Methods and systems for enhancing electronics cooling using thermoelectric elements

Abstract

The present application discloses an immersion cooling system and a method for immersion cooling information technology (IT) components. The immersion cooling system includes a coolant tank containing a liquid coolant and at least partially immersed in the liquid coolant: 1) a thermoelectric cooling (TEC) element, the TEC element being coupled to an information technology (IT) component mounted on the IT equipment; and 2) a heat sink coupled to the TEC element, wherein the TEC element is configured to transfer heat generated by the IT component to the liquid coolant via the heat sink.

Description

Technical Field

[0001] Embodiments of the present disclosure generally relate to cooling electronic devices using thermoelectric cooling (TEC) elements. Background Art

[0002] Thermal management for a data center that includes several active electronic racks is critical to ensuring proper performance of the servers and other information technology (IT) equipment operating in the racks (e.g., performing IT data processing services). However, without proper thermal management, the thermal environment (e.g., temperature) within the rack may exceed a thermal operating threshold, which may result in adverse consequences (e.g., server failure, etc.). One way to manage the thermal environment is to cool the IT equipment using chilled air. The chilled air is recirculated using a cooling unit that extracts the heat captured by the chilled air. One commonly used cooling unit is a computer room air conditioning (CRAC) unit, which is a device that draws in hot exhaust air and supplies chilled air to the data center to maintain the thermal environment of the data center. CRAC is an air cooling unit that is widely used in existing air-cooled data centers, and there are many other types of solutions for air-cooled data centers. In addition, most existing data centers are air cooled.

[0003] Recently, data centers have been deploying more high-power density electronic racks, where more high-density chips are packed more tightly together to provide more processing power. Especially due to the development of artificial intelligence (AI) and cloud-based services, high-performance and high-power density processors, such as control processing units (CPUs) and graphics processing units (GPUs), are needed. Cooling these high-density racks by maintaining an appropriate thermal environment can be a problem for existing cooling systems (such as CRAC units). For example, although a CRAC unit can maintain a thermal environment with more conventional racks (or lower density racks), it may not be able to effectively cool high-power density racks because high-power density racks may generate heat loads at higher rates due to the higher density of electronic devices. In some cases, liquid cooling becomes a more efficient and viable cooling solution in situations with high power density or high heat flux.

[0004] On the other hand, immersion cooling (which involves immersing electronic devices at least partially in a dielectric solution) is a feasible solution for high-density electronic devices. However, implementing immersion cooling is challenging. For example, since the working medium is no longer air, a customized heat sink structure is required to obtain good thermal performance. In addition, unlike air cooling, immersion cooling depends on several factors, such as the material properties of a specific dielectric fluid and the height at which immersion cooling is performed. Changing the position (e.g., height) or using different fluids may result in a significant performance degradation when immersion cooling. Summary of the Invention

[0005] A first aspect of the present application provides an immersion cooling system. The system includes a coolant tank containing liquid coolant and containing, at least partially immersed in the liquid coolant, 1) a thermoelectric cooling (TEC) element coupled to an IT component mounted on information technology (IT) equipment, and 2) a heat sink coupled to the TEC element. The TEC element is configured to transfer heat generated by the IT component to the liquid coolant via the heat sink.

[0006] A second aspect of the present application provides a method for immersion cooling of information technology (IT) components. The method includes determining a power input to an active IT component mounted on an IT device, the active IT component being contained within a coolant tank having a liquid coolant, wherein mounted on the IT component is a stack of thermoelectrically cooled TEC elements and a heat sink, wherein the TEC element is disposed between the IT component and the heat sink, wherein the stack is at least partially immersed in the liquid coolant. The method also includes determining whether the TEC element will operate according to a first operating mode or a second operating mode. The method also includes calculating a critical temperature difference of the TEC element in response to determining that the TEC element will operate according to the first operating mode; determining a temperature difference and a voltage of the TEC element based on the power input, determining whether the temperature difference is greater than the critical temperature difference, and applying a voltage across the TEC element in response to determining that the temperature difference is greater than the critical temperature difference.

[0007] A third aspect of the present application provides a non-transitory machine-readable medium having instructions stored therein, wherein when the instructions are executed by a processor, the processor is caused to perform the method according to the first aspect of the present application.

[0008] A fourth aspect of the present application provides a computer program product, comprising a computer program, wherein the computer program implements the method according to the first aspect of the present application when executed by a processor.

[0009] The immersion cooling system of the present application can adjust the thermal performance of the cooling system based on changing cooling requirements, and cool high-power-density racks more efficiently and practically. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Various aspects are shown by way of example and not limitation in the figures of the accompanying drawings, wherein like reference numerals indicate like elements. It should be noted that references to "one" or "an" aspect of the present disclosure are not necessarily to the same aspect, and that they mean at least one. In addition, for the sake of brevity and to reduce the total number of drawings, a given figure may be used to illustrate features of more than one aspect, and not all elements in a figure may be required for a given aspect.

[0011] Figure 1is a block diagram illustrating an example of a thermoelectric cooling (TEC) element according to one embodiment.

[0012] Figure 2 is a block diagram illustrating an example of an immersion cooling system including a TEC element according to one embodiment.

[0013] Figure 3 is a block diagram illustrating an immersion cooling system including a cooling distribution unit (CDU) according to one embodiment.

[0014] Figure 4 is a flow chart of a process performed by an immersion cooling system to operate in an efficiency mode, according to one embodiment.

[0015] Figure 5 An example of an efficiency profile diagram illustrating a maximum coefficient of performance (COP) of a TEC element is shown, according to one embodiment.

[0016] Figure 6 is a flow chart of a process performed by an immersion cooling system to operate in a performance mode, according to one embodiment.

[0017] Figure 7 An example of a performance profile diagram according to one embodiment is shown.

