Computer Cooling System And Method of Use
A reliable, leak-tolerant liquid cooling system with a backup air-cooling system for computers is provided. The system may use a vacuum pump and a liquid pump and/or an air compressor in combination to provide negative fluid pressure so that liquid does not leak out of the system near electrical components. Alternatively, the system can use a single vacuum pump and a valve assembly to circulate coolant. The system distributes flow and pressure with a series of pressure regulating valves so that an array of computers can be serviced by a single cooling system. The system provides both air and liquid cooling so that if the liquid cooling system does not provide adequate cooling, the air cooling system will be automatically activated. The heat may be removed from the building efficiently with a cooling tower. A connector system is provided to automatically evacuate the liquid from the heat exchangers before they are disconnected. Various turbulators are also provided, as well as a system and method for optimizing the heat transfer characteristics of a heat exchanger to minimize total energy requirements.
The present application claims priority as a continuation of U.S. patent application Ser. No. 13/410,558 filed on Mar. 2, 2012 which is a non-provisional of U.S. Patent Application Ser. No. 61/595,989 filed on Feb. 7, 2012, U.S. patent application Ser. No. 13/410,558 also claims priority as a non-provisional of U.S. Patent Application Ser. No. 61/451,214 filed on Mar. 10, 2011, U.S. patent application Ser. No. 13/410,558 also claims priority as a continuation in part to U.S. patent application Ser. No. 13/308,208 filed on Nov. 30, 2011 which is a non-provisional of U.S. Patent Application Ser. No. 61/422,564 filed on Dec. 13, 2010, and U.S. patent application Ser. No. 13/410,558 also claims priority as a continuation-in-part to U.S. patent application Ser. No. 12/762,898 filed on Apr. 19, 2010. The full disclosure of each of these references is herein incorporated by reference.2.0 TECHNICAL FIELD
The present invention relates to systems and methods for cooling computer systems.3.0 BACKGROUND
Arrays of electronic computers, such as are found in data centers, generate a great deal of heat. An example Central Processing Unit of a computer (“CPU”) generates over 100 watts of heat and has a maximum case temperature of about 60 C. An example rack of 88 CPUs may generate 9 kW of heat. The outdoor temperature at a hot urban location might be 45 C, so even in hot environments heat can still theoretically flow away from the higher temperature computer and toward the lower temperature outside environment. Accordingly, no refrigeration of computers should be required, theoretically. Nonetheless, the standard way to keep data centers cool is to use expensive and relatively inefficient vapor-compression refrigeration systems at least part of the time. These conventional cooling or “air conditioning” systems often use more power that the computers themselves, all of which is discharged to the environment as waste heat. These systems use air as the heat transfer medium, and it is due to the low heat capacity and low thermal conductivity of air that refrigeration must be used to remove the heat generated by multiple air heat exchangers. Removing heat generated by heat exchangers is also referred to as overcoming the thermal resistance of the heat exchangers. Some operators use evaporation of cooling liquid to cool cooling liquid-to-air heat exchangers that cool computers, and this is more thermally efficient than refrigeration, but the computers run hotter, reducing their reliability, decreasing their efficiency and making the data center uncomfortable for human occupants. Water is used as the cooling liquid or coolant throughout this disclosure, but it will be known to those in art that other coolants may be used. The cooling liquid may consist essentially of water, including tap water, or may comprise one or more perfluorocarbons or avionics cooling liquids. The cooling liquid may flow over a plated surface.
Water has approximately 4000 times more heat capacity than air of the same volume, so water is a theoretically ideal heat transfer agent for direct heat transfer from heat generating components. Other cooling liquids offer similar performance. Liquid cooling is recognized as a thermally efficient way to cool computer CPUs due to their high concentration of power and heat generation in a small space, but the rest of a computer's electronics generate heat at a lower rate and temperature, so air-cooling is appropriate for much of the associated hardware. Current systems may use liquid cooling to move the heat from the CPU to a radiator mounted close to the CPU, or they may use an air-to-liquid heat exchanger to remove heat from the computer enclosure and heat-up liquid in the heat exchangers. These systems suffer from the high thermal resistance and bulkiness of air-to-liquid or liquid-to-air heat exchangers. Other systems use a chilled cooling liquid loop to cool the computer, but these systems require complex and expensive connectors and plumbing to connect the server to the building cooling liquid supply while insuring that no leaks occur, which may be devastating in or near a computer. Accordingly, operators of server systems are rightly concerned about leaks and reliability of cooling liquid-cooled computers. Furthermore, chillers require a large amount of power. Additionally, for operation in a data center, servers, particularly blade servers, need to be compact. Therefore, what is needed is a compact cooling solution adaptable for up to a large number of computers, that combines and balances air-cooling capacity for low-intensity heat sources with cooling liquid-cooling capacity for high-intensity heat sources while using a minimum amount of cooling liquid flow, and that is reliable, leak-free and low in power consumption.4.0 SUMMARY
The present system addresses these issues and more by providing in various example embodiments an efficient and compact heat exchanger for a CPU utilizing liquid under negative pressure to minimize chances of leakage, with an air-cooling backup system. Also provided is a cooling solution that integrates with an air-cooled heat sink for backup and utilizes only the minimum amount of water necessary to provide adequate cooling for each heat-generating element. Various embodiments further provide systems and methods to cool the CPU, the server and the data center with liquid in an optimal manner, by cooling the CPU to reduce leakage current, removing heat from the data center by means of the air cooled portion of the CPU heat exchanger, and utilizing an outdoor evaporative cooling system or dry cooler with part time evaporative cooling system that eliminates the need for a chiller in the liquid cooling system. Additionally, provided is a system and method for disconnecting and reconnecting liquid-cooled heat exchangers without losing any water. Heat exchangers employing efficiency-increasing turbulators are also provided.
