Answers are grounded in this paper and cite their sources.
This analysis reveals that machine learning models outperform traditional methods for predicting boiling heat transfer coefficients in various refrigerants.
This study extensively discusses the use of machine learning to predict boiling heat transfer coefficients in piping and compares it to traditional prediction-related formulas. A comprehensive database of 1532 experimental conditions, refrigerants, and pipe surface structures was developed for the study. The machine learning model incorporates seven input parameters including pipe information, operating conditions, and refrigerant type. Three traditional correlation models were evaluated using these datasets, demonstrating prediction accuracies of 69.12%, 39.62%, and 40.73% respectively, with a margin of error within ±20%. The boiling heat transfer coefficients were also predicted using four machine learning models including Convolutional Neural Network (CNN), Support Vector Regression (SVR), Random Forest (RF) and Radial Neural Network (RBF). Five sets of training and prediction were performed by the cross-validation method. The RF model exhibited the highest prediction performance, achieving an R2 value of 0.894, a Mean Relative Deviation (MRD) of 5.65%, and a Mean Absolute Relative Deviation (MARD) of 14.40%. In comparison, traditional models demonstrated relatively poor performance, particularly under varying operating conditions. These findings highlight the great potential of machine learning in solving complex nonlinear two-phase flow heat transfer problems.
No takes yet. Share an insight, caveat, or question.
Zhu et al. (2025) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: