Data center and procedures

Abstract

According to various embodiments, a data center (151) can comprise: a computer system (104) arranged in a room (102), wherein the computer system (104) comprises several heat sinks (104k) and several processors, each processor being thermally coupled to one of the several heat sinks (104k); at least one heat pump (222) arranged in the room (102); at least one gas-liquid heat exchanger (1041) arranged in the room (102); a hot liquid circuit (152h) which couples the several heat sinks (104k) to the gas-liquid heat exchanger (1041), wherein the hot liquid circuit (152h) further comprises a hot liquid connection (952) on a wall of the room (102); and a cold liquid circuit (152k) which couples the heat pump (222) with the gas-liquid heat exchanger (1041);wherein the heat pump (222) is configured to extract thermal energy from the cold liquid circuit (152k) and supply it to the hot liquid circuit (152h).

Description

[0001] Data centers are generally used to provide large amounts of computing power in a centralized manner, often employing thousands of processors. The resulting high volume of waste heat necessitates a suitable cooling concept to dissipate it economically and environmentally friendly. Traditionally, the focus has been solely on the efficiency of heat removal.

[0002] According to various embodiments, it has been recognized that conventional air-based and / or hybrid (water / air-based) cooling concepts are difficult to operate year-round, as they rely on the availability of cool fresh air, which can only provide the required cooling capacity during the summer months in a few climate zones. Furthermore, direct air exchange with the environment results in increased maintenance and energy consumption due to the necessary filters. If additional heat transfer with system separation to an external circuit occurs, the temperature of the medium to be cooled decreases further, thus increasing the requirement for a low-temperature heat sink. Various implementations of such hybrid cooling concepts are described in WO 2018 / 201425 A1 and US 2009 / 0133866 A1.US patent 2014 / 0 069 111 A1 describes a hybrid cooling concept based on thermoelectrically enhanced air and liquid cooling.

[0003] If no natural heat sink is available throughout the year, a compression refrigeration system is conventionally used.

[0004] With this system, condensation can occur at a higher temperature level, but this requires additional space due to the need for an outdoor unit that transfers the heat from the refrigerant to the outside air. Furthermore, the coefficient of performance (COP) of refrigeration systems decreases at high outside temperatures, resulting in increased electrical energy consumption.

[0005] Both conventional systems use air as the heat transfer medium. Due to air's low density and heat storage capacity, the amount of heat that can be transported is limited. Increasing the airflow rate and lowering the inlet temperature are only technically feasible and economically viable to a certain extent. Therefore, the power density of conventional systems is typically limited to less than 3 kW (kilowatts) per rack, or at least to significantly less than 80 kW per rack.

[0006] According to various embodiments, the task is solved to provide an improved infrastructure for dissipating thermal energy from a data center.

[0007] In various configurations, a data center, for example, for use as a containerized data center, is provided with an improved infrastructure for dissipating thermal energy (also known as cooling infrastructure) from the data center. This cooling infrastructure makes it possible to operate the data center year-round and / or even in hot climates. Accordingly, a method for operating the data center is provided.

[0008] The cooling infrastructure consists of two water circuits with different temperature levels, coupled by a heat pump. During operation, heat is extracted from the air within the room using the lower temperature level (i.e., it is cooled). This heat is then raised to a higher temperature level by the heat pump and fed into the higher temperature level, from which the heat is dissipated. The higher temperature level makes it easier to release the heat or put it to efficient use.

[0009] According to various embodiments, a data center (also referred to as a data center) can comprise: a computer system arranged in a room, wherein the computer system has several heat sinks and several processors, each processor being thermally coupled to one of the several heat sinks; at least one heat pump arranged in the room; at least one gas-liquid heat exchanger arranged in the room; a hot liquid circuit which couples the several heat sinks to the gas-liquid heat exchanger, wherein the hot liquid circuit further comprises a hot liquid connection on a wall of the room; and a cold liquid circuit which couples the heat pump to the gas-liquid heat exchanger; wherein the heat pump is configured to extract thermal energy from the cold liquid circuit and supply it to the hot liquid circuit.

[0010] To accommodate continuous growth, container data centers can be used, where the components of a data center are housed in a container. These containers can be prefabricated and pre-installed by manufacturers, thus enabling the efficient modular construction of larger data centers.

[0011] Depending on the specific design, the space can be set up as a container. In this case, the data center can also be referred to as a container data center. The effect of the container shape is greater mobility and modularity of the data center.

[0012] According to various embodiments, a container data center (container data center) can comprise the following: a container and a computing system arranged in the container, wherein the computing system comprises several heat sinks and several processors, each processor being thermally coupled to one of the heat sinks; wherein the container comprises at least one heat pump and a gas-liquid heat exchanger; a hot liquid circuit which couples the several heat sinks to the gas-liquid heat exchanger, wherein the hot liquid circuit further comprises a hot liquid connection on a wall (e.g., partition wall) of the container; a cold liquid circuit which couples the at least heat pump to the air-liquid heat exchanger; wherein the at least heat pump is configured to extract thermal energy from the cold liquid circuit and supply it to the hot liquid circuit.

[0013] They show Fig. 1 and Fig. 9 a RZ according to different embodiments in a schematic assembly diagram; Fig. 2 a method according to different embodiments in a schematic flowchart; Fig. 3 a container data center according to various embodiments in a schematic assembly diagram; Fig. 4 a data center according to different embodiments in a schematic supply diagram; Fig. 5 to 8 each a RZ according to different embodiments in various detailed assembly diagrams; Fig. 10 and Fig. 11 each a container data center according to different embodiments in a schematic perspective view; and Fig. 12 to 14 each describe a method according to different embodiments in a schematic flowchart.

[0014] The following detailed description refers to the accompanying drawings, which form part thereof and illustrate specific embodiments in which the invention can be implemented. In this context, directional terminology such as "top," "bottom," "front," "back," "anterior," "rear," etc., is used with reference to the orientation of the described figure(s). Since components of embodiments can be positioned in a number of different orientations, the directional terminology serves only for illustration and is in no way limiting. It is understood that other embodiments may be used and structural or logical modifications may be made without deviating from the scope of protection of the present invention.It is understood that the features of the various exemplary embodiments described herein can be combined with one another, unless specifically stated otherwise. The following detailed description is therefore not to be interpreted in a limiting sense, and the scope of protection of the present invention is defined by the appended claims.

[0015] Within the scope of this description, the terms "connected," "connected," and "coupled" are used to describe both direct and indirect connections, direct or indirect connections, and direct or indirect couplings. In the figures, identical or similar elements are designated with identical reference numerals where appropriate. According to various embodiments, the term "coupled" or "coupling" can be understood as a connection and / or interaction (e.g., mechanical, hydrostatic, fluid-conducting, thermal, and / or electrical), e.g., direct or indirect. Several elements can, for example, be coupled to one another along an interaction chain, along which the interaction (e.g., energy) can be transferred. For example, two coupled elements can exchange an interaction with each other, e.g.,A mechanical, hydrostatic, thermal, and / or electrical interaction. Depending on the specific embodiment, "coupled" can be understood as a mechanical (e.g., physical) coupling, for example, by means of direct physical contact. A coupling can be designed to transmit a mechanical interaction (e.g., force, torque, etc.).

[0016] Control can be understood as the intentional influencing of a system. The system's state can be changed according to a specification (also called the target state). Regulation can be understood as control, but with the additional benefit of counteracting changes in the system's state caused by disturbances. Visually, control can have a forward-directed control path, thus implementing a sequential control that converts an input variable (e.g., the specification) into an output variable. However, the control path can also be part of a control loop, thus implementing a regulation system. In contrast to pure forward control, regulation involves a continuous influence of the output variable on the input variable, which is effected by the control loop (feedback). In other words, regulation can be used as an alternative or in addition to control.Alternatively or additionally to control, regulation can be implemented. In regulation, the actual state of the controlled variable (e.g., determined based on a measured value) is compared with a reference value (e.g., the target state of the controlled variable). The controlled variable can then be influenced by a manipulated variable (using an actuator) to minimize the deviation of the actual state of the controlled variable from the reference value. The state of the system can be represented, for example, by at least one physical parameter of the system. Control and / or regulation can be performed by a control device, which can be implemented separately from the computer system or as part of it. The control device can be configured to carry out the procedure described herein.