[0018] Figure 8 is a flow chart of a process performed by an immersion cooling system according to one embodiment. DETAILED DESCRIPTION

[0019] Several aspects of the present disclosure will now be explained with reference to the accompanying drawings. Whenever the shapes, relative positions, and other aspects of the components described in a given aspect are not explicitly defined, the scope of the present disclosure is not limited to the components shown, which are intended for illustrative purposes only. In addition, although many details are set forth, it should be understood that some aspects can be practiced without these details. In other examples, well-known circuits, structures, and techniques are not shown in detail to avoid obscuring the understanding of this specification. In addition, unless the meaning is clearly to the contrary, all ranges set forth herein are deemed to include the endpoints of each range.

[0020] References in the specification to "one embodiment" or "an embodiment" mean that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present disclosure. The phrase "in one embodiment" appearing in various places in the specification is not necessarily all referring to the same embodiment.

[0021] The present disclosure solves the problems described herein by providing an immersion cooling system that deploys one or more thermoelectric cooling (TEC) elements (or thermoelectric elements), the TEC elements being configurable to operate in one or more operating modes. The immersion cooling system includes a coolant tank containing TEC elements mounted to electronic devices, particularly information technology (IT) components (e.g., processors or chips) mounted on IT equipment (e.g., printed circuit boards (PCBs)), all of which are at least partially immersed in a liquid coolant. During operation, the TEC elements are configured to transfer heat generated by the IT components (e.g., via a heat sink mounted on the TEC elements) to the liquid coolant. To control the TEC elements, the cooling system may also include a TEC controller configured to draw power from a power source (e.g., from a PCB or an external power source) and control heat transfer by applying an input voltage across the TEC elements. In addition, the TEC controller may be configured to operate in various operating modes to adapt to the changing cooling requirements of the IT components. For example, the more processing power an IT component uses, the more heat it may generate, and therefore the IT component may require more cooling. The TEC controller can operate in a first operating mode (e.g., an efficiency mode) in which the TEC element is driven with minimal power consumption, while all (or most) the generated heat is transferred away from the IT component and the temperature of the component is (e.g., maintained) below a maximum allowable temperature threshold. As another example, the TEC controller can operate in a second operating mode (e.g., a performance mode) in which the TEC element is driven such that the temperature of the IT component is the lowest usable temperature. In particular, this mode provides the lowest temperature that the IT component can achieve when the TEC element is driven by the TEC controller. Thus, the TEC controller can adjust the thermal performance of the cooling system based on the changing cooling requirements of the IT components.

[0022] According to one embodiment, an immersion cooling system includes a coolant tank having a liquid coolant and containing 1) a thermoelectric cooling (TEC) element at least partially immersed in the liquid coolant, the TEC element coupled to an information technology (IT) component mounted on the IT equipment, and 2) a heat sink coupled to the TEC element, wherein the TEC element is configured to transfer heat generated by the IT component to the liquid coolant via the heat sink.

[0023] In one embodiment, the immersion cooling system further includes a TEC controller mounted on the IT equipment and electrically coupled to the TEC element, wherein the TEC controller is configured to draw power from the IT equipment and to control heat transfer from the TEC element by applying an input voltage across the TEC element. In some embodiments, the TEC controller is configured to sense power input to the IT component and, based on whether the power input to the IT component exceeds a power threshold, to operate in one of a plurality of operating modes. In one embodiment, in response to the power input being less than the power threshold, the TEC controller is configured to operate in a first one of several operating modes, in which the input voltage is a minimum voltage that, when applied across the TEC element, causes the TEC element to transfer all heat generated by the IT component to a heat sink while maintaining a temperature of the IT component below a maximum temperature threshold; and in response to the power input being greater than the power threshold, the TEC controller is configured to operate in a second one of several operating modes, in which the applied input voltage causes the TEC element to transfer all heat generated by the IT component to the heat sink while maintaining a temperature of the IT component at a minimum usable temperature.

[0024] In one embodiment, the TEC element and the heat sink are an integrated unit.In some embodiments, a thermal interface material (TIM) is disposed between at least one of: 1) the IT component and the TEC element, and 2) the TEC element and the heat sink.

[0025] According to another embodiment, a method for immersion cooling of information technology (IT) components includes determining a power input to an active IT component mounted on IT equipment, the active IT component contained within a coolant tank having a liquid coolant, wherein mounted on the IT component is a stack of a thermoelectric cooling (TEC) element and a heat sink, wherein the TEC element is disposed between the IT component and the heat sink, wherein the stack is at least partially immersed in the liquid coolant; determining whether the TEC element is to operate according to a first operating mode or a second operating mode; and responsive to determining that the TEC element is to operate according to the first operating mode, calculating a critical temperature difference of the TEC element, determining the temperature difference and a voltage of the TEC element based on the power input, determining whether the temperature difference is greater than the critical temperature difference, and responsive to determining that the temperature difference is greater than the critical temperature difference, applying a voltage across the TEC element.

[0026] In one embodiment, determining the temperature difference and the voltage includes identifying a maximum coefficient of performance (COP) of the TEC element associated with the temperature difference and the voltage using the power input. In some embodiments, the voltage is a first voltage, and the method further includes, in response to determining that the temperature difference is less than a critical temperature difference, determining a second voltage based on the critical temperature difference; and applying the second voltage across the TEC element.

[0027] In one embodiment, the voltage is a first voltage, and the method further includes, in response to determining that the TEC element is to operate according to the second operating mode, determining an operating current for the TEC element associated with a lowest available temperature of the IT component based on the power input; determining whether the lowest available temperature is less than a maximum temperature of the IT component; and in response to determining that the lowest available temperature is less than the maximum temperature of the IT component, applying a second voltage associated with the operating current across the TEC element. In some embodiments, the method further includes, in response to determining that the lowest available temperature is greater than the maximum temperature of the IT component, deactivating the active IT component.