Provided in various embodiments is a system for cooling one or more electrical devices inside a building, comprising: one or more liquid coolant-containing heat exchangers thermally coupled to one or more electrical devices and each having a liquid input port and a liquid output port and containing liquid coolant at below atmospheric pressure; a liquid coolant-containing chamber in fluid communication with the liquid output port of the heat exchanger(s), the chamber containing liquid coolant and gas at a pressure at least as low as the pressure of the liquid coolant in the heat exchanger(s); a vacuum pump in vacuum communication with the gas in the chamber; a fluid pump with a fluid intake port in fluid communication with the liquid coolant in the chamber and a fluid output port in fluid communication with liquid coolant in an evaporative cooler operating at substantially atmospheric pressure and located at least partially outside the building; the evaporative cooler in fluid communication with the liquid input port of the heat exchanger(s); wherein the fluid pump in combination with the vacuum pump cause the liquid coolant to flow from the chamber through the evaporative cooler and the heat exchanger(s) and back to the chamber. Alternatively, the optional evaporative cooler or other external cooling means can be in a separate loop not in fluid communication with the electronics-mounted heat exchanger system, which may transfer heat to the external cooling loop via an additional water-to-water (liquid-to-liquid) or other heat exchanger.
Also provided in various embodiments is a system for cooling at least one electrical device inside a building, comprising: one or more liquid coolant-containing heat exchangers thermally coupled to a first electrical device and having a liquid input port and a liquid output port and containing liquid coolant at below atmospheric pressure; a system of first and second chambers comprising: a first liquid coolant-containing chamber in one-way fluid communication with the liquid output port of the heat exchanger, the first chamber containing liquid coolant and gas; a second liquid coolant-containing chamber in one-way fluid communication with the liquid output port of the heat exchanger, the second chamber containing liquid coolant and gas; a vacuum pump switchably in vacuum communication with the gas in the first and second chambers; an a source of higher pressure air switchably in pressure communication with the gas in the first and second chambers; the liquid coolant in the first and second chambers in one-way fluid communication with liquid coolant in an evaporative cooler operating at substantially atmospheric pressure and located at least partially outside the building; the evaporative cooler in fluid communication with the liquid input port of the heat exchanger; wherein the vacuum pump and the higher pressure air source coordinates with the system to serially pressurize and depressurize the first and second chambers and thereby cause the liquid coolant to flow substantially steadily from heat exchanger through the system of first and second chambers to the evaporative cooler and back to the heat exchanger. Once again, the optional evaporative cooler or other external cooling means can be in a separate loop not in fluid communication with the electronics-mounted heat exchanger system, which may transfer heat to the external cooling loop via an additional water-to-water or other heat exchanger.
In any of the systems the liquid coolant-containing heat exchangers may comprise one or more turbulators, and may also be thermally coupled to the atmosphere adjacent the electrical device, where a fan may urge circulation of the atmosphere adjacent the liquid coolant-containing heat exchangers. A vacuum accumulator may be in fluid communication with and between the evaporative cooler and the heat exchangers. The turbulator may be located in a heat exchanger tube and configured to force the liquid coolant to flow in a path having a length more than twice the largest dimension of the heat exchanger tube, or may be configured to reduce the cross-sectional area of the flow path of the liquid coolant to less than 50% of the cross-sectional area of the heat exchanger tube. The turbulator may define a conical helix flow path for the liquid coolant, may direct a jet of liquid coolant against a surface proximate one of the electrical devices, may define a rectangular cross-section helical liquid coolant flow path, a round cross-section helical liquid coolant flow path, a rectangular cross-section single-entry helical liquid coolant flow path, a rectangular cross-section double-entry helical liquid coolant flow path, a round cross-section single-entry helical liquid coolant flow path, a round cross-section single-entry helical liquid coolant flow path, or a round cross-section double-entry helical liquid coolant flow path. It may also define a helical path in which the direction of the helix reverses periodically, for example from left handed to right handed. For purposes of this aspect of the disclosure a square cross-section is considered a special case of a rectangular cross-section, i.e., one where the sides are the same length. Systems are provided wherein a portion of the liquid coolant flows axially over the outer surface of the turbulator, thereby causing swirl and turbulence in the flow path and increasing the heat transfer effectiveness of the turbulator.
Also provided are systems comprising: a connector releasably connecting the liquid coolant-containing heat exchanger to the chamber, the connector adapted to release the liquid coolant-containing heat exchanger from the chamber only when substantially all of the liquid coolant has been evacuated out of the heat exchanger. For example, provided is a supply valve in removable fluid communication with the liquid input port of the heat exchanger; a return valve in removable fluid communication with the liquid output port of the heat exchanger; wherein the supply valve is actuatable to open the liquid input port of the heat exchanger to atmospheric pressure air that is at a higher pressure than the water inside the heat exchanger and thereby evacuate the water from inside the heat exchanger; the supply valve and return valve being constructed to close and disconnect the heat exchanger from the system after the water is evacuated from inside the heat exchanger. A passive latching system is also provided. The latching system may include a mechanical delay in order to prevent premature disconnection.
Provided in various systems is a liquid level sensor located in the chamber and providing an output based on the level of the liquid in the chamber, the fluid pump being adapted to operate in response to the output of the fluid level sensor. In other embodiments, provided are fluid level sensors located in both the first and second chambers and providing first and second outputs, respectively, based on the respective levels of the liquid in the chambers, the vacuum pump and the pressure pump each being adapted to operate in response to one or both of the first and second outputs and to maintain the fluid levels in the chambers within predetermined ranges.
Systems may further comprise a vacuum regulator in vacuum communication with the vacuum pump and adapted to maintain a pressure in at least a portion of the system less than atmospheric pressure. Also provided may be a filter in fluid communication with the liquid coolant-containing heat exchanger and adapted to prevent debris from entering the liquid coolant-containing heat exchanger or valves. Additionally provided is a pressure regulator in fluid communication with the liquid coolant-containing heat exchanger, the pressure regulator adapted to provide a constant pressure differential across the liquid coolant-containing heat exchanger. A dome loaded, spring biased regulator, as is known in the art may accomplish this.
