Flow boiling in mini/micro channels is a very effective technique for meeting the high dissipating requirements of thermal management systems. However, accurate prediction of heat transfer coefficients remains an elusive task because of the complex fluid and thermal behavior in these two-phase systems. In this study, a machine learning based approach for predictingheat transfer for saturated flow boiling in mini/micro channels is proposed. A consolidated database of 16953 data points for flow boiling heat transfer in mini/micro-channels is amassed from 50 sources that includes 16 working fluids, reduced pressures of 0.0046 - 0.77, hydraulic diameters of 0.15mm - 6.5mm, mass velocities of 19 < G < 1608kg/m(2)s, liquid-only Reynolds numbers of 27 - 55270 and flow qualities of 0 - 1. An Artificial Neural Network (ANN) model is developed based on the universal consolidated database that was split into training data and test data, and used to predict the saturated flow boiling heat transfer coefficients. An optimization is conducted and ANN model architecture is selected which consists of dimensionless input parameters: Bd, Bo, Co, Fr-g, Fr-go, Fr-f, Fr-fo, Pr-g, Pr-f, Re-g, Re-go, Re-f, Re-fo, Su(g), Su(f), We(g), We(go), We(f), and We(fo), and hidden layers (75,70,60,50,30,20,10) that predicts the test data with an MAE of 14.3%. The ANN model is superior to universal correlations for saturated flow boiling heat transfer at predicting the test data, even predicting individual databases with high accuracy. The robustness of the ANN model was tested by excluding databases from the training datasets altogether and utilized to predict these excluded databases. The ANN model did extremely well when a working fluid data was included in the training dataset, and poorly when a working fluid data was excluded from training dataset. The use of a universal ANN model utilizing a consolidated database can become an extremely useful tool when it comes to predicting heat transfer coefficients for saturated flow boiling in mini/micro channels. (C) 2019 Elsevier Ltd. All rights reserved.
