Different flow patterns formed between horizontal tubes in falling film heat exchangers have an impact on the heat and mass transfer performance. To accurately predict the transitions of different inter-tube flow patterns of two-dimensional finned tubes under isothermal conditions, this study designs and establishes an experimental system for falling film flow between horizontal tubes. Based on the experimental results, it is clarified that tube spacing ( s ) and fin structure (fin height h , fin pitch p , and fin tip t ) affect the transitions of inter-tube flow patterns. Therefore, this study adopts a machine learning approach to establish an inter-tube flow pattern prediction model that comprehensively considers the influencing factors of tube spacing and fin structure. The results show that, in the ranges of 6.56 × 10 7 ≤ Ga m ≤ 2.74 × 10 10 , 0.32 ≤ h /( p - t ) ≤ 1.79 and 2.19 ≤ s / ξ ≤ 5.49, the prediction deviations for the four falling film flow patterns are within ± 20 %, with at least 90 % falling within ± 20 % for each pattern. For the same falling film working fluid, the present model outperforms the correlation for all four flow patterns, reducing the RMSE and MAE by 28.3 %–65.4 % and 17.8 %–39.6 %, respectively. For different working fluids, the prediction deviations of the present model remain within ± 20 % for all four flow patterns, whereas those of the correlation exceed ± 20 % for the D-DC and DC-C patterns, demonstrating its stronger generalization ability. The research results of this study provide a new reference for the development of subsequent falling film flow pattern prediction models.
No takes yet. Share an insight, caveat, or question.
Chen et al. (2026) studied this question.