A method is provided of modifying a non-liquid cooled electrical device heat exchanger with fins extending from a base to become liquid cooled, comprising the steps of: removing at least a portion of one or more of the fins and thereby making accessible a portion of the base; and affixing liquid cooling tubing having an input port and an output port to at least a portion of the exposed base.
A method is also provided of disconnecting a heat exchanger from a system for cooling at least one electrical device, as described herein where the system comprises: providing such a system, and actuating the supply valve and opening the liquid input port of the heat exchanger to atmospheric pressure air that is at a higher pressure than the cooling liquid inside the heat exchanger; evacuating the cooling liquid from inside the heat exchanger; closing the supply valve and the return valve; and disconnecting the heat exchanger from the system after the cooling liquid is evacuated from inside the heat exchanger. This method may also apply to systems with a plurality of heat exchangers.
Further provided is a method of minimizing the energy needed to cool heat-generating electronics inside a cabinet having a higher than ambient temperature, comprising the steps of: providing a heat exchanger comprising: a thermally conductive base adapted to thermally couple to the heat-generating electronics; a plurality of thermally conductive fins extending outward from the base; and one or more cooling liquid pathways thermally coupled to the base and the fins; balancing the thermal load of the heat generating electronics and the ambient air inside the cabinet by positioning the one or more cooling liquid pathways relative to the base and the fins; thermally coupling the heat exchanger to the heat generating electronics; and providing a source of cooling liquid to the one or more cooling liquid pathways. This method may also comprise the steps of: providing a fan and locating the fan so that it causes air to flow across one or more of the fins; and balancing the thermal load of the heat generating electronics and the ambient air inside the cabinet by: positioning the one or more cooling liquid pathways relative to the base and the fins in further view of the heat transfer effect of the fan; and adjusting the speed of the fan.
Also provided is a system that uses one vacuum pump to circulate coolant under negative pressure. The system includes a pump connected to a vacuum line such that the pump creates a pressure of less than atmospheric on the vacuum line. The vacuum line, along with a pressurized line, is connected to a valve assembly, and that assembly is connected to a first and second fluid chamber. A coolant circuit is provided that allows coolant to circulate through the first and second chambers, through a primary heat exchanger and through an electrical device heat exchanger. The circulation is accomplished through a controller that operates the valve assembly. The circuit may also have a reservoir, various pressure and temperature sensors, and other valves and nozzles to optimize the system. The controller operates the valve assembly by substantially alternating between (a) actuating the valve assembly to create a higher pressure in the first coolant chamber relative to the second coolant chamber, thus emptying coolant from the first coolant chamber and drawing coolant into the second coolant chamber; and (b) actuating the valve assembly to create a higher pressure in the second coolant chamber relative to the first coolant chamber, thus emptying coolant from the second coolant chamber and drawing coolant into the first coolant chamber. The system may also optionally have a coolant recovery devices so as to minimize the maintenance of the system.
Other aspects of the invention are disclosed herein as discussed in the following Drawings and Detailed Description.
The invention can be better understood with reference to the following figures. The components within the figures are not necessarily to scale, emphasis instead being placed on clearly illustrating example aspects of the invention. In the figures, like reference numerals designate corresponding parts throughout the different views and/or embodiments. It will be understood that certain components and details may not appear in the figures to assist in more clearly describing the invention.
Following is a non-limiting written description of example embodiments illustrating various aspects of the invention. These examples are provided to enable a person of ordinary skill in the art to practice the full scope of the invention without having to engage in an undue amount of experimentation. As will be apparent to persons skilled in the art, further modifications and adaptations can be made without departing from the spirit and scope of the invention, which is limited only by the claims.6.1 Example Negative Pressure System Designs
Referring to the example liquid cooling system 100 shown in
In the example embodiment shown in
The cooled cooling liquid 12 is preferably moved through the heat exchanger 1 under a pressure that is less than the local atmospheric pressure. In certain embodiments the entire system 100 runs at a low absolute pressure, so that any leaks are of air into the system 100, rather than cooling liquid 12 out of the system 100. One potential issue with cooling liquid-cooled negative pressure systems is that at low absolute pressures, cooling liquid may boil. For example, at 50 C, water boils at 4 in Hg absolute, so the pressure in water-based systems cannot get that low. Accordingly, this limits the potential pressure drop available to each heat exchanger 1 to the difference between the vapor pressure of the warmest cooling liquid 12 within the system 100 and the local absolute atmospheric pressures. Maximum pressure drops available for each heat exchanger 1 are thus substantially less than one atmosphere. The remainder of the available pressure drop must be used for plumbing to and from the heat exchangers 1 and the pump 10, including head loss, elevation changes, and increases in flow resistance due to fouling.
The plumbing 4, 5, etc. to and from the computer/server/CPU heat exchangers 1 may be designed for unusually low pressure drop, so as to keep the total pressure drop of the system 100 within the aforesaid limits. This may be accomplished in certain embodiments by using, for example, simple surgical tubing or similar light-duty material with large-radius bends and low-pressure-drop fittings, which would not work with conventional high-pressure systems. Conventional high-pressure systems typically use heavier-duty plumbing with sharp bends and large pressure-drop interfaces, which combine to create systems having too much overall pressure drop to work as described herein.
Alternatively, the plumbing 5, etc. to the computer/server/CPU heat exchangers may be high pressure plumbing supplied by an additional pump (not shown), with a pressure regulator 3 to reduce the pressure to below atmospheric as the cooling liquid 12 gets close to the electronics. For the return plumbing 4, etc., larger pipes may be required for the flow of air and cooling liquid, as air will be introduced to the system as computers/servers are removed or replaced. Local air removal systems (not shown) may be used in order to prevent the return plumbing 4, etc., from getting too large. Such systems may use local vacuum pumps, plumbing to a central vacuum pump, or float actuated drain valves and multiple compartments, as in U.S. Pat. No. 4,967,832 to Porter, published Nov. 6, 1990, the full disclosure of which is incorporated herein by reference.