[0017] The electrical power supplied to a data center is essentially converted 100% into thermal energy (also referred to simply as heat). The majority of this heat is typically generated within the servers by semiconductor components. Other heat sources in the data center include the electrical supply lines, heat losses in the (hot water) pipes, all electrical consumers, as well as fuses and electrical components in the circuit. Consumers also include all control and regulation units, valves, dampers, control electronics, lighting, network equipment, fire protection systems, UPS, batteries, accumulators (especially during discharge and charging processes), access control systems, and all security technology, including intrusion detection.

[0018] According to various designs, it has been recognized that it is not economically viable or even technically difficult to implement liquid cooling for all of the heat sources listed above. Consequently, their waste heat is dissipated by other means. If the waste heat is not dissipated, the temperature within a hermetically sealed space rises continuously, often far exceeding the permissible operating limits. To counteract this, the data center can incorporate a water-to-water chiller that feeds a warm water circuit to cool the heat contained in the air. The water-to-water chiller operates as a heat pump, meaning that heat is transferred from the chiller's condenser into the warm water circuit. This eliminates the need for a large outdoor unit and also enables the waste heat to be reused.The heat sink can, for example, be kept at a constant temperature level all year round, which makes it possible to select the chiller and its refrigerant in such a way that high efficiency can be achieved at this operating point.

[0019] According to various embodiments, a secondary air system can be provided, comprising one or more fans. The secondary air system can be configured to draw air from the room and return it to the room, for example, without exhausting the air from the room or exchanging it with other air, thus creating a closed airflow within the room. Each fan can, for example, provide an air volume in the range of 1000 to 11000 m³ / h. 3 / h (cubic meters per hour) at a pressure of 50 to 200 Pa (Pascals), for example an air flow rate in the range of 9000 to 11000 m³ 3 / h at a pressure of 75 to 175 Pa, or for example an air flow rate in the range of 9800 to 10200 m³ / h 3 / h at a pressure of 100 to 150 Pa. A fan operating in this range has proven to be particularly energy-efficient. For example, the fan can deliver an air flow of 10,000 m³ / h. 3 / h at a pressure of 100 to 150 Pa.

[0020] The computer system in each room contains one or more computing units (e.g., a series of racks) designed for high-density support of numerous processors and / or storage media. Processors can include server processors (CPUs), graphics processing units (GPUs), cryptographic processors, ASICs, FPGAs, TPUs (tensor processing units), or cryptocurrency mining hardware. Storage media can be mechanical hard disk drives (HDDs) or solid-state drives (SSDs).

[0021] The data center can have multiple racks (e.g., processor racks and / or shelves) for the high-density accommodation of a large number of processors. In other words, the data center enables a particularly efficient arrangement of processors, which on the one hand allows for high processor density, but on the other hand also ensures adequate cooling of the processors. In various configurations, the processors are arranged on corresponding processor cards, e.g., on electronic circuit boards with the electronics necessary for the processor, which are held in place by the racks. Processor cards can, for example, be graphics cards.The high-density arrangement corresponds to an arrangement that allows, in various embodiments, at least 8, at least 10, at least 12, or at least 16 processors, for example processor cards, to be arranged per meter in at least one row in a first frame and a second frame.

[0022] The computer system can consist of multiple racks (e.g., between 5 and 16 racks) housing server hardware. Each rack can, for example, accommodate a server according to a 19-inch standard or a 21-inch standard (e.g., the so-called "Open Compute" standard). The computer system can be configured to be primarily liquid-cooled.

[0023] In various embodiments, the racks are designed to accommodate processors in multiple horizontal rows spanning the entire height of the room. For example, suitable shelving units, such as those for processor racks and / or servers, can be used, with the horizontal rows filling the entire interior height of the room. This allows for a particularly high processor density and a large number of processors.

[0024] In various configurations, at least 30 processors or processor cards can be installed per m². 2 (square meters), for example at least 60 processors or processor cards per m² 2 e.g. at least 100 processors or processor cards per m² 2 e.g. at least 150 processors or processor cards per m² 2can be arranged. For example, such an arrangement of processors, which corresponds to a high-density arrangement, can be achieved by a corresponding arrangement in several horizontal rows.

[0025] To supply the computer system, the data center can have multiple supply paths with a supply interface configured to deliver at least one medium to the data center from an external source. This medium can be, for example, a cooling fluid (e.g., a coolant and / or a cooling gas), electrical energy, and / or a communication signal (e.g., a network signal). Each supply path can be configured to interact functionally with the computer system, passing on and / or exchanging the supplied medium with it. The set of supply and disposal paths within the room can also be referred to as the infrastructure. Depending on the type of medium (cooling fluid, electrical energy, and / or communication signal), the infrastructure can be called cooling infrastructure, power supply infrastructure (e.g., power supply infrastructure), or telecommunications infrastructure.For example, the cooling infrastructure can be set up to extract thermal energy from the computer system along the supply path (also known as cooling). Optionally, supply-critical paths or room components can be redundantly configured.

[0026] Redundancy refers to the presence of functionally identical or comparable resources in a technical system, not all of which are typically required during trouble-free operation. Functional redundancy can manifest itself in the design of multiple parallel supply paths required for operation, ensuring uninterrupted operation in the event of a failure of one supply path or during maintenance. Optionally, the redundant supply paths can be spatially separated from each other, for example, by protective walls and / or physical distance (e.g., by placing them on opposite sides of a room), to provide additional redundancy.

[0027] Redundancy of an element used for the operation of the computer system (e.g., a supply path, a component thereof, or a processor) can be understood, for example, as the existence of at least one functionally identical or comparable copy of the element, and the configuration of the element and its copy allowing switching between them, e.g., without interrupting the operation of the computer system. The element and its copy can then be intuitively configured to be redundant with each other (also referred to as a mutually redundant pair).

[0028] Switching between two mutually redundant elements (e.g., from a primary power supply to a redundant power supply) can be automated, for example, if a malfunction is detected in the active element. The malfunction might be deemed critical, meaning it could lead to a complete or partial failure of the computer system. The switchover can be performed, for example, by means of a failover switch (e.g., a transfer switch), and can be automated.

[0029] Pre-certification may, for example, require that the data center has at least partially redundant infrastructure. Alternatively or additionally, for example, if only some of the typical data center components are located in the room itself (e.g., transformers, generators, and the uninterruptible power supply (UPS) can be centralized and / or located outside of it), the redundant components and supply paths can fulfill at least some of the requirements for pre-certification (e.g., the room can have two redundant electrical sub-distribution boards and two supply paths), so that pre-certification essentially certifies the expected availability if the requirements are also met outside the room (e.g., classified as "supports availability class x," where x = 1, 2, 3, or 4).One effect of pre-certification is that the containers do not need to be certified again after they have been transported to the data center site. This allows the data center to be put into operation at its destination more quickly and with less effort.

[0030] Depending on the specific implementation, redundancy can be configured as N+1 redundancy. N+1 redundancy means that the computer system requires a maximum of (e.g., exactly) N power supply paths for operation, with at least N+1 power supply paths available in the room. The N+1 power supply path can be configured as a passive standby power supply path. If one of the N power supply paths fails or requires maintenance, its function can be taken over by the N+1 power supply path, for example, without interrupting the operation of the computer system. If two of the N power supply paths fail, this can result in a complete or partial failure of the computer system (corresponding, for example, to availability class "VK 3" according to DIN EN 50600 or "Tier 3" according to the North American Uptime Institute standard). This could be mitigated by using a higher level of redundancy, for example, by implementing parallel redundancy.Parallel redundancy requires at least 2·N supply paths, e.g. 2·(N+1) supply paths (corresponds, for example, to availability class “VK 4” according to DIN EN 50600 or “Tier 4” according to the North American standard of the Uptime Institute).

