scholarly journals Preliminary experimental of GPU immersion-cooling

2019 ◽  
Vol 93 ◽  
pp. 03003 ◽  
Author(s):  
Nugroho Agung Pambudi ◽  
Husin Bugis ◽  
Ilham Wahyu Kuncoro ◽  
Nova Dany Setiawan ◽  
Miftah Hijriawan ◽  
...  

A typical information technology system takes around 40% of the total energy used in cooling the system. There are three major classifications of cooling system and they are: water cooling, close loop liquid cooling, and immersion cooling systems. Immersion cooling has been observed to be the latest trend in cooling systems for IT devices. It is a cooling procedure that is carried out through the immersion of all computer components in a dielectric coolant. This research examined the cooling process of GPU using this immersion method. Mineral oil, because of its high dielectric strength, is used as a medium fluid. The temperature difference between the use of fan and immersion cooling was then measured using a benchmark software. The result showed that the immersion cooling produced a lower GPU temperature compared to the conventional fan. The working temperature of the GPU with the use of immersion method was 70°C while it was 80°C with the conventional fan method.

Author(s):  
Uschas Chowdhury ◽  
Manasa Sahini ◽  
Ashwin Siddarth ◽  
Dereje Agonafer ◽  
Steve Branton

Modern day data centers are operated at high power for increased power density, maintenance, and cooling which covers almost 2 percent (70 billion kilowatt-hours) of the total energy consumption in the US. IT components and cooling system occupy the major portion of this energy consumption. Although data centers are designed to perform efficiently, cooling the high-density components is still a challenge. So, alternative methods to improve the cooling efficiency has become the drive to reduce the cooling cost. As liquid cooling is more efficient for high specific heat capacity, density, and thermal conductivity, hybrid cooling can offer the advantage of liquid cooling of high heat generating components in the traditional air-cooled servers. In this experiment, a 1U server is equipped with cold plate to cool the CPUs while the rest of the components are cooled by fans. In this study, predictive fan and pump failure analysis are performed which also helps to explore the options for redundancy and to reduce the cooling cost by improving cooling efficiency. Redundancy requires the knowledge of planned and unplanned system failures. As the main heat generating components are cooled by liquid, warm water cooling can be employed to observe the effects of raised inlet conditions in a hybrid cooled server with failure scenarios. The ASHRAE guidance class W4 for liquid cooling is chosen for our experiment to operate in a range from 25°C – 45°C. The experiments are conducted separately for the pump and fan failure scenarios. Computational load of idle, 10%, 30%, 50%, 70% and 98% are applied while powering only one pump and the miniature dry cooler fans are controlled externally to maintain constant inlet temperature of the coolant. As the rest of components such as DIMMs & PCH are cooled by air, maximum utilization for memory is applied while reducing the number fans in each case for fan failure scenario. The components temperatures and power consumption are recorded in each case for performance analysis.


Author(s):  
Sandu Constantin ◽  
Dan Brasoveanu

Abstract The thermal efficiency of gas turbines is critically dependent on the temperature of burnt gases at turbine inlet, the higher this temperature the higher the efficiency. Stochiometric combustion would provide maximum efficiency, but in the absence of an internal cooling system, turbine blades cannot tolerate gas temperatures that exceed 1300 K. Therefore, for this temperature, the thermal efficiency of turbine engine is 40% less than theoretical maximum. Conventional air-cooling techniques of turbine blades allow inlet temperatures of about 1500 K on current operating engines yielding thermal efficiency gains of about 6%. New designs, that incorporate advanced air-cooling methods allows inlet temperatures of 1750–1800 K, with a thermal efficiency gain of about 6% relative to current operating engines. This temperature is near the limit allowed by air-cooling systems. Turbine blades can be cooled with air taken from the compressor or with liquid. Cooling systems with air are easier to design but have a relatively low heat transfer capacity and reduce the efficiency of the engine. Some cooling systems with liquid rely on thermal gradients to promote re-circulation from the tip to the root of turbine blades. In this case, the flow and cooling of liquid are restricted. For best results, cooling systems with liquid should use a pump to re-circulate the coolant. In the past, designers tried to place this pump on the engine stator and therefore were unable to avoid high coolant losses because it is impossible to reliably seal the stator-rotor interface. Therefore it was assumed that cooling systems with liquid could not incorporate pumps. This is an unwarranted assumption as shown studying the system in a moving frame of reference that is linked to the rotor. Here is the crucial fact overlooked by previous designers. The relative motion of engine stator with respect to the rotor is sufficient to motivate a cooling pump. Both the pump and heat exchange system that is required to provide rapid cooling of liquid with cold ambient air, could be located within the rotor. Therefore, the entire cooling system can be encapsulated within the rotor and the sealing problem is circumvented. Compared to recent designs that use advanced air-cooling methods, such a liquid cooling system would increase the thermal efficiency by 8%–11% because the temperatures at turbine inlet can reach stoichiometric levels and most of the heat extracted from turbine during cooling is recuperated. The appreciated high reliability of the system will permit a large applicability in aerospace propulsion.