[0028] In one embodiment, determining whether the TEC element is to operate according to the first operating mode or the second operating mode includes selecting the first operating mode when the power input of the active IT component is less than a power threshold, and selecting the second operating mode when the power input of the active IT component is greater than the power threshold. In some embodiments, determining whether the TEC element is to operate according to the first operating mode or the second operating mode includes selecting one of the first operating mode and the second operating mode in response to user input.

[0029] According to another embodiment, a non-transitory machine-readable medium has instructions stored therein that, when executed by a processor, cause the processor to perform operations similar to those performed in the methods described herein.

[0030] In one embodiment, as used herein, "coupling" one component (or element) to another component may refer to "fluidically" coupling the two components so as to allow a fluid, such as a coolant or liquid coolant, to flow between the two components. For example, coupling a first tube to a second tube may couple the two tubes together and also allow fluid to flow from the first tube to the second tube.

[0031] Figure 1is a block diagram illustrating an example of a TEC element (or TEC device) according to one embodiment. The diagram shows a TEC element 1 having (at least) two semiconductors 4 (a p-type semiconductor and an n-type semiconductor) with (at least) different electron densities (e.g., n-type>p-type). The semiconductors can be arranged as columns placed in thermal parallel to each other and electrically in series via an electrical conductor 3 connected to a thermal conductor 2. Each end of the semiconductor that is not connected to the electrical conductor 3 is connected to a separate conductor. Specifically, the p-type semiconductor is connected to the electrical conductor 5, while the n-type semiconductor is connected to another electrical conductor 9 that is separate from the electrical conductor 5. The electrical conductors 5 and 9 are both connected to another terminal conductor 6. In addition, the electrical conductor 5 is electrically coupled to the negative terminal of the voltage source 7 and the electrical conductor 9 is electrically coupled to the positive terminal of the voltage source (e.g., via one or more wires 8). In one embodiment, the voltage source can be an external voltage source (e.g., separate from the TEC element 1). In another embodiment, the voltage source and the TEC element can be coupled to each other. For example, the two elements can be part of an electronic assembly (e.g., both mounted on a PCB board). In some embodiments, the voltage source can be part of one or more controllers, such as Figure 2 The TEC controller 12 is shown.

[0032] In one embodiment, any of the components of the TEC element can be made of any (known) material. For example, thermal conductors 2 and 6 can be made of a ceramic material, which can conduct heat but not electricity. As another example, electrical conductors 3, 5, and / or 9 can be made of any conductive material (e.g., copper, etc.).

[0033] The TEC element is arranged to move heat from one surface to another at the expense of additional external power. Specifically, when a voltage source 7 applies a voltage (e.g., 10 volts) across two electrical conductors 5 and 9, a current I (e.g., 10 amps) flows through the semiconductor, causing the TEC element to transfer heat Q0 from the "cold side" to the "hot side." Thus, when current flows, a temperature difference is generated by the semiconductor 4, as shown by T0-T1. The amount of heat ultimately removed from the hot side is Q1, which may be greater than Q0 because Q1 may include Q0 and any heat Q' generated by the TEC element when the TEC is activated. In other words, Q1=Q0+Q'. Thus, during operation as part of a cooling system as described herein, the thermal conductor 2 may be coupled to (e.g., mounted on) an IT component such that the TEC element draws Q0 generated by the IT component from the component and into the thermal conductor 6. Thus, heat is drawn from the IT component to cool the IT component when the IT component is activated and the TEC element is operational.

[0034] Figure 2is a block diagram illustrating an example of an immersion cooling system including a TEC element according to one embodiment. In particular, the figure illustrates an immersion cooling system 10 (which may be referred to as a cooling system hereinafter) that is configured to immerse cool one or more IT components coupled to (or mounted on) one or more IT devices 11. As shown, the system includes IT devices 11, a TEC controller 12, IT components 13, a TEC element 1, a radiator 15, a coolant tank 18, and a coolant 19. In one embodiment, the system may include more or fewer elements (or components). For example, the system may include one or more IT devices, each of which includes one or more mounted elements, as shown here. As another example, the system may not include a coolant tank or coolant. In that case, the cooling system may be air-cooled. This article describes more about alternative air cooling systems.

[0035] In one embodiment, IT equipment 11 may be any element (such as a PCB as described herein) arranged to house one or more IT components and / or any type of electronic components (such as TEC controller 12). In some embodiments, the IT equipment may be a blade server arranged to be (at least partially) immersed in a liquid coolant and / or arranged to be mounted in an electronics server rack.

[0036] The IT device 11 may include one or more IT components 13 (e.g., a central processing unit or CPU, a graphics processing unit (GPU), memory, and / or storage devices). The IT components may perform data processing tasks, wherein the IT components may include software installed in a storage device, loaded into a memory, and executed by one or more processors to perform the data processing tasks. In one embodiment, the IT device may include a host server (referred to as a host node) coupled to one or more computing servers (also referred to as computing nodes, such as CPU servers and GPU servers). The host server (having one or more CPUs) is typically connected to a client via a network (e.g., the Internet) to receive requests for specific services (such as storage services) (e.g., cloud-based storage services (such as backup and / or restore)) to execute applications to perform certain operations (e.g., image processing, deep data learning algorithms or modeling as part of a software as a service or SaaS platform). In response to the request, the host server assigns the task to one or more performance computing nodes or computing servers (having one or more GPUs) managed by the host server. The performance computing server (e.g., IT component 14) performs the actual task, which may generate heat during operation. As described herein, this heat needs to be transferred away from the component to ensure that it does not overheat, otherwise it may cause a failure.

