New correlation of heat transfer coefficient for saturated flow boiling in smooth helically coiled tubes

Abstract

Flow boiling heat transfer in smooth helically coiled tubes (HCTs) is widely used in many industrial sectors, such as nuclear reactors, refrigeration, and heat pump systems. It is important to predict accurately the heat transfer coefficient (HTC) of saturated flow boiling in smooth HCTs, and the prediction accuracy of the existing correlations needs to be improved. For such needs, this paper presents the work developing a new HTC correlation, with a systematic strategy combining the parameter identification, the least squares regression, the machine learning methods, the genetic algorithm, and the error distribution analysis. For developing the new HTC correlation, a database containing 1423 experimental data points of saturated flow boiling heat transfer in smooth HCTs was compiled from 16 published articles, involving four working fluids of water, R134a, R123, and R407C. It is far larger than the largest counterpart, which only contains 1035 data points from 13 published articles and involves three working fluids, and thus it has a potential to enhance greatly the applicability of the new correlation to be developed. Based on the database, 23 existing correlations were evaluated, and a new correlation was proposed. The comparison results based on the database show that the new correlation has much higher prediction accuracy than the best-performing existing one. It has a mean absolute deviation (MAD) of 20.0% and a coefficient of determination (R²) of 0.83, while the latter only has an MAD of 26.1% and a R² of 0.73.