Performance prediction and optimization of cross-flow indirect evaporative cooler by regression model based on response surface methodology

Abstract

In recent years, indirect evaporative cooling has rapidly developed with high-accuracy numerical models. As the application of this technology expands from hot-arid areas to hot-humid regions, there is still a lack of regression models of the cross-flow indirect evaporative cooler (IEC) that can be used in different climate regions. Regression models can not only improve prediction efficiency but also be helpful for engineering design. In this study, the regression models of the cross-flow IEC were established based on the response surface methodology (RSM). Eight essential factors, including the inlet air properties, geometric size, and operating parameters, were determined as the input factors, while five indicators were selected as the output responses. The central composite design was employed to generate the matrix for the RSM-based model, and the matrix response data were obtained from an established numerical IEC model validated by the experimental results. The effects of the single and interactive factors are analyzed for each response. Furthermore, the developed models are evaluated by comparing the anticipated results with the on-site measurement data in a real project, and then the multi-objective optimization is conducted for the prediction of IEC performances in five typical cities of China. In summary, the regression models can forecast the cross-flow IEC in a more straightforward approach, which may also assist the design and optimization.