Static Reduced-Order Model of a 2D Axisymmetric Counterflow Wet Cooling Tower: Source-Term Modeling and Non-Dimensional Analysis

Wet cooling towers are widely used for low-energy thermal management and ventilation support; however, high-fidelity simulations are computationally expensive for large design studies. This work develops a physics-based static reduced-order model for a two-dimensional axisymmetric counterflow wet cooling tower derived from computational fluid dynamics (CFD) simulations coupled with a user-defined source-term formulation for heat and mass transfer in the fill region. A design of experiments based on advanced Latin hypercube sampling generated 210 configurations, of which 168 valid simulations were retained. The active inputs included tower diameter, fill height, inlet air mass flow rate, inlet air temperature, inlet humidity ratio, inlet water mass flow rate, and inlet water temperature, while the cooling range and evaporation rate were selected as target outputs. Five surrogate families were compared by cross-validation. Kriging was statistically most accurate, with RCV2 values of 0.9999 and 0.9998 for the cooling range and evaporation rate, respectively. Second-order quadratic polynomial models were selected as the engineering reduced order model (ROM) because they capture non-linear boundary curvatures with accuracy, achieving RCV2≥0.9989 and root mean square errors of 0.0426 K and 0.00042 kg/s while preserving an explicit, directly implementable algebraic form. Sensitivity analysis indicated that the inlet water temperature and air mass flow rate are dominant factors within the sampled domain.