基于机器学习的改型歧管式微通道热沉流动传热特性多目标优化

doi:10.11949/0438-1157.20251212

摘要/Abstract

摘要：

针对超高热通量电子设备的散热需求，开展数值模拟研究了一种改型微通道热沉的流动与传热特性，分析了通道高度和宽度对其综合性能的影响。建立了遗传算法优化的最小二乘支持向量回归预测模型，采用多目标粒子群算法对热沉的几何构型进行优化，并借助灰色关联分析-多准则妥协解排序法与熵权TOPSIS方法筛选出全局最优设计。研究结果表明：相较于本文设定的原始歧管式微通道热沉基准模型（几何参数：h_m1=25 μm、h_m2=25 μm、w_m1=200 μm、w_m2=200 μm），优化方案通过增强二次涡流效应有效降低了峰值温度，使得Nusselt数（Nu）提升20.2%，压降（Δp）下降10.2%。本文所提出的优化策略为高热通量管理提供了新思路，在保证性能显著提升的同时，较大幅度降低了传统参数化方法所需的计算成本，可为新型高效微通道热沉的研发提供理论指导。

关键词: 微通道, 传热, 流动, 深度学习, 优化

Abstract:

To address the heat dissipation requirements of ultra-high heat flux electronic devices, a numerical study was conducted to investigate the flow and heat transfer characteristics of a modified microchannel heat sink, and the influence of channel height and width on its comprehensive performance was analyzed. A prediction model based on the Genetic Algorithm-optimized Least Squares Support Vector Machine was established, the Multi-Objective Particle Swarm Optimization algorithm was employed to optimize the geometric configuration of the heat sink, and the Grey Relational Analysis-VlseKriterijumska Optimizacija Kompromisno Resenje (GRA-VIKOR) and entropy-weighted Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) methods were utilized to screen out the globally optimal design. The results show that: compared with the original MMCHS benchmark model established in this study (geometric parameters: h_m1=25 μm, h_m2=25 μm, w_m1=200 μm, w_m2=200 μm), the optimized scheme effectively reduces the peak temperature by enhancing the secondary vortex effect, resulting in a 20.2% increase in the Nusselt number (Nu) and a 10.2% decrease in the pressure drop (Δp). The optimization strategy proposed in this study provides new insights for high-heat-flux thermal management; while ensuring a significant improvement in performance, it greatly reduces the computational cost required by traditional parametric methods and can offer theoretical guidance for the development of novel and high-efficiency microchannel heat sinks.

Key words: microchannels, heat transfer, flow, deep learning, optimization

中图分类号:

TK124

汤松臻, 张飞杨, 晏稷, 张牧樵, 郭明. 基于机器学习的改型歧管式微通道热沉流动传热特性多目标优化[J]. 化工学报, DOI: 10.11949/0438-1157.20251212.

Songzhen TANG, Feiyang ZHANG, Ji YAN, Muqiao ZHANG, Ming GUO. Multi objective optimization of flow and heat transfer characteristics of modified manifold microchannel heat sink based on machine learning[J]. CIESC Journal, DOI: 10.11949/0438-1157.20251212.

图/表 15

图1 改型MMCHS的结构示意图

Fig. 1 Schematic diagram of the modified MMCHS external structure

图2 本研究MMCHS结构抽取的流体域

Fig. 2 The fluid domain extracted from the MMCHS structure in this study

表1 模型MMCHS参数总结

Table 1 Summary of model MMCHS geometric parameters

名称	数值（μm）	名称	数值（μm）
L	5000	H	525
W	500	h_in	250
w_in	200	h_out	250
w_out	200	h_m1	25
w_m1	200	h_m2	25
w_m2	200	-	-

图3 MMCHS网格无关性分析

Fig. 3 MMCHS grid independence test

图4 微通道内平均温度的实验验证[23, 33]。

Fig. 4 Experimental verification of average temperature within microchannels.

