Design Of Microchannel Heat Sink Research Articles

A Multivariable Numerical Investigation of Wavy-Based Microchannel Heat Sink Geometry Towards Optimal Thermal Performance

Abstract This research provides a comprehensive multivariable comparative investigation of the effect of various microchannel configurations on their thermal performance. Three-dimensional fluid flow and heat transfer simulations are performed with different arrangements of the channel's width tapering and cross-sectional aspect ratio with an emphasis on the synergistic proven effects of geometrical parameters in innovatory combinations. Results confirm that wavy channels are significantly superior to straight channels in terms of thermal performance due to the creation of secondary flow (Dean Vortices), which improves the processes of advective mixing and, therefore, overall heat transfer characteristics with minimal pumping power penalty. Width-tapering of wavy channels also shows better thermal resistance than untapered wavy channels producing almost 10 percent thermal resistance improvement. The study indicates a significant dependency of thermal performance on the cross-sectional aspect ratio of the channel which suggests that there are ideal tapering and aspect ratio conditions. An innovative wavy-tapered microchannel heat sink has been introduced, featuring an optimal parametric configuration and directionally alternating coolant flow. This design results in additional thermal resistance improvement by 15% and significantly improves substrate temperature distribution uniformity. Overall, the results demonstrate the superior potential of the configuration for high-end electronics cooling tasks. These results provide interpretive insight into microchannel heat sink design and optimization.

The influence of micro-channel shape filled by nanofluid on heat transfer characteristics: Numerical and experimental study

The integration of nanofluid effects and channel shape effects in a heat sink, which exhibits both variable and constant cross-section, has gained significant traction as an efficient cooling method for thermal devices, particularly microelectronic devices. This research presents an experimental and numerical analysis to compare the performance of microchannel heat sink designs (straight, zigzag, wavy, and circular cavities). In addition, the study dealt with the use of pure water and nanofluid (CuO-H2O) with volumetric concentrations of (0.01, 0.02, and 0.03) as coolants. COMSOL Multiphysics was used for numerical analysis to simulate and solve the problem of fluid and heat flow in 3D. The bottom wall of the four microchannels is subjected to a steady heat flux of 170 kW/m2. The simulations were only performed within the laminar domain, encompassing a spectrum of Reynolds numbers ranging from 50 to 150. The influence on the microchannel's wall temperature, thermal resistance, pressure drop, and friction factor is exhibited. According to the findings, the wavy and zigzag microchannel heat sink cooled by nanofluid displays higher performance in terms of heat transmission and dissipation in comparison to the heat sink that was cooled by distilled water. as evidenced by a 12% increase in mussel number at volume concentrations of 0.03%.

Field synergy analysis on thermal-hydraulic behavior of phase change slurry in porous microchannel heat sink with graded porosity configuration

In this study, a novel design of porous microchannel heat sink (MCHS) with graded porosity configuration is proposed in the hope of attaining lower flow and thermal resistances than constant porosity (CP) configuration. The efficacy of new scheme is validated by utilizing a three-dimensional solid-fluid conjugate model based on the finite volume method, which considers the flow and heat transfer of microencapsulated phase change material slurry (MPCMS) inside the porous fins. It is reported that the thermal performance of porous MCHS with constant and graded porosity configurations is significantly better than that of non-porous MCHS. The feasibility of new design is verified by comparing the hydrothermal behavior of multiple graded porosity MCHSs and CP configuration for different fin designs (involving the number of graded equidistant fins N, graded length ratio R, and fin-to-pitch width ratio β) and operating conditions (including Reynolds number Re, mass fraction cm, and inlet velocity uin). The comprehensive performance improvement mechanism of MPCMS in graded porosity microchannels is discussed through the analysis of flow and thermal details, field synergy angle (θ) and field synergy number (Fc), and the overall performance of new design is evaluated by performance evaluation factor (PEF). Results indicate that the six types of porous MCHS with stepwise varying porosity exhibit lower thermal resistance than CP mode, and better thermal behavior and larger PEF can be obtained at N = 2 and β = 0.7. Smaller and larger R should be chosen for stepwise increasing porosity (SIP) and stepwise decreasing porosity (SDP) modes, respectively, to achieve higher heat transfer enhancement while avoiding greater friction losses. The Fc value of porous MCHS with various porosity configurations is about two orders of magnitude less than 1 and shows a monotonically reducing trend with Re, and the synergistic effect deteriorates at high Re. Compared to the CP mode, a low-θ “drift” toward the low porosity region of porous fin occurs within the graded porosity MCHS, which decreases and increases in the downstream and upstream directions of the “drift”, respectively. Among the novel MCHS designs, the z-directional SIP mode acquires the best overall performance due to the excellent synergy between the velocity and temperature fields, and its PEF consistently exceeds 1.13 and reaches up to 1.33. Generally speaking, the fin porosity of graded porosity MCHS near the channel cavity, channel top, and channel inlet should be greater for superior comprehensive performance.

Study on the efficient heat transfer mechanism of microchannel pin-fin arrays under low pumping power

The efficient heat transfer with low pressure loss has long been a critical challenge in the microchannel heat sinks (MCHS) design. Further study is required for the quantitative analysis of the longitudinal vortex heat transfer enhancement mechanism. In this research, a microchannel heat transfer experiment and corresponding simulations are conducted. Based on the common Reynolds number range in engineering applications of MCHS, the underlying causes of heat transfer enhancement and pressure loss increase in MCHS with respect to Reynolds numbers is clarified. The results show that when the Reynolds number is 400, the cylindrical wake reduces the synergistic angle between the velocity and the temperature gradient at the streamwise gap, which greatly enhances the heat transfer performance of MCHS. In addition, when the Reynolds number is 900, the early separation of the cylindrical wake vortex leads to a significant reflux near the center of the spanwise gap, which significantly increases the pressure loss. A Different Reynolds number Efficiency Evaluation Criterion (DEEC) is proposed for evaluating the performance of MCHS. It is found that the flow state of cylindrical wake with Reynolds number of 400 has the best heat transfer efficiency. The concept of introducing longitudinal vortices into the pin–fin array can serve as a key guideline in the heat transfer enhancement optimization design of MCHS.

On the reticulate pattern and heat transfer performance of the topologically optimized microchannel heat sink

With the increasing integration of electronic devices, the rapid heat dissipation has become a key challenge for performance improvement. In this work, a 2D topology optimization for the channel pattern of a microchannel heat sink is conducted, and a novel reticulate pattern of optimized design of microchannel heat sink is found based on the theoretical and numerical analysis. This pattern has tertiary ranked reticulate channels resulting from the topology optimization, which are highly related to the convection heat transfer of fluid and the conduction in solid. The reticulate channels could significantly optimize the total number of channels in the heat sink, and accordingly regulate the flow distribution among the channels. The profiles of the temperature gradients and velocity exhibit high similarity in all reticulate channels, and lead to a synergistic enhancement of heat transfer. A full scale 3D simulation shows that the reticulate microchannel heat sink has better heat transfer performance than the other optimized heat sinks. With pressure drop of 1 kPa, the heat flux for the reticulate microchannel heat sink can reach up to 405 W/cm2, and its performance evaluation criteria (PEC) can reach up to 3.7.

Optimal Design of Microchannel Heat Sink with Staggered Fan-Shaped Cavities and Ribs

In the present study, a three-dimensional numerical simulation is conducted to study the characteristics of fluid flow and heat transfer in a newly designed microchannel heat sink with fan-shaped cavities and ribs (MC-FCFR). The Reynolds number ranges from 147 to 736. A straight rectangular microchannel (MC-RC) is used as the baseline. The geometric parameters include the relative spacing θ, relative width α, and relative height β of the fan-shaped cavities and ribs. The overall performance index (OPI) of the MC-FCFR is evaluated in terms of the friction factor and Nusselt number. The OPI for the MC-FCFR with θ=3.56, α=1.63, and β=0.5 reaches 1.70.