Each server or computer with a liquid heat exchanger 1 may have an inlet pressure regulator 3 and an outlet pressure regulator (not shown) in order to maintain a desired pressure drop across the CPU heat exchanger 1. Each CPU may have a temperature sensor (not shown), and an increase in temperature over the inlet cooling liquid temperature may indicate a problem with the heat exchanger 1. A temperature sensor, such as a thermistor, may be used to measure the inlet cooling liquid temperature. Flow meters, such as a rotameter, thermal mass flow sensor or turbine meter with a digital readout (not shown), may also be used to monitor the flow. The filter 9 may be used after the cooling tower 11 and before the heat exchanger 1 to prevent clogging of the passages in the heat exchanger 1. Chemical additives may be used to prevent fouling of the heat exchanger 1 with biological films and to prevent corrosion. The internal passages of the heat exchanger 1 may be plated or anodized to prevent corrosion.
The cooling liquid chamber 6 is preferably at lower pressure than the heat exchanger on the device being cooled 1. This can be accomplished by keeping the chamber 6 at a lower elevation than the heat exchanger 1 or by means of a check valve with a given cracking pressure or a pressure regulator (see, e.g., check valves 38 and 49 in
For the fluid pump 10, a seal-less centrifugal pump with a magnetic drive may be used, as well as a solenoid pump with an internal fluidic check valve, such as described in U.S. Pat. No. 1,329,559 to Tesla, published on Feb. 3, 1920, the full disclosure of which is incorporated herein by reference. In addition, a system may be required to prime the pump 10, as is known in the art of pumps. For example, this may be accomplished by turning off the liquid pump 10 and allowing fluid 12 to flow back through the pump 10. A flow actuated shuttle valve in the pump output (not shown) may be at a default off position allowing the vacuum pump to suck fluid into the chamber 6. Once the liquid pump 10 is primed and the level sensor 7 activated, the liquid pump 10 may then turn on and pump the fluid out of the chamber 6 and into the cooling tower 11. A pump 10 with low net positive suction head (NPSH) is preferred, so that the cooling liquid does not cavitate at the inlet of the pump 10. The fluid pumps 10 and vacuum pumps 8 for the system 100 may be selected to be reliable and have a long life. They also may provide a steady pressure on the suction side, and a low pressure on the outlet, in order to deliver flow to the cooling tower 11. One example design for maximum operational life would be to use a dual chamber pump such as described in, for instance, U.S. Pat. No. 7,611,333 B1 to Harrington, published on Nov. 3, 2009, the full disclosure of which is incorporated herein by reference, due to the very low NPSH required and due to its ability to reject bubbles from the inlet flow. Such a pump, when driven by a vacuum pump and an air compressor, may provide a very low inlet pressure and an independent output pressure. This type of pump may be fitted with additional backup vacuum pumps and compressors (not shown) connected with check valves so that any single point failure would not cause a system-wide failure. In addition, the check valves and pressurization and vacuum valves and controls may include redundant units (not shown). A condenser and automatic drain system may be required to capture any coolant vapor and droplets which may be pumped out by said vacuum pump.
Although a computer or server or server rack with a liquid heat exchanger 1 is described, systems such as system 100 maybe used to cool any electronic component. Although water is described in various embodiments, any coolant 12 may be used instead of or in addition to water. Although the system 100 is described as using cooling liquid 12 for evaporation and for cooling, a liquid-to-liquid heat exchanger may be used to transfer heat from an evaporator 11 to a closed system (not shown) so that any coolant 12 may be used to interact with the hot components such as CPUs, such as a non-corrosive or non-conductive coolant. This may be used in the case of evaporative coolers 11 that use salt water or reclaimed water, for example. In this way, the coolant used for the computer heat exchangers 1 may be separate from the cooling used for other systems. Then the heat can be transferred from one system to another using, for instance, a plate type heat exchanger in a separate cooling loop. For low temperature operation, as in Northern latitudes, a radiator (not shown), fan 14 and glycol system may be used to reject the heat while preventing freezing of the coolant 12. A mister system can evaporatively pre-cool the air going into the radiators (dry coolers) for use during occasional hot days. Since CPUs can get up to 60 C, cooling liquid 12 can be heated to 50 C and still be used to cool the CPUs. The cooling liquid used for cooling the computers may be kept at a temperature higher than the dew point of the air in the data center to prevent condensation on the plumbing or the heat exchangers.
Referring now to the example liquid cooling system 800 shown in
The cooling liquid flows into the chamber 6 until the level sensor 41 indicates that the chamber 6 is nearly full. Then the valve 34 opens, connecting the vacuum pump 8 with the auxiliary chamber 56 and lowering the pressure of auxiliary chamber 56 so that cooling liquid may flow into it from the extraction pipe 4 through check valve 38. Once flow of cooling liquid is established into both chambers 6 and 56, valve 44 shuts and valve 43 opens, connecting chamber 6 with the pressure pump 53 and thereby pressurizing main chamber 6 so that cooling liquid flows through check valve 48 and into the cooling tower 11. Then the level in chamber 6 reaches a low level, as indicated from level senor 42, at which time the valve 43 shuts. Then the valve 44 opens and flow is again established under suction into the main chamber 6, at which time the auxiliary chamber vacuum valve 34 is shut and the valve 33 is opened connecting chamber 56 with the pressure pump 53 and forcing cooling liquid out of chamber 56 through check valve 39 until the level in the chamber 56 reaches the low level sensor 32. Under normal operation level sensor 31 would not be activated because the system is designed so that the flow out of the chambers 6, 56 is higher than the flow into the chambers 6, 56, so that the auxiliary chamber 56 is never completely full, thereby allowing for the flow through the heat exchangers 1 to be steady while the flow to the cooling tower 11 is intermittent. Accordingly, the level sensor 31 can be used to indicate if there is a system failure. The pressure and vacuum levels can be monitored by the pressure pump 53 and the vacuum pump 8 using the pressure sensors 54 and 15. The entire system can be controlled by a computer or by a logic circuit or any other suitable means. Floats 51 may be used to sense the levels in the chambers 6, 56 and reduce evaporation of the cooling liquid 12 in the chambers 6, 56.
