Cooling system configurations for data centers

Abstract

Disclosed herein are systems and methods that include a data hall comprising a plurality of computer devices and an air-based cooling system comprising a plurality of fan units / modules and / or a liquid-based cooling system comprising at least one coolant distribution unit / module, wherein each of the fan units / modules and the at least one coolant distribution unit / module can be fluidly connected to one another and can be in a form factor that allows for addition and / or removal from a data center.

Description

Attorney Docket No.: 355532000440COOLING SYSTEM CONFIGURATIONS FOR DATA CENTERSCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Application No. 63 / 755,170, filed February 6, 2025, the entire contents of which are incorporated herein by reference.FIELD

[0002] The present disclosure relates generally to systems and methods for data center configurations. More specifically, the present disclosure relates to systems and methods of cooling system configurations for data centers.BACKGROUND

[0003] Current techniques for cooling data centers steal power off of the primary lineups that power the IT load and dedicate that power to mechanical load for cooling the data centers.

[0004] In addition, current fan systems used to cool data centers are often huge units that are on the perimeter of a data hall and / or part of the infrastructure of the data center. Thus, it is difficult (or not even possible) to remove and / or replace these fans from the data center.SUMMARY

[0005] Data centers are rapidly transitioning from predominantly air-cooled IT to a mix of high-density liquid-cooled racks and conventional air-cooled racks. This transition is driven by the increasing prevalence of high-performance computing workloads, including artificial intelligence (Al), machine learning (ML), and accelerator-based computer platforms, which can often exceed the practical cooling limits of traditional air-based systems. These high-performance computing workloads can drive large, uncertain variations in liquid vs. air-cooled rack fractions over the lifetime of a data center; frequent re-balancing of capacity between cloud services and Al services; and / or high rack densities that may not be served by air cooling alone.

[0006] Conventional data center mechanical designs are typically based on a fixed or narrowly bounded assumption regarding the ratio of air-cooled to liquid-cooled IT load. Data centers are typically engineered with a specific combination of air handling capacity, liquid coolant distribution units (CDUs), pumping capacity, and heat rejection equipment sized for that assumed ratio. Once constructed, altering the air-to-liquid cooling ratio generally requires:1MF-366724995Attorney Docket No.: 355532000440reconfiguration / reengineering of mechanical rooms and header configurations; deployment of additional CRAHs, CDUs, or pumping systems; modification of control systems; and / or significant construction activity, often accompanied by operational risk or downtime. As a result, such fixed-ratio designs can frequently lead to stranded or underutilized mechanical capacity when actual IT deployment deviates from initial assumptions. This inefficiency increases capital expenditures, operating costs, and power usage effectiveness (PUE), while reducing sellable IT capacity. In addition, it can be challenging to support customer- specific cooling requirements that can differ by data hall or data center requirements.

[0007] Current industry solutions frequently employ separate and largely independent systems for air and liquid cooling, such as: fan wall or CRAH units dedicated to air cooling; and standalone CDUs or liquid cooling skids sized for specific liquid-cooled deployments. While some original equipment manufacturers (OEMs) offer fan wall and CDU products from a single vendor, these components are not designed as a coordinated, interoperable family of modules that can be numerically recomposed to adjust air-to-liquid cooling ratios. Instead, air and liquid systems are generally treated as distinct product lines, each optimized for a static design point. As a result, adjusting the air-to-liquid ratio typically requires adding or replacing large subsystems rather than rebalancing standardized unit / module counts. In addition, late-stage configuration changes, close to project handover, are difficult or often impractical and / or mechanical designs are often bespoke and non-repeatable across different data centers.

[0008] Accordingly, Applicant discovered a cooling architecture that can enable flexible, repeatable, and / or late-binding configuration of air and liquid cooling capacity using standardized, interoperable units / modules. Specifically, disclosed herein is a modular cooling system architecture for data centers that can enable configurable and reconfigurable air-to-liquid cooling ratios through the use of standardized, interoperable mechanical units / modules. The cooling systems disclosed herein can have a modular architecture that can selectively provide air-based cooling, liquid-based cooling, or combinations thereof, and that can enable configurable and reconfigurable air-to-liquid cooling capacity ratios over the lifecycle of a data center facility.

[0009] In some embodiments, the cooling systems disclosed herein can include one more fan units / modules, one or more coolant distribution units / modules, and / or one or more control & pump units / modules. In some embodiments, one or more of the various unit / module types can be standardized with respect to physical form factor, hydraulic interfaces (i.e., fluid connections), electrical interfaces (i.e., electrical connections), and / or control protocols, such that the units / modules are interoperable and combinable in varying quantities. For example, a 2MF-366724995Attorney Docket No.: 355532000440desired air-to-liquid cooling ratio for a data hall can be achieved by selecting a numeric combination of fan units / modules, coolant distribution units / modules, and / or control & pump units / modules.

[0010] Advantageously, the disclosed architecture can allow the air-to-liquid cooling ratio to be modified by changing unit / module counts while leaving primary piping headers, mechanical rooms, and / or heat rejection equipment substantially unchanged. This can enable late-stage configuration changes, improved lifecycle flexibility, and / or repeatable deployment across multiple sites. In addition, mixed configurations and ratios of these units / modules can be adjusted as more precise customer requirements emerge.

[0011] In some embodiments, the modular cooling system architecture can provide a formalized modular architecture that can define interoperable air cooling, liquid distribution, and / or pumping / control modules; can provide numerically defined configurations (e.g., 0-1-1, 3-0-1, 5-1-1) that can directly map to cooling capacity, redundancy behavior, and / or air-to-liquid ratios; and / or can provide the ability to reconfigure cooling capacity late in the project lifecycle by changing only module counts and placement while leaving primary headers, roof equipment, and / or utility interfaces substantially unchanged.

[0012] The systems and methods disclosed herein can include air-based cooling system units / modules and / or liquid-based cooling system units / modules that can be in a form factor and / or size that one can remove, replace, and / or add to the data center without disrupting the data center and without requiring significant (or any) reconfiguration of other cooling system units that are not being removed or replaced. In some embodiments, the fan units / modules that can make up the air-based cooling system and / or the coolant distribution units / modules that can make up the liquid-based cooling system can be in a form factor that allows for addition and / or removal of the units in the data hall / center.

[0013] The systems and methods disclosed herein can also have dedicated lineup(s) for powering the IT load of data centers and a separate dedicated lineup(s) for powering the cooling system mechanical loads. For example, a data center can have a dedicated mechanical lineup(s) to power the cooling system or one or more cooling subsystems (e.g., cooling equipment and house loads). This can enable the data hall IT load to be powered in a standardized manner independent of the cooling technology / system selected. For example, depending on the location (e.g., a desert, humid location, etc.) of the data center, the cooling technology (e.g., indirect evaporative cooling via dry cooler with wetted media pads, air-cooled chillers, adiabatic chillers, and / or others) and therefore the mechanical load of the cooling technology can differ as different cooling systems can have different power requirements. To standardize 3MF-366724995Attorney Docket No.: 355532000440the design of the data hall independent of the outside cooling system, a dedicated lineup(s) can be used to power the outside cooling system. In addition, this can prevent stealing / losing any power from the primary lineup(s) that power the IT load of the data center. In some embodiments, the internal mechanical load of a data hall can be powered from the regular lineup(s) powering the IT load.