Jet Impingement Cooling of a Microchannel Heat Sink with Microgroove

The present study works on opportunities in improving the design of micro-channel heatsink (MCHS). The present study focuses on investigating numerically the effects of adding grooves at the MCHS which subject to jet impingement cooling. Commercial software ANSYS Fluent is used and realizable k-epsilon model is adopted to conduct a parametric study on the width and depth of rectangular longitudinal grooves at a constant heat flux of 250 W/cm2 applied at the base of MCHS. Two type of channel designs with grooves i.e. center-groove and side-groove were created and investigated numerically. Results show that addition of grooves generally give improvements in cooling performance and reducing the pressure drop. Some designs of side-grooved channels and center-grooved channels improve the temperature uniformity. The size of the groove affects the flow within the grooves and therefore affect the cooling performance.

Thermal-hydraulic characterization of manifold microchannel heat sink with diverging channels and uniform heating

This study presents the design and experimentation of a manifold microchannel heat sink with diverging channels, employing ammonia as the working fluid to evaluate its thermal–hydraulic performance under uniform heating conditions. The impact of heat flux, inlet subcooling, saturation temperature, and outlet thermodynamic quality were investigated. Under single-phase conditions, the heat sink exhibited low sensitivity to variations in heat flux with constant thermal resistance and pressure drop. However, with the onset of boiling, the heat transfer capability and pressure drop rapidly increased. A lower inlet subcooling boosted boiling inside the heat sink, leading to decreased thermal resistance and increased pressure drop. Under near-saturated inlet conditions, a higher saturation temperature enhanced heat transfer owing to smaller bubble nucleation and detachment size. The outlet thermodynamic quality for engineering applications should be moderately low because it provides a balance between high heat transfer capability and low pressure drop. Moreover, the results demonstrated excellent thermal–hydraulic performance, with a heat flux of up to 636.9 W/cm2 efficiently dissipated under uniform heating conditions while maintaining the heating surface temperature and pressure drop below 79.3 °C and 150.3 kPa, respectively.

Numerical study on infrared detectors cooling by multi-stage thermoelectric cooler combined with microchannel heat sink

Lightweight, noiseless, and constant low-temperature refrigeration is crucial for cooled infrared (IR) detectors, especially short-wave-infrared (SWIR) detectors. However, some SWIR detectors, such as HgCdTe photovoltaic detectors are working below 210 K. For this application, multi-stage thermoelectric coolers (TEC) are preferable to single-stage thermoelectric coolers because they can generate larger temperature differences. This paper presents a refrigeration design that integrates SWIR and TEC into a vacuum Dewar package and combines them with a microchannel heat sink (MHS) to improve the detection performance and better responsiveness of SWIR. Two different refrigeration models, SWIR + TEC + AHS (air-cooled heat sink) and SWIR + TEC + MHS, were established to evaluate the refrigeration performance of the proposed SWIR + TEC + MHS scheme using the finite element method. The influences of multi-stage TEC leg length and current, inlet water temperature, and water velocity on the refrigeration performance of the TEC are analyzed. The results show that the SWIR + TEC + MHS design scheme has the best comprehensive performance and minimum volume. The influences of all the factors are obvious while the current has the most important effect on the performance. The temperature of SWIR reached 201.69 K at standard working conditions. It shows that the cooling scheme can not only meet the needs of flexible deployment of SWIR but also provide better thermal security. The proposed cooling design has an excellent prospect for short-wave-infrared detectors.

Numerical analysis of various shapes of lozenge pin-fins in microchannel heat sink