[0031] Single-path supply routes without component duplication can, however, comply with "VK 1" or "Tier 1", whereby, in addition to the availability classes, further structural engineering requirements (e.g., burglar protection, fire protection, etc.) may be defined in the standards. The data center, for example, can be built according to the "VK 2" or "Tier 2" standard or according to the "VK 3" or "Tier 3" standard.

[0032] In various embodiments, the room has a hermetically sealed (e.g., airtight) and / or thermally insulated interior (also referred to as an air circulation zone) in which the air is located and / or flows. This prevents air exchange between the air circulation zone and the exterior of the room. For example, the air may be in physical contact with the data center, one or more pieces of infrastructure, and / or the heat exchanger. The room can optionally be configured to release the hermetic seal in response to a disaster (e.g., a fire). However, the hermetic seal can also be maintained (e.g., unidirectionally) so that a gas extinguishing system located inside the room is more effective. For example, computer equipment may be located within this space.

[0033] This section describes various fluid circuits, each circulating a different fluid. The fluid can act as a cooling medium, meaning its flow can transfer thermal energy. The fluid circuits are categorized according to their temperature level (in other words, the temperature of the fluid) as hot fluid or hot fluid circuits and cold fluid or cold fluid circuits. The fluid can be water, for example, or be composed of water. However, other fluids can also be used, such as oil, petroleum, ethanol, or a liquid mixture of several fluids. Water has a very high specific heat capacity at a low density, which improves its heat dissipation capacity.

[0034] In various configurations, the data center is equipped with a chilled liquid circuit, the thermal energy of which (also referred to as heat) is transferred to the hot liquid circuit via a heat pump. One effect of this heat transfer is that the waste heat generated by computing operations can be supplemented by the waste heat from the air-cooled components of the room. This allows the majority (e.g., more than approximately 80%, 90%, or 95%) of the electrical power consumed in the data center to be absorbed by the hot liquid in the form of heat. This improves the data center's energy efficiency.

[0035] In various configurations, the temperature level of the hot liquid circuit can enable its use in district heating and cooling networks. These networks operate at temperatures in the range of 80-130°C. Waste heat from the data center can thus be fed into such a network, for example, via an interface that incorporates one or more additional heat pumps. Using the mechanisms described above, a very energy-efficient return temperature increase can be achieved in the return line of the district heating and cooling network, thereby significantly reducing the energy required (usually fossil fuels) for the subsequent increase to the target temperature.

[0036] Depending on the specific configuration, a data center is provided for outdoor use (also known as an outdoor data center). The data center can be based on a container and / or a modular design. Furthermore, the outdoor data center can be a completely self-contained system, insulated from the outside to minimize environmental influences such as thermals, humidity, and particles, thus increasing its potential applications. For example, the outdoor data center can be deployed largely independent of the climate zone.

[0037] For example, the proportion of heat that the server(s) dissipate into the fluid (e.g., water) relative to the server's total electrical power consumption can be greater than approximately 30%, e.g., approximately 50% or more, for example, in the range of approximately 80% to approximately 100%. Water as a cooling medium increases the maximum heat dissipation capacity (also known as heat dissipation power).

[0038] The power consumption of the computer system can be, for example, at least approximately 100 kW (kilowatts), approximately 150 kW or more, approximately 200 kW or more, approximately 250 kW or more, or approximately 500 kW or more. The power supply infrastructure can be configured in various ways to provide the computer system with an electrical power output of at least its power consumption (e.g., at least twice it). The cooling infrastructure can be configured in various ways to provide a heat dissipation capacity of at least its power consumption (e.g., at least twice it) by means of a hot liquid circuit.

[0039] The data center can have at least one server rack (also referred to simply as a rack) whose computing equipment is primarily liquid-cooled. The primary waste heat can be fed into the main supply (the so-called hot line) and can be transferred to the external interface (also referred to as the hot liquid connection) of the data center for waste heat recovery or highly energy-efficient cooling (e.g., free air cooling and / or hybrid cooling). A heat pump can provide a low-temperature coolant level (the chilled liquid) to maintain the necessary operating temperatures of the electronic components. The heat pump uses electrical power to cool the chilled liquid to this low temperature level (a relatively low temperature), which can then be transferred to the air within the room via a gas-liquid heat exchanger.The heat pump, powered by electrical energy, generates thermal output within the pump that can be fed back into the hot fluid circuit. This allows, for example, with additional thermal insulation, nearly 100% of the generated waste heat to be captured in the hot fluid and reused. The heat pump can optionally be configured with redundancy to prevent server downtime in the event of a heat pump failure.

[0040] According to various embodiments, a closed data center (e.g., comprising a container or other modular unit) is provided with predominantly liquid-cooled server components in combination with at least two supplementary and redundant heat pumps. This data center is largely independent of its environment, requiring very little maintenance and offering long periods of trouble-free operation.

[0041] According to various embodiments, at least one of the two heat pumps can be a controllable heat pump. This allows the room temperature to be controlled according to a setpoint, e.g., a different temperature is set for maintenance than for normal operation.

[0042] Optionally, a water-side control system can be implemented, which, for example, controls a shut-off valve downstream of a heat exchanger on the consumer side (i.e., the room outlet side) to maintain a predefined temperature (also referred to as the outlet temperature) of the hot fluid. If the temperature of the hot fluid drops, the control valve can reduce the flow rate of the hot fluid through the heat exchanger. This ensures that exactly the same amount of energy is dissipated on the water side as is converted into heat energy in the room.

[0043] According to various embodiments, the data center can comprise a room and, within it, at least one (single-row or multi-row) server array, extending, for example, parallel to the long side of the room (e.g., a container) or perpendicular to it. Optionally, the room can be subdivided by partition walls into an electrical compartment, a hydraulic compartment, a network compartment, and a server compartment. Optionally, at least one heat pump can be provided for each server array (e.g., computing unit).

[0044] Fig. Figure 1 illustrates a data center 151 according to various embodiments in a schematic diagram. The data center 151 can have a computing unit 104 inside 102i of a space 102 (for example, a container 102), which has a plurality (e.g., at least 10, at least 100, or at least 1000) of processors, each of which is coupled to a heat sink 104k. Each heat sink 104k can have a thermally conductive material (e.g., copper) through which a cavity is penetrated. Thermally conductive can be understood as a body and / or material having a thermal conductivity of more than approximately 200 (or 300) watts per meter and Kelvin (W / m·K). Each of the heat sinks 104k can be in thermally conductive contact with one or more processors. A thermally conductive contact can have a thermal resistance of less than 1 K / W and / or be dielectrically configured.The thermally conductive contact can be provided, for example, by means of a polymer.

[0045] Inside 102i of space 102, a heat pump 222 and a gas-liquid heat exchanger 1041 can be arranged. A heat pump 222 (also referred to as a heat engine) can be understood as a device configured to extract thermal energy from its colder input side 222e (also referred to as the cold side) and supply this energy (e.g., together with the drive energy) to its warmer output side 222a (also referred to as the hot side). In various embodiments, the heat pump has a fluid circuit 222k in which a suitable fluid (e.g., a fluid that can reach both a liquid and a gaseous state in the circuit 222k, e.g., propane) circulates. Components of the fluid circuit 222k include, for example, an evaporator, an expansion valve, a condenser, and a compressor.The fluid is selected, for example, such that it evaporates in the evaporator due to the heat input from the inlet side 222e and, after compression (in the compressor), condenses again in the condenser, releasing heat to the outlet side 222a. The throttle is used, for example, to enable controlled expansion of the fluid.