Also shown in system 1800 is a flow sensor 1830. The flow sensor 1830 may include a self heated thermistor or RTD, such that if the liquid coolant stops flowing, or the coolant is too hot, the fan 1840 is turned on to high speed. This could be accomplished by flowing a known current through a thermistor such that in still coolant, and under 25 C ambient conditions, the thermistor temperature rose to 35 C. A comparator circuit could detect the voltage decrease associated with the temperature rise and a MOSFET could be switched on to control the speed of the fan 1840. Under air cooling conditions the power to fan 1840 would typically be on all the time, but under liquid-cooled conditions, the power to the fan 1840 could be pulse width modulated at 10-500 Hz to slow down the fan 1840 but not allow it to stop. The controller for the fan 1840 is represented by unit 1850. These features are applicable to any of the present systems.
Any leakage in the system may be detected by monitoring the cycle time of a pump 8 used to remove air from the systems 100, 800, 800′, 1800, etc. If the pump 8 is cycling on too often, then a leak is indicated. The leak may be discovered by pulling a vacuum on each heat exchanger 1 and measuring the decrease in vacuum over time. A simple hand operated vacuum pump may be used for this type of testing.
Systems 100, 800, 800′, 1800, etc. may use a pump with a chamber (not shown) to supply fluid to all the heat exchangers 1. During a shutdown procedure, the pump may evacuate the system; purge it with air and store the fluid until such time as the liquid cooling system is reactivated. During a reactivation procedure, the pump control system may apply a vacuum or a pressure to the system, check to see if the fluid system loses vacuum or pressure and then start pumping again, based on the rate of change of the system pressure.6.2 Example Dry-Disconnect Systems
For example, the computers/servers with liquid heat exchangers 1 may be connected to the pumping system using a connector 1600 such as shown in
Then, once the fluid 12 is evacuated from the bottom chamber 1670 to a predetermined level, a larger leak opens up, the bottom spool valve 1660 drops all the way to the bottom of bottom chamber 1670, and the valve 1600 is closed or sealed from both the supply 1620 and return 1630 lines. The valve 1600 may be latched in the closed position until it is reconnected to a server 1, at which point both spools 1650, 1660 rise and the supply and return lines 1620, 1630 flow freely and the bottom chamber 1670 is refilled. The valve 160 may also be held in the intermediate position (i.e., with top spool valve 1650 closed while bottom spool valve 1660 remains open) by the negative pressure which will be present until the server 1 is purged of liquid 12. For example, a spring loaded diaphragm or piston (not shown) could hold the valve in the intermediate position until the negative pressure was reduced, as it would be once the server 1 was completely vented of liquid. The valve 1600 may also be triggered by pressure differences created with an orifice or venturi, which differences would be higher when flowing liquid than when flowing gas, as is known in the art of fluid mechanics.
Each computer or server or server rack with a liquid heat exchanger 1 may be connected with the present dry disconnect system that allows for the automatic draining of the heat exchanger 1 as described above. Such connectors may include supply and return flows. Supply and return flows may be coaxial, in order to allow for a small interconnect. The system is preferably designed to remove all of the cooling liquid from inside each heat exchanger subsystem 1 such as a CPU, server or server rack during the disconnection process. For example, if the heat exchanger 1 contains one cc of cooling liquid 12, and the flow rate is 150 cc/minute of cooling liquid, then it will take less than 1 second to drain the cooling liquid out of the computer or server or server rack with a liquid heat exchanger 1. As the cooling liquid 12 is replaced by air, the flow resistance of the heat exchanger decreases, so the process may happen in less than 0.5 seconds.6.3 Example Turbulator Designs
With reference to
The heat exchanger 1 may incorporate a helical flow pattern for the cooling liquid 12 to put a long path into a short passage to increase heat transfer. This helical flow passage may have multiple starts and paths, as shown in
The rod and cylinder may be square, cylindrical, conical, triangular, hexagonal, or any other appropriate shape. The rod or other turbulator structure may be designed so that some of the cooling liquid 12 flows over the edge 1004 of flow passages 1005 in an axial direction, for instance directly from a proximal end 1002 to a distal end 1003 of the turbulator 1001 shown in
For example, referring to the embodiment shown in
Referring to the example embodiment shown in
In the example embodiment shown in
The design of turbulators shown in
A thermodynamic model of these competing thermal resistances is shown in
The fan 1840 that is typically connected to the CPU heat exchanger may also be used to cool the interior of the computer by transferring heat from the air inside the computer to the cooling liquid 12 so that other components within the server enclosure may be cooled with or without the use of external air flow—i.e., the computer may be sealed. The speed of the fan 1840 may be adjusted to remove additional heat from the air inside the server enclosure of the data center as required to minimize the overall power consumption of the data center. The overall power consumption versus fan speed may be determined based on the power consumption of the air conditioning system versus temperature in the data center and the power consumption of the CPU 1820 versus its temperature. The CPU 1820 uses additional power depending on the temperature of the processor due to leakage currents, with the leakage currents increasing exponentially with the processor at the higher temperature range. For example, CMOS-based processors use more energy as the temperature of the processor goes up, due to leakage currents. Also, the air conditioning system of the data center uses additional power depending on the temperature of the data center and the building heat removal requirements. This increase is generally linear; with higher temperatures requiring proportionally higher air conditioning power. By controlling and selecting the optimal speed of the CPU fan 1840, the flow rate of liquid 12 through the heat exchanger 1, and the position of the liquid heat exchanger tubes 1010, 1130, 1320, 1420, 1440 in the overall assembly consisting of a base 1110 and fins 1120, the overall power required for the data center can be decreased. Examples of these relative flows of heat between the various components are depicted by the wavy arrows in
With further reference to
An example heat exchanger design started with an existing air-cooled system. In order to provide the best cooling with minimum volume and input power, a spiral cooling channel with a Reynolds number just above the laminar limit was used. This is believed to provide the best cooling with a reasonably sized channel that can pass contamination.
