石油化工设备技术 ›› 2026, Vol. 47 ›› Issue (4): 13-18,56.doi: 10.3969/j.issn.1006-8805.2026.04.003

• 静设备 • 上一篇    下一篇

基于交叉验证的高斯代理模型的CW型原表面换热通道优化设计

王海波1,马一鸣1,孙海生1,马金伟1,詹勇2,宋豪2,郭凯刚2,魏利平2,3   

  1. 1. 上海蓝滨石化设备有限责任公司,上海 201500;
    2. 西北大学化工学院,陕西 西安 710069;
    3. 西北大学陕北能源先进化工利用技术教育部工程研究中心,陕西 西安 710069
  • 收稿日期:2025-02-27 修回日期:2026-04-26 接受日期:2026-06-28 出版日期:2026-07-17 发布日期:2026-07-17
  • 通讯作者: 魏利平 E-mail:wanghaibo@lanpec.com
  • 作者简介:王海波,男,1994年毕业于甘肃工业大学换热器设计与制造专业,学士,主要从事压力容器设计与新产品开发工作,正高级工程师。
  • 基金资助:
    国家自然科学基金委员会(批准号:22278332)资助的课题;陕西省科学技术厅(批准号:2023-YBGY-317)资助的课题;陕西省教育厅(批准号:23JP170)资助的课题

Design Optimization of CW-Type Plain Surface Heat Transfer Channel Based on Cross-Validation Gaussian Surrogate Model

Wang Haibo1, Ma Yiming1, Sun Haisheng1, Ma Jinwei1, Zhan Yong2, Song Hao2, Guo Kaigang2, Wei Liping2,3   

  1. 1. Shanghai Lanbin Petrochemical Equipment Co., Ltd., Shanghai, 201500;
    2. School of Chemical Engineering, Northwest University, Xi'an, Shaanxi, 710069;
    3. Ministry of Education Engineering Research Center for Shaanbei Energy Advanced Chemical Utilization Technology, Northwest University, Xi'an, Shaanxi, 710069
  • Received:2025-02-27 Revised:2026-04-26 Accepted:2026-06-28 Online:2026-07-17 Published:2026-07-17
  • Contact: Wei Liping E-mail:wanghaibo@lanpec.com

摘要: 交叉波纹型(cross-wavy channels,简称CW型)原表面换热器流动通道结构的优化设计对提高换热能力、降低流动阻力具有重要意义,然而传统设计方法难以全面考虑高维结构参数和操作参数的综合影响,因此迫切需要开发一种新型优化设计方法。文章构建了一个较为完备的数据库,综合考虑了结构参数和操作参数的影响,并开发了高精度的交叉验证的高斯代理模型,结合多目标优化遗传算法进行优化设计。该方法可以有效筛选出高传热系数和低摩擦系数的通道结构和操作参数,可为CW型原表面换热器的设计运行提供重要理论依据。

关键词: 高斯回归, 遗传算法, 双目标优化, 原表面换热器, 数据库

Abstract: Cross-wavy channel (CW-type) flow geometry optimization in plain surface heat exchangers plays a critical role in enhancing heat transfer performance while reducing flow resistance. However, traditional design methods struggle to comprehensively account for the integrated effects of high-dimensional structural and operational parameters; therefore, there is an urgent need to develop a novel optimization methodology. A comprehensive database was constructed, integrating the influences of structural and operational parameters, and a high-precision, cross-validation Gaussian surrogate model was developed to enable optimization design via a multi-objective optimization genetic algorithm. The method can effectively identify channel geometries and operational parameters featuring high heat transfer coefficients and low friction coefficients, providing a critical theoretical basis for the design and operation of CW-type plain surface heat exchangers.

Key words: Gaussian process regression (GPR), genetic algorithm, bi-objective optimization, plain surface heat exchanger, database