# Multi-Objective Optimization and Performance Evaluation of Rhombic Pin-Fin Microchannel Heat Sinks with Diverse Manifold Configurations

## TL;DR

This paper evaluates and optimizes rhombic pin-fin microchannel heat sinks with different manifold designs to improve thermal performance and temperature uniformity in electronic devices.

## Contribution

The study introduces and compares new manifold microchannel heat sink designs with rhombic pin-fins, achieving significant improvements in thermal performance and uniformity.

## Key findings

- A trapezoidal manifold design shows better thermal performance and temperature uniformity compared to conventional microchannel heat sinks.
- Optimal performance evaluation criterion (PEC) is achieved with specific geometric parameters for microchannel heat sinks with and without manifolds.
- Multi-objective optimization using Kriging models improves thermal performance by 34.05% and temperature uniformity by 18.6%.

## Abstract

In response to the increasingly severe heat dissipation challenges in electronic devices, three types of manifold microchannel heat sinks (MMC) incorporating rhombic pin-fins were proposed. Under the constraint that the maximum temperature of the heat source surface remains below 343.15 K, numerical comparisons with a conventional straight rectangular microchannel heat sinks (MCHS) reveal that the design featuring a trapezoidal manifold exhibits superior comprehensive thermal performance and improved temperature uniformity. Furthermore, the influence of rhombic pin-fin geometry on thermal performance was investigated for both MCHS with and without the trapezoidal manifold under varying mass flow rates. Results show that for the MCHS without a manifold, performance evaluation criterion (PEC) reaches its maximum when the inlet angle of the rhombic pin-fin is 120°, the side length is 0.17 mm, and the pin-fin height is 0.18 mm. In contrast, for the MCHS with the trapezoidal manifold, optimal PEC is achieved at an inlet angle of 110°, a side length of 0.18 mm, and a pin-fin height of 2.2 mm. Additionally, a multi-objective optimization was conducted using the Latin hypercube sampling method. Three objective functions—maximum temperature (Tmax), thermal performance (PEC), and temperature uniformity (σT)—were considered. A total of 150 sample points were used to train Kriging surrogate models for the rhombic pin-fin MCHS with trapezoidal manifold. The optimization results demonstrate a 34.05% enhancement in thermal performance and an 18.6% improvement in temperature uniformity.