The enhancement of pool boiling heat transfer by copper-particle surface coatings is experimentally investigated, using the wetting dielectric fluid FC-72. In one technique, loose copper particles are placed on a heated copper surface to provide additional vapor nucleation sites in the cavities formed at particlesurface and particle-particle contact points, thereby enhancing boiling performance over a polished surface. This 'free-particle' technique is benchmarked against the more traditional technique of sintering a fixed layer of copper particles to the surface to enhance boiling heat transfer performance. The effect of particle size on the heat transfer performance is studied for particle diameters ranging from 45 µm to 1000 µm at a constant coating layer thickness-to-particle diameter ratio of approximately 4. The parametric trends in the boiling curve and the critical heat flux are compared between the two techniques, and the dominant boiling mechanisms influencing these trends are compared and contrasted. High-speed visualizations are performed to qualitatively assess the boiling patterns and bubble departure size/distribution, and thus corroborate the trends observed in the boiling curves. The measured wall superheat is significantly lower with a sintered coating compared to the free-particle layer for any given particle size and heat flux. Performance trends with respect to particle size, however, are remarkably similar for both enhancement techniques, and an optimum particle size of ~100 µm is identified for both free particles and sintered coatings. The free-particle technique is shown to offer a straightforward method to screen the boiling enhancement trends expected from different particulate layer compositions that are intended to be subsequently fabricated by sintering.

A three-dimensional numerical model is developed and validated to study the effect of geometric parameters such as microchannel depth and width, manifold depth, and manifold inlet and outlet lengths on the performance of a manifold microchannel (MMC) heat sink. The manifold arrangement used to distribute the flow through alternating inlet and outlet pairs greatly reduces the pressure drop incurred in conventional fluid supply arrangements due to its shorter flow paths, while simultaneously enhancing the heat transfer coefficient by limiting the growth of thermal boundary layers. The computational analysis is performed on a simple unit-cell model to obtain an optimized design for uniform thermal boundary conditions, as well as on a porous-medium model to obtain a complete system-level analysis of multiple microchannels across one manifold. The porous-medium approach can be further modified to analyze the performance under asymmetrical heating conditions. Along with conventional deterministic optimization, a probabilistic optimization study is performed to identify the optimal geometric design parameters that maximize heat transfer coefficient while minimizing pressure drop for an MMC heat sink. In the presence of uncertainties in the geometric and operating parameters of the system, this probabilistic optimization approach yields a design that is robust and reliable, in addition to being optimal. Such an optimization analysis provides a quantitative estimate of the allowable uncertainty in input parameters for acceptable uncertainties in the relevant output parameters. The approach also yields information such as the local and global sensitivities which are used to identify microchannel width and manifold inlet length as the critical input parameters to which the outputs are most sensitive. The deterministic analysis shows that the heat transfer performance of the MMC heat sink is optimal at a manifold inlet to outlet length ratio of 3. A comparison between the deterministic and probabilistic optimization approaches is presented for the unit-cell model. A probabilistic optimization study is performed for the porous-medium model and the results thus obtained are compared with those of the unit-cell model for a uniform heat flux distribution.

Immersion cooling strategies often employ surface enhancements to improve the pool boiling heat transfer performance. Sintered particle/powder coatings have been commonly used on smooth surfaces to reduce the wall superheat and increase the critical heat flux (CHF). However, there is no unified understanding of the role of coating characteristics on pool boiling heat transfer enhancement. The morphology and size of the particles affect the pore geometry, permeability, thermal conductivity, and other characteristics of the sintered coating. In turn, these characteristics impact the heat transfer coefficient and CHF during boiling. In this study, pool boiling of FC-72 is experimentally investigated using copper surfaces coated with a layer of sintered copper particles of irregular and spherical morphologies for a range of porosities (∼40–80%). Particles of the same effective diameter (90–106 μm) are sintered to yield identical coating thicknesses (∼4 particle diameters). The porous structure formed by sintering is characterized using microcomputed tomography (μ-CT) scanning to study the geometric and effective thermophysical properties of the coatings. The boiling performance of the porous coatings is analyzed. Coating characteristics that influence the boiling heat transfer coefficient and CHF are identified and their relative strength of dependence analyzed using regression analysis. Irregular particles yield higher heat transfer coefficients compared to spherical particles at similar porosity. The coating porosity, pore diameter, unit necking area, unit interfacial area, effective thermal conductivity, and effective permeability are observed to be the most critical coating properties affecting the boiling heat transfer coefficient and CHF.

Conventional microchannel heat sinks provide good heat dissipation capability but are associated with high pressure drop and corresponding pumping power. The use of a manifold system that distributes the flow into the microchannels through multiple, alternating inlet and outlet pairs is investigated here. This manifold arrangement greatly reduces the pressure drop incurred due to the smaller flow paths, while simultaneously increasing the heat transfer coefficient by tripping the thermal boundary layers. A three-dimensional numerical model is developed and validated, to study the effect of various geometric parameters on the performance of the manifold microchannel heat sink. Apart from a deterministic analysis, a probabilistic optimization study is also performed. In the presence of uncertainties in the geometric and operating parameters of the system, this probabilistic optimization approach yields an optimal design that is also robust and reliable. Uncertainty-based optimization also yields auxiliary information regarding local and global sensitivities and helps identify the input parameters to which outputs are most sensitive. This information can be used to design improved experiments targeted at the most sensitive inputs. Optimization under uncertainty also provides a quantitative estimate of the allowable uncertainty in input parameters for an acceptable uncertainty in the relevant output parameters. The optimal geometric design parameters with uncertainties that maximize heat transfer coefficient while minimizing pressure drop for fixed input conditions are identified for a manifold microchannel heat sink. A comparison between the deterministic and probabilistic optimization results is also presented.

Efficient and compact cooling technologies play a pivotal role in determining the performance of high performance computing devices when used with highly parallel workloads in supercomputers. The present work deals with evaluation of different cooling technologies and elucidating their impact on the power, performance, and thermal management of Intel® Xeon Phi™ coprocessors. The scope of the study is to demonstrate enhanced cooling capabilities beyond today’s fan-driven air-cooling for use in high performance computing (HPC) technology, thereby improving the overall Performance per Watt in datacenters. The various cooling technologies evaluated for the present study include air-cooling, liquid-cooling and two-phase immersion-cooling. Air-cooling is evaluated by providing controlled airflow to a cluster of eight 300 W Xeon Phi coprocessors (7120P). For liquid-cooling, two different cold plate technologies are evaluated, viz, Formed tube cold pates and Microchannel based cold plates. Liquidcooling with water as working fluid, is evaluated on single Xeon Phi coprocessors, using inlet conditions in accordance with ASHRAE W2 and W3 class liquid cooled datacenter baselines. For immersion-cooling, a cluster of multiple Xeon Phi coprocessors is evaluated, with three different types of Integrated Heat Spreaders (IHS), viz., bare IHS, IHS with a Boiling Enhancement Coating (BEC) and IHS with BEC coated pin-fins. The entire cluster is immersed in a pool of Novec 649 (3M fluid, boiling point 49 °C at 1 atm), with polycarbonate spacers used to reduce the volume of fluid required, to achieve target fluid/power density of ∼ 3 L/kW. Flow visualization is performed to provide further insight into the boiling behavior during the immersion-cooling process. Performance per Watt of the Xeon Phi coprocessors is characterized as a function of the cooling technologies using several HPC workloads benchmark run at constant frequency, such as the Intel proprietary Power Thermal Utility (PTU), and industry standard HPC benchmarks LINPACK, DGEMM, SGEMM and STREAM. The major parameters measured by sensors on the coprocessor include total power to the coprocessor, CPU temperature, and memory temperature, while the calculated outputs of interest also include the performance per watt and equivalent thermal resistance. As expected, it is observed that both liquid and immersion cooling show improved performance per Watt and lower CPU temperature compared to air-cooling. In addition to elucidating the performance/watt improvement, this work reports on the relationship of cooling technologies on total power consumed by the Xeon-Phi card as a function of coolant inlet temperatures. Further, the paper discusses form-factor advantages to liquid and immersion cooling and compares technologies on a common platform. Finally, the paper concludes by discussing datacenter optimization for cooling in the context of leakage power control for Xeon-Phi coprocessors.