[0046] A heat exchanger 1041 can be understood as a device that transfers thermal energy from a first fluid stream to a second fluid stream. The heat exchanger 1041 can, for example, be passively configured, so that the heat flow is essentially caused only by the temperature difference between the first and second fluid streams. In a gas-liquid heat exchanger 1041 (also referred to simply as heat exchanger 1041), the first fluid can be a liquid and the second fluid a gas. For example, the heat exchanger 1041 can have a pipe 164 through which the liquid flows and a plurality of fins 166 through which the gas (e.g., air) flows. The fins can be thermally coupled to the pipe 164, for example, by extending the pipe 164 transversely through the fins. The heat exchanger 1041 can, for example, be passively configured.

[0047] Furthermore, the data center 151 (e.g., within room 102) can have a hot liquid circuit 152h and a cold liquid circuit 152k. A liquid circuit (e.g., the cold liquid circuit 152k or the hot liquid circuit 152h) can be understood here as a system designed to circulate a liquid along a flow path 155 through room 102. The liquid circuit can be coupled to at least one component (e.g., the heat pump 222) of the data center 151 such that the flow path 155 passes through this component, i.e., the liquid is supplied to and extracted from it. For example, several components (e.g., the heat pump 222 and the heat exchanger 1041) can be connected in series with respect to the flow path 155. Alternatively or additionally, several components (e.g. the heat sinks 104k) can be connected in parallel to each other with respect to the flow path 155.A pump in the fluid circuit can serve as a reference point, for example.

[0048] The hot liquid connection 952 can have an inlet flange 952e and an outlet flange 952a. This allows the liquid from the hot liquid circuit 152h to be extracted from chamber 102 (through the outlet flange) and returned to chamber 102 (through the inlet flange). The flow path at the hot liquid connection 952 can optionally be interrupted or branched.

[0049] The cold liquid circuit 152k generally has no connection or fewer connections on the outside of space 102 than the hot liquid circuit 152h.

[0050] The hot liquid circuit 152h can be coupled to the hot side 222a of the heat pump 222. The cold liquid circuit 152k can be coupled to the cold side 222e of the heat pump 222.

[0051] The cold liquid circuit 152k can connect the heat exchanger in series with the heat pump 222. The hot liquid circuit 152h can connect the multiple cooling sinks 104k in series with the heat pump 222.

[0052] For the sake of clarity, the following discussion refers to water as the hot and cold liquid. However, in principle, other liquids or mixtures of liquids can also serve as the hot or cold liquid. Water has a high heat capacity and can therefore provide high efficiency.

[0053] The data center 151 can optionally have several such rooms 102, as described below. The components that are thermally and / or fluid-conductively coupled to each other via the chilled liquid circuit 152k and the hot liquid circuit 152h can be part of the cooling infrastructure 114.

[0054] In various embodiments, the heat pump 222 is a high-temperature heat pump. Depending on the heat pump, heat can then be extracted from the cold liquid circuit 152k at a temperature (of the cold liquid) of, for example, at least 30°C, at least 40°C, at least 50°C, or at least 60°C, and this heat can be raised to a higher temperature level of, for example, at least 50°C, at least 60°C, at least 70°C, or 85°C, at which point it is supplied to the hot liquid circuit 152h.

[0055] According to various embodiments, the room 102 can be configured as a container 102. In this case, the walls of the room can be the container walls, the floor of the room the container floor, and the ceiling of the room the container ceiling. The following refers, among other things, to a room 102 that is configured, for example, as a container 102. The description can also apply analogously to a differently configured room 102, e.g., a room in a building and / or a room with stone walls.

[0056] Fig. Figure 2 illustrates a method 200 according to various embodiments in a schematic flowchart for handling multiple containers 102 of a container data center 151. Each container 102 can have a housing 1102g, which may have adjoining (e.g., four) side walls, a roof, and a floor that enclose the interior of the container. Optionally, several side walls 102s of the container 102 can be substantially fully opened (also referred to as loose walls 102s). The container 102 can furthermore have one or more infrastructures 702 inside for distributing a medium. For the sake of simplicity, the following may refer more generally to an infrastructure 702 of the container 102, whereby what is described for the infrastructure 702 can apply to the energy infrastructure, the cooling infrastructure, and / or the telecommunications infrastructure (e.g., by analogy).The same can also apply, by analogy, to one or more than one of the other infrastructures 702 of container 102.

[0057] The infrastructure 702 of each container 102 can be individually pre-certified 110 with regard to the reliability of the computer system 104. Optionally, multiple infrastructures 702, or the entire container as part of a data center, or the container with additional technical containers, can be pre-certified.

[0058] Pre-certification 110 can clearly illustrate the level of fault tolerance of the computer system. For example, fault tolerance (also referred to as availability) can be greater than 95%, e.g., at least approximately 98.97%, at least approximately 99%, at least approximately 99.9% (also referred to as high fault tolerance), at least approximately 99.99% (also referred to as very high fault tolerance), at least approximately 99.999%. Fault tolerance can be classified according to the certification type, i.e., divided into classes (also referred to as availability classes).

[0059] For example, pre-certification according to DIN EN 50600 (from 2013, e.g., DIN EN 50600-1 from 2013, or DIN EN 50600-2-2 from 2014, or DIN EN 50600-2-3 from 2015) and / or according to the American Tier classification (e.g., from 2015) can be used. Other (e.g., commercial) certification types can also be used, such as Bitkom certification (e.g., according to the Bitkom Guidelines 2013) or InfraOpt certification (from 2017). Depending on the certification type or availability class, various pre-certification requirements may need to be met, for example, at least N+1 redundancy (or 2·N redundancy) of the 702 infrastructure.

[0060] Method 200 may, in 101, comprise: providing at least one (i.e., one or more than one) container 102. The provision of the 101 may optionally, in 103, comprise: relocating the at least one container 102, e.g., by land, water, and / or air. Method 200 may, for example, in 103 comprise: arranging several containers 102 relatively close to one another such that any two containers 102 of the multiple containers 102 are arranged directly adjacent to each other. For example, these can be arranged with at least two (e.g., end-facing or longitudinal) loose walls 102s facing each other. Method 200 may, for example, in 105 comprise: opening one of the multiple loose walls 102s of the container or each container 102, which, for example, faces another container 102 of the multiple containers 102. Each container 102 can be configured in such a way that, upon opening the loose wall 102s, the pre-certification of the container 102 (e.g.whose energy infrastructure 106) is maintained. For this purpose, each loose wall 102s of container 102 can, for example, be free of elements that affect pre-certification, e.g., elements that influence compliance with the requirements of the pre-certification. This makes it possible to achieve, clearly, that after connecting the interiors of the multiple containers 102, no certification of container 102 is necessary. This accelerates the deployment of the data center 151, which contains the multiple containers 102.

[0061] The procedure 200 can include in 107: coupling of at least one container 102 to each other and / or to at least one external supply arrangement 202. The supply arrangement 202 or each supply arrangement 202 can have one or more than one supply structure (also referred to as supply), e.g. optionally a telecommunications supply 202t, optionally a power supply 202z, optionally a hot water supply 202k (more generally a hot liquid supply) and / or optionally a gas extinguishing supply 202f. The coupling of an infrastructure 702 can more generally be effected by means of a supply interface 722, to which the infrastructure has corresponding connections.

[0062] For example, the telecommunications infrastructure has several network lines that connect the supply interface 722 to the computer system 104 to connect the numerous processors to a local and / or global network (e.g., the Internet). Similarly, the cooling infrastructure 114 has several pipes (e.g., supply and return pipes) that connect the supply interface 722 to the computer system, allowing thermal energy to be extracted from the interior of the container.