For example, if a 140 watt CPU is to be cooled with water, and an 18 degrees F. (10 degrees C.) temperature rise can be accepted, then a flow rate of 220 cc/minute would be needed based on the heat capacity and mass flow rate of water. Next, rocket science was employed to develop a nozzle cooling system, which in rocket science is done with an array of tubes that cool the nozzle and preheat the fuel on the way to the combustion chamber. The goal there is to adjust the length and diameter of the parallel tube array to get the optimum cooling for a given flow rate. In the present case, the water outlet temperature and the heat sink temperature are desired to be within 1 degrees C. of each other. So a fluid path was selected with a Reynolds number slightly higher than 2100, so that the flow was turbulent, but the pressure drop was not too high. In this example two helical flow passages were used, 0.055 inch (1.4 mm) in diameter. This system was analyzed using empirical heat transfer equations for flow in a tube, modeled using computation flow dynamics (CFD), and tested with a Xeon processor running stress software. The thermal resistance heat sink to water, based on the temperature of the water into the heat sink, was 0.04 watt/degrees C. with 230 cc/minute flow rate per CPU. A similar heat sink design with coolant passages in the base is shown in
The test heat sink worked exactly as modeled, but when the flow was increased, it was discovered that it could actually remove heat from the entire system. A stack of three DL380 servers was run at idle power levels in an insulated box and the heat sinks were able to remove all the heat (700 watts) from the computers. In this case the ambient air was 107 degrees F. (42 degrees C.) and the coolant inlet was at 76 degrees F. (22 degrees C.).
Additionally, a test was conducted with a 2 kW rack of servers in an office environment at 75 degrees F. (24 degrees C.) ambient. The servers were either air-cooled or water-cooled using an outdoor miniature cooling tower with water at 65 degrees F. (18 degrees C.). The temperature data is shown in
The RAM temperatures were lower with liquid cooling because the RAM chips were located downstream of the heat exchanger. Assuming a typical data center power distribution of 56% Servers, 30% HVAC, 5% UPS and 6% other, the total power required for the original air cooled system would be 3.6 kW (Server Power divided by 0.56). Using liquid cooling allows 1 kW to bypass the HVAC system and go directly outdoors, saving HVAC power. And this has a multiplying energy savings effect, since it takes more than 1 kW of energy for an HVAC system to remove 1kW of heat. It also saved 10% of the server input power due to lower fan power and because the processors required less power at lower temperatures. The liquid pump and cooling tower fan used only 50 watts. This reduced the overall power consumption based on typical data center power distribution to 2.9 kW, a total power reduction of approximately 20%. The power reduction is diagrammed in
Accordingly, the combination of an air cooled heat sink modified for redundant liquid cooling, a negative pressure system to prevent leaks, and a connector that automatically purges the coolant adds up to a system that offers a path from the current air-cooled technology to the liquid cooled data center of the future, without having to modify the building. The present liquid cooled and air cooled heat sink system reverses the thermodynamics of traditional systems so that the heat sink removes heat from the CPUs and the server interior and the data center in general in order to reduce the HVAC loads and fan power by a large margin.6.7 Single Vacuum Pump Cooling System
In previously described embodiments, a separate vacuum pump and circulation pump are used to circulate the coolant throughout the system at negative pressure. The embodiment shown in
Turning in detail to
In one embodiment, both the main chamber 2206 and the auxiliary chamber 2208 are connected to the reservoir 2218, such that the coolant can travel in one direction from the main/aux chamber to the reservoir, the one direction travel being accomplished by the use of check valves 2220. The reservoir 2218 is where the coolant is drawn from for circulation to the electronic equipment, shown as servers 2222, and the removal of heat from that equipment through the use of an electronic equipment heat exchanger 2223. The reservoir 2218 connects to the primary heat exchanger 2224 (this can be a liquid-liquid exchanger or an air-liquid exchanger) reducing the temperature of the coolant prior to circulating the coolant via cold manifold 2226 to the servers 2222, and returning the heated coolant via hot manifold 2228 back to the main and auxiliary chambers (2206 and 2208). The coolant from the hot manifold 2228 travels only in one direction to the main and auxiliary chambers (2206 and 2208), the one direction travel being accomplished by the use of check valves 2230. The travel of the coolant throughout the system 2200 is also referred to herein as the coolant circuit.
Alternatively the main chamber 2206 and the auxiliary chamber 2208 can be connected directly to the primary heat exchanger 2224, completely obviating the need for the reservoir 2218. The reservoir 2218, however, is helpful is equalizing the negative pressure through the system 2200, such that the flow of coolant is more constant and less pulsating. Also the reservoir 2218 allows the system 2200 to hold more coolant, minimizing the possibility that the system 2200 will run dry.
The system 2200 may also have redundant valves and pumps to reduce the chance of shutdown to a negligible level. One such redundancy system may have two vacuum pumps running at 50% capacity, such that if one fails the other ramps up to cover the load. This redundancy also imbues the system 2100 with enough vacuum capacity to work with one server completely open to air.
The system 2200 may also have several sensors, filters and structures to help optimize its performance. For example, the reservoirs 2218 may include level sensor to make certain that there is sufficient cooling liquid in the system to meet the demands of the electronic equipment. Filters may be placed throughout the system to remove debris that could interfere with the valves and negatively affect performance. A set of temperature (2240) and pressure (2242) sensors may be placed on the cool manifold and a set on the hot manifold to detect the temperature and pressure difference of the coolant. All the information from these sensors may be feed to the controller 2232. If for example, the system detects insufficient coolant, the system may open the fill valve 2234 to add more coolant and alert the system operator that the coolant level was low. If the pressure sensors detect an abnormal pressure drop, this could signal that there is a leak in the system and the system would alert the operator. Because the system operates under negative pressure, the leak would not expose the computer equipment to the cooling liquid, but rather would introduce air into the system and potentially reduce the efficiency of the system in cooling the computer equipment. To reduce the ability of a leak to compromise the cooling efficiency of the system a novel set of valves and nozzles are used on the hot and cold manifolds, and discussed in greater detail with reference to
Other valves may be used to further optimize the system. For example, test valve 2236 may be used when the system is first turned on. Test valve 2236 should remain closed until the system detects at the various pressure sensors that the appropriate amount of negative pressure has been reached and maintained. This prevents the system from being activated with leaks present and prevents coolant from circulating to the electronic equipment under atmospheric or near atmospheric pressure, such that a leak would actually cause coolant to spill. Purge valve 2238 may be used to purge the system of coolant when the system is turned off. Again, this prevents coolant from remaining in the electronic equipment plumbing under atmospheric or near atmospheric pressure, such that a leak would actually cause coolant to spill.