图5 微通道内温度差值的文献验证[12, 32]。

Fig. 5 Literature verification of temperature difference within microchannels.

图6 GA优化LSSVR模型流程图

Fig. 6 Flowchart of GA optimized LSSVR model

图7 数据驱动LSSVR模型超参数优化结果

Fig. 7 Hyperparameter optimization results for data-driven LSSVR model

图8 LSSVR和GA-LSSVR模型测试集预测精度比较

Fig. 8 Comparison of prediction accuracy between LSSVR and GA-LSSVR models based on test data

图9 LSSVR和GA-LSSVR模型训练集预测值与真实值比对

Fig. 9 Comparison between the real value and predicted value of the training set of LSSVR and GA-LSSVR models

图10 MOPSO算法求解所得帕累托前沿解

Fig. 10 The MOPSO algorithm solves for the Pareto front solution

表2 选取的验证点比对情况总结

Table 2 Summary of selected validation points

模型参数		原模型	A点	B点	C点
h_m1 (μm)		25	78.57	29.75	20
h_m2 (μm)		25	95.49	99.59	95.46
w_m1 (μm)		200	122.66	75.94	40.33
w_m2 (μm)		200	219.63	219.88	207.86
Δp	预测值	-	38310	43866	49304
Δp	真实值	48373	38289	43903	49144
误差（％）		-	0.05	0.08	0.33
Nu	预测值	-	5.981	7.025	7.255
Nu	真实值	5.841	5.961	7.018	7.181
误差（％）		-	0.34	0.1	1.03

图11 熵权TOPSIS和GRA-VIKOR的多准则决策用于帕累托前沿解

Fig. 11 Multi-criteria decision making of Entropy Weighted TOPSIS and GRA-VIKOR for Pareto front solution points