Author(s):  
John D. Bernardin ◽  
Walter C. Barkley ◽  
Jack Gioia ◽  
Pilar Marroquin

This paper discusses the design, analysis, and testing of a Water Cooling System (WCS) for a Drift Tube Linear (DTL) Particle Accelerator structure at the Los Alamos Neutron Science Center (LANSCE). The DTL WCS removes large amounts of dissipated electrical energy in a very controlled manner to maintain a constant temperature of the large structure. First, the design concept and method of water temperature control is discussed. Second, the layout of the water cooling system, including the selection of plumbing components and instrumentation is presented. Next, the development of a numerical nodal network model, used to size the plumbing, pump, control valves, and mixing tank (heat exchanger), is discussed. Finally, empirical pressure, flow rate, and temperature data from a functioning DTL water cooling system are used to assess the water cooling system performance and validate the numerical model.


2019 ◽  
Vol 100 ◽  
pp. 00080
Author(s):  
Lukasz Scislo ◽  
Nina Szczepanik-Scislo

The aim of the research was to describe the effects that air and water cooling systems can have on the dynamic behaviour of precise machines. Although much thought is paid to vibration isolation of the ground and isolation of individual effects of machines operating close to each other, it is often forgotten to model or to measure the effects that the ventilation or the machines cooling systems have on the machine itself. This can be especially important for high precision machines used for experimental research and medical equipment. The article shows the effects of ventilation and cooling system on the induction of additional resonant frequencies of the system for the high precision machine like a linear collider. This kind of machine requires special environmental conditions to assure proper beam stability. Due to the dynamic behaviour of typical machines, the presence of the new high amplitude frequencies in the 0-100 Hz range is very dangerous for its stability of work. In the case of high precision machines, it is not only a cause of not optimal working conditions but very often is a cause of serious problems.


Author(s):  
Kanchan M. Kelkar ◽  
Suhas V. Patankar

Liquid cooling is used for thermal management of electronics in defense, power, medical, and computer applications due to the increasing power density and the desire for compact packaging. The objectives in the design of these systems are to create a sufficient amount of total flow and to appropriately distribute the flow so as to maintain the electronic component temperatures at the desired level. The technique of Flow Network Modeling (FNM) is ideally suited for the analysis of flow distribution and heat transfer in liquid-cooling systems. The FNM technique uses overall flow and thermal characteristics to represent the behavior of individual components. Therefore, solution of conservation equations over the network enables efficient prediction of the flow rates, pressures, and temperatures in a complete liquid-cooling system. This article describes the technical basis of the FNM technique and illustrates its application in the design of a distributed-flow cold plate and of a complete water-cooled system. The study demonstrates the utility of the FNM technique for rapid and accurate evaluation of different design options and the ensuing productivity benefits in the design of liquid cooling systems.


Processes ◽  
2021 ◽  
Vol 9 (6) ◽  
pp. 1060
Author(s):  
David Žurovec ◽  
Lucie Jezerská ◽  
Jan Nečas ◽  
Jakub Hlosta ◽  
Jan Diviš ◽  
...  

Cooling is an important process during the production of pellets (as post-treatment). The pellet cooling process significantly impacts the quality of the pellets produced and the systematic use of energy. However, the cooling systems currently in use sometimes encounter technical problems, such as clogging of the perforated grids (sieves), the discharge hopper, or pellet degradation may occur. Therefore, a prototype of a new pellet cooling system using a vibrating feeder was tested. The aim of the study is to present a new variation of pellet cooling system using spiral vibration cooler as a possible solution next to a counterflow cooler. The presented system was tested (critically evaluated and discussed) in two design variants. The first variant consists in cooling by chaotic movement of the pellets. The second is then in combination with the chaotic movement of the pellets together with the action of intense air flow using specially placed air hoses. All tests involved pelletization of rapeseed straw. It was found that both cooling system variants could, realistically, be used. However, the variant with an intense air flow was more energy-intensive, a factor which is, however, offset by the higher quality of the pellets. No negative impact of vibrations to pellets quality was occur. Studies provide insight into new usable technologies that do not reduce the efficiency of the process as a result of grate clogging.