Abstract Higher density heat flux is the major cause of damage to the electronic component; therefore, cooling such components are of the utmost importance to operate in a safe zone and to increase their life. For this purpose, Microchannel heat sinks (MHSs) are among the most practical methods for dissipating unwanted heat. In this regard, the novel lozenge-shaped pin-fins in the flow passage of the microchannel heat sink (MHS) have been designed and proposed to achieve higher cooling performance. Aspect ratios (λ = 0.30, 0.39, 0.52, 0.69, 1.00) of several lozenge-shaped pin-fins have been used into the design of MHS to investigate their impact on heat transmission and fluid flow characteristics. A three-dimensional model of MHS with a lozenge-shaped has been generated and simulated numerically in the following range of Reynolds numbers, starting from 100 to 900. Heat transmission and flow characteristics have been presented and discussed in detail. It has been found that introducing lozenge-shaped pin-fins in MHS has greatly improved cooling performance. The highest improvement in Nusselt number has been observed when aspect ratio (λ) of lozenge-shaped pin-fins was 1.00. The Nusselt number have been varied in the following ranges of 6.96–12.34, 6.97–12.72, 7.01–13.62, 7.09–14.43, and 7.12–15.26 at λ = 0.30, λ = 0.39, λ = 0.52, λ = 0.69, and λ = 1.0, respectively. In addition, a study of the thermohydraulic performance of the proposed lozenge-shaped pin-fins in the MHS found that this design is an effective means of lowering operating temperature.

Analysis and optimization of thermo-hydraulic characteristics of phase change slurry in wavy MCHS with porous fins based on RSM and desirability methods

In this study, the hydromechanical and diathermanous characteristics of microencapsulated phase change material (MPCM) slurry in wavy microchannel heat sink (MCHS) with porous fins are investigated by using two–phase mixture and Dacry–Brinkman–Forchheimer models. Based on the finite volume method (FVM) and response surface methodology (RSM), 54 cases considering 6 structural parameters are carried out. The correlations of Nusselt number (Nu) and friction factor (f) in straight and wavy microchannels with phase change slurry and porous fin designs are creatively proposed and unified. The maximum deviations between the actual and predicted values of Nu and f are 13.24% and 14.21%, respectively. In addition, the effects of radiator's parameters on Nu and f are emphatically discussed and analyzed by using three-dimensional surface diagram. It reveals that within the range of parameters tested, wavy amplitude (A) has the most obvious effect on responses, while channel aspect ratio (α) has the least impact on them. Moreover, the interactive effects between the five structural parameters are significant and complex, as one factor influences the impact of another on responses. Finally, the optimal parameters are obtained by desirability method with the goal of maximum Nu and minimum f. The results indicate that the overall performance of optimized configuration improved by 48.6% over the initial design. Our results may provide helpful guidance for the optimization design of MCHS, as significant improvement in heat transfer coefficient is expected to save substantial operating costs.

Experimental and numerical investigation of a single-phase microchannel flow under axially non-uniform heat flux

The design of microchannel based heat sinks for advanced electronic component continues to receive research interest, especially with the ever-increasing density of advanced chips with concomitant increase in heat generation. One aspect of microchannel heat sink design objectives is the optimum employment of non-uniform heat dissipation through the channel length. Two dominant schools in the performance optimization of non-uniform heat sinks are: 1) entropy minimization and 2) thermal resistance minimization. These two approaches usually result in contradictory conclusions: one recommending the use of increasing heat flux, the other recommending a decreasing heat flux. The current study seeks to explain this discrepancy by experimentally and numerically studying the heat transfer mechanisms in a single microchannel tube under the effect of non-uniform heat flux. This study presents an experimental framework capable of generating a wide range of pre-specified heat flux profiles over a single microchannel, mimicking actual heat flux profiles observed in thermo-electronic devices. We employ three flux profiles: 1) hotspots located at different locations with uniform background heat flux, 2) linearly ascending heat flux, and 3) linearly descending heat flux. The results show good agreement between numerical simulations and experimental measurements and provide insights into the microchannel tube's fundamental heat transfer mechanisms. Based on these insights, the study provides design guidelines to enhance microchannel heat sink performance under non-uniform axial heat fluxes. Finally, the discrepancy between entropy and heat resistance minimization approaches is explained.

