Deep learning strategies for critical heat flux detection in pool boiling

Abstract Image-based deep learning (DL) models are employed to enable the detection of critical heat flux (CHF) based on pool boiling experimental images. Most machine learning approaches for pool boiling to date focus on a single dataset under a certain heater surface, working fluid, and operating conditions. For new datasets collected under different conditions, a significant effort in re-training the model or developing a new model is required under the assumption the new dataset has a sufficient amount of data. This research is to explore strategies of DL adapting to new datasets with limited data available. The insights gained could help improve the practicality and reliability of DL for boiling regime studies. Specifically, convolutional neural networks (CNN) and transfer learning (TL) are studied. Using a base model trained and tested for one public dataset (DS1), CNN and TL models are trained with a small portion of a new public dataset (DS2) and tested for the rest of DS2. Results show that TL outperforms CNN by having much higher accuracy and a much lower false negative rate for scarce data (less than5% DS2). When 1% DS2 is used for re-training in CNN versus fine-tuning in TL, the TL model can detect the CHF with an accuracy of 94.79% and a false negative rate of 0.0997, compared with the CNN model with an accuracy of 85.10% and a false negative rate of 0.3237. To further demonstrate the advantage of TL over CNN, an in-house dataset (DS3) is acquired. With less than 0.05% DS3 being used, the TL model can detect the CHF with an accuracy of 95.31% and a false negative rate of 0.0016, compared with the CNN model with an accuracy of 85.91% and a false negative rate of 0.1263. It is observed that TL has much higher robustness than CNN by having more consistent results and smaller standard deviations over multiple trials using stratified random sampling from both DS2 and DS3. Besides, the training time for TL is significantly lower than CNN when limited data used in the re-training and fine-tuning for both DS2 and DS3. These results demonstrate the ability of TL for handling data scarcity in pool boiling applications with potentials for real-time implementations.

[1]  Geoffrey F. Hewitt,et al.  Visualisation and modelling studies of churn flow in a vertical pipe , 2001 .

[2]  Christopher Joseph Pal,et al.  Brain tumor segmentation with Deep Neural Networks , 2015, Medical Image Anal..

[3]  S. Kandlikar.  A Theoretical model to predict pool boiling CHF incorporating effects of contact angle and orientation , 2001 .

[4]  Xi-Qiao Feng,et al.  Deep neural network method for predicting the mechanical properties of composites , 2019, Applied Physics Letters.

[5]  Qiang Yang,et al.  A Survey on Transfer Learning , 2010, IEEE Transactions on Knowledge and Data Engineering.

[6]  J. Weibel,et al.  A coupled wicking and evaporation model for prediction of pool boiling critical heat flux on structured surfaces , 2019, International Journal of Heat and Mass Transfer.

[7]  Konstantinos Gryllias,et al.  Mechanical fault diagnosis using Convolutional Neural Networks and Extreme Learning Machine , 2019, Mechanical Systems and Signal Processing.

[8]  Hyunsoo Yoon,et al.  A Feature Transfer Enabled Multi-Task Deep Learning Model on Medical Imaging , 2019, Expert Syst. Appl..

[9]  Ah Chung Tsoi,et al.  Face recognition: a convolutional neural-network approach , 1997, IEEE Trans. Neural Networks.

[10]  Gustavo M. Hobold,et al.  Visualization-based nucleate boiling heat flux quantification using machine learning , 2019, International Journal of Heat and Mass Transfer.

[11]  Karthik Duraisamy,et al.  Prediction of aerodynamic flow fields using convolutional neural networks , 2019, Computational Mechanics.

[12]  Gustavo M. Hobold,et al.  Automatic detection of the onset of film boiling using convolutional neural networks and Bayesian statistics , 2019, International Journal of Heat and Mass Transfer.

[13]  J. Lienhard,et al.  Hydrodynamic Prediction of Peak Pool-boiling Heat Fluxes from Finite Bodies , 1973 .

[14]  S. Jun,et al.  Pool Boiling Heat Transfer Enhancement of Water Using Brazed Copper Microporous Coatings , 2016 .

[15]  Deqing Wang,et al.  Aligning Domain-Specific Distribution and Classifier for Cross-Domain Classification from Multiple Sources , 2019, AAAI.

[16]  Guanghui Su,et al.  Analysis of CHF in saturated forced convective boiling on a heated surface with impinging jets using artificial neural network and genetic algorithm , 2011 .

[17]  Teresa Wu,et al.  Deep Residual Inception Encoder–Decoder Network for Medical Imaging Synthesis , 2020, IEEE Journal of Biomedical and Health Informatics.

[18]  S. Kandlikar.  Review and Projections of Integrated Cooling Systems for Three-Dimensional Integrated Circuits , 2014 .

[19]  Limiao Zhang,et al.  Percolative Scale-Free Behavior in the Boiling Crisis. , 2018, Physical review letters.

[20]  Moncef Gabbouj,et al.  Real-Time Motor Fault Detection by 1-D Convolutional Neural Networks , 2016, IEEE Transactions on Industrial Electronics.

[21]  Behzad Vaferi,et al.  Estimation of pool boiling heat transfer coefficient of alumina water-based nanofluids by various artificial intelligence (AI) approaches , 2018 .

[22]  Guanghui Su,et al.  Applications of ANNs in flow and heat transfer problems in nuclear engineering: A review work , 2013 .

[23]  Xudong Song,et al.  A flow feature detection method for modeling pressure distribution around a cylinder in non-uniform flows by using a convolutional neural network , 2020, Scientific Reports.

[24]  T. Poinsot,et al.  Training convolutional neural networks to estimate turbulent sub-grid scale reaction rates , 2018, Combustion and Flame.