[0014] Liquid-based cooling systems (e.g., coolant distribution units / modules (CDUs or CDMs)) or air-based cooling systems (e.g., fan units / modules) can be products that can be readily replaced, removed, and / or added over the lifespan of a data center. In some embodiments, during the planning and permitting stage of a data center, the data center can obtain permits for both 100% air and 100% liquid cooling for the data center. For example, the permit can initially say the data center will have 200% capacity (100% air, 100% liquid), and then within a given time period (e.g., 6-8 weeks) before the data center is finished, a certain amount of air-based cooling system units / modules and / or liquid-based cooling system units / modules can be omitted from the permit set such that the exact configuration desired is obtained. This workflow can be beneficial as it can reduce the permitting time down to a few weeks as compared to a few months.

[0015] In some embodiments, a data center system includes a data hall comprising a plurality of computer devices; a heat rejection system comprising at least one first heat exchanger, wherein the heat rejection system is outside the data hall; and a cooling system comprising: a plurality of fan units, wherein each fan unit comprises at least one fan and at least one second heat exchanger and each fan unit is configured to transfer heat from air from the data hall to a primary coolant fluid; and / or a coolant distribution unit comprising at least one third heat exchanger, wherein the coolant distribution unit is configured to transfer heat from heated secondary coolant fluid from the data hall to primary coolant fluid; and a control & pump unit comprising one or more pumps, wherein the control & pump unit is configured to receive the heated primary coolant fluid from each fan unit and / or the coolant distribution unit and send the heated primary coolant fluid to the heat rejection system. In some embodiments, each fan unit and the coolant distribution unit are in a form factor that allows for addition and / or removal from the data center system. In some embodiments, the at least one first heat exchanger is an air-to-liquid heat exchanger, the at least one second heat exchanger is an air-to-liquid heat exchanger, and the at least one third heat exchanger is a liquid-to-liquid heat exchanger. In some embodiments, each fan unit has the same length, height, and depth. In some embodiments, the coolant distribution unit has at least one of the same length, height, and depth as each fan unit. In some embodiments, the coolant distribution unit has a length twice as long 4MF-366724995Attorney Docket No.: 355532000440as the length of each fan unit. In some embodiments, the cooling system is in a mechanical area of a data center separated from the data hall. In some embodiments, the mechanical area is separated from the data hall by at least one wire mesh barrier. In some embodiments, the data center system includes a second cooling system comprising: a second plurality of fan units, wherein each of the second plurality of fan units comprises at least one fan and at least one second heat exchanger and each of the second plurality of fan units is configured to transfer heat from air from the data hall to primary coolant fluid; and / or a second coolant distribution unit comprising at least one third heat exchanger, wherein the second coolant distribution unit is configured to transfer heat from heated secondary coolant fluid from the data hall to primary coolant fluid, wherein the first cooling system is on a first side of the data hall and the second cooling system is on a second side of the data hall opposite the first side. In some embodiments, the first cooling system is parallel to the second cooling system. In some embodiments, the first cooling system is a first wall or array and the second cooling system is a second wall or array. In some embodiments, the cooling system comprises the plurality of fan units and the coolant distribution unit, and the control & pump unit is configured to receive the heated primary coolant fluid from each fan unit and from the coolant distribution unit and send the heated primary coolant fluid to the heat rejection system. In some embodiments, the plurality of computer devices are powered by at least one first lineup; and the cooling system is powered by at least one second lineup different from the at least one first lineup.

[0016] In some embodiments, a method includes transferring heat from a data hall comprising a plurality of computer devices to air; flowing the heated air to a plurality of fan units, wherein each fan unit comprises at least one first heat exchanger and at least one fan; transferring heat from the heated air to a primary coolant fluid in each of the fan units; flowing the heated primary coolant fluid from each fan unit to at least one pump of a control & pump unit; transferring the heated primary coolant fluid from the control & pump unit to a heat rejection system comprising at least one second heat exchanger; and transferring heat from the heated primary coolant fluid to air outside the data hall. In some embodiments, the method includes transferring heat from the data hall to a secondary coolant fluid and flowing the heated secondary coolant fluid to a coolant distribution unit comprising at least one third heat exchanger. In some embodiments, the method includes transferring heat from the heated secondary coolant fluid to primary coolant fluid in the coolant distribution unit. In some embodiments, the method includes flowing the heated primary coolant fluid from the coolant distribution unit to the heat rejection system. In some embodiments, the method includes flowing the heated primary coolant fluid from the coolant distribution unit to the at least one 5MF-366724995Attorney Docket No.: 355532000440pump of the control & pump unit. In some embodiments, the at least one first heat exchanger is an air-to-liquid heat exchanger, the at least one second heat exchanger is an air-to-liquid heat exchanger, and the at least one third heat exchanger is a liquid-to-liquid heat exchanger. In some embodiments, the method includes transferring cooled primary coolant fluid from the heat rejection system to the plurality of fan units, the coolant distribution unit, and / or the control & pump unit.

[0017] In some embodiments, the first wall or array can include a first plurality of fan units and / or the second wall or array comprises a second plurality of fan units. In some embodiments, the first wall or array comprises a first coolant distribution unit and / or the second wall or array comprises a second coolant distribution unit. In some embodiments, the first wall or array comprises a first coolant distribution unit and a first plurality of fan units and the second wall or array comprises a second coolant distribution unit and a second plurality of fan units.

[0018] In some embodiments, the fan unit and / or the coolant distribution unit are combined into integrated cooling units and these integrated cooling units are arranged to provide N+l redundancy such that N integrated cooling units are sufficient to meet the data hall’s cooling load. In some embodiments, the data center system includes a reserve cooling system, distinct from the cooling system, comprising an additional integrated cooling unit comprising an additional fan unit(s) and / or coolant distribution unit(s).

[0019] In some embodiments, a data center includes at least one data hall comprising a plurality of computer devices, wherein the plurality of computer devices is powered by at least one first lineup; and at least one cooling system, wherein the at least one cooling system is powered by at least one second lineup different from the at least one first lineup. In some embodiments, the at least one cooling system is located outside of the data hall. In some embodiments, the data hall comprises internal mechanical equipment powered by the at least one first lineup.

[0020] In some embodiments, a data center system comprises a data hall comprising: an airbased cooling system comprising a plurality of fan units; and a liquid-based cooling system comprising a plurality of coolant distribution units (CDUs), wherein each of the fan units and CDUs are in a form factor that allows for addition and / or removal from the data hall.

[0021] A person of ordinary skill in the art of data center mechanical design can appreciate that the disclosed systems, methods, and architectures are compatible with conventional chilled water plants, adiabatic dry coolers, indirect evaporative systems, and / or other known heat rejection technologies.6MF-366724995Attorney Docket No.: 355532000440

[0022] In some embodiments, any of the features of any of the embodiments, systems, and methods described above and / or described elsewhere herein may be combined, in whole or in part, with one another. Additional advantages will be readily apparent to those skilled in the art from the following figures and detailed description. The aspects and descriptions herein are to be regarded as illustrative in nature and not restrictive.BRIEF DESCRIPTION OF THE FIGURES

[0023] A better understanding of the features and advantages of the present disclosure will be obtained by reference to the following detailed description that sets forth illustrative embodiments, in which the principles of the disclosure are utilized, and the accompanying figures of which:

[0024] FIG. 1 depicts an exemplary data center system in accordance with some embodiments disclosed herein.

[0025] FIG. 2 illustrates an exemplary overhead cross-sectional view of a data center in accordance with some embodiments disclosed herein.

[0026] FIG. 3 illustrates an example of a fan unit / module in accordance with some embodiments disclosed herein.