Extremely high heat flux dissipation and hotspots removal with nature-inspired single-phase microchannel heat sink designs

A novel pyramid thermal dissipation unity with conduction–convection coupling thermal dissipation and nature-inspired hotspot removal channel heat sinks is proposed to degrade and dissipate ultra-high heat flux with the order of 103 W cm−2 by using single-phase coolant. A 3D conjugate thermal and flow numerical model is used to validate the feasibility of the proposed ultra-high heat flux dissipation method. The pyramid thermal dissipation unit consists of a spiral tube embedded with etched lotus leaf vein or snowflake-shaped channels. This configuration enables the dissipation of heat fluxes ranging from 1000 to 1500 W cm−2 while maintaining safe operating temperatures for the chip using a single-phase coolant. By utilizing high thermal conductivity materials, the proposed dissipation unit achieves chip working temperatures of 89.89 °C for 1500 W cm−2 and 66.58 °C for 1000 W cm−2. Furthermore, the new design allows for a reduction in the minimum required thermal conductivity for materials to 700 W m−1 K−1 at a pump power of 7.05 W, while still operating within the chip's safe temperature limit of ≤ 120 °C. This expanded range of thermal conductivity values provides more options for selecting suitable materials to fabricate the dissipation unit.

Multi-objective topology optimization and thermal performance of liquid-cooled microchannel heat sinks with pin fins

The optimized structures for the liquid-cooled microchannel heat sinks with different pin-fin arrays were designed via using the topology optimization method for the better performance, including minimization of the flow energy dissipation and the average temperature of the bottom surface, aiming to facilitate the more efficient design of microchannel heat sink for electronic chips’ cooling. The heat transfer and flowing performance was simulated with steady-state incompressible Navier-Stokes equation and the energy equation. Based on the rational approximation of material properties (RAMP) method, the mathematical optimization model for heat sink was established according to the multi-objective function. The Galerkin finite element method (GFEM) was employed to calculate the flow and heat transfer system. Furthermore, the globally convergent method of moving asymptotes (GCMMA) was adopted to deal with the mathematical optimization problem as the optimization method. The entropy generation analysis and comprehensive heat transfer factor were employed to evaluate the performance of the conjugate heat transfer problem of 3D microchannel heat sinks with various pin-fin structures. The calculation results reveal that the multi-objective optimal structure has good heat transfer performance, and the temperature distribution of the optimized structure is more uniform. As the same pumping power, the comprehensive heat transfer efficiency of the optimized inline and staggered pin-fin arrays increases by 31.20% and 32.18% compared to the smooth rectangular channel. Compared with the cylindrical pin-fin configuration with the same heat transfer area, the entropy generation rates of the optimized inline and staggered pin-fin configurations decrease by 4.67% and 4.25%, respectively. The results can provide theoretical basis and application support for the electronic chip cooling technology.

Local hydrothermal characteristics and temperature uniformity improvement of microchannel heat sink with non-uniformly distributed grooves

In order to improve the temperature uniformity of bottom walls, a novel design of microchannel heat sink with From-Sparse-To-Dense embowed grooves (MC-FSTD) was proposed as well as a comparative one with From-Dense-To-Sparse embowed grooves (MC-FDTS). The local hydrothermal characteristics and temperature distribution variance of bottom wall of microchannel heat sinks were numerically investigated for Reynolds number (Re) ranging from 100 to 800. The results showed that the denser grooves led to lower flow velocity, lower local friction factor and higher local pressure as well as the lower bottom wall temperature (Tw) and higher local Nusselt number (Nuz). The local Nusselt number and the bottom wall temperature of MC-FSTD were 2.17 times higher and 9.1K lower than that of MC-SC at z = 10 mm, respectively. MC-FSTD also had the most uniform temperature of the bottom wall and the corresponding temperature variance was 6.42 times lower than that of MC-SC at Re = 795.38. The heat transfer entropy generation rate and the total entropy generation rate of MC-FSTD were also the lowest with an acceptable friction entropy generation rate. Finally, the effects of the decreasing magnitude (Li) of the distance between two adjacent grooves on the local hydrothermal characteristics and temperature uniformity of MC-FSTD were explored and the total entropy generation rate was further reduced to 5.22 × 10−4 W/K at Li = 0.015 mm and Re = 795.38.

