Impact of pin-fin shape on thermal-hydraulic performance under condensation of flat-tube heat exchangers: numerical and machine learning approaches

Flat-tube heat exchangers are widely used in HVAC, automotive, and electronics cooling systems, yet the absence of extended surfaces in conventional designs limits air-side heat transfer, especially under humid conditions where condensate films increase drag and suppress convection. Pin fins, although well established in heat sinks and internal flows, have not been systematically studied for flat-tube exchangers under humid conditions. This study presents a high-fidelity multiphase numerical investigation of eighteen pin–fin geometries (circular, square, and triangular in inline and staggered arrangement) across Re = 250–4000, using a coupled VOF and species transport model available in ANSYS Fluent. Flow development was analyzed through temperature, pressure, and velocity contours and streamlines, while thermal–hydraulic performance was evaluated with Colburn “ j ” and Fanning factor “ f ”, surface goodness, compactness-adjusted metrics, and fixed pumping-power per unit area criteria. Results showed that staggered layouts significantly improved thermal effectiveness compared to inline. Heat transfer for staggered geometries increased by 65.2 %, 93 % and 76.4 %, while friction losses increased by 39.0 %, 93 % and 40.8 %, for circular, square, and triangular pin fin shapes, respectively. Compactness-adjusted rankings consistently identified GP₁₂, GP₅, GP₁₇, and GP₁₈ as top performers, balancing vortex-driven enhancement with acceptable drag. To generalize predictions, nine algorithms for “ f ” and eighteen for “ j ” were trained; artificial neural networks achieved accuracy above 95 %, outperforming regression methods. This work delivers the compactness-adjusted, pumping-power-normalized ranking of wet air-side pin–fin surfaces. The presented regression correlation can be used to predict the pin fin performance in terms of “ j ” and “ f ” factor.