[25]  N. Miljkovic,et al.  Water immersion cooling of high power density electronics , 2020 .

[26]  Oral Büyüköztürk,et al.  Deep Learning‐Based Crack Damage Detection Using Convolutional Neural Networks , 2017, Comput. Aided Civ. Infrastructure Eng..

[27]  Chao Yang,et al.  A Survey on Deep Transfer Learning , 2018, ICANN.

[28]  Nam Dinh,et al.  Data-driven modeling for boiling heat transfer: Using deep neural networks and high-fidelity simulation results , 2018, Applied Thermal Engineering.

[29]  Taghi M. Khoshgoftaar,et al.  A survey of transfer learning , 2016, Journal of Big Data.

[30]  Mehrzad Shams,et al.  Optimization of subcooled flow boiling in a vertical pipe by using artificial neural network and multi objective genetic algorithm , 2017 .

[31]  S. Jun,et al.  Effect of heater orientation on pool boiling heat transfer from sintered copper microporous coating in saturated water , 2016 .

[32]  Giovanni Montana,et al.  Deep neural networks for anatomical brain segmentation , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[33]  Gokhan Pekcan,et al.  Vibration‐based structural condition assessment using convolution neural networks , 2018, Structural Control and Health Monitoring.

[34]  Sebastian Thrun,et al.  Dermatologist-level classification of skin cancer with deep neural networks , 2017, Nature.

[35]  O. A. Gobyzov,et al.  Combustion Regime Monitoring by Flame Imaging and Machine Learning , 2018 .

[36]  Matthew L. Dering,et al.  A Convolutional Neural Network Model for Predicting a Product's Function, Given Its Form , 2017 .

[37]  V. Carey.  Liquid-Vapor Phase-Change Phenomena: An Introduction to the Thermophysics of Vaporization and Condensation Processes in Heat Transfer Equipment, Third Edition , 2020 .

[38]  G. Falcone,et al.  Axial development of annular, churn and slug flows in a long vertical tube , 2013 .

[39]  O. T. Olakoyejo,et al.  Thermal Energy Processes in Direct Steam Generation Solar Systems: Boiling, Condensation and Energy Storage , 2019, Front. Energy Res..

[40]  Christopher E. Brennen,et al.  Fundamentals of Multiphase Flow , 2005 .

[41]  Ivan Laptev,et al.  Learning and Transferring Mid-level Image Representations Using Convolutional Neural Networks , 2014, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[42]  M. El-Genk.  Immersion cooling nucleate boiling of high power computer chips , 2012 .

[43]  N. Zuber.  Hydrodynamic aspects of boiling heat transfer (thesis) , 1959 .

[44]  V. Dhir.  Mechanistic Prediction of Nucleate Boiling Heat Transfer-Achievable or a Hopeless Task? , 2006 .

[45]  S. Graham,et al.  Pool boiling characteristics and critical heat flux mechanisms of microporous surfaces and enhancement through structural modification , 2017 .

[46]  Yoshua Bengio,et al.  Understanding the difficulty of training deep feedforward neural networks , 2010, AISTATS.

[47]  Hui Xiong,et al.  A Comprehensive Survey on Transfer Learning , 2019, Proceedings of the IEEE.

[48]  Jing Li,et al.  SD-CNN: a Shallow-Deep CNN for Improved Breast Cancer Diagnosis , 2018, Comput. Medical Imaging Graph..

[49]  S. M. You,et al.  Pool boiling heat transfer of a copper microporous coating in borated water , 2020 .

[50]  S. Jun,et al.  Enhancement of Pool Boiling Heat Transfer in Water Using Sintered Copper Microporous Coatings , 2016 .

[51]  Henri Fenech,et al.  Heat transfer and fluid flow in nuclear systems , 1981 .

[52]  Yang Yu,et al.  A novel deep learning-based method for damage identification of smart building structures , 2018, Structural Health Monitoring.

[53]  S. Graham,et al.  Pool boiling enhancement using vapor channels in microporous surfaces , 2019, International Journal of Heat and Mass Transfer.

[54]  Rafael Garcia-Dias,et al.  Convolutional neural networks , 2020 .

[55]  Hui Li,et al.  Digital mammographic tumor classification using transfer learning from deep convolutional neural networks , 2016, Journal of medical imaging.

[56]  U. Rajendra Acharya,et al.  Application of deep transfer learning for automated brain abnormality classification using MR images , 2019, Cognitive Systems Research.

[57]  M. Bucci,et al.  Online, quasi-real-time analysis of high-resolution, infrared, boiling heat transfer investigations using artificial neural networks , 2019 .

[58]  Nick C Fox,et al.  The clinical use of structural MRI in Alzheimer disease , 2010, Nature Reviews Neurology.

[59]  Teresa Wu,et al.  A dual-mode deep transfer learning (D2TL) system for breast cancer detection using contrast enhanced digital mammograms , 2019, IISE Transactions on Healthcare Systems Engineering.

[60]  Gustavo M. Hobold,et al.  Machine learning classification of boiling regimes with low speed, direct and indirect visualization , 2018, International Journal of Heat and Mass Transfer.

[61]  M. McCarthy,et al.  Role of wickability on the critical heat flux of structured superhydrophilic surfaces. , 2014, Langmuir : the ACS journal of surfaces and colloids.

[62]  P. A. Lyakhov,et al.  Application of the residue number system to reduce hardware costs of the convolutional neural network implementation , 2020, Math. Comput. Simul..

[63]  Shuiwang Ji,et al.  Deep Convolutional Neural Networks for Multi-instance Multi-task Learning , 2015, 2015 IEEE International Conference on Data Mining.