[0027] FIG. 4 illustrates an example of a coolant distribution unit / module in accordance with some embodiments disclosed herein.

[0028] FIG. 5 illustrates an example of a control & pump unit / module in accordance with some embodiments disclosed herein.

[0029] FIG. 6 illustrates three different exemplary flexible cooling configurations (A, B, C) in accordance with some embodiments disclosed herein.

[0030] FIG. 7 illustrates an example of a cooling wall / array with a coolant distribution unit / module in accordance with some embodiments disclosed herein.

[0031] FIG. 8 illustrates an example of cooling wall / array with fan units / modules in accordance with some embodiments disclosed herein.

[0032] FIG. 9 illustrates an example of a cooling wall / array with a combination of fan units / modules and a coolant distribution unit / module in accordance with some embodiments disclosed herein.

[0033] FIG. 10 illustrates an example of a smaller or subset cooling wall / array with fan units / modules in accordance with some embodiments disclosed herein.

[0034] FIG. 11 illustrates an exemplary cross section of a data center system in accordance with some embodiments disclosed herein.7MF-366724995Attorney Docket No.: 355532000440

[0035] FIG. 12 illustrates conventional and (N+l)+l cooling design redundancy in accordance with some embodiments disclosed herein.

[0036] FIG. 13A illustrates a first exemplary flow diagram for a fan unit / module in accordance with some embodiments disclosed herein.

[0037] FIG. 13B illustrates a second exemplary flow diagram for a fan unit / module in accordance with some embodiments disclosed herein.

[0038] FIG. 14A illustrates a first exemplary flow diagram for a coolant distribution unit / module in accordance with some embodiments disclosed herein.

[0039] FIG. 14B illustrates a second exemplary flow diagram for a coolant distribution unit / module in accordance with some embodiments disclosed herein.

[0040] FIG. 15 illustrates a flow diagram for a cooling system or subsystem (e.g., a cooling wall / array) of a data center in accordance with some embodiments disclosed herein.

[0041] In the Figures, like reference numerals refer to like components unless stated differently herein.DETAILED DESCRIPTION

[0042] As stated above, Applicant discovered a data center system that can utilize a cooling architecture that can enable flexible, repeatable, and / or late-binding configuration of air and / or liquid cooling capacity using standardized, interoperable units / modules. Disclosed herein are fan units / modules making up the air-based cooling system and / or liquid cooling units / modules making up the liquid-based cooling system that can be in a form factor that allows for addition and / or removal of the units / modules in the data center without disrupting data center operations. In addition, disclosed herein are systems and methods that can have dedicated lineup(s) for powering the IT load of data center and a separate dedicated lineup(s) for powering the cooling system mechanical loads.

[0043] Systems and methods disclosed herein may include data centers designed to prioritize safety, simplicity, and / or long-term operational efficiency. An exemplary data center may be structured to support long-term operation with components making up the data center and / or underground utilities designed for extended operation. A data center can be a single building or structure or include multiple buildings / structures and / or other components or facilities. For example, components may be designed to last 50 or more years. To increase environmental sustainability, a data center may be designed to include biodiverse landscaping, use of municipal recycled water, implementation of energy- and water-efficient practices, use of low global warming potential refrigerants, and / or selection of construction materials that minimize 8MF-366724995Attorney Docket No.: 355532000440embodied carbon dioxide, such as low-carbon concrete and / or low-carbon steel. Sustainability efforts may also include the management of construction waste, energy consumption during construction, and / or water usage during construction.

[0044] An exemplary data center may include one or more data halls that in turn may include independent sections or areas referred to as phases. A data hall can be a standalone building or structure or a portion of a building or structure. A phase may include server racks, electrical lineups transferring power to the server racks, and / or cooling units / modules transferring heat away from the server racks. In some embodiments, a data hall can include at least one computer device, including for example one or more servers, storage devices, processors, network communication devices (e.g., switches, routers, firewalls, etc.), etc. In some embodiments, a data hall can include a plurality of networked computer devices. In some embodiments, a data hall phase may include at least one computer device and / or a plurality of networked computer devices. In some embodiments, an electrical lineup may refer to a configuration of equipment supplying power to a designated portion of a data center. In some embodiments, the configuration of equipment can include a set of electrical distribution equipment supplying power to the designated portion of the data center. In some embodiments, an electrical lineup can include, for example, switching components, cabling, and / or power distribution units. For example, a phase may include at least one electrical lineup, at least two electrical lineups, at least three electrical lineups, at least four electrical lineups, at least five electrical lineups, at least six electrical lineups, at most six electrical lineups, at most five electrical lineups, at most four electrical lineups, at most three electrical lineups, at most two electrical lineups, and / or at most one electrical lineup.

[0045] Phases of data halls may form independent structural and / or operational units, allowing incremental expansion without impacting already commissioned sections. Following construction of the core and shell of an exemplary data center building, the installation and / or fit-out of mechanical, electrical, and / or plumbing systems may be performed on a per-phase basis without affecting the operation of existing phases. An exemplary phase may include at least one aisle, at least two aisles, at least four aisles, at least six aisles, at least eight aisles, at most eight aisles, at most six aisles, at most four aisles, at most two aisles, and / or at most one aisle. An exemplary phase may include at least one row, at least two rows, at least four rows, at least six rows, least eight rows, at least ten rows, at least 12 rows, at most 12 rows, at most ten rows, at most eight rows, at most six rows, at most four rows, at most two rows, and / or at most one row. In some implementations, a first-built phase of an exemplary data center may include more rows than one or more later-built phases.9MF-366724995Attorney Docket No.: 355532000440

[0046] A data center may include an electrical system based on a standard power capacity of one or more of the electrical lineups transferring power to server racks and / or other IT equipment. For example, said electrical lineups may have a standard power capacity of at least 0.5 MW, at least 1 MW, at least 1.5 MW, at least 2 MW, at least 2.5 MW, at least 2.75 MW, at least 3 MW, at least 3.5 MW, at least 4 MW, at most 4 MW, at most 3.5 MW, at most 3 MW, at most 2.75 MW, at most 2.5 MW, at most 2 MW, at most 1.5 MW, at most 1 MW, and / or at most 0.5 MW.

[0047] Use of standard electrical lineups forming a phase of a data hall may enable incremental data center construction as mentioned above. For example, a phase may include one or more electrical lineups transferring power to one or more server racks and / or other IT equipment. Electrical lineups may additionally or alternatively transfer power to mechanical loads, for example mechanical loads corresponding to facility cooling. Additionally or alternatively, a phase may include one or more reserve electrical lineups to provide redundancy to electrical lineups transferring power to server racks and / or other IT equipment and / or to electrical lineups transferring power to mechanical loads. The number of electrical lineups transferring power to mechanical loads may vary depending on the cooling technology selected for a specific site.

[0048] In some embodiments, a data hall may have dedicated electrical lineup(s) separate from electrical lineup(s) for powering the mechanical load (e.g., cooling system(s) and / or subsystem(s) that include cooling units / equipment). This can allow for the standardization of a data hall of a data center. For example, this can enable the data hall IT load to be powered the same regardless of the cooling technology or system(s) selected. In some embodiments, the IT load can be the power consumed by any and / or all servers, storage, and networking equipment and the mechanical load can be the power consumed by any and / or all systems that support the IT load, for example, the cooling system(s) and / or subsystem(s). In some embodiments, any mechanical load located outside of the data hall (e.g., cooling systems) can be powered by its own electrical lineup(s) distinct and independent from electrical lineup(s) for the data hall. In some embodiments, the mechanical load within the data hall can be powered from the electrical lineup(s) powering the IT load.