[0037] The TEC element 1 is arranged to be coupled to an IT component 13. The figure described herein shows the TEC element being mounted on the IT component. Coupled to the TEC element is a heat sink 15, which includes a base 16 and one or more heat sinks 17. Specifically, the base is arranged to be coupled to the TEC element (e.g., mounted on the TEC element). Thus, the IT component, the TEC element, and the heat sink are arranged in a stack. In this example, each of the elements can be a separate element that can be removably coupled within the stack. For example, the TEC element can be removed from the top of the IT component and the heat sink can be separated from the TEC element. As a result, elements can be added to / removed from the stack as needed (e.g., in the event of a component failure). In another embodiment, at least some of the elements can be packaged together as an integrated unit. For example, the heat sink (e.g., the base of the heat sink) and the TEC element can be an integrated unit.

[0038] The TEC element 1 is configured to transfer heat (e.g., at least a portion of the heat) (e.g., Q0) generated by (e.g., active) IT components to one or more fins 17 of a heat sink 15. For example, in this figure, the cold side (e.g., Figure 1 The thermal conductor 2 shown can be coupled to the IT component 13, and the hot side of the TEC element (e.g., Figure 1 The heat conductors 6 shown can be coupled to a heat sink 15 to transfer heat away from the IT components, thereby reducing their temperature. As described herein, the heat transferred to the heat sink can then be discharged (or transferred) to the coolant 19 flowing around and / or between the heat sinks. More information on operating a TEC element is described herein.

[0039] In one embodiment, disposed between the components (at least some of these components) is a thermal interface material (TIM) 14, which can be a thermally conductive pad, thermally conductive tape, and / or thermally conductive paste. Specifically, the TIM is disposed between 1) the IT component 13 and the TEC element 1, and 2) the TEC element and the heat sink 15 (the heat sink's base 16). In one embodiment, the TIM 14 may not be disposed between at least some of the components. For example, the TEC element and the heat sink may be directly coupled to each other. As another example, as described herein, the TIM may not be disposed between two components that form an integrated unit, as is the case when the TEC element and the heat sink are an integrated unit.

[0040] The TEC controller 12 can be a dedicated processor, such as an application-specific integrated circuit (ASIC), a general-purpose microprocessor, a field-programmable gate array (FPGA), a digital signal controller, or a collection of hardware logic structures (e.g., filters, arithmetic logic units, and a dedicated state machine). In one embodiment, the TEC controller can be a circuit having a combination of analog components (e.g., resistors, capacitors, inductors, etc.) and / or digital components (e.g., logic-based components such as transistors, etc.). The TEC controller can also include memory. As shown, the TEC controller and IT component 13 are separate components coupled to the IT equipment 11 (e.g., mounted on the IT equipment 11). In one embodiment, the TEC controller 12 can be part of the TEC element 1 (e.g., within a container). In another embodiment, the TEC controller can be part of (or integrated within) the IT component 13. In another embodiment, the TEC controller can be one of (or part of) another IT device (e.g., contained within the coolant tank 18 or in another location, such as mounted in an electronics rack).

[0041] The coolant tank 18 is a container designed to hold (or contain) a coolant and one or more IT devices. In one embodiment, the tank can be formed of one or more materials, such as plastic, metal, etc. In one embodiment, the tank can have one or more openings (e.g., at the top of the tank) to allow a technician to add or remove IT equipment and / or replenish the coolant within the tank. As shown, the tank contains the coolant 19 and the remaining components described herein (e.g., TEC elements, IT components, IT equipment, TEC controller, and radiator), with the remaining components at least partially immersed in the coolant. In one embodiment, the (liquid) coolant 19 can be any type of thermally conductive dielectric liquid. In another embodiment, the coolant can be a non-toxic fluid. In some embodiments, the coolant can be designed and operated to have a high boiling point (e.g., above the threshold operating temperature (e.g., maximum temperature threshold) of at least some of the main components of the IT equipment).

[0042] As shown, a TEC controller 12 is connected (e.g., electrically coupled) to the TEC element 1 via one or more wires 8. The TEC controller is configured to activate (or operate) the TEC element so that the TEC element can (e.g., begin) transferring (at least a portion of) the heat generated by the IT components to (one or more of) the fins 17 of the heat sink 15, as described herein, and ultimately dissipated into the coolant 19. In one embodiment, the TEC controller can include or be electrically coupled to a power source, which can be part of the IT device 11 or an external power source. The TEC controller can be configured to draw power from the power source (e.g., from the IT device) and can be configured to provide an input current to the TEC element to activate the element. In particular, the TEC controller can be configured to control the heat transfer of the TEC element by applying an input voltage across the TEC element. Thus, upon activation, heat generated by the IT components is absorbed by the TEC element and transferred to the heat sink. The heat is then transferred to the fins of the heat sink and then discharged into the coolant 19.

[0043] In one embodiment, the TEC controller is configured to operate in one of several operating modes. Specifically, as described herein, the TEC controller can operate in an efficiency mode or a performance mode. In these modes, the TEC controller can adjust the input voltage applied across the TEC element, thereby controlling heat transfer, to adapt to the changing cooling requirements of the IT component. The cooling requirements may vary based on the amount of processing power (or power input) the IT component is performing in order to perform one or more computing operations. As an example, the TEC controller can operate in an efficiency mode while the power input to the IT component is below a power threshold. In this mode, the IT component is cooled by the TEC element while the amount of power overhead required by the TEC controller to drive the TEC element is minimized. On the other hand, when the power exceeds the power threshold, the TEC controller can switch to a performance mode, which ensures that the IT component remains as cool as possible. This article describes more about two of these modes and what determines the mode in which the TEC controller will operate.

[0044] Figure 3is a block diagram illustrating an immersion cooling system including a cooling distribution unit (CDU) according to one embodiment. In particular, the figure shows a coolant tank 18 coupled to a CDU 25, thereby creating a heat exchange loop (e.g., a secondary heat exchange loop), as described herein. Contained within the coolant tank is an IT device 11 including four TEC elements 1a-1d, each of which is mounted on at least one IT component (not shown), which is mounted on the IT device. Additionally, although not shown, each of the TEC elements may include one or more heat sinks mounted thereon. In one embodiment, the IT device may include more or fewer TEC elements (e.g., based on having more than four IT components). For example, when the IT device has eight IT components, the cooling system 20 may include eight TEC elements, one TEC element for each of the IT components. Also shown is that each of the TEC elements is coupled to a TEC controller 12 via one or more wires 8. In one embodiment, the TEC controller may be configured to control each of the TEC elements independently of each other.

