   more » « less
4. Multimodal Characterization of Steady-State and Transient Boiling Heat Transfer

   https://doi.org/10.1115/HT2023-106015  

   Pandey, Hari; Dunlap, Christy; Williams, Amanda; Marsh, Jackson; Hu, Han (September 2023, American Society of Mechanical Engineers)

   Abstract Boiling is a high-performance heat dissipation process that is central to electronics cooling and power generation. The past decades have witnessed significantly improved and better-controlled boiling heat transfer using structured surfaces, whereas the physical mechanisms that dominate structure-enhanced boiling remain contested. Experimental characterization of boiling has been challenging due to the high dimensionality, stochasticity, and dynamicity of the boiling process. To tackle these issues, this paper presents a coupled multimodal sensing and data fusion platform to characterize boiling states and heat fluxes and identify the key transport parameters in different boiling stages. Pool boiling tests of water on multi-tier copper structures are performed under both steady-state and transient heat loads, during which multimodal, multidimensional signals are recorded, including temperature profiles, optical imaging, and acoustic signals via contact acoustic emission (AE) sensors, hydrophones immersed in the liquid pool, and condenser microphones outside the boiling chamber. The physics-based analysis is focused on i) extracting dynamic characteristics of boiling from time lags between acoustic-optical-thermal signals, ii) analyzing energy balance between thermal diffusion, bubble growth, and acoustic dissipation, and iii) decoupling the response signals for different physical processes, e.g., low-to-midfrequency range AE induced by thermal expansion of liquids and bubble ebullition. Separate multimodal sensing tests, namely a single-phase liquid test and a single-bubble-dynamics test, are performed to reinforce the analysis, which confirms an AE peak of 1.5 kHz corresponding to bubble ebullition. The data-driven analysis is focused on enabling the early fusion of acoustic and optical signals for improved boiling state and flux predictions. Unlike single-modality analysis or commonly-used late fusion algorithms that concatenate processed signals in dense layers, the current work performs the fusion process in the deep feature domain using a multi-layer perceptron regression model. This early fusion algorithm is shown to lead to more accurate and robust predictions. The coupled multimodal sensing and data fusion platform is promising to enable reliable thermal monitoring and advance the understanding of dominant transport mechanisms during boiling. 

   more » « less
5. Experimental Characterization and Server-Level Demonstration of Confined, Direct-on-Silicon Two-Phase Jet-Impingement Cooling With Additively Manufactured Porous-Wick-Assisted Phase Separation

   https://doi.org/10.1115/1.4071262  

   Yogi, Ketan; Sahu, Gopinath; Rajeev, Sidharth; Lattupalli, Harish Kumar; Wang, Keyu; Chen, Zhengwei; Yang, Yunchun; Schiffres, Scott N; Rangarajan, Srikanth; Sammakia, Bahgat; et al (June 2026, Journal of Electronic Packaging)

   Two-phase jet impingement cooling is a promising solution for high-heat-flux semiconductor thermal management, as it combines strong convective heat transfer with boiling to remove large heat loads at relatively low flow rates and pressure drops. However, practical deployment is hindered by challenges including inconsistent boiling initiation on smooth surfaces, surface dry-out, vapor-induced flow instabilities, and premature critical heat flux (CHF). Excessive vapor generation within confined geometries can disrupt flow uniformity, causing temperature oscillations and unstable operation. To address these challenges, this work presents a confined, direct-on-silicon two-phase jet impingement cooling approach incorporating a porous-wick-assisted phase separation mechanism. The engineered porous wick enhances nucleate boiling and enables in situ phase separation at the boiling surface. Integrated with a custom three-path manifold, the design routes separated liquid and vapor streams, minimizing vapor accumulation within the confined region and suppressing two-phase instabilities. The porous wick is directly printed onto the silicon substrate using advanced additive manufacturing, eliminating the need for a thermal interface material (TIM) and its associated thermal resistance. Thermal–hydraulic characterization using a low-surface-tension dielectric fluid demonstrates that wick-assisted phase separation reduces thermal resistance by 23–29% compared to configurations without phase separation. Extended testing over more than 200 h of continuous operation confirms stable thermal performance and indicates strong potential for long-term reliability. System-level validation is demonstrated in a 1 U server equipped with an NVIDIA V100 GPU (graphics processing unit) incorporating a direct-on-silicon printed wick. 

   more » « less