Using proper orthogonal decomposition to solve heat transfer process in a flat tube bank fin heat exchanger

Advances in Geo-Energy Research

article

Ye Wang¹  Xintong Xia¹  Yi Wang²  Liangbi Wang³  Wenting Hu¹ 

[1] School of Environmental and Municipal Engineering, Lanzhou Jiaotong University;National Engineering Laboratory for Pipeline Safety, MOE Key Laboratory of Petroleum Engineering, Beijing Key Laboratoryof Urban Oil and Gas Distribution Technology, China University of Petroleum;Key Laboratory of Railway Vehicle Thermal Engineering

关键词: Proper orthogonal decomposition (POD);    POD reduced-order model;    flat tube bank fin heat exchanger;    interpolation method;    reconstructed temperature field;   

DOI  :  10.26804/ager.2017.03.03

来源: Yandy Scientific Press

【 摘 要 】

Proper orthogonal decomposition (POD) reduced-order model can save computing time by reducing the dimension of physical problems and reconstructing physical fields. It is especially suitable for large-scale complex problems in engineering, such as ground heat utilization, sea energy development, mineral exploitation, multiphase flow and flow and heat transfer with complex structure. In this paper, the POD reduced-order model was used to calculate the heat transfer in a flat tube bank fin heat exchanger. The calculating results of the finite volume method (FVM) were adopted as the snapshot samples. Singular value decomposition method was used to decompose the samples to obtain a series of bases and corresponding coefficients on sampling conditions. With these coefficients, interpolation method was used to calculate the coefficients on predicting conditions. And the physical field has been reconstructed using the bases and the interpolated coefficients directly. In the calculation of heat transfer unit of flat tube fin heat exchanger, air-side Reynolds number, transverse tube spacing and the fin spacing were chosen as the variables. The results obtained by the POD method are in good agreement with the results calculated by the FVM. Moreover, the POD reduced-order model presented in this paper is more advantageous in comparison with the FVM in terms of accuracy, suitability, and computational speed.

【 授权许可】

Unknown