Author(s):  
Mayumi Ouchi ◽  
Yoshiyuki Abe ◽  
Masato Fukagaya ◽  
Haruhiko Ohta ◽  
Yasuhisa Shinmoto ◽  
...  

Energy consumption in data center has been drastically increasing in recent years. In data center, server racks are cooled down by air conditioning for the whole room in a roundabout way. This air cooling method is inefficient in cooling and it causes hotspot problem that IT equipments are not cooled down enough, but the room is overcooled. On the other hand, countermeasure against the heat of the IT equipments is also one of the big issues. We therefore proposed new liquid cooling systems which IT equipments themselves are cooled down and exhaust heat is not radiated into the server room. For our liquid cooling systems, three kinds of cooling methods have been developed simultaneously. Two of them are direct cooling methods that the cooling jacket is directly attached to heat source, or CPU in this case. Single-phase heat exchanger or two-phase heat exchanger is used as cooling jackets. The other is indirect cooling methods that the heat generated from CPU is transported to the outside of the chassis through flat heat pipes, and condensation sections of the heat pipes are cooled down by liquid. Verification tests have been conducted by use of real server racks equipped with these cooling techniques while pushing ahead with five R&D subjects which constitute our liquid cooling system, which single-phase heat exchanger, two-phase heat exchanger, high performance flat heat pipes, nanofluids technology, and plug-in connector. As a result, the energy saving effect of 50∼60% comparing with conventional air cooling system was provided in direct cooling technique with single-phase heat exchanger.


Author(s):  
Yoshiyuki Abe ◽  
Mayumi Ouchi ◽  
Masato Fukagaya ◽  
Takashi Kitagawa ◽  
Haruhiko Ohta ◽  
...  

Energy utilization in data centers, especially cooling systems for server racks, needs extensive improvement. The present authors proposed advanced cooling network systems for data centers, and R & D activities have been conducted under the so-called Green IT Project sponsored by NEDO (New Energy and Industrial Technology Development Organization). In the present concept, CPUs in servers are cooled down by either direct liquid cooling system or heat pipes with liquid cooling systems in the condensation region. The liquid cooling systems are integrated in each server rack and among server racks. A series of studies on both single phase and two phase narrow channel heat exchangers, high performance heat pipes with self-rewetting fluids and nanofluids for heat transfer enhancement are ongoing. In addition, a prototype server rack with the cooling network systems is also under development toward commercial products. This paper reports the updated status of the present R & D.


Author(s):  
Roger Schmidt ◽  
Mike Ellsworth ◽  
Madhu Iyengar ◽  
Gary New

The increased focus on green technologies and energy efficiency coupled with the insatiable desire of IT equipment customers for more performance has driven manufacturers to deploy liquid cooling technologies for cooling IT equipment. In April, 2008 IBM announced the world’s fastest UNIX server using unique water cooling technologies. The first system (eleven racks) was shipped and installed at The National Center for Atmospheric Research (NCAR) in Boulder, Colorado in May, 2008 to be used for research on climate change and severe weather. The POWER6 575 system utilizes water cooling to cool the processors as well as a water cooled rear door to remove 80% of the 60 kW heat load generated by each rack. This cluster of eleven racks with this highly efficient cooling system achieves a performance of 76 teraflops within a small footprint. This paper will describe the infrastructure required to support this cluster of racks including the chilled water cooling loop, water thermal storage tanks, and the 480 VAC power distribution system. The time response of the various potential failure modes in the water cooling system will also be described.


Author(s):  
Cheng Chen ◽  
Emad Samadiani ◽  
Bahgat Sammakia

Demands for higher computational speed and miniaturization have already resulted in extremely high heat fluxes in microprocessors. Fractal tree-shaped microchannel liquid cooling systems are novel heat transfer enhancement systems to keep the temperature of the microprocessors in a safe range. Due to the complexity of these systems, their full field numerical modeling for simulation of the flow and temperature fields is too time consuming and costly, particularly to be used within iterative optimization algorithms. In this paper, a quick but still accurate compact modeling approach based on Flow Network Modeling (FNM) is introduced for analysis of the flow filed in fractal microchannel liquid cooling systems. The compact method is applied to a representative fractal microchannel cooling system and the obtained velocity and flow rate distribution are validated against a full Computational Fluid Dynamics (CFD)-based model for three different designs. The compact model shows good agreement with the CFD results and robustness on different designs, while requiring much less computational capability and time. Afterwards, the compact model is used for optimization of the geometry of the fractal cooling system to achieve maximum flow rate and uniform flow distribution among the channels for a fixed pressure drop.


Sign in / Sign up

Export Citation Format

Share Document