图12 MMCHS底部温度分布

Fig. 12 Temperature distribution at the bottom of MMCHS

图13 不同微通道纵切面速度分布

Fig. 13 Velocity distribution of microchannel in various cut planes

参考文献 42

[1]	Pedram M, Nazarian S. Thermal modeling, analysis, and management in VLSI circuits: principles and methods[J]. Proceedings of the IEEE, 2006, 94(8): 1487-1501.
[2]	Agostini B, Fabbri M, Park J E, et al. State of the art of high heat flux cooling technologies[J]. Heat Transfer Engineering, 2007, 28(4): 258-281.
[3]	Tuckerman D B, Pease R F W. High-performance heat sinking for VLSI[J]. IEEE Electron Device Letters, 1981, 2(5): 126-129.
[4]	Sharma C S, Tiwari M K, Zimmermann S, et al. Energy efficient hotspot-targeted embedded liquid cooling of electronics[J]. Applied Energy, 2015, 138: 414-422.
[5]	He Z Q, Yan Y F, Zhang Z E. Thermal management and temperature uniformity enhancement of electronic devices by micro heat sinks: a review[J]. Energy, 2021, 216: 119223.
[6]	Jiang M X, Pan Z L. Optimization of microchannel heat sink based on multi-objective particle swarm optimization algorithm for integrated circuit chips cooling[J]. Numerical Heat Transfer, Part B: Fundamentals, 2025, 86(4): 840-858.
[7]	Li J B, Zhang T Y, Li Z D, et al. Multi-objective parameter optimization design of tapered-type manifold/variable cross-section microchannel heat sink[J]. Applied Thermal Engineering, 2024, 251: 123587.
[8]	Fakhri M, Rezaee B, Pakzad H, et al. Facile, scalable, and low-cost superhydrophobic coating for frictional drag reduction with anti-corrosion property[J]. Tribology International, 2023, 178: 108091.
[9]	Chen H P, Zhang T S, Gao Q, et al. Thermal management enhancement of electronic chips based on novel technologies[J]. Energy, 2025, 316: 134575.
[10]	Chen C W, Li F, Wang X Y, et al. Improvement of flow and heat transfer performance of manifold microchannel with porous fins[J]. Applied Thermal Engineering, 2022, 206: 118129.
[11]	Chen C W, Zhang J Z, Wang X Y, et al. Optimizing hydrothermal performance of manifold microchannels: a study on geometric dimensionless parameters[J]. International Communications in Heat and Mass Transfer, 2025, 164: 108979.
[12]	Drummond K P, Back D, Sinanis M D, et al. A hierarchical manifold microchannel heat sink array for high-heat-flux two-phase cooling of electronics[J]. International Journal of Heat and Mass Transfer, 2018, 117: 319-330.
[13]	Kang H Z, Mei X S, Xu K D, et al. Investigation of heat transfer performance of the manifold microchannel heat sink with different interface configurations[J]. International Communications in Heat and Mass Transfer, 2024, 159: 107807.
[14]	Yang J S, Cheng K Y, Zhang K, et al. Numerical study on thermal and hydraulic performances of a hybrid manifold microchannel with bifurcations for electronics cooling[J]. Applied Thermal Engineering, 2023, 232: 121099.
[15]	Siddiqui M U, Siddiqui O K, Al-Sarkhi A, et al. A novel heat exchanger design procedure for photovoltaic panel cooling application: an analytical and experimental evaluation[J]. Applied Energy, 2019, 239: 41-56.
[16]	Lin Y H, Luo Y, Li W, et al. Single-phase and two-phase flow and heat transfer in microchannel heat sink with various manifold arrangements[J]. International Journal of Heat and Mass Transfer, 2021, 171: 121118.
[17]	Pan Y H, Zhao R, Nian Y L, et al. Numerical study on heat transfer characteristics of a pin–fin staggered manifold microchannel heat sink[J]. Applied Thermal Engineering, 2023, 219: 119436.
[18]	Cheng J P, Xu H S, Tang Z G, et al. Multi-objective optimization of manifold microchannel heat sink with corrugated bottom impacted by nanofluid jet[J]. International Journal of Heat and Mass Transfer, 2023, 201: 123634.
[19]	Harpole G M, Eninger J E. Micro-channel heat exchanger optimization[C]//1991 Proceedings, Seventh IEEE Semiconductor Thermal Measurement and Management Symposium. February 12-14, 1991, Phoenix, AZ, USA. IEEE, 2002: 59-63.
[20]	Rabiee A, Ahmadian-Elmi M, Hajmohammadi M R, et al. Multi-objective optimization of rectangular microchannel heat sink based on entropy generation and hydro-thermal performance using NSGA-II algorithm[J]. International Communications in Heat and Mass Transfer, 2023, 149: 107140.
[21]	Zhao S Q, Yan L M, Goyal V, et al. Artificial neural network-based optimization of baffle geometries for maximized heat transfer efficiency in microchannel heat sinks[J]. Case Studies in Thermal Engineering, 2023, 49: 103331.
