Investigations of the curvature-driven effect on falling-film heat and mass transfer via numerical and machine-learning methods

作者：Jiale Huang, Zhuohang Zhang, Ding Zhang, Rong Jiang, Shouxiang Xie, Tian Hua Jiang, Zihan Huang, Trevor Hocksun Kwan, Jielin Luo · 发表于：International Communications in Heat and Mass Transfer · 年份：2025 · DOI：10.1016/j.icheatmasstransfer.2025.110045 · 被引用次数：5 · 研究领域：Fluid Dynamics and Thin Films、Heat Transfer and Optimization、Nanofluid Flow and Heat Transfer

Heat exchangers with curved structure are widely used in energy-intensive industries because of their compact design and high thermal efficiency. However, the curved characteristic brings up complexity and difficulty in accurately simulating heat transfer performance, while the influence of curvature has not yet been fully understood. In this study, a three-dimensional numerical model is developed to examine falling film flow and heat transfer characteristics on curved tube, with the corresponding variations in heat transfer coefficient (HTC) analyzed. Results show that curvature leads to circumferential asymmetry in liquid film thickness distribution, where the inner-side films are thicker than the outer-side films due to centrifugal force effects, while the axial thickness distribution remains nearly unaffected. Accordingly, curvature exhibits a directional dependence on HTC, with little axial influence but significant circumferential effects. The optimal curvature is explored to obtain the highest HTC improvement, and further increases in curvature lead to the reduction of HTC because of the occurrence of dry patches on outer-side and accumulation effect on inner-side. The fundamental relationships between other key parameters and HTC are also discussed under different curvatures. Additionally, an artificial neural network (ANN) model is established to predict heat transfer performance of curved tube. The ANN model demonstrates reliable performance prediction with a root m...