[0063] Fig. Figure 3 illustrates a container-RZ 151 according to various embodiments in a schematic assembly diagram 300. In the container-RZ 151, the space 102 or each space 102 can be configured as a container 102.

[0064] Each container (e.g., an ISO container) used by the data center can be designed or manufactured in accordance with ISO standard 668. This standardizes and simplifies the transport of containers by ship, rail, and truck. Depending on the configuration, the container can have an external length of 13.716 m (45 ft), 12.192 m (40 ft, e.g., as a standard container or shipping container), 9.125 m (30 ft), 6.058 m (20 ft, e.g., as a standard container or shipping container), 2.991 m (10 ft), 2.438 m (8 ft), or 1.968 m (6 ft), an external height of 2.591 m (e.g., as a standard container) or 2.896 m (also known as a high-cube container), and an external width of 2.438 m. For example, a so-called 20ft container has an external length of 6,058 m, an external height of 2,591 m, and an external width of 2,438 m. A so-called 40ft container (e.g., a 40ft HC container) has an external length of 12,192 m and an external height of 2,591 m.896 m and an external width of 2,438 m. For example, the container could have external dimensions (length x width x height) of 6,058 m x 2,438 m x 2,896 m. The container could have internal dimensions (length x width x height of the interior) of 5,853 m x 2,342 m x 2,697 m.

[0065] The casing 1102g of the container 102 can have at least two (e.g., three) loose walls 102s, optionally one or more fixed walls 112s. The fixed wall 112s can, for example, run along a longitudinal dimension of the container 102 and / or be arranged on a longitudinal side of the container 102. The fixed wall 112s and the second loose wall 102s (also referred to as longitudinal side walls) can be arranged opposite each other. Furthermore, the first loose wall 102s and the third loose wall 102s (also referred to as end walls) can be arranged opposite each other.

[0066] In one embodiment, each substantially fully opening side wall (also referred to as a loose wall) of the container is designed as a multi-panel, foldable, or demountable wall. Alternatively or additionally, the loose wall can be designed so that it can be reclosed in accordance with ISO standard 668, and / or so that the container cannot be distinguished from other containers. In this configuration, shipping the container via standard freight forwarders and routes is thus easily facilitated using trucks, trains, and ships.

[0067] Container 102 can also have one or more than one infrastructure 702 (e.g., the energy infrastructure 106, the cooling infrastructure 114 and / or the telecommunications infrastructure).

[0068] Furthermore, the housing of container 102 can have several partition walls 102z, as described in more detail below. Each of the partition walls can, for example, have thermal insulation and / or hermetically seal the interior 102i of container 102.

[0069] Between the computer system 104 and the first and third partition walls 102s, for example, an end-facing partition wall 102z can be arranged. The computer system 104 and / or the supply lines 7021 of the infrastructure 702 can be arranged between the two end-facing partition walls 102z. Each end-facing partition wall 102z can optionally have a door opening 712 in which, for example, a door 712t (also referred to as a personnel door 712t) can be arranged. The personnel door 712t can be a security door. The security door 712t can be lockable, thermally insulated, and / or fire-resistant and / or smoke-tight.

[0070] Furthermore, the infrastructure 702 can have at least one pair (e.g., two pairs) of mutually redundant supply paths 702u, each pair of which connects the supply interface 722 to the computer system 104. For example, each computing unit 104a, 104b of the computer system 104 can be connected to a pair of mutually redundant supply paths 702u. For this purpose, the computing unit 104a, 104b can be configured, for example, to switch between a pair of mutually redundant infrastructure connections 704n (e.g., per computing unit 104a, 104b) of the computer system 104. Alternatively or additionally, the infrastructure 702 can be configured to switch between the mutually redundant supply paths 702u of a pair. The switching can be carried out, for example, by means of an automatic failover switch.

[0071] For example, each supply interface 722 can have at least one pair of mutually redundant connections 412a, 412b, of which a first connection 412a is connected to the computer system 104 (e.g., each computing unit 104a, 104b) by means of at least one first supply line 7021, and of which a second connection 412b is connected to the computer system 104 (e.g., each computing unit 104a, 104b) by means of at least one second supply line 7021. Each of the supply paths can, for example, have several supply lines and / or a distribution unit that connects the several supply lines to the supply interface 722.

[0072] Optionally, an additional partition wall 102z can be arranged against the fixed wall 112s and support one or more components of the container 102, such as the infrastructure 702, a user interface, or similar. The additional partition wall 102z allows the fixed wall 112s to remain unchanged and / or provides additional thermal insulation for the container 102.

[0073] One or more than one medium can be provided locally by means of the supply arrangement 202, e.g. hot water, a low voltage 400 V (alternating current - AC) generated from a medium voltage (by means of a transformer), an uninterruptible current (e.g. by means of UPS), an optional extinguishing gas.

[0074] The infrastructure 702 can be set up to meet the requirements of availability class 2 or higher (e.g. availability class 3) with regard to the reliability of the computer system 104 and be pre-certified accordingly, e.g. according to Tier and / or according to DIN EN 50600.

[0075] The housing structure (e.g., fixed wall) of container 102 can be made of steel and may optionally include one or more personnel doors. The container frame can have four corner steel beams and their horizontal steel connecting beams (and optionally the floor structure), which adjoin the partition walls 102z.

[0076] The energy infrastructure 106 can include two separate sub-distribution units 106u (UVs) and / or separate cable trays to meet requirements such as availability class 2 (e.g., Tier 2) and above (e.g., UPS power supply A and B). The cable trays can run continuously through the entire container 102 in the raised floor to provide, for example, two redundant power supply paths from the power supply interface 722 on opposite ends of the container 102. Each power supply path or UV 106u can be configured to provide a power output of at least 250 kW (kilowatts) or less. For example, the cross-section of the supply lines in each power supply path can be configured to provide the power output at either 220 V (volts) or 110 V.Alternatively or additionally, each supply path of the energy infrastructure 106 can be set up to provide a power of approximately 250 kW or more per 6 meters of longitudinal extension of the container 102 and / or to provide a total of approximately 500 kW or more (e.g. with less or no redundancy).

[0077] Fig. Figure 4 illustrates a data center 151 according to various embodiments in a schematic supply diagram 400 with a schematic redundancy pairing 901. The data center 151 can have a supply arrangement 202 and the room 102. However, the room 102 can also be provided without the supply arrangement 202. The supply arrangement 202 can have one or more supply modules for supplying the data center 151, wherein each supply module can optionally have a container in which the corresponding supply structure is arranged. The supply interface 722 can be arranged on a wall of the data center 151, e.g., on an intermediate wall of the container 102.

[0078] The supply arrangement 202 can optionally be coupled with a medium-voltage main distribution board 604 and / or an optional fuel supply 606 (e.g., supplying gas or diesel). The supply arrangement 202 can include several modularly provided supply devices (then also referred to as modules), e.g., a transformer 612, a power generator 220, a low-voltage main distribution board 218, an uninterruptible power supply 216, a standard power distribution board 616, a hot water supply 618, a cooling tower 620, and / or a heat pump system 422. In addition to the heat pump system 422, e.g., as an alternative to the cooling tower 620, a district heating or cooling line 222f can be connected to supply the waste heat for use.

[0079] The cooling infrastructure 114 (e.g., comprising air conditioning 1041) can optionally have at least one pair of mutually redundant supply paths, each supply path having a hot water connection 952 (e.g., having flanges) at the supply interface 722. The power infrastructure 106 can optionally have at least one pair of mutually redundant supply paths, each supply path having at least one sub-distribution device 106u (UV) and / or at least one power supply connection 916 at the supply interface 722.

[0080] The telecommunications module 202t can, for example, have two mutually redundant telecommunications connections 214. Accordingly, the telecommunications infrastructure can have at least one pair of mutually redundant supply paths, each of which has at least one network line and / or one network connection at the supply interface 722 (e.g., via telecommunications interface 924s). Each computing unit 104a, 104b of the computer system 104 can optionally be coupled to each pair of mutually redundant supply paths of the telecommunications infrastructure 914, the power infrastructure 106, and / or the cooling infrastructure 114.