[0045] In addition, the IT device includes a baseboard management controller (BMC) 31 communicatively coupled to the TEC controller. The BMC is configured to communicate with one or more IT components installed on the IT device 11. For example, the BMC may be communicatively coupled to one or more sensors, such as temperature sensors, in each of the IT components. The BMC may be configured to receive temperature readings (e.g., as signals) from the temperature sensors and may be configured to transmit the temperature readings (e.g., one or more associated with each of the IT components) to the TEC controller 12. As another example, the BMC may sense the temperature of the coolant 19 from one or more temperature sensors within the coolant tank 18 and / or from the CDU. In another embodiment, the BMC is configured to detect (sense) the power input of each of the IT components (e.g., by measuring the input voltage and / or input current of the IT component). The BMC is configured to transmit this data (e.g., temperature reading, power reading, etc.) of at least one of the IT components of the IT device to the TEC controller 12. In another embodiment, the TEC controller may communicate directly with one or more of the IT components and / or one or more sensors, as described herein.

[0046] As shown, the coolant tank 18 includes a supply port 21, a return port 22, and a perforated rack 30. Supply port 21 is an inlet arranged to be coupled to a CDU 25 via a supply line 23. Port 22 is an outlet arranged to be coupled to the CDU via a return line 24. In one embodiment, the ports can be configured to be removably coupled to their respective lines. For example, the ports can be connectors, such as drip-free blind-mating quick disconnects. In this case, each of the ports can be a socket-type drip-free connector, while their respective lines include a plug-type end arranged to connect to the port connector.

[0047] The CDU 25 includes a heat exchanger 27, a liquid pump 28, and a CDU controller 26 (e.g., a pump). In one embodiment, the CDU may include other components such as a liquid reservoir, a power supply, and monitoring sensors. Note that the CDU may be any type of commercially available or custom CDU.

[0048] Heat exchanger 27 can be a liquid-to-liquid heat exchanger. In one embodiment, heat exchanger 27 can include (or be part of) one or more heat exchange circuits. For example, heat exchanger 27 can include a heat exchange circuit (e.g., a primary heat exchange circuit) having a cooling source 29, where cooling liquid is received from the source (e.g., via a supply line) and warmed liquid is returned to the source (e.g., via a return line). In one embodiment, the cooling source can be an IT liquid cooling water system or any other type of cooling liquid source.

[0049] In one embodiment, the coolant circulating within the primary heat exchange loop is warmed due to heat transfer from within the secondary heat exchange loop. The secondary heat exchange loop operates as follows. Coolant 19 is supplied to the coolant tank via supply port 21 by the CDU (e.g., pump 28 of the CDU) (via supply line 23). The coolant 19 flows through a porous rack 30 (which may be designed to increase the fluid flow rate within the tank). As the coolant moves upward through the tank, heat generated by the IT components is transferred to the flowing coolant through their respective TEC elements 1c and 1d (via radiators). The heat transfer warms the coolant, which then continues upward to cool the IT components where the TEC elements 1a and 1b are mounted. In one embodiment, IT components mounted below the TEC controller 12 may have lower temperature requirements than IT components mounted above the controller. The warmed coolant travels through return port 22 and reaches heat exchanger 27 via return line 24. The heat exchanger cools the warmed coolant by transferring the heat contained therein to the cooling liquid flowing through the primary heat exchange loop (from cooling source 29). The heat exchanger provides the cooled coolant to pump 28, which then recirculates the cooled coolant through the secondary loop.

[0050] Figure 4 、 Figure 6 and Figure 8 is a flow chart including processes performed by a cooling system to operate in one of several operating modes. In particular, at least one of the processes may be performed by the TEC controller 12 (and / or the BMC 31) of any of the (immersion) cooling systems illustrated and described herein (e.g., by one or more algorithms executed thereon).

[0051] Figure 4 is a flow chart of a process performed by an immersion cooling system to operate in an efficiency mode, according to one embodiment. In particular, the diagram illustrates a process 40 in which TEC controller 12 may operate in efficiency (e.g., a first operation) in which a minimum voltage (e.g., an input voltage) is applied across TEC element 1, which causes the TEC element to transfer all heat generated by IT component 13 to heat sink 15 (ultimately discharged to coolant 19), while the IT component temperature remains below a maximum threshold temperature.

[0052] Process 40 begins by determining the power input of the active IT components on which the TEC elements and heat sinks are mounted (at block 41). For example, the BMC 31 can sense the power drawn by the IT components as they perform one or more computing operations. The BMC can transmit the power data to the TEC controller 12. Process 40 calculates the critical temperature difference δT of the TEC element 1. c (at box 42). Figure 2 In the equation, the heat transfer mechanics can be expressed as

[0053] T hs,base -T liquid,in =(Q0+Q')*R hs (1)

[0054] Among them, T hs,base is the temperature on the base 16 of the heat sink 15, T liquid,in is the temperature of the coolant 19 at the radiator (eg, the inlet of the radiator) and R hs is the thermal resistance of the heat sink. As described herein, Q0 is the heat generated by the IT component 13, and Q' is the heat generated by the TEC element, where the heat is to be removed (e.g., discharged to the coolant through the heat sink). Two heat values ​​can be generated from the power input of the two devices. For example,

[0055] Q=R*I 2 (2)

[0056] Where R is the resistance and I is the current of the power input. Therefore, Q0 is the heat generated based on the power input of the IT component, and Q' is the heat based on the power input used by the TEC controller 12 to drive the TEC element. By defining the temperature difference as

[0057] δT=T Component –T hs,base (3)

[0058] Where T Component is the temperature of the IT component 13 and T hs,base is the temperature of the base 16 of the heat sink, and T is replaced by δT hs,base After that, the temperature of the IT components is obtained as

[0059] T Component =(Q0+Q')*R hs +δT+T liquid,in (4)

[0060] In one embodiment, when the TEC element 1 is active (eg, transferring heat away from the IT component), δT<0.