[0049] For example, FIG. 1 illustrates a data center system 100 that includes at least one data hall 101 and at least one cooling system 102 for the data hall. The at least one data hall 101 can be any data hall, phase, and / or phases of a data hall disclosed herein. The at least one cooling system 102 (and 102a) can be any cooling system, unit / module, equipment, and / or portion thereof disclosed herein. In some embodiments, the at least one data hall 101 can have at least one lineup 103a for powering the at least one IT load 104 (e.g., computer components,10MF-366724995Attorney Docket No.: 355532000440servers, etc.). In some embodiments, the at least one cooling system 102 is outside the at least one data hall 101. In some embodiments, the at least one cooling system can have at least one lineup 103b for powering the at least one cooling system 102. In some embodiments, the at least one lineup 103 a and at least one lineup 103b are independent and distinct from one another such that no electrical power is shared between the at least one lineup 103 a and the at least one lineup 103b.

[0050] In some embodiments, the data hall can include at least one mechanical load 102a. In some embodiments, at least one mechanical load 102a can be powered from the at least one lineup 103a that powers the at least one IT load 104. In some embodiments, the at least one mechanical load 102a can be a cooling system or subsystem and / or a portion of the at least one cooling system 102. In some embodiments, the at least one mechanical load 102a can be a portion of the at least one cooling system 102 such that the at least one mechanical load 102a can be powered from the at least one lineup 103b that powers the at least one cooling system 102.

[0051] An exemplary electrical lineup may use a bus duct system for power distribution, with the bus duct current rating determined based on factors such as voltage drop constraints. For example, an electrical lineup may include an aluminum bus duct rated for 4000 A. Each server rack or server rack grouping may receive one or more separate power feeds, for example each may receive two power feeds for redundancy. Power from a utility substation or other mediumvoltage source may be routed through distribution equipment (e.g. switchgear) to one or more transformers. Said one or more transformers may use oil for cooling and / or insulation and may have built-in automatic fault protection. Said transformers may be arranged in a looped system with a single open point in each loop to prevent direct connections between different power sources, thereby improving reliability and / or fault isolation.

[0052] In some embodiments, an exemplary data center may include a facility cooling system configuration designed to increase energy efficiency by leveraging opportunities for free cooling and reducing reliance on mechanical cooling systems. Different areas of a data center including, for example, data halls, electrical rooms, and / or offices, may have distinct cooling setpoints to optimize efficiency. For example, using a higher data hall aisle temperature setpoint may extend free cooling operation. Airflow distribution within an exemplary data hall may be based on a flooded room approach. For example, air may be delivered throughout a data hall and / or to one or more aisles from one or both sides of the one or more aisles. An exemplary facility cooling system may define a maximum allowable duration for which a data11MF-366724995Attorney Docket No.: 355532000440center facility may operate outside acceptable temperature and / or humidity ranges before triggering an error escalation.

[0053] An exemplary air-based cooling system may be designed to maintain server air intake temperatures within a normal operating range of 10 °C to 35 °C. An air-based cooling system may interface with one or more servers within a data hall to reduce the risk of thermal degradation. For example, a cooling system may set a maximum server air intake temperature above which server degradation may occur, for example a system may set a maximum temperature between 35 °C and 45 °C. A cooling system may throttle the processors of one or more servers if server air intake temperature is between 40 °C and 45 °C, and / or may shutdown one or more servers if intake temperature exceeds 45 °C to 50 °C. An air-based cooling system may additionally maintain humidity levels below 60% to 75% relative humidity, and cooling system airflow may be limited to 170 CFM per server rack kW. All of the setpoints described above (e.g., temperature, pressure, etc.) for an air-based cooling system can all be adjusted to meet different data halls or IT loads. In other words, the values in this paragraph are purely exemplary and non-limiting.

[0054] An exemplary liquid-based cooling system may include specifications for cooling liquid filtration levels, pressure limits, and / or flow rates. The cooling liquid or coolant used in such a cooling system may be water, PG25, and / or a similar heat transfer fluid. The filtration of a server cooling system may include a particle filter, for example a 50-micron filter. A cooling system may set the minimum server liquid inlet temperature to the dew point plus 2°C, and / or may set the maximum server liquid inlet temperature to 28 °C to 34 °C. A cooling system may further set the maximum server liquid outlet temperature to 59 °C. A cooling system may set the maximum allowable server rack flow rate to 4.0 LPM per server rack kW, the minimum server rack flow rate to 0.7 LPM per server rack kW, the minimum supply pressure for filling the system to 35 PSI, and / or the maximum server inlet pressure to 100 PSI. In some embodiments, the maximum fluid volume for a system may be 40 L. All of the setpoints described above (e.g., temperature, pressure, etc.) for a liquid-based cooling system can all be adjusted to meet different data halls or IT loads. In other words, the values in this paragraph are purely exemplary and non-limiting.

[0055] In some embodiments, the specifications for liquid-based cooling systems can assume constant liquid flow. This flow may be set using flow setters (e.g., pressure independent, manually operated adjustable flow setters) supplying a server rack from a main row liquid manifold. A liquid-based cooling system may involve constant pressure and variable flow in which case flow setters may be replaced by pressure-independent control valves. This change 12MF-366724995Attorney Docket No.: 355532000440in control approach may also involve support from installed coolant distribution units / modules (CDU / Ms).

[0056] In some embodiments, the data center(s) or data hall(s) may include liquid-based server cooling systems (e.g., coolant distribution units or modules (CDUs or CDMs)) and / or air-based cooling systems (e.g., fan units / modules, fan wall arrays). In some embodiments, a cooling system can include one or more fan units / modules, one or more coolant distributions units / modules, and / or one or more control & pump units / modules (CPU / M). Each unit / module can perform a distinct function and / or be designed to interoperate with the others via standardized interfaces such as fluid connections (for coolant) and / or electrical connections. In some embodiments, each unit / module can be tested individually before assembly in the data center. In some embodiments, a combination of one or more of these units / modules can cool the IT of the data hall. In some embodiments, any unit / module can be fluidly and / or electrically connected to any other unit / module. For example, in some embodiments, the one or more fan units / modules can be fluidly and / or electrically connected to the one or more coolant distribution units / modules. In some embodiments, a fan unit / module can be fluidly and / or electrically connected to another fan unit / module. In some embodiments, the one or more fan units / modules can be fluidly and / or electrically connected to the one or more control & pump units / modules. In some embodiments, the one or more coolant distribution units / modules can be fluidly and / or electrically connected to the one or more control & pump units / modules. In some embodiments, a coolant distribution unit / module can be fluidly and / or electrically connected to another coolant distribution unit / module. In some embodiments, a control & pump unit / module can be fluidly and / or electrically connected to another control & pump unit / module. In some embodiments, the one or more fan units / modules, one or more coolant distribution units / modules, and / or the one or more control & pump units / modules can be fluidly connected such that primary coolant fluid in a cooling system can flow in and out of each of the one or more fan units / modules, one or more coolant distribution units / modules, and / or the one or more control & pump units / modules.

