EUVIMEDSkopri EUVIMED
Open access · CC BY-NC 4.0

Multi-Objective Optimization of Manifold Microchannel Heat Sink Structures Using a Gaussian Process Surrogate and NSGA-II

Weifu Huang, Wenyan Xue

Academic Journal of Applied Sciences · 2026-11-20 · 4 (1) · 6-14

DOI: 10.54097/fdaate22 ↗

At a glance

Journal article in Academic Journal of Applied Sciences, published online on 2026-11-20. Written by Weifu Huang and 1 co-authors. 10 references. The full text is freely available (open access, CC BY-NC 4.0).

AstrattDettalji biblijografiċiCiteRelated studies

Abstract (original language)

Thermal resistance, pressure drop and wall-temperature uniformity of a manifold microchannel heat sink are controlled by the same three structural parameters; they conflict, the mapping is strongly nonlinear, and a study of this kind rarely holds more than a few dozen samples, too few for a search that calls a solver at every step. The paper develops a general procedure from such a sample set to a design decision. Semi-empirical forms first fix the physical magnitudes of the three indices, and the calibration decides which of them an analytical expression can carry: pressure drop can (coefficient of determination 0.8994), thermal resistance cannot (0.6229), because two huge fitted coefficients cancel and leave the parameters unidentifiable. Surrogates are then compared on the same samples, Gaussian process regression against two tree ensembles, to establish the accuracy ceiling inside the sampled domain and to cut the cost of one objective evaluation from a simulation to a matrix product. NSGA-II searches the non-dominated front on that surrogate, TOPSIS returns an equal-weight compromise, and a Latin hypercube sweep over the weight simplex tests the stability of the recommendation against preference change. The procedure yields a design with no pin fins, a depth ratio of 4.48 and four rows, at 36.4% of the pressure drop of the thermal-resistance optimum for 4.4% more resistance, and that design stays optimal under 71 of 100 random weight vectors. The procedure does not depend on the specific geometry and transfers to other small-sample structural design problems of the same parameter–performance type.

Cite

APA

Huang, W., & Xue, W. (2026). Multi-Objective Optimization of Manifold Microchannel Heat Sink Structures Using a Gaussian Process Surrogate and NSGA-II. Academic Journal of Applied Sciences, 4(1), 6-14. https://doi.org/10.54097/fdaate22

Vancouver

Huang W, Xue W. Multi-Objective Optimization of Manifold Microchannel Heat Sink Structures Using a Gaussian Process Surrogate and NSGA-II. Academic Journal of Applied Sciences. 2026;4(1):6-14. doi:10.54097/fdaate22

BibTeX
@article{Huang2026,
  author = {Huang, Weifu and Xue, Wenyan},
  title = {Multi-Objective Optimization of Manifold Microchannel Heat Sink Structures Using a Gaussian Process Surrogate and NSGA-II},
  journal = {Academic Journal of Applied Sciences},
  year = {2026},
  volume = {4},
  number = {1},
  pages = {6--14},
  doi = {10.54097/fdaate22}
}
Start for free →