LAUSR

Abstract Two-phase flow occurs frequently in the field of thermal engineering. The heat transfer performances of gas-liquid two-phase flows play a crucial role in the structure design, material selection and optimization operation of equipments. It is highly desired to acquire the convective heat transfer coefficient of gas-liquid two-phase flows. Different from common experimental technologies and numerical computation methods, in this paper the least square support vector machine (LSSVM) method is developed to predict the heat transfer coefficient of two-phase flows from the limited number of experiment data. The group search optimizer (GSO) algorithm is employed to find the optimal hyper-parameters in the LSSVM method to improve the prediction quality. Numerical experiment results validate that the LSSVM method has a good generalization ability, and can deal with the complicated nonlinear problems with small samples and high accuracy. Furthermore, the use of the LSSVM method to predict the heat transfer coefficients of two-phase flows can significantly reduce the experiment period and test materials, and decrease the cost. As a result, an efficient method is introduced for the prediction of the heat transfer coefficient of gas-liquid two-phase flows.