[22]	Sui Z G, Lin H S, Sun Q, et al. Multi-objective optimization of efficient liquid cooling-based battery thermal management system using hybrid manifold channels[J]. Applied Energy, 2024, 371: 123766.
[23]	Sang S M, Liu P, Jin Y, et al. Multi-factor impact mechanism of the hydrothermal performance of manifold microchannel heat sinks based on multi-method collaborative optimization[J]. International Communications in Heat and Mass Transfer, 2025, 164: 108931.
[24]	Shanmugam M, Sirisha Maganti L. Multi-objective optimization of parallel microchannel heat sink with inlet/outlet U, I, Z type manifold configuration by RSM and NSGA-II[J]. International Journal of Heat and Mass Transfer, 2023, 201: 123641.
[25]	Fu L Z, Zhao M C, Wu P, et al. Geometric multi-objective optimization of a microchannel-pinfin hybrid heat sink[J]. International Journal of Thermal Sciences, 2025, 211: 109711.
[26]	Hajialibabaei M, Saghir M Z, Dincer I, et al. Optimization of heat dissipation in novel design wavy channel heat sinks for better performance[J]. Energy, 2024, 297: 131155.
[27]	Sarangi S, Bodla K K, Garimella S V, et al. Manifold microchannel heat sink design using optimization under uncertainty[J]. International Journal of Heat and Mass Transfer, 2014, 69: 92-105.
[28]	Chen B Q, Zhang C Y, Xu Y Q, et al. Performance improvement of Z-type manifold microchannel heat sink with novel fin-arrangements[J]. Heat Transfer Engineering, 2025, 46(8): 661-676.
[29]	Yang M, Cao B Y. Multi-objective optimization of a hybrid microchannel heat sink combining manifold concept with secondary channels[J]. Applied Thermal Engineering, 2020, 181: 115592.
[30]	史晓军, 李珊, 魏亚东, 等. 纳米流体矩形微通道热沉结构参数多目标优化[J]. 西安交通大学学报, 2018, 52(5): 56-61, 132.
	Shi X J, Li S, Wei Y D, et al. Multi-objective optimization on the geometrical parameters of a nanofluid-cooled rectangular microchannel heat sink[J]. Journal of Xi'an Jiaotong University, 2018, 52(5): 56-61, 132.
[31]	Zhang J Z, Xu J J, Chen J X, et al. Multi objective optimization of manifold microchannel heat sink with staggered microchannels[J]. International Communications in Heat and Mass Transfer, 2024, 159: 108106.
[32]	Pu X J, Zhao Z C, Sun M K, et al. Numerical study on temperature distribution uniformity and cooling performance of manifold microchannel heat sink[J]. Applied Thermal Engineering, 2024, 237: 121779.
[33]	Zhou J H, Chen X M, Zhao Q, et al. Flow thermohydraulic characterization of hierarchical-manifold microchannel heat sink with uniform flow distribution[J]. Applied Thermal Engineering, 2021, 198: 117510.
[34]	Suykens J A K, Vandewalle J. Least squares support vector machine classifiers[J]. Neural Processing Letters, 1999, 9(3): 293-300.
[35]	Ding S F, Su C Y, Yu J Z. An optimizing BP neural network algorithm based on genetic algorithm[J]. Artificial Intelligence Review, 2011, 36(2): 153-162.
[36]	Xu G, Luo K, Jing G X, et al. On convergence analysis of multi-objective particle swarm optimization algorithm[J]. European Journal of Operational Research, 2020, 286(1): 32–38.
[37]	Liu D S, Tan K C, Huang S Y, et al. On solving multiobjective Bin packing problems using evolutionary particle swarm optimization[J]. European Journal of Operational Research, 2008, 190(2): 357-382.
[38]	Luo Y Q, Yuan X G, Liu Y J. An improved PSO algorithm for solving non-convex NLP/MINLP problems with equality constraints[J]. Computers & Chemical Engineering, 2007, 31(3): 153-162.
[39]	Baranidharan B, Liu J, Mahapatra G S, et al. Group decision on rationalizing disease analysis using novel distance measure on Pythagorean fuzziness[J]. Complex & Intelligent Systems, 2024, 10(3): 4373-4395.
[40]	Tajally A, Babakhani B, Jeyzanibrahimzade E, et al. Sustainable supplier selection and order allocation problem considering the agility and resilience dimensions: a novel multi-stage data-driven decision-making approach[J]. International Journal of Systems Science: Operations & Logistics, 2025, 12(1): 2458756.
[41]	Cheng X, Zhao H, Zhang Y Q, et al. A study on site selection of pumped storage power plants based on C-OWA-AHP and VIKOR-GRA: a case study in China[J]. Journal of Energy Storage, 2023, 72: 108623.
[42]	Theilig K, Vollmer M, Lang W, et al. Multi-criteria decision-making for energy building renovation: Comparing exterior wall structures with the AHP, ANP, utility analysis, and TOPSIS[J]. Building and Environment, 2025, 280: 113075.