Experimental study of topology optimized, additively manufactured microchannel heat sinks designed using a homogenization approach

Topology optimization generates complex geometry heat sink designs having intricate features well-suited for fabrication by additive manufacturing. In particular, a homogenization approach to topology optimization creates microchannel heat sink designs wherein the material porosity distribution corresponds to microstructures of varying dimensions, thereby eliminating the need to penalize the porosities to achieve binarized solid/liquid designs. This approach is inherently able to take into consideration the capabilities and limitations of available additive manufacturing processes as the types of microstructures can be user-defined. The current study is the first to fabricate and experimentally test microchannel heat sink designs that are topology optimized using the homogenization approach. To this end, a multi-objective optimization is performed to generate a series of Pareto optimal designs that minimize pressure drop and thermal resistance. The effect of the grid resolution is investigated. The resulting topology optimized designs are found to have different geometries and performances with different grid cell sizes because this impacts the physical dimensions of the microstructures. A series of pin fin arrays with different grid cell sizes and fin thicknesses are fabricated using Direct Metal Laser Sintering of AlSi10Mg. The additively manufactured surfaces have high roughness which creates connections between pin fins with small gap sizes and creates tortuous flow paths. The smallest pin fin that consistently survived the fabrication process is ∼0.25 mm thick and hence series of topology optimized microchannel heat sinks with a 0.5 mm grid cell size are additively manufactured. The pressure drop and thermal resistance of the fabricated heat sinks are experimentally characterized, and the measured results lie on the Pareto optimality curve predicted by the topology optimization algorithm. This work demonstrates that the homogenization approach to topology optimization generates high-performance microchannel heat sink designs that can be achieved using available additive manufacturing processes.

Topological optimization of flow-shifting microchannel heat sinks

Efficient thermal management is required for electronic devices that generate multiple different spatially nonuniform thermal workloads during operation. Conventional heat sinks route coolant to all potential heat-generating regions of an electronic component which may result in unnecessarily broad dispersion of the coolant to regions that are sometimes inactive. To fully utilize flow for multiple potential thermal workloads, a “flow-shifting” design approach is proposed. A flow-shifting heat sink has multiple inlets, where the flow path from each inlet is optimized for cooling of a specific workload or heat map. Only the inlet corresponding to the active operating workload is open at a given time, while the rest of the inlets are kept closed until the operating workload changes, allowing a majority of the flow to be properly utilized for cooling the active heat map. This has the potential to allow many different optimal heat sinks to be simultaneously encoded into a single structure but brings forth a complex design problem to optimize the internal fluid flow pathways. A multi-objective topology optimization algorithm is implemented in this work for flow-shifting heat sink design generation. Microchannel heat sink designs are generated for a demonstration case involving two workloads. Designs optimized for flow-shifting between two inlets are compared to a benchmark that is optimized with only a single inlet. Pareto optimality curves, as well as the associated heat sink designs, are created that weigh between the thermal resistances of the two workloads. The flow-shifting designs are predicted to have lower thermal resistance under both workloads compared with every benchmark heat sink design generated along the Pareto optimality curve. Investigation of two designs selected from the Pareto curve showed that the flow-shifting heat sink achieved 10.7% and 6.8% lower thermal resistance at the two workloads relative to the benchmark. The flow-shifting approach is superior as it utilizes more flow rate for cooling the active heated region. The flow-shifting approach thereby allows for fixed heat sink structures that can be topologically optimized for many different possible heat maps and actively respond to changes in workload.

Entropy generation analysis on thermo-hydraulic characteristics of microencapsulated phase change slurry in wavy microchannel with porous fins

• Combined design enables smaller temperature gradient and more uniform temperature. • Larger h m and lower ΔP can be concurrently gained in wavy MCHS with porous fins. • Thermal entropy is three orders of magnitude larger than the frictional entropy. • Thermal entropy increases with heat flux but the frictional entropy is opposite. • Lower irreversible losses can be achieved at the cost of higher power consumption. Recent advancement in micro/nano-scale electronic systems, the need for efficient heat dissipation has become more rigorous, and traditional thermal management solutions are facing enormous challenges. In this study, a novel combined design of wavy microchannel heat sink (MCHS) with porous fins and microencapsulated phase change material (MPCM) suspension is developed and numerically studied to achieve a balance between flow and heat transfer behaviors. The equivalent heat capacity method is adopted to deal with the phase change process of microcapsule particles in steady and laminar states, and the Brinkman–Darcy–Forchheimer model based on the volume-averaged method is employed to characterize the fluid flow in porous fins. A three-dimensional solid-fluid conjugate model is established based on the finite-volume method to investigate the effects of different coolant types (water and MPCM suspension), geometric parameters (wavy amplitude, wavelength, and channel width ratio), and working conditions (inlet flow velocity, heat flux, and slurry concentration) on the thermo-hydraulic properties and entropy generation of wavy MCHS. The results reveal that the combination of porous fins and MPCM suspension presents smaller temperature gradient and more uniform temperature distribution than conventional design. Larger heat transfer coefficient and lower pressure drop can be simultaneously obtained in wavy MCHS with porous fins than in solid fin configuration. The thermal entropy generation is three orders of magnitude larger than the frictional entropy generation, and the Bejan number is close to 1 for different microchannel configurations. Thermal entropy generation increases with increasing heat flux but the frictional entropy generation decreases. Lower irreversible losses can be achieved at the expense of greater pumping power consumption caused by increasing inlet velocity and suspension concentration. The superior effect of wavy MCHS with porous fins using MPCM suspension as coolant has been observed in all cases.

Machine learning based surrogate models for microchannel heat sink optimization

Microchannel heat sinks are an efficient cooling method for semiconductor packages. However, to properly cool increasingly complex and thermally dense circuits, microchannel designs should be improved and expanded on. In this paper, microchannel designs with secondary channels and with ribs are investigated using computational fluid dynamics and are coupled with a multi-objective optimization algorithm to determine and propose optimal solutions based on observed thermal resistance and pumping power. A workflow that combines Latin hypercube sampling, machine learning-based surrogate modeling and multi-objective optimization is proposed. Random forests, gradient boosting algorithms and neural networks were considered during the search for the best surrogate. We demonstrated that tuned neural networks can make accurate predictions and be used to create an acceptable surrogate model. Optimized solutions show a negligible difference in overall performance when compared to the conventional optimization approach. Additionally, solutions are calculated in one-fifth of the original time. Generated designs attain temperatures that are lower by more than 10% under the same pressure limits as a convectional microchannel design. When limited by temperature, pressure drops are reduced by more than 25%. Finally, the influence of each design variable on the thermal resistance and pumping power was investigated by employing the SHapley Additive exPlanations technique. Overall, we have demonstrated that the proposed framework has merit and can be used as a viable methodology in microchannel heat sink design optimization.

Thermal performance enhancement and optimization of double-layer microchannel heat sink with intermediate perforated rectangular fins

In the present work, three-dimensional numerical investigation has been carried out to comprehend the thermohydraulic performance of novel design of double-layer microchannel heat sinks (DL MCHS). Conventional DL MCHS is modified by incorporating intermediate rectangular fin with holes of four distinct shapes namely triangular, rectangular, circular and square having equal perimeter. Series of numerical simulations have been performed for the Reynolds number (Re) ranging from 100 to 400, and uniform heat flux varied from 500 to 2000 kW/m2. Single-phase liquid water with varying thermophysical properties is used as cooling medium. Predicted numerical results confirm that compared to conventional configuration; modified DL MCHS has significantly augmented the heat transfer rate causing average Nusselt number to be increased by ≈ 45–60%. But they also encounter higher pressure drop due to massive flow restrictions prevailing in such configurations. Nevertheless, overall thermal performance of the modified DL MCHS is better than the conventional design. Among all the considered cases, heat sinks with circular and triangular holes consistently show superior thermal performance compared to square and rectangular holes, but it is very marginal of the order of ≈ 2–5%. Analysis of coolant flow behavior reveals that intermediate fins with holes facilitate better fluid mixing and provides staggered flow passages in addition to thinner boundary layer which keeps on re-evolving. All these favourable flow behaviors led to improve the heat transfer rate in modified DL MCHS. Furthermore, optimization study of the heat sink with circular holes identify that a configuration with hole radius (Rh) = 0.286 mm and number of holes (Nh) = 15 is superior which exhibits maximum heat transfer rate ≈51–64% higher than conventional DL MCHS.