[0057] In some embodiments, the one or more fan units / modules, one or more coolant distribution units / modules, and / or the one or more control & pump units / modules can be fluidly connected to at least one heat rejection system. In some embodiments, the one or more fan units / modules, one or more coolant distribution units / modules, and / or the one or more control & pump units / modules can be fluidly connected to the heat rejection system such that primary coolant fluid in a cooling system can flow in and out of the one or more fan units / modules, one or more coolant distribution units / modules, the one or more control & pump units / modules,13MF-366724995Attorney Docket No.: 355532000440and / or the heat rejection system. In some embodiments, the one or more fan units / modules can be fluidly and / or electrically connected to the at least one heat rejection system. In some embodiments, the at least one heat rejection system can be fluidly and / or electrically connected to the one or more control & pump units / modules. In some embodiments, the one or more coolant distribution units / modules can be fluidly and / or electrically connected to the at least one heat rejection system. In some embodiments, a heat rejection system can be fluidly and / or electrically connected to another heat rejection system.

[0058] In some embodiments, fluidly connected can mean arranged such that a fluid can flow, directly or indirectly, between components. In some embodiments, electrically connected can mean arranged such that electrical power, electrical current, and / or electrical signals can be transmitted, directly or indirectly, between components, including through one or more intermediate components.

[0059] FIG. 2 illustrates an overhead view of a data center 200. In some embodiments, the data center can include a data hall 101 (which can be any data hall disclosed herein) that can include a plurality of computer devices 203 (which can be any computer devices disclosed herein). In some embodiments, the data center can include a cooling system such as cooling system 102. In some embodiments, the cooling system can include one or more cooling subsystems such as cooling subsystems 102a and 102b. In some embodiments, the data center can have multiple stories such that there can be an upper and lower area of the data hall as shown in FIG. 11. In addition, FIG. 11 also illustrates four different cooling systems / subsystems. In some embodiments, the cooling systems (and cooling subsystems) can be separated from the data center by a barrier (e.g., wire-mesh barrier) or wall. In some embodiments, this separator can still allow cooling fluid (e.g., cool air from a fan unit / module and / or cool liquid from a coolant distribution unit / module) to enter and / or heated fluid (e.g., heated air and / or heated liquid) to exit the data hall. In some embodiments, a data center can have cooling systems or subsystems on opposite sides of the data hall. In some embodiments, the cooling systems or subsystems can be parallel to one another. Examples of cooling systems or subsystems on opposite sides of a data hall and parallel to one another are shown at least in FIGS. 2, 6, and 11.

[0060] In some embodiments, the cooling systems and subsystems can include one or more fan units / modules, one or more coolant distribution units / modules, and / or one or more control & pump units / modules. Together these units / modules can deliver both air cooling (via one or more fan units / modules) and / or liquid cooling (via one or more coolant distribution units / modules) from a shared mechanical plant. In some embodiments, these units / modules 14MF-366724995Attorney Docket No.: 355532000440can be dimensioned and cataloged such that combinations can be numerically planned for each project.

[0061] In some embodiments, a fan unit / module can be configured to provide air-based cooling to a data hall. In some embodiments, a fan unit / module can be configured to provide air-based cooling to some of the plurality of computers in the data hall. In some embodiments, a fan unit / module can be an indoor air cooling unit configured to remove heat from the data hall air.

[0062] FIG. 3 provides an example of a fan unit / module 300. In some embodiments, the fan unit / module can include at least one fan. In some embodiments, the fan unit / module can include a plurality of fans. As shown in FIG. 3, the fan unit / module 300 includes six fans 301. In some embodiments, the fan unit / module can include at least one heat exchanger (e.g., an air-to-liquid heat exchanger and / or a cooling coil). In some embodiments, the fan unit / module can be a fan coil unit / module. In some embodiments, a fan unit / module can be an indoor fan-wall style unit / module configured to provide air cooling to the data hall. In some embodiments, a fan unit / module can be a fan wall or fan array unit. In some embodiments, the fan unit / module can reject that heat to a circulating coolant (i.e., primary coolant fluid).

[0063] In some embodiments, the one or more fans can be configured to move air across a heat exchanger such as a heat exchange coil or other air-to-liquid heat exchanger. In some embodiments, the heat exchanger can be thermally coupled to a liquid coolant loop (i.e., primary coolant fluid network). In some embodiments, the fan unit / module can include supply and / or return fluid connections for coupling to the control & pump unit / module. In some embodiments, a fan unit / module can move air to pick up heat from the data hall and push it across a heat exchanger, while a coolant loop (primary coolant fluid network) can carry that heat away from the data center. In some embodiments, the at least one fan of the fan unit / module can draw in hot fluid from the data hall (e.g., hot exhaust air from IT racks) and can pass it through at least one heat exchanger (e.g., air-to-liquid heat exchanger) to transfer heat from the hot fluid to a coolant fluid (e.g., primary coolant fluid) in the heat exchanger resulting in cooler fluid / air (i.e., temperature of fluid is less than the hot fluid from the data hall). As explained herein, this heated coolant fluid can reject the heat to a heat rejection system / equipment located outside the data center (e.g., on the roof of a data center).

[0064] For example, FIGS. 13A and 13B illustrate fluid flow diagrams 1300A and BOOB for a fan unit / module 300. As shown, the fan unit / module can include at least one heat exchanger 1304 and at least one fan 301. The fan unit / module can be configured to receive heated fluid 1309 from the data hall or IT load (e.g., hot exhaust air). In some embodiments, the at least one fan can move a heated fluid 1309 (e.g., hot exhaust air from data hall or IT load) through 15MF-366724995Attorney Docket No.: 355532000440at least one heat exchanger (e.g., air-to-liquid heat exchanger or cooling coil) to transfer heat from the heated fluid to a coolant fluid (e.g., primary coolant fluid) in the at least one heat exchanger resulting in a cooler fluid 1310 (e.g., cool air) that can be sent to the data hall for cooling the data hall or IT load. In some embodiments, the heated fluid can be hot exhaust air from the data hall or IT load. In some embodiments, the cooler fluid can be cold supply air that goes back to the data hall or IT load for cooling. In some embodiments, each fan unit can be configured to transfer heat from air from the data hall to a primary coolant fluid.

[0065] The fan unit / module can be configured to receive a primary coolant fluid (e.g., a liquid coolant) 1307. This primary coolant fluid can be the coolant fluid for the primary cooling systems or subsystems disclosed herein. In some embodiments, the primary coolant fluid can be water, PG25, and / or a similar heat transfer fluid. In some embodiments, the fan unit / module can receive the primary coolant fluid from a cold primary coolant source. In some embodiments, the fan unit / module can receive the primary coolant fluid from a heat rejection system 1303 (as shown in FIG. 13B). In some embodiments, the fan unit / module can receive the primary coolant fluid from a control & pump unit 500 (as shown in FIG. 13A).

[0066] In some embodiments, the primary coolant fluid enters the fan unit / module colder than when it leaves the fan unit / module. In some embodiments, the at least one heat exchanger of the fan unit / module can be configured to receive the primary coolant fluid. As the heated fluid passes through the at least one heat exchanger (e.g., as hot exhaust air passes over the air-to-liquid heat exchanger), heat from the heated fluid can transfer to the primary coolant fluid forming heated primary coolant fluid 1308. In some embodiments, a control & pump module 500 can be configured to receive heated primary coolant fluid from a fan unit / module. In some embodiments, the one or more pumps 1306 of the control & pump module can be configured to pump the heated primary coolant fluid to the heat rejection system 1303.

