Multi-objective optimization of dew point indirect evaporative coolers for data centers

Weichao Yan,Xin Cui,Min Zhao,Xiangzhao Meng,Chuanjun Yang,Yu Zhang,Yilin Liu,Liwen Jin

DOI: https://doi.org/10.1016/j.applthermaleng.2024.122425

IF: 6.4

2024-01-21

Applied Thermal Engineering

Abstract:This study introduces a novel counter-flow dew point indirect evaporative cooler (DPIEC) tailored for data center cooling. A numerical model for DPIEC was developed in COMSOL and validated experimentally, followed by the formulation of regression models using response surface methodology. Eight critical design parameters (five operational and three geometrical) were linked to three performance evaluation indexes (outlet primary air temperature T 1, out , coefficient of performance COP, and cooling capacity per unit volume Q v ). The magnitude of parameter influence on the evaluation indexes was assessed. Employing the three regression models as objective functions, a genetic algorithm, implemented in MATLAB, was adopted to perform multi-objective optimization for DPIEC under four climatic conditions, yielding an array of optimal parameter combinations. The results underscored the robust predictive accuracy of the established regression models, with R 2 and Adeq Precision exceeding 0.96 and 53.7, respectively. The ideal point method achieved design solutions of the DPIEC with T 1, out below 21.6 °C, COP above 61.3, and Q v over 15.5 kW/m 3 . Compared to the original design, the optimized design featured reductions of over 13.84 %, elevations of over 23.50 %, and enhancements of over 104.10 % in these three indexes, respectively. In addition, the confirmation study of the Pareto front proved the reliability of the obtained DPIEC design reference.

energy & fuels,engineering, mechanical,thermodynamics,mechanics