[0081] The RZ 151 according to supply diagram 400 can, for example, meet the requirements of availability class 3.

[0082] Fig. 5, Fig. 6, Fig. 7, Fig. Figures 8 illustrate an RZ 151 (e.g., Container-RZ 151) according to various embodiments in different detailed assembly diagrams 500 to 800.

[0083] According to various embodiments, a compact, enclosed, highly energy-efficient, and as low-maintenance as possible fail-safe data center 151 (e.g., a mobile container data center 151) is enabled. For this purpose, a cooling infrastructure 114 is provided, which allows the generated heat to be absorbed in the liquid and thus separates the computing technology from the environment.

[0084] The thermal and mechanical separation of the computing units 104a and 104b from the environment allows for extremely energy-efficient operation and great flexibility in implementation across a wide range of climates and under diverse weather conditions. Furthermore, the use of the cooling infrastructure 114 enables the energy-efficient utilization of the generated heat for industrial processes or district heating of buildings.

[0085] The data center 151, for example its hot liquid cooling system, has one or more hot liquid circuits 152h, each of which supplies the server racks 602 of the computer system 104. Optionally, one or more heat exchangers 1041 can be arranged either at the housing level or at the rack level. To extract the waste heat released into the air, one or more heat pumps 222 can be used.

[0086] Components of a liquid circuit (e.g., the hot liquid circuit 152h and / or the cold liquid circuit 152k) can be: several supply lines 7021 (e.g., pipelines), one or more than one valve 702b (e.g., ball valve) and / or one or more than one pump 702p. Optionally, the fluid circuit can include one or more of the following components: one or more than one flow sensor 702s, one or more than one quick coupling 552, one or more than one flow actuator 554, one or more than one expansion tank 556, one or more than one pressure relief valve 558, one or more than one pressure sensor 560, one or more than one flange 566, one or more than one vent 562, one or more than one temperature sensor 654, one or more than one flow sensor 668, one or more than one filling valve 670, and / or one or more than one check valve 672.Each pair of supply lines 7021 that provides a flow path through the heat pump 222 can provide a supply path.

[0087] Each 552 quick coupling can optionally have a valve (e.g., a check valve) that stops the flow when the quick coupling is uncoupled.

[0088] By means of one or more partition walls 102z, several (e.g., spatially and / or thermally) separated (e.g., gas-separated) areas can be provided, e.g., a first area 662 in which at least one heat pump 222 is located, e.g., a second area 664 in which the computer system 104 is located, e.g., a third area 666 in which the UV or the network is located, e.g., a fourth area 762 which provides an airlock. Optionally, a heat exchanger 1041 can be arranged in each of these areas.

[0089] The heat exchanger 1041 can, for example, be part of a secondary air system, which also includes at least one fan. The fan can be configured to blow air from the relevant area through the heat exchanger 1041 and into the area. Secondary air, which is drawn from a room or area and, after cooling, returned to the same room or area, is to be distinguished from recirculated air.

[0090] The fluid supply of the supply arrangement 202 can, for example, include one or more additional heat pumps 422.

[0091] Optionally, instead of a single heat pump 222, two mutually redundant heat pumps 222 can be used. Alternatively or additionally, instead of a single heat pump 222, two cascaded heat pumps 222 can be used. For example, two pairs of heat pumps 222 can be configured redundantly, with the heat pumps 222 of each pair cascaded together. In the case of two cascaded heat pumps 222, the hot side of one heat pump can be coupled to the cold side of the other heat pump, e.g., by means of a fluid circuit 222k. The heat pumps 222 cascaded in this way can act together like a larger heat pump 222, providing a greater temperature difference. In the case of two redundant heat pumps 222, their hot sides can be coupled to each other, and their cold sides can be coupled to each other, e.g., each by means of the decoupled fluid circuit 152h, 152k.

[0092] One or more heat pumps 222 can make it possible to remove heat from the air without having to use a conventional air conditioner, which, for example, requires an exhaust duct to the outside.

[0093] The heat pump 222, or any (e.g., at least one) can optionally be controllable (also referred to as modulating), meaning that its pumping capacity can be set and / or regulated. For example, the pumping capacity can depend on an operating point of the heat pump 222, whereby the operating point of the heat pump 222 can be changed by means of an actuator (e.g., a power controller). For example, the modulating heat pump 222 can be provided with several operating points, e.g., at least three (e.g., at least ten), or its operating point can be continuously changed.

[0094] Alternatively (e.g., if regulation is not possible) or additionally, a hot fluid return system or a cold buffer storage tank can be coupled with the heat pump 222. Alternatively or additionally, the heat pump 222 can be operated in a pulsed mode, so that its pumping output is calculated as an average over several cycles.

[0095] The chilled liquid circuit 152k (also referred to as chilled liquid network or chilled liquid piping) can extract heat from at least one heat exchanger 1041. For example, one or more heat exchangers 1041 can be located inside or next to (e.g., above) a server rack 602. For example, one or more gas-liquid heat exchangers 1041 can be located in a wall (e.g., floor, ceiling, and / or side wall) of room 102. For example, one or more gas-liquid heat exchangers 1041 can be configured as a so-called side cooler (i.e., next to the racks) or alternatively or additionally as a so-called front-door cooler or so-called back-door cooler.

[0096] If one or more gas-liquid heat exchangers 1041 are arranged above the racks 602 (e.g., above each rack 602), this simultaneously enables separation (visually, gas separation) between the hot aisle 706 and the cold aisle 708. The cold aisle 708 can be located at the front of the servers, while the hot aisle can be located at the rear of the servers.

[0097] If the data center 151 has exactly one heat pump 222 or two redundant heat pumps 222, the chilled liquid circuit 152k can include a fresh water heat exchanger, which provides additional cooling for the servers. In this case, the chilled liquid circuit 152k can be cooled alternatively or additionally to the heat pump 222 by means of the fresh water heat exchanger, which is supplied with fresh water (for example, with a temperature between 5 and 15°C) from outside room 102. The fresh water can be taken, for example, from a body of water, a municipal water supply, or groundwater. For system separation, a heat exchanger can then be installed in the chilled liquid circuit 152k, through which the chilled liquid flows alternatively or additionally to the heat pump 222.

[0098] On the output side, the RZ 151 (e.g., its container) essentially has the following interfaces: a power connection 916, a cold liquid connection 954 (e.g., for water with a temperature between 4°C and 60°C), and a hot liquid connection 952 (e.g., for water with a temperature between 35°C and 95°C). The hot liquid connection 952 can, for example, receive water with a temperature in the range of approximately 45°C to approximately 55°C and discharge water with a temperature in the range of approximately 60°C to approximately 70°C or higher.

[0099] One or more heat pumps 222 and the processors of the computer system 104 feed thermal energy into the hot liquid circuit 152h, so that a heat flow from the heat pump(s) 222 and from the computer system out of room 102 is provided. The feed into the heat pump 222 can be connected upstream of the servers 104s, downstream of the servers 104s, or in parallel with the servers. According to various embodiments, the heat flow can be controlled and / or regulated.

[0100] The regulation can be implemented, for example, in such a way that the cooling of server 104s and all other components of data center 151 (e.g., within room 102) is highly reliable and has priority.

[0101] The control can be achieved, for example, by setting the room temperature (e.g., air temperature) to a value (setpoint) that is lower than the maximum operating temperature of all components used within room 102. A room temperature (RT) that is as high as possible leads to the energy-efficient operation of the heat pumps 222, meaning that a high coefficient of performance (COP) is achieved. Alternatively or additionally, a high RT minimizes the heat losses through all water pipes of the hot liquid circuit 152h. For example, heat can flow from the room air into the colder liquid pipes of the hot liquid circuit 152h and / or the cold liquid circuit 152k.