[0067] In some embodiments, the heat rejection system can be configured to receive the heated primary coolant fluid (from a control & pump module). In some embodiments, the heat rejection system is outside the data hall and / or data center (such as on the roof of a data center). In some embodiments, the heat rejection system can include at least one heat exchanger 1307a such as a dry cooler, cooling tower, chiller (e.g., air cooled chiller), etc., or combinations thereof. In some embodiments, the at least one heat exchanger of the heat rejection system can be an air-to-liquid heat exchanger. In some embodiments, the heat rejection system can reject or dump the heat out of the cooling system. For example, in some embodiments, ambient or outside air 1311 can blow through or across the heat exchanger such that heat from the heated primary coolant fluid transfers to the ambient or outside air 1311 forming heated ambient or 16MF-366724995Attorney Docket No.: 355532000440outside air 1312. This can cause the heated primary coolant fluid to cool such that it can be used again in the cooling systems or subsystems disclosed.

[0068] In some embodiments, a fan unit / module can include standardized supply and return connections for coupling to a control & pump unit / module and / or other fan units / modules. In some embodiments, a fan distribution unit / module can expose standardized hydraulic and control interfaces to the control & pump unit(s) / module(s) and to facility headers, allowing fan units / modules to be added, removed, or repositioned without substantial system redesign.

[0069] In some embodiments, the at least one heat exchanger in the fan unit / module may be a microchannel coil heat exchanger, fin-and-tube coil heat exchanger, or other suitable heat exchanger. In some embodiments, a fan unit / module utilizes a sensible-only cooling coil that can reduce air temperature without reducing humidity. In some embodiments, the heat exchanger of a fan unit / module can be configured to achieve a reduced approach temperature relative to conventional air handling equipment, thereby improving thermal efficiency. For example, high-efficiency coil design (e.g., microchannel) can achieve about a 4°F approach temperature vs. about 11 °F in typical OEM equipment, enabling smaller mechanical reserves and more IT load per site. Specifically, a coil(s) of a fan unit / module can allow the supplied air temperature to operate within approximately four degrees Fahrenheit of the circulating coolant temperature. This is what can enable efficient high-temperature cooling loops. In some embodiments, fans in the fan unit / module can draw warm air from the return plenum and supply cooled air into the data hall. In some embodiments, the cooled air is supplied through a secure barrier or wall (e.g., wire-mesh barrier), thereby maintaining physical separation between the data hall (i.e., IT space) and mechanical areas / regions (that can contain cooling systems / equipment) while ensuring efficient airflow. In some embodiments, the mechanical areas / regions can be a room or phase of a data center.

[0070] In some embodiments, a fan unit / module can be rated for a known air-cooling capacity. In some embodiments, as discussed herein, multiple fan units / modules can be deployed in a data center and the total air-cooling capacity can be determined by the number of installed fan units / modules. In some embodiments, the configurable number of fan units / modules deployed can depend on the desired air-cooling capacity. In some embodiments, each fan unit / module can be physically connected to an adjacent fan unit / module, adjacent coolant distribution unit / module, and / or control & pump unit / module. In some embodiments, these units / modules can be configured to be stacked on top of one another.

[0071] In some embodiments, the fan unit / module can have a length, depth, and height. FIG.3 illustrates the height (HF), depth (DF), and length (LF) of the fan unit / module 300. In some 17MF-366724995Attorney Docket No.: 355532000440embodiments, each fan unit / module can have the same length, depth, and / or height as the length, depth, and / or height of another fan unit / module. In some embodiments, each fan unit / module can have the same length, depth, and height as the length, depth, and height as the other fan units / modules.

[0072] In some embodiments, a coolant distribution unit / module can be configured to provide liquid cooling to the data hall. In some embodiments, a cooling distribution unit / module can provide liquid cooling to at least some of the plurality of computer devices in the data hall. In some embodiments, a cooling distribution unit / module can be configured to provide / deliver liquid coolant to at least some computer devices in the data hall. In some embodiments, a coolant distribution unit / module can transfer heat from a secondary fluid network to a primary fluid network. In some embodiments, a coolant distribution unit / module can be configured to distribute liquid coolant (e.g., secondary coolant fluid) to the data hall (e.g., to liquid-cooled IT equipment) and transfer heat to a heat rejection system (in a primary fluid network).

[0073] FIG. 4 provides an example of a coolant distribution unit / module 400. In some embodiments, the coolant distribution unit / module can have a length, depth, and height. FIG.4 illustrates the height (HL), depth (DL), and length (LL) of the coolant distribution unit / module 400. In some embodiments, at least one of the height, depth, and length of a coolant distribution unit / module is the same as at least one of the height, depth, and length of a fan unit / module or each fan unit / module. In some embodiments, the height and / or depth of a coolant distribution unit / module is the same as the height and / or depth of a fan unit / module or each fan unit / module. In some embodiments, the length of a coolant distribution unit / module is larger than the length of a fan unit / module or each fan unit / module. In some embodiments, the length of a coolant distribution unit / module is 2, 3, or 4 times as long as the length of a fan unit / module or each fan unit / module. For example, length LL of FIG. 4 is double the length of LF of FIG. 3. By having at least one of the height, length, or depth of a coolant distribution unit / module the same as the height, length, or depth a fan unit / module or each fan unit / module, the coolant distribution units / modules and fan units / modules can be in form factor and / or size that one can remove, replace, and / or add to the data center without disrupting the data center and without requiring significant (or any) reconfiguration of other cooling system units / modules that are not being removed or replaced.

[0074] In some embodiments, a coolant distribution unit / module can provide the liquidcooling capability required for liquid-cooled computer devices (e.g., IT hardware) in the data hall. In some embodiments, a coolant distribution unit / module can include one or more liquid-to-liquid heat exchangers; valves, pumps, manifolds, and distribution piping for supplying 18MF-366724995Attorney Docket No.: 355532000440coolant to at least some computer devices in the data hall; and / or sensors configured to measure / monitor temperature, pressure, and / or flow. In some embodiments, the liquid-to-liquid heat exchanger can be a plate-frame heat exchanger. In some embodiments, a coolant distribution unit / module can include at least one pump, at least one filtration, at least one heat exchanger (e.g., a plate heat exchanger), and at least one expansion tank. In some embodiments, heat generated by computer chips can be directly transferred to coolant (e.g., secondary coolant fluid) circulating through a plurality of cold plates in an IT rack that ties back to a fluid network circulated by at least one pump to at least one heat exchanger (liquidliquid heat exchanger) in a coolant distribution unit / module to transfer the IT rack heat into another coolant system (primary fluid / cooling network) and eventually to heat rejection system(s) outside the data center (e.g., on the roof).

[0075] For example, FIGS. 14A and 14B illustrate fluid flow diagrams 1400A and 1400B for a coolant distribution unit / module 400. As shown, the coolant distribution unit / module can include at least one heat exchanger 1408 and at least one pump 1407. In some embodiments, the coolant distribution unit / module can be configured to receive a primary coolant fluid (e.g., liquid coolant) 1307. As previously discussed with respect to the fan unit / module above, this primary coolant fluid can be the coolant fluid for the primary cooling networks or subsystems disclosed herein and the primary coolant fluid can be water, PG25, and / or a similar heat transfer fluid. Thus, the primary coolant fluid previously discussed with respect to the fan unit / module can be the same primary coolant fluid for the coolant distribution units / modules.