[0061] To determine δT c , T componnet =T max (i.e., the maximum (allowable) temperature of the IT component (e.g., before overheating), which can be predefined (e.g., by the manufacturer of the IT component)) and Q' = 0, which means that the critical temperature represents the minimum requirement for the thermal performance of the TEC element in the immersion cooling system. As a result, two formulas are derived. First, if T max <Q0*R hs +T liquid,in ,but

[0062] δT c =|T max –Q0*R hs –T liquid,in | (5)

[0063] Therefore, if T max <Q0*R hs +T liquid,in , then the system cannot handle the heat load Q0 when the TEC element is not activated. However, if T max ≥Q0*R hs +T liquid,in ,but

[0064] δT c =0 (6)

[0065] Therefore, the TEC controller 12 can calculate δT from the above equation (5) or equation (6) based on the inequalities mentioned herein (any of the inequalities). c .

[0066] The process 40 determines the temperature difference δT* of the TEC element 1 and the first voltage U* of the TEC element based on the power input of the IT component (at block 43). In one embodiment, δT* may be Figure 1 The temperature difference T0–T1 is shown as a negative temperature value between two conductive surfaces. The cold side in contact with the IT component has a lower temperature than the hot side in contact with the heat sink. In some embodiments, the TEC controller can minimize power consumption by determining a maximum coefficient of performance (COP), where the COP is equal to the heat load removed from the IT component (e.g., Q0) divided by the power input to the TEC element. Therefore, as the COP increases, less power input is required to drive the TEC element to remove the same amount of heat generated by the IT component. In one embodiment, to determine the δT* and U* corresponding to the maximum COP, the TEC controller can determine an efficiency profile, such as Figure 5 shown.

[0067] Figure 5 An example of an efficiency profile graph 50 illustrating the maximum coefficient of performance (max COP) of a TEC element is shown, according to one embodiment. Specifically, graph 50 illustrates the amount of heat removed from IT components (and / or generated by the TEC element when activated) relative to a temperature difference. Additionally, the graph illustrates a maximum COP curve, where any point along the curve represents the minimum power expenditure required to provide a specific temperature difference based on the amount of heat removed. The graph also illustrates a voltage line, U*, which represents the (e.g., constant) input voltage to the TEC element. Points along the voltage line represent how a constant voltage applied across the TEC varies as the amount of heat generated by the IT component (to be removed by the TEC element) varies (along the y-axis), while the temperature difference across the TEC element also varies (along the x-axis). As shown, as the amount of heat generated by the IT component (to be removed by the TEC element) increases, the temperature difference decreases at the same voltage. In one embodiment, graph 50 is a predefined graph. In another embodiment, the graph is component-based (e.g., IT components and TEC elements). Therefore, different components may have different efficiency graphs.

[0068] To determine δT* and U*, the TEC controller determines at what point the maximum COP curve intersects Q0 (e.g., by drawing a horizontal line from Q0). The point where the horizontal line intersects the maximum COP defines δT* and U*, where the voltage is a predefined line passing through that point. In one embodiment, if the horizontal line from Q0 intersects the maximum COP more than once (e.g., twice), δT* and U* can be defined at the intersection point with the lowest temperature difference (in absolute value).

[0069] In another embodiment, to determine these values, the TEC controller 12 may use the power input of the IT component (e.g., thermally) determined to perform a table lookup on a data structure that correlates the amount of heat removed, the temperature difference, the maximum COP, and the input voltage. Specifically, the TEC controller may use Q0 as input to perform the table lookup and, in return, receive the table outputs δT* and U* associated with the maximum COP at Q0.

[0070] return Figure 4 , the process 40 determines whether the (determined) temperature difference is greater than a critical temperature difference (at decision block 44). Specifically, the TEC controller 12 determines whether the absolute value of |δT*| is greater than δT c If so, the process 40 applies a first voltage across the TEC element 1 (at block 45). In particular, the TEC controller applies U* across the TEC element. However, if the determined temperature difference is not greater than the critical temperature difference, the process determines a second voltage U*' based on the critical temperature difference (at block 46). In one embodiment, the TEC controller may determine U*' based on the efficiency pattern map 50 (and / or data structure), as described herein. For example, to determine U*', the TEC controller may identify a point along a horizontal line from Q0 that is (-δT in the graph 50). c ) temperature difference intersects. Once the intersection point is found, the TEC controller can determine U*' associated with that point (e.g., based on a table lookup). In another embodiment, the TEC controller can use the critical temperature difference to perform a table lookup of a data structure that associates critical temperature differences with voltage values. Once determined, process 40 applies a second voltage across the TEC element (at block 47).

[0071] Figure 6 is a flow chart of a process performed by an immersion cooling system to operate in a performance mode, according to one embodiment. In particular, the figure illustrates a process 60 in which the TEC controller 12 may operate in a performance mode (e.g., a second mode of operation) in which all heat generated by active IT components is transferred to a heat sink by the TEC element while the temperature of the IT components is at a minimum usable temperature, i.e., the immersion cooling system (the TEC element) may set the minimum usable temperature of the IT components while the TEC element is operating.

[0072] Process 60 begins by determining the power input of the active IT components on which the TEC elements and heat sinks are mounted (at block 61). In one embodiment, the controller may execute the Figure 4 Similar operations are performed in block 41 to determine the power input of the active IT components.