The components of the system 2200 encompassed by the box 2244 may be sufficiently small to be installed as a rack mount device in a traditional server tower. Further those components may be placed on a tray such that any leaks that may occur in the rack-mounted unit would be captured by the tray and would not impact any of the server equipment.
The operation of the system will now be described with reference to
After the vacuum pump 2602, a second coolant recovery device 2604 may be placed. Here the second coolant recovery device 2604 is a muffler/condenser. Although not shown, the pressurized line 2204 may be vented to atmospheric pressure. When vented in this fashion, the system 2200 still operates and the pressurized line 2204 would be at atmospheric pressure and would be pressurized as compared to the vacuum line 2202. As the vacuum pump 2602 evacuates the air, it will still have some humidity, and if that moisture in the air is not captured, then the system will require frequent coolant addition. The second coolant recovery device 2604, condenses the coolant out of the air, allowing the coolant to collect at the bottom of the device 2604. Here the coolant recovery device 2604 is a muffler, but other coolant recovery devices include, and are not limited to, air/water separators and thermoelectric devices that condense any moisture out of the air. In fact, a thermoelectric device may be used to condense moisture out of the atmosphere in order to make up for any coolant loss. The device 2606 is connected to main chamber 2206 via the condensation return line 2250, such that when the main chamber 2206 is under vacuum, there is a pressure differential of about 25 to 30 in Hg. A float valve 2606 may be used, such that the valve 2606 is closed until a sufficient amount of coolant has collected at the bottom of the device 2604. Once the valve 2606 opens, the pressure differential pushes the recovered coolant to the main chamber 2206. Of course a pump could also be used to pump the coolant back to the system. For example, a piston pump, gear pump or peristaltic pump would be suitable.6.8 Flow Control in the Event of a Single Point Gross Leak
In the event that one of the servers has a damaged liquid cooling system and is leaking air into the system through a completely broken coolant conduit, the rest of the system should still operate provided that the leakage rate into both sides of the liquid cooling system is controlled.
On the return side (valve 2715), the air flow must not be so excessive as to reduce the pump effectiveness substantially. One way to achieve this goal is to use a Venturi, which has a low pressure drop when flowing coolant at the nominal flow rate, and limits the flow of air into the system to that which can flow through the minimum diameter of the Venturi at the speed of sound. For example, a Venturi with a throat of 0.05 inches, will have a pressure drop of approximately 1 in Hg at 300 cc/min coolant flow rate, and it will flow approximately ½ standard cubic feet per minute of air at 20 in Hg vacuum. A large-scale system designed to cool 100 kW of servers may flow approximately 35 gal/min of coolant. A completely open sever line would therefore represent approximately 17% of the overall volume flow rate, and the system could still efficiency cool the electronic equipment.6.9 Leak Detection
Leak detection can also be included in the systems previously described. Detecting leaks is important because it can lower the efficiency of the system. To detect a leak of air into the system, the flow rate of air and coolant back into the system should be measured. The flow rate of coolant may be measured by measuring the time it takes to fill one of the chambers (i.e., 2206, 2208) because the volume of those chambers is already known. Placing a level sensor in the chamber (see
The flow rate of air may be measured by a flowmeter (see
For example, if the vacuum pump has a displacement of 0.05 liters per revolution, and it spins at 1500 RPM, then it should flow 75 liters per minute. There are some losses and leaks within the vacuum pump, but they are repeatable and known for a given pump. Therefore given the RPM of the pump, and the pressure at the inlet, then the mass flow rate can be determined using the ideal gas law. The flow rate can also be measured at the vacuum pump outlet by means of a thermal mass flowmeter, or a venturi, orifice or other flow meter. The flow meter should be calibrated for humid air. The temperature and pressure of the water entering and leaving the cooling system is known, and this can be used to predict the amount of air that is dissolved in the water. It can also be used to predict the amount of water vapor in the gas above the water. The maximum amount of air that can be evolved from the water is the difference between the amount that could be dissolved at the temperature and pressure of the water leaving the cooling system, and the amount of air that could be in the water returning to the system. The controller could use a look up table to calculate the two amounts and the difference that would be expected at the system air outlet. If the amount of air in the water returning to the pump from the servers is excessive, then an alarm could activated. A humidity sensor could be used to determine how much of the gas evolved is water vapor, or testing could be used to determine the range of outlet humidity. For example, in a typical system with a single vacuum pump the maximum vacuum achieved with no air leaks and a flow of 1 gal/min of water might be 24 in Hg. If there is a leak of 0.03 ft3/min of air into the system then the maximum vacuum will be only 20 in Hg, and this would indicate a leak or a vacuum pump failure.
The invention has been described in connection with specific embodiments that illustrate examples of the invention but do not limit its scope. Various example systems have been shown and described having various aspects and elements. Unless indicated otherwise, any feature, aspect or element of any of these systems may be removed from, added to, combined with or modified by any other feature, aspect or element of any of the systems. As will be apparent to persons skilled in the art, modifications and adaptations to the above-described systems and methods can be made without departing from the spirit and scope of the invention, which is defined only by the following claims. Moreover, the applicant expressly does not intend that the following claims “and the embodiments in the specification to be strictly coextensive.” Phillips v. AHW Corp., 415 F.3d 1303, 1323 (Fed. Cir. 2005) (en banc).