[0076] In some embodiments, a coolant distribution unit / module can be fluidly connected to the data hall (e.g., IT racks with direct to chip cooling and / or cold plates). In some embodiments, the coolant distribution unit / module can be configured to receive heated secondary coolant fluid (e.g., liquid coolant) 1411 from the data hall 101 such as from computer devices with direct to chip cooling and / or cold plates. This secondary coolant fluid can be the coolant fluid for a secondary cooling network 1405 disclosed herein. In some embodiments, the secondary coolant fluid can be water, PG25, and / or a similar heat transfer fluid. In some embodiments, heat from this heated secondary coolant fluid can be transferred to the primary coolant fluid in the at least one heat exchanger of the coolant distribution unit / module, thereby cooling the heated secondary coolant fluid. This cooled secondary coolant fluid 1412 can then be sent / returned to the data hall for cooling computer devices. For example, this cold secondary coolant fluid can be supplied to cold plates inside an IT racks to transfer chip heat to the secondary coolant fluid. The heated secondary coolant fluid from the data hall can then be transferred back to the coolant distribution unit / module. In some embodiments, the one or 19MF-366724995Attorney Docket No.: 355532000440more pumps of a coolant distribution unit / module can pump the secondary coolant fluid into the at least one heat exchanger and / or out of the coolant distribution unit / module to the data hall.

[0077] In some embodiments, the at least one heat exchanger (e.g., a liquid-to-liquid heat exchanger) of the coolant distribution unit / module can transfer heat from the heated secondary coolant fluid to the primary coolant fluid. In some embodiments, the coolant distribution unit / module can receive the primary coolant fluid from a cold primary coolant source. In some embodiments, the coolant distribution unit / module can receive the primary coolant fluid from a heat rejection system 1303 (as shown in FIG. 14B). In some embodiments, the coolant distribution unit / module can receive the primary coolant fluid from a control & pump unit / module 500 (as shown in FIG. 13A).

[0078] In some embodiments, the primary coolant fluid enters the coolant distribution unit / module colder than when it leaves the coolant distribution unit / module. In some embodiments, the at least one heat exchanger of the coolant distribution unit / module can be configured to receive the primary coolant fluid. In some embodiments, the at least one heat exchanger of the coolant distribution unit / module can be configured to receive the heated secondary coolant fluid. As the heated secondary coolant fluid passes through the at least one heat exchanger, heat from the heated secondary coolant fluid can transfer to the primary coolant fluid forming heated primary coolant fluid 1308. In some embodiments, a control & pump module 500 can be configured to receive heated primary coolant fluid from a coolant distribution unit / module. In some embodiments, the one or more pumps 1306 of the control & pump module can be configured to pump the heated primary coolant fluid to the heat rejection system 1303.

[0079] In some embodiments, the heat rejection system can be configured to receive the heated primary coolant fluid (e.g., from a control & pump module and / or from a coolant distribution unit / module). In some embodiments, the heat rejection system can include at least one heat exchanger 1307a such as a dry cooler, cooling tower, chiller (e.g., air cooled chiller), etc., or combinations thereof. In some embodiments, the heat rejection system can reject or dump the heat out of the cooling system (e.g., primary cooling network). For example, in some embodiments, ambient or outside air 1311 can blow through or across the heat exchanger such that heat from the heated primary coolant transfers to the ambient or outside air forming heated ambient or outside air 1312. This can cause the heated primary coolant fluid to cool such that it can be used again in the cooling systems or subsystems disclosed.20MF-366724995Attorney Docket No.: 355532000440

[0080] In some embodiments, a coolant distribution unit / module can support roughly 1.3-2.6 megawatts of heat rejection, depending on coolant supply temperatures between about 88-108°F. In some embodiments, a system can include a dedicated reserve coolant distribution unit / module. In some embodiments, the dedicated reserve coolant distribution unit / module can be connected through a switchover cabinet, enabling coolant flow to be redirected without interrupting service if a coolant distribution unit / module is removed or taken offline. In some embodiments, the dedicated reserve coolant distribution unit / module can be located in mechanical areas / regions of the data center. In some embodiments, a mechanical area / region can be a dedicated reserve coolant distribution unit / module.

[0081] In some embodiments, a coolant distribution unit / module can include standardized supply and return connections for coupling to a control & pump unit / module and / or fan units / modules. In some embodiments, a coolant distribution unit / module can include standardized supply and return connections for coupling to a fan unit / module. In some embodiments, a coolant distribution unit / module can include standardized distribution connections for coupling to computer device-level cooling interfaces (e.g., rack-level cooling interfaces). In some embodiments, a coolant distribution unit / module can expose standardized hydraulic and control interfaces to the control & pump unit(s) / module(s) and to facility headers, allowing the coolant distribution unit / module to be added, removed, or repositioned without substantial system redesign.

[0082] In some embodiments, a coolant distribution unit / module can be rated for a known liquid cooling capacity. The total liquid cooling capacity of a data center can be determined by the number of installed coolant distribution units / modules.

[0083] In some embodiments, a control & pump unit can be configured to control the air-based cooling system(s) and / or the liquid-based cooling system(s). In some embodiments, a control & pump unit / module can be configured to provide pumping and / or control for circulating primary coolant fluid between the fan unit(s) / module(s), the coolant distribution unit(s) / module(s), and / or heat rejection system. In some embodiments, a control & pump unit / module can be configured to receive heated primary coolant fluid from one or more fan units / modules and / or one or more coolant distribution units / modules. In some embodiments, a control & pump module can be configured to provide primary pumping, control, and / or hydraulic coupling (i.e., fluid connection) between the fan unit(s) / module(s), the coolant distribution unit(s) / module(s), and / or heat rejection system. In some embodiments, a control & pump unit / module can function as the integrating control and pumping unit / module that connects the coolant distribution unit(s) / module(s), the fan unit(s) / module(s), and the external 21MF-366724995Attorney Docket No.: 355532000440heat-rejection system, ensuring that heat generated by the data hall (e.g., IT equipment in data hall) can be reliably transferred out of the data center. For example, in some embodiments, heated primary coolant fluid from a fan unit / module and / or a coolant distribution unit / module can be sent to a control & pump unit / module to be moved (e.g., pumped) to an external heat rejection system. In some embodiments, all heated primary coolant fluid from all fan units / modules and / or all coolant distribution units / modules can be sent to one or more control & pump units / modules to be moved (e.g., pumped) to an external heat rejection system. In some embodiments, a cooling system or subsystem can have one heated primary coolant fluid outlet to the heat rejection system from a control & pump unit / module and / or one cooled primary coolant fluid inlet from the heat rejection system to the various units / modules. In such embodiments, the amount of piping can be drastically reduced rather than having each unit / module send their heated primary coolant fluid to the heat rejection system.