[0073] Process 60 determines the lowest available temperature T for the TEC element and the IT component based on the power input. lowThe associated operating current I* (at block 62). In particular, given the constants Q0 and T liquid,in , taking the derivative of the TEC current in equation (4), we get

[0074]

[0075] Where I is the operating current of the TEC element and R TEC is the resistance of the TEC element. In one embodiment, This corresponds to a decrease in IT component temperature as the operating current of the TEC element increases. In another embodiment, This corresponds to the increase in IT component temperature as the operating current of the TEC element increases. The lowest available temperature of the IT component is obtained and I can be solved for (where I = I*), which represents the performance mode.

[0076] In one embodiment, the TEC controller 12 is configured to determine T by analyzing the performance pattern map. low and I*, which corresponds to the lowest usable temperature of the IT component, such as Figure 7 shown.

[0077] Figure 7 An example of a performance pattern graph 70 according to one embodiment is shown. The graph 70 shows the temperature of the IT component 13 relative to the current of the TEC element. In particular, the graph includes a performance curve that is a graphical representation of equation (7) above. In another embodiment, the performance graph (or more particularly, the performance curve) can be a predefined graph. In another embodiment, the graph can be based on a determined power input of the IT component. For example, the TEC controller can retrieve (e.g., from a local memory) the graph 70, where the graph is associated with the determined power input. In other words, the TEC controller can store several graphs 70 in the local memory, each graph being associated with a different power input. Once the performance curve graph 70 is determined (and retrieved), the TEC controller determines the lowest point of the performance curve, which represents the T low and I*, as shown.

[0078] In some embodiments, T may be determined based on performing a table lookup on a data structure. low and I*, a data structure that relates the lowest usable temperature of the IT component to the input current of the TEC element.

[0079] return Figure 6, process 60 determines whether the lowest available temperature is less than the maximum temperature of the IT component (at decision block 63). Specifically, the TEC controller determines the maximum available temperature of the IT component, which may be a predetermined temperature (e.g., determined by the manufacturer of the IT component), and compares this value to the lowest available temperature. The TEC controller can therefore retrieve this value from (e.g., local) memory. If so, process 60 applies a voltage such that the determined operating current I* is supplied to TEC element 1 (at block 64). However, if the lowest available temperature is greater than the maximum temperature of the IT component, process 60 deactivates the activated IT component (at block 65). In one embodiment, TEC controller 12 deactivates the IT component because, if the lowest available temperature is greater than the maximum temperature, the TEC controller may not be able to adequately cool the IT component, which could ultimately lead to component overheating and failure.

[0080] As described herein, a TEC controller can operate in one of several operating modes. In one embodiment, the TEC controller 12 can be configured to determine which mode to operate in. Specifically, the TEC controller 12 can be configured to determine whether the TEC element is to operate (e.g., transfer heat away from the corresponding IT component) based on whether the TEC controller is to operate in the first operating mode or the second operating mode. For example, the TEC controller can be configured to operate in one of the operating modes based on whether the power input to the IT component exceeds a power threshold (e.g., a predefined threshold). For example, the TEC controller can be configured to select the first operating mode when the power input to the active IT component is less than the power threshold (which can be a predefined threshold). Conversely, the TEC controller can be configured to select the second operating mode when the power input to the active IT component is greater than the power threshold. Thus, the TEC controller can be configured to operate in either mode in response to changes in power input. In one embodiment, the TEC controller can change (or adapt) the operating mode based on real-time power readings of the IT component. In another embodiment, the TEC controller can be configured to operate in one of the operating modes based on user input. For example, the TEC controller may be configured to receive user input (eg, via a user interface such as a desktop computer to which the TEC controller is communicatively coupled) and, in response, select one of the operating modes based on the user input.

[0081] Figure 88 is a flow chart of a process performed by an immersion cooling system according to one embodiment. Process 80 begins by determining a power input to an active IT component mounted on an IT device, the active IT component contained within a coolant tank having a liquid coolant, wherein mounted on the IT component is a stack of thermoelectrically cooled TEC elements and a heat sink, wherein the TEC element is disposed between the IT component and the heat sink, and wherein the stack is at least partially immersed in the liquid coolant (at block 81). Process 80 determines whether the TEC element will operate according to a first operating mode or a second operating mode (at block 82). In response to determining that the TEC element will operate according to the first operating mode, process 80 calculates a critical temperature difference for the TEC element, determines the temperature difference and a voltage for the TEC element based on the power input, determines whether the temperature difference is greater than the critical temperature difference, and in response to determining that the temperature difference is greater than the critical temperature difference, applies a voltage across the TEC element (at block 83).

[0082] The cooling systems described so far are immersion cooling systems that are arranged to cool electronic devices using TEC elements that are at least partially immersed in a coolant. In one embodiment, any of the cooling systems described herein can be deployed as an air-cooled cooling system. For example, Figure 2 The cooling system 10 shown can be air-cooled so that the heat transferred to the fins 17 of the heat sink 15 by the TEC element 1 is discharged to the air flowing over and around the fins. In another embodiment, any of the equations shown so far can be modified for an air-cooled solution. For example, in equation (5), the temperature of the ambient air, T, can be used to determine the critical temperature difference. air,in (e.g., the air in the room housing the cooling system), rather than using the temperature of the coolant, T liquid,in .in addition, Figure 3 The liquid pump 28 of the illustrated CDU 25 may be one or more fans controlled by the CDU and configured to push (or pull) ambient air onto one or more heat sinks.

[0083] Some embodiments may vary the processes described herein. For example, at least some of the specific operations in the process may not be performed in the exact order shown and described. Specific operations may not be performed as a continuous series of operations, and different specific operations may be performed in different embodiments. In another embodiment, some operations may be optional and therefore may not be performed. For example, the operations performed in block 61 may be optional because equation (7) defines the heat of the IT component, Q0, as a constant that is discarded by the derivative.