1. A turbulator for use in a liquid flow passageway of a coolant-containing heat exchanger that is adapted to transfer heat from an electrical component to the coolant, the passageway having a cross-sectional area, the turbulator comprising:
- a core that is substantially concentric to the passageway, the core having a cross-sectional area; and
- a ridge structure connected to the core, the ridge structure radiating away from the core, and the ridge structure defining a flow path, wherein the flow path has a length more than twice the largest dimension of the passageway;
- wherein the cross-sectional area of the core is at least 20% of the cross sectional area of the passageway.
2. The turbulator of claim 1, wherein the ridge structure is adapted to allow leakage of coolant over the ridge structure sufficient to induce swirling of coolant, wherein the swirling of coolant is substantially perpendicular to flow path.
3. The turbulator of claim 1 wherein the flow path defines a shape selected from a group consisting a helix, a conical helix, a rectangular cross-section helix, a round cross-section helix, a rectangular cross-section single-entry helix, a rectangular cross-section double-entry helix, a round cross-section single-entry helix and a round cross-section double-entry helix.
4. The turbulator of claim 1 wherein the turbulator is adapted to direct a jet of coolant against a surface proximate the electrical component.
5. The turbulator of claim 1 wherein the turbulator and passageway define a second turbulence-inducing liquid flow path when the turbulator is placed in the passageway, the second turbulence-inducing liquid flow path being at least 25% shorter than the first flow path.
6. The turbulator of claim 5 wherein the second flow path causes swirling of the liquid in the first flow path.
7. A liquid cooling system for cooling an electrical component, comprising:
- a coolant containing heat exchanger adapted to transfer heat from the electrical component to the liquid, the heat exchange comprising coolant flow passageway having a cross-sectional area;
- a turbulator disposed of in the passageway comprising: a core that is substantially concentric to the passageway, the core having a cross-sectional area; and a ridge structure connected to the core, the ridge structure radiating away from the core, and the ridge structure defining a flow path, wherein the flow path has a length more than twice the largest dimension of the passageway; wherein the cross-sectional area of the core is at least 20% of the cross sectional area of the passageway.
8. The system of claim 7, the heat exchanger further comprising a base plate thermally coupled to the component and a plurality of fins extending from the base plate.
9. The system of claim 7, the heat exchanger further comprising a base plate thermally coupled to the device;
- a heat pipe thermally coupled to the base plate; and
- the heat pipe thermally coupled to a plurality of fins.
10. The system of claim 7, wherein the ridge structure is adapted to allow leakage of coolant over the ridge structure sufficient to induce swirling of coolant, wherein the swirling of coolant is substantially perpendicular to flow path.
11. The system of claim 7, wherein the flow path defines a shape selected from a group consisting a helix, a conical helix, a rectangular cross-section helix, a round cross-section helix, a rectangular cross-section single-entry helix, a rectangular cross-section double-entry helix, a round cross-section single-entry helix and a round cross-section double-entry helix.
12. The system of claim 7, further comprising:
- a vacuum pump adapted to propel the coolant through the passageway at less than ambient pressure.
13. The system of claim 12, further comprising:
- a pressure sensor in fluid communication with the heat exchanger and adapted to take a pressure reading of the coolant; and
- a controller connected to the pressure sensor adapted to signal an alert if the pressure reading is outside a normal operable range.
14. The system of claim 12, further comprising:
- a pressure sensor in fluid communication with the heat exchanger and adapted to take a pressure reading of the coolant;
- a valve in fluid communication with the heat exchanger; and
- a controller connected to the pressure sensor and the valve, the controller adapted to open the valve to allow the flow of coolant into the heat exchanger when the pressure reading is within a normal operable range.
15. The system of claim 7, further comprising:
- a vacuum pump adapted to remove the coolant from the heat exchanger when the heat exchanger is removed from the system.
16. A method of minimizing the energy needed to cool heat-generating components inside a cabinet having a higher than ambient temperature, comprising the steps of:
- providing a heat exchanger comprising: a thermally conductive base adapted to thermally couple to the heat-generating components; a plurality of thermally conductive fins extending outward from the base; and one or more coolant pathways thermally coupled to the base and the fins;
- balancing the thermal load of the heat generating components and the ambient air inside the cabinet by positioning the one or more coolant pathways relative to the base and the fins;
- thermally coupling the heat exchanger to the heat generating components; and
- providing a source of coolant to the one or more coolant pathways.
17. The method of claim 16, further comprising the steps of:
- providing a fan and locating the fan so that it causes air to flow across one or more of the fins; and
- balancing the thermal load of the heat generating components and the ambient air inside the cabinet by: positioning the one or more coolant pathways relative to the base and the fins in further view of the heat transfer effect of the fan; and adjusting the speed of the fan.
18. The method of claim 16, wherein the one or more pathways is a coolant flow passageway having a cross-sectional area;
- a turbulator disposed of in the passageway comprising: a core that is substantially concentric to the passageway, the core having a cross-sectional area; and a ridge structure connected to the core, the ridge structure radiating away from the core, and the ridge structure defining a flow path, wherein the flow path has a length more than twice the largest dimension of the passageway; wherein the cross-sectional area of the core is at least 20% of the cross sectional area of the passageway.
19. The method of claim 18, wherein the ridge structure is adapted to allow leakage of coolant over the ridge structure sufficient to induce swirling of coolant, wherein the swirling of coolant is substantially perpendicular to flow path.
20. The system of claim 18, wherein the flow path defines a shape selected from a group consisting a helix, a conical helix, a rectangular cross-section helix, a round cross-section helix, a rectangular cross-section single-entry helix, a rectangular cross-section double-entry helix, a round cross-section single-entry helix and a round cross-section double-entry helix.
Filed: Apr 13, 2015
Publication Date: Aug 6, 2015
Inventor: Steve Harrington (Cardiff, CA)
Application Number: 14/685,524