Study on the prediction and optimization of flow mal-distribution in printed circuit heat exchangers based on machine learning

Jianxin Qiao, Shuangqing Chen, Shenghui Liu, Junjie Fei, Xiaoliang Zhu, Minyun Liu, Houjun Gong, Ruohan Zheng and Yanping Huang

Energy, 2024, vol. 313, issue C

Abstract: Printed circuit heat exchanger has become a new choice for the recuperators and intermediate heat exchangers of supercritical carbon dioxide nuclear power system due to its advanced power density and ability to operate under high temperature and high pressure conditions. The flow mal-distribution among the channels of PCHE significantly affects its heat transfer efficiency. This study focuses on the prediction and optimization of flow mal-distribution in the PCHE. A simulation framework of manifolds-flow channel system which consists of virtual channels that connecting the upstream and downstream manifolds is established to analyze the variations of flow characteristics in the PCHE as geometric parameters change. The study indicates that with channel array dimension, channel length and inlet header diameter increase, the flow distribution improves, but the flow resistance increases. When the channel array dimension is 7, within the range of design variables, as the flow non-uniformity decreases from 0.314 to 0.021, the total pressure drop increases from 1256.9Pa to 6950.3Pa. Then, a prediction scheme for the flow mal-distribution in the PCHE based on dimensionless number analysis and machine learning methods is established, resulting in a neural network response surface that reflects flow characteristics and velocity distribution consisting of 162 design points. Finally, to reduce flow mal-distribution in the heat exchanger design process, the response surface model is used to generate flow characteristics for each design point, and the combination of NSGA-II and TOPSIS method is applied to get pareto front and sequence the performance of each design point, leading to a heat exchanger subcomponent design scheme. Compared to the design point on the Pareto front with the maximum flow non-uniformity, the application of TOPSIS method enables the PCHE to reduce flow non-uniformity by 93.39 %, while the pressure drop only increases by 32.67 %.

Keywords: Printed circuit heat exchanger; Flow mal-distribution; Machine learning; Hydraulic optimization (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544224038076
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:313:y:2024:i:c:s0360544224038076

DOI: 10.1016/j.energy.2024.134029

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().