[0084] In some embodiments, a control & pump unit can be configured to provide heat from the data hall to the heat rejection system (that can be outside the data center or data hall). For example, FIG. 15 illustrates a flow diagram 1500 for a data center having a cooling system (e.g., a cooling wall / array) in accordance with some embodiments disclosed herein. In some embodiments, FIG. 15 illustrates a flow diagram for the cooling wall / array 700 of FIG. 9. As previously described, the fan units / modules 300 can reject heat from an air-based heat load from a data hall and a cooling distribution unit / module 400 can reject heat from the liquidbased heat load from a data hall. As previously explained, the one or more fan units / modules can be fluidly connected to a control & pump module and a coolant distribution unit / module can be fluidly connected to a control & pump module (which can be the same control & pump module 500 shown in FIG. 15). As such, both the air-based heat and liquid-based heat from the data hall can be transferred into a primary coolant fluid network such that the heat can be released to a heat rejection system. In some embodiments, this primary coolant fluid network can be a primary coolant loop. In other words, this primary coolant fluid network can take heat from exhaust air from a data hall and transfer it to primary coolant fluid 1307 using the at least one heat exchangers 1304 of fan units / modules 300 and / or can take heat from heated secondary coolant fluid 1411 from a data hall and transfer it to primary coolant fluid 1307 using the at least one heat exchanger 1408 of coolant distribution unit / module 400. The coolant fluid network between the coolant distribution unit / module and the data hall can be referred to as the secondary cooling fluid 1405. In some embodiments, the at least one heat exchanger in the coolant distribution unit / module can separate the secondary coolant fluid network from the primary coolant fluid network. In some embodiments, in the secondary coolant fluid network,22MF-366724995Attorney Docket No.: 355532000440heat from the data hall (e.g., data hall IT heat load) can be transferred to the secondary coolant fluid 1412 to form the heated secondary coolant fluid 1411. As shown in FIG. 15, the control & pump module can be configured to receive the heated primary coolant fluid from the fan units / modules 300 and the coolant distribution unit / module 400. In some embodiments, the control & pump module can be configured to receive some but not all or all of the heated primary coolant fluid from the fan units / modules and / or coolant distribution units / modules. The control & pump module can be fluidly connected to a heat rejection system 1303 such that the control & pump module can send the heated primary coolant fluid (via one or more pumps 1306) to the heat rejection system. In some embodiments, the one or more fan units / modules can be fluidly connected to a heat rejection system to receive primary coolant fluid. In some embodiments, the one or more coolant distribution units / modules can be fluidly connected to a heat rejection system to receive primary coolant fluid. In some embodiments, control & pump units / modules can be fluidly connected to a heat rejection system to receive primary coolant fluid. In such an embodiment, the control & pump module can send the primary coolant fluid to the one or more fan units / modules and / or one or more coolant distribution units / modules to receive the heat from the data hall in such units / modules.

[0085] In some embodiments, a control & pump unit / module can provide power (and building management control) and control to the one or more coolant distribution units / modules and / or the one or more fan units / modules. In some embodiments, a control & pump unit / module can allow 0-7 fan units / modules and / or coolant distribution units / modules to be connected and controlled from a single source.

[0086] In some embodiments, a control & pump unit / module can house primary pumps and / or control components that move heat from the one or more fan units / modules and / or one or more coolant distribution units / modules to external heat-rejection systems (e.g., roof chillers or adiabatic dry chillers). In some embodiments, a control & pump unit / module can be designed to support variable- speed pump control and predictive algorithms. In some embodiments, a control & pump unit / module provides pumping and control functionality for circulating coolant between the fan unit(s) / module(s), coolant distribution unit(s) / module(s), and / or heat rejection system. In some embodiments, a control & pump unit / module can include one or more pumps configured to circulate coolant; variable speed drives associated with the pumps; control hardware and software configured to regulate flow and pressure; and / or interfaces for coupling to heat rejection equipment such as chillers or dry coolers. In some embodiments, a control & pump module can include one or more variable- speed pumps; control hardware and software that manages pump operation based on demand, efficiency, and redundancy requirements;23MF-366724995Attorney Docket No.: 355532000440and / or hydraulically couple indoor units / modules (e.g., fan units / modules and / or coolant distribution units / modules) to exterior / outside heat rejection system / equipment (e.g., rooftop chillers, dry coolers, and / or other heat rejection devices).

[0087] In some embodiments, a control & pump module may be configured to support multiple fan units / modules and / or coolant distribution units / modules (simultaneously), depending on the selected configuration and / or capacity requirements. In some embodiments, the control & pump unit / module distributes primary coolant fluid from external heat-rejection equipment into the primary cooling network. In some embodiments, a control & pump unit / module can include at least one pump sized at half of peak flow, providing N+l redundancy and stable coolant flow under all operating conditions. In some embodiments, a control & pump unit / module can include expansion tank(s), air separation, automatic makeup, and / or pressureindependent valves, enabling precise flow and pressure control in air-cooling modes, liquidcooling modes, or any combination thereof.

[0088] FIG. 5 provides an example of a control & pump unit / module 500. In some embodiments, the control & pump unit / module can have a length, depth, and height. FIG. 5 illustrates the height (He), depth (De), and length (Lc) of the control & pump unit / module 500. In some embodiments, at least one of the height, depth, and length of a control & pump unit / module is the same as at least one of the height, depth, and length of a fan unit / module or each fan unit / module. In some embodiments, the height, length, and / or depth of a control & pump unit / module is the same as the height, length, and / or depth of a fan unit / module or each fan unit / module. In some embodiments, the height, length, and depth of a control & pump unit / module is the same as the height, length, and depth of a fan unit / module or each fan unit / module.

[0089] In some embodiments, at least one of the height, depth, and length of a coolant distribution unit / module is the same as at least one of the height, depth, and length of a control & pump unit / module. In some embodiments, the height and / or depth of a coolant distribution unit / module is the same as the height and / or depth of a control & pump unit / module. In some embodiments, the length of a coolant distribution unit / module is larger than the length of a control & pump unit / module. In some embodiments, the length of a coolant distribution unit / module is 2, 3, or 4 times as long as the length of a control & pump unit / module. For example, length LL of FIG. 4 is double the length of Lc of FIG. 5. By having at least one of the height, length, or depth of a control & pump unit / module the same as the height, length, and / or depth a fan unit / module and the same as the height, length, and / or depth of a coolant distribution unit / module, the control & pump units / modules, fan units / modules, and coolant 24MF-366724995Attorney Docket No.: 355532000440distribution units / modules can be in form factors and / or size that one can remove, replace, and / or add to the data center without disrupting the data center and without requiring significant (or any) reconfiguration of other cooling system units / modules that are not being removed or replaced.

[0090] In some embodiments, each fan unit / module, coolant distribution unit / module, and control & pump unit / module can be designed in a modular form factor that allows it to be installed, removed, or reconfigured within the data center as cooling requirements change. In modem data centers, operators normally manage a wide range of environmental variables: air and coolant temperatures, humidity, airflow distribution, water quality, power stability, and climate-related challenges such as wildfire smoke or sudden humidity swings. To manage these conditions reliably, colocation providers, almost universally avoid introducing outside air directly into the IT equipment space. Instead, they typically utilize fully enclosed recirculation systems that can isolate the data hall (e.g., IT hall) from outdoor air. This approach can stabilize thermal conditions, reduce contamination and filter loading, prevent weather intrusions, minimize noise to surrounding areas, and / or eliminate large outside-air plenums — often reducing building height by approximately five feet per floor.

[0091] Industry standards such as ASHRAE continue to influence acceptable temperature and humidity ranges. While hyperscale operators may adopt wider allowable environmental envelopes, colocation facilities typically cannot do so due to the diverse customer hardware hosted within them. In some embodiments, a data center can be designed to deliver air and liquid coolant temperatures near 88 °F (31 °C) or higher, which can maximize efficiency without restricting customer workloads. Operating at lower temperatures can consume more power and, in some systems, more water, despite offering little benefit, as modern IT hardware supports higher operating temperatures. The data center systems and architecture disclosed herein can maintain these high-temperature conditions efficiently even under challenging ambient conditions, matching or exceeding the performance of competing systems while using significantly less water and energy.

[0092] The design of the cooling systems disclosed herein can emphasize standardization, adaptability, and / or operational efficiency. For example, various data centers can use the same indoor cooling platform, composed of modular assemblies that support both air-cooled and liquid-cooled IT loads. These assemblies can be configured or reconfigured at any point in the project