[0084] As previously explained, embodiments of the present disclosure may be (or include) a non-transitory machine-readable medium (such as a microelectronic memory) having stored thereon instructions that program one or more data processing components (generally referred to herein as "processors") to perform cooling operations, such as selecting and operating in one of several operating modes, and controlling (or driving) one or more TEC elements based on the selected operating mode, as described herein. In other embodiments, some of these operations may be performed by specific hardware components containing hard-wired logic. Alternatively, these operations may be performed by any combination of programmed data processing components and fixed hard-wired circuit components.

[0085] In the foregoing description, embodiments of the present disclosure have been described with reference to specific exemplary embodiments thereof. It will be apparent that various modifications may be made thereto without departing from the broader spirit and scope of the present disclosure as set forth in the appended claims. Accordingly, the description and drawings are to be regarded as illustrative rather than restrictive.

[0086] While certain aspects have been described and shown in the drawings, it should be understood that these aspects are merely illustrative of the present disclosure and not limiting thereof, and that the present disclosure is not limited to the particular constructions and arrangements shown, as various other modifications may be made by one of ordinary skill in the art. Accordingly, the description should be regarded as illustrative and not restrictive.

[0087] In some aspects, the disclosure may include language such as “at least one of [element A] and [element B]”. The language may refer to one or more of the elements. For example, “at least one of A and B” may refer to “A”, “B”, or “A and B”. In particular, “at least one of A and B” may refer to “at least one of A and at least one of B” or “at least one of A or B”. In some aspects, the disclosure may include language such as “[element A], [element B], and / or [element C]”. The language may refer to any one of the elements or any combination thereof. For example, “A, B, and / or C” may refer to “A”, “B”, “C”, “A and B”, “A and C”, “B and C”, or “A, B, and C”.

Claims

1. An immersion cooling system comprising: a coolant tank having a liquid coolant and containing 1) a thermoelectric cooling (TEC) element at least partially immersed in the liquid coolant coupled to an IT component mounted on the information technology (IT) equipment, and 2) a heat sink coupled to the TEC element, wherein the TEC element is configured to transfer heat generated by the IT components to the liquid coolant via the heat sink; The immersion cooling system further includes a TEC controller mounted on the IT device and electrically coupled to the TEC element, wherein the TEC controller is configured to draw power from the IT device and is configured to control heat transfer of the TEC element by applying an input voltage across the TEC element; wherein the TEC controller is configured to sense a power input to the IT component and is configured to operate in one of a plurality of operating modes based on whether the power input to the IT component exceeds a power threshold; wherein, in response to the power input being less than a power threshold, the TEC controller is configured to operate in a first operating mode of a plurality of operating modes, wherein the input voltage is a minimum voltage that, when applied across the TEC element, causes the TEC element to transfer all heat generated by the IT component to the heat sink while the temperature of the IT component remains below a maximum temperature threshold, In response to the power input being greater than the power threshold, the TEC controller is configured to operate in a second one of the plurality of operating modes, wherein the applied input voltage causes the TEC element to transfer all heat generated by the IT component to the heat sink while the temperature of the IT component is at a minimum usable temperature.

2. The immersion cooling system of claim 1 , wherein the TEC element and the heat sink are an integrated unit.

3. The immersion cooling system of claim 1, wherein a thermal interface material (TIM) is disposed between at least one of: 1) the IT component and the TEC element, and 2) the TEC element and the heat sink.

4. A method for immersion cooling of information technology (IT) components, the method comprising: determining a power input of an active IT component mounted on the IT equipment, the active IT component contained within a coolant tank having a liquid coolant, wherein mounted on the IT component is a stack of thermoelectrically cooled TEC elements and a heat sink, wherein the TEC elements are disposed between the IT component and the heat sink, wherein the stack is at least partially immersed in the liquid coolant; determining whether the TEC element will operate according to a first operating mode or a second operating mode; as well as In response to determining that the TEC element is to operate according to the first operating mode, Calculate the critical temperature difference of the TEC element; Determine the temperature difference and voltage across the TEC element based on the power input, Determine if the temperature difference is greater than a critical temperature difference, and In response to determining that the temperature difference is greater than the critical temperature difference, a voltage is applied across the TEC element. 5 . The method of claim 4 , wherein determining the temperature difference and the voltage comprises identifying a maximum coefficient of performance (COP) of the TEC element associated with the temperature difference and the voltage using the power input.

6. The method of claim 4 , wherein the voltage is a first voltage, wherein the method further comprises, in response to determining that the temperature difference is less than a critical temperature difference, determining a second voltage based on the critical temperature difference; and A second voltage is applied across the TEC element.

7. The method of claim 4 , wherein the voltage is a first voltage, wherein the method further comprises, in response to determining that the TEC element is to operate according to the second operating mode, determining, based on the power input, an operating current for the TEC element associated with a lowest usable temperature of the IT component; Determine if the lowest available temperature is less than the maximum temperature of the IT component; as well as In response to determining that the lowest available temperature is less than the maximum temperature of the IT component, a second voltage associated with the operating current is applied across the TEC element. 8 . The method of claim 7 , further comprising, in response to determining that the lowest available temperature is greater than the maximum temperature of the IT component, deactivating the active IT component.

9. The method of claim 4, wherein determining whether the TEC element is to operate according to the first operating mode or the second operating mode comprises: When the power input of the active IT component is less than a power threshold, selecting a first operating mode; as well as When the power input of the active IT components is greater than a power threshold, the second operating mode is selected.

10. The method of claim 4, wherein determining whether the TEC element is to operate according to the first operating mode or the second operating mode comprises selecting one of the first operating mode and the second operating mode in response to a user input.

11. A non-transitory machine-readable medium having instructions stored therein, which, when executed by a processor, cause the processor to perform the method according to any one of claims 4 to 10.

12. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 4 to 10.

Images