[0102] The control system also allows the room temperature to be lowered further for maintenance purposes, creating moderate working conditions. For example, a room temperature of less than 35°C can be provided or set up for maintenance.

[0103] The RZ 151 can optionally be equipped with the 762 airlock to provide greater independence from external influences, such as weather conditions like snowfall and / or sandstorms. One or more RZ 151 containers can be installed, for example, inside a building or outdoors.

[0104] The space 102 (e.g., the housing 1102g of container 102) can have a closed, thermally separating shell, which may, for example, contain or be formed from an insulating material (e.g., fibers and / or foam). For example, the housing 1102g of container 102 may contain or be formed from a bioplastic substitute, such as WPC. The housing 1102g may, for example, be prefabricated, e.g., according to an ISO container standard.

[0105] The frame of the computer system 104 does not necessarily have to include racks or server cabinets, but can, for example, consist of only a frame structure and optionally accommodate the power and water distribution systems for the computer system 104. The interior of room 102 can serve as a cold air supply via a heat exchanger 1041, which can be coupled to the cold side of the heat pump(s) 222 by means of the chilled liquid circuit 152k and thus act as a heat source for the heat pump(s) 222. This ensures that lossy electrical components are adequately cooled and the resulting heat is absorbed in the cooling medium of the chilled liquid circuit 152k. This enables subsequent, very efficient heat dissipation, e.g., further use of the heat in industrial processes or the heating of buildings via district heating.

[0106] Fig. Figure 9 illustrates an RZ 151 according to various embodiments in a schematic diagram 900. The RZ 151 can have exactly one heat pump 222 in diagram 900. The heat exchangers 1041 can be part of passive heat sinks.

[0107] Fig. 10 and Fig. Figures 11 illustrate a container-RZ 151 according to different embodiments in a schematic perspective view 1000, 1100, highlighting components of the hot liquid circuit 152h and the cold liquid circuit 152k.

[0108] Fig. Figure 12 illustrates a method 1200 according to various embodiments in a schematic flowchart.

[0109] Method 1200 can include: in 1201, introducing a hot liquid into the hot liquid circuit by means of the hot liquid connection; in 1203, heating the hot liquid in the hot liquid circuit by supplying thermal energy to the hot liquid from the several cooling sinks and from the at least one heat exchanger; in 1205, removing the heated hot liquid from the hot liquid circuit by means of the hot liquid connection.

[0110] Fig. Figure 13 illustrates a method 1300 according to various embodiments in a schematic flowchart.

[0111] Method 1200 can include: in 1301, first transfer of thermal energy from air to a first liquid (also referred to as a cold liquid), wherein the air is arranged in a room in which a computer system is further arranged (and, for example, a heat pump is arranged), wherein the computer system and / or the heat pump are, for example, in contact with the air; in 1303, second transfer of thermal energy from the first liquid to a second liquid (also referred to as a hot liquid) which has a higher temperature than the first liquid, for example, by means of the heat pump; in 1305, third transfer of thermal energy from the computer system (e.g., its processors) to the second liquid; and in 1307, transport of the second liquid along a flow path that passes through the room.

[0112] The transport process can consist of: removing the second liquid from the space and then re-entering the space along the flow path. The flow path can, for example, be closed. The cold liquid can, for instance, be guided (e.g., only) along an additional flow path located within the space.

[0113] The first transfer, the second transfer and / or the third transfer can, for example, take place within room 102.

[0114] Fig. Figure 14 illustrates a method 1400 according to various embodiments in a schematic flowchart, which includes, for example, method 1200 and / or method 1300.

[0115] Method 1400 may include in 1401: Controlling and / or regulating the temperature (also referred to as hot temperature) of the hot fluid, e.g., according to a setpoint. Controlling and / or regulating the hot temperature may include: Increasing the flow rate of hot fluid through at least one heat pump when the hot temperature meets the setpoint (e.g., is equal to or greater than a target temperature); and / or decreasing the flow rate of hot fluid through the at least one heat pump when the hot temperature does not meet the setpoint.

[0116] Alternatively or additionally, method 1400 in 1403 can include: controlling and / or regulating the temperature (also referred to as room temperature) of the air within the room. Controlling and / or regulating the room temperature can include: increasing the electrical power consumption of at least one heat pump when the room temperature meets the setpoint (e.g., is equal to or greater than a target temperature); and / or decreasing the electrical power consumption when the room temperature does not meet the setpoint. The electrical power consumption can be changed, for example, by means of an actuator of the heat pump. The actuator can, for example, set one or more operating parameters of the heat pump, which define the operating point of the heat pump. The operating parameter can, for example, be the pumping capacity of the heat pump. The at least one heat pump can thermally couple the hot fluid with the cold fluid.

[0117] Alternatively or additionally, the room temperature control and / or regulation can include: controlling the chilled liquid circuit 152k according to a setpoint temperature (e.g., of the air within room 102); switching the setpoint temperature between maintenance mode and normal operation, where the setpoint temperature in maintenance mode is less than 35°C and the setpoint temperature in normal operation is greater than approximately 35°C (e.g., 40°C, 45°C, 50°C, or 55°C). The actual temperature can be measured, for example, by one or more temperature sensors. Switching can be triggered, for example, by input via a user interface, remotely, and / or in response to the detection of a person within room 102.

[0118] The following are various examples that refer to what has been described previously and depicted in the figures.

[0119] Example 1 is a data center, exhibiting: A computer system arranged in a room, wherein the computer system comprises several heat sinks and several processors, each processor being thermally coupled to one of the heat sinks; at least one heat pump arranged in the room; at least one gas-liquid heat exchanger arranged in the room; a hot liquid circuit which couples the several heat sinks to the gas-liquid heat exchanger, the hot liquid circuit further comprising a hot liquid connection on a wall of the room; and a cold liquid circuit which couples the heat pump to the gas-liquid heat exchanger; wherein the heat pump is configured to extract thermal energy from the cold liquid circuit and supply it to the hot liquid circuit.

[0120] Example 2 is the data center according to Example 1, where the room is set up as a container, for example, where the container is set up as a container according to ISO standard 668.

[0121] Example 3 is the data center according to one of Example 1 or 2, further comprising: a hot liquid supply outside the room, wherein the hot liquid supply is coupled to the hot liquid connection.

[0122] Example 4 is the data center according to Example 3, where a closed flow path through the hot liquid circuit is provided by means of the hot liquid supply.

[0123] Example 5 is the data center according to one of Examples 1 to 4, wherein a flow path is provided from the hot liquid connection through the multiple cooling sinks and the heat exchanger back to the hot liquid connection by means of the hot liquid circuit.

[0124] Example 6 is the data center according to one of Examples 1 to 5, further comprising: a secondary air system within the room, which is set up to provide an airflow to the gas-liquid heat exchanger, e.g. by means of a fan.

[0125] Example 7 is the data center according to Example 6, where the airflow circulates only within the room.

[0126] Example 8 is the data center according to one of Examples 1 to 7, wherein the gas-liquid heat exchanger is set up to exchange thermal energy between a gas within the room and the chilled liquid circuit.

[0127] Example 9 is the data center according to one of Examples 1 to 8, wherein a heat pump or at least one heat pump with a modulating operating point is installed.

[0128] Example 10 is the data center according to Example 9, wherein the heat pump has an actuator which modulates the operating point, wherein the actuator can switch between at least two states (e.g. at least three states and / or stepless), according to which the operating point is provided.

[0129] Example 11 is the data center according to one of Examples 1 to 10, further comprising: an energy supply infrastructure within the room to supply the computer system with electrical energy.

[0130] Example 12 is the data center according to Example 11, wherein the room's power supply infrastructure is pre-certified with regard to the reliability of the computer system; and / or wherein the gas-liquid heat exchanger is set up to extract thermal energy from the power supply infrastructure (e.g. by means of an airflow within the room).

[0131] Example 13 is the data center according to one of Examples 1 to 12, wherein a cooling infrastructure of the room, which includes the heat pump, the gas-liquid heat exchanger, the hot liquid circuit and the cold liquid circuit, is pre-certified with regard to the reliability of the computer system.

[0132] Example 14 is the data center according to one of Examples 1 to 13, wherein the hot liquid circuit has at least one expansion tank, for example a membrane expansion tank.

[0133] Example 15 is the data center according to any one of Examples 1 to 14, further comprising: a failover switch, wherein the at least one heat pump comprises a pair of heat pumps configured redundantly to each other; wherein the failover switch is configured to switch to the other heat pump of the pair in response to a failure of one heat pump of the pair.

[0134] Example 16 is the data center according to one of Examples 1 to 15, wherein the room is arranged such that an air circulation area of ​​the room, adjacent to the computer system and / or the heat exchanger, is hermetically sealed against the outside of the enclosure.

[0135] Example 17 is the data center according to Example 16, wherein the housing has several side walls surrounding the interior, each of which has thermal insulation (e.g., having an insulating material).

[0136] Example 18 is the data center according to one of Examples 1 to 17, wherein the hot liquid circuit is set up such that in operation a hot liquid temperature of the hot liquid circuit (i.e. a temperature of the hot liquid) is greater than a cold liquid temperature of the cold liquid circuit (i.e. a temperature of the cold liquid).

[0137] Example 19 is the data center according to Example 18, wherein the hot liquid temperature delivered by the hot liquid circuit via the hot liquid port is in a range of approximately 55°C to approximately 80°C; and / or wherein the hot liquid temperature received by the hot liquid circuit via the hot liquid port is in a range of approximately 5°C to approximately 55°C; and / or wherein the cold liquid temperature is in a range of approximately 10°C to approximately 50°C.

[0138] Example 20 is the data center according to one of Examples 1 to 19, further comprising: a control device which is configured to control a flow of the hot liquid circuit according to the hot liquid temperature and / or the at least one heat pump according to the cold liquid temperature, wherein, for example, the control device is configured to switch the cold liquid temperature between at least two temperatures, wherein, for example, the control device is configured to switch the cold liquid temperature according to an operating state (e.g., between normal operation and maintenance operation) of the data center.

[0139] Example 21 is the data center according to one of Examples 1 to 20, wherein the data center has multiple racks, each rack holding a plurality of processors.

[0140] Example 22 is the data center according to one of Examples 1 to 21, wherein each rack is set up such that more than 30 processors per square meter of floor space of the computer system (e.g. the rack) are held.

[0141] Example 23 is the data center according to one of Examples 1 to 22, where the heat extraction capacity of the hot liquid circuit is greater than the maximum electrical power consumption of the data center.

[0142] Example 24 is the data center according to one of Examples 1 to 23, wherein the cold liquid circuit is hermetically sealed to the outside of the room.

[0143] Example 25 is a method for operating a data center according to one of Examples 1 to 24, comprising the method of: introducing a hot liquid into the hot liquid circuit by means of the hot liquid connection; heating the hot liquid in the hot liquid circuit by supplying thermal energy to the hot liquid from the several cooling sinks and from the at least one heat exchanger; removing the heated hot liquid from the hot liquid circuit by means of the hot liquid connection.

[0144] Example 26 is the method according to Example 25, wherein the temperature of the hot liquid being discharged from the hot liquid circuit is in a range of approximately 55°C to approximately 80°C; and / or wherein the temperature of the hot liquid being discharged into the hot liquid circuit is in a range of approximately 5°C to approximately 55°C.

[0145] Example 27 is a method comprising: first, transferring thermal energy from air to a first liquid, wherein the air is arranged in a room in which a computer system is also arranged (and, for example, a heat pump is arranged); second, transferring thermal energy from the first liquid to a second liquid which has a higher temperature than the first liquid, for example, by means of the heat pump; third, transferring thermal energy from the computer system to the second liquid; removing the second liquid from the room and returning the second liquid to the room along a closed flow path.

Claims

[1] Data center (151), comprising: • a computer system (104) arranged in a room (102), the computer system (104) having a plurality of heat sinks (104k) and a plurality of processors, each processor being thermally coupled to a heat sink (104k) of the plurality of heat sinks (104k); • at least one heat pump (222) arranged in the room (102); • at least one gas-liquid heat exchanger (1041) arranged in the space (102); • a hot liquid circuit (152h) which couples the plurality of heat sinks (104k) to the gas-liquid heat exchanger (1041), wherein the hot liquid circuit (152h) further comprises a hot liquid connection (952) on a wall of the space (102); and • a cold liquid circuit (152k) which couples the heat pump (222) to the gas-liquid heat exchanger (1041); • wherein the heat pump (222) is configured to extract thermal energy from the cold liquid circuit (152k) and supply it to the hot liquid circuit (152h). [2] Data center (151) according to claim 1, wherein the room (102) is configured as a container. [3] Data center (151) according to claim 2, wherein the container is configured as a container according to ISO standard 668. [4] Data center (151) according to one of claims 1 to 3, further comprising: a hot liquid supply (202k) outside the space (102), wherein the hot liquid supply (202) is coupled to the hot liquid connection (952). [5] Data center (151) according to claim 4, wherein a self-contained flow path (155) through the hot liquid circuit (152h) is provided by means of the hot liquid supply (202k). [6] Data center (151) according to one of claims 1 to 5, wherein a heat pump (222) of the at least one heat pump (222) is configured with a modulatable operating point. [7] Data center (151) according to one of claims 1 to 6, further comprising: an energy supply infrastructure (106) within the room (102) for supplying the computer system (104) with electrical energy, wherein the gas-liquid heat exchanger (1041) is arranged, wherein the energy supply infrastructure of the room (102) is pre-certified with regard to the reliability of the computer system (104). [8] Data center (151) according to one of claims 1 to 7, wherein a cooling infrastructure (114) comprising the heat pump (222), the gas-liquid heat exchanger (1041), the hot liquid circuit (152h) and the cold liquid circuit (152k) is pre-certified with regard to a fail-safety of the computer system (104). [9] Data center (151) according to one of claims 1 to 8, further comprising: • a failure switch, • wherein the at least one heat pump (222) comprises a pair of heat pumps (222) arranged redundantly to one another; • wherein the failure switch is configured to switch to the other heat pump (222) of the pair in response to a failure of one heat pump (222) of the pair. [10] Data center (151) according to one of claims 1 to 9, wherein the room (102) is arranged such that an air circulation area (102i) of the room (102), to which the computer system (104) and / or the heat exchanger (1041) are adjacent, is hermetically sealed from an exterior of the room (102). [11] Data center (151) according to one of claims 1 to 10, wherein the hot liquid circuit (152h) is configured such that, during operation, a hot liquid temperature of the hot liquid circuit (152h) is greater than a cold liquid temperature of the cold liquid circuit (152k). [12] Data center (151) according to claim 11, further comprising: a control device which is configured to control a flow of the hot liquid circuit (152h) according to the hot liquid temperature and / or the at least one heat pump (222) according to the cold liquid temperature. [13] Data center (151) according to claim 12, wherein the control device is arranged to switch the cold liquid temperature between at least two temperatures. [14] Data center (151) according to claim 13, wherein the control device is configured to switch the cold liquid temperature according to an operating state of the data center (151). [15] Method (1300), comprising: • Transferring (1301) thermal energy from air to a first liquid, wherein the air is arranged in a space (102) in which a computing system (104) is further arranged; • Transferring (1303) thermal energy from the first liquid to a second liquid having a higher temperature than the first liquid; • Transferring (1303) thermal energy from the computing system (104) to the second fluid; • Transporting (1307) the second liquid along a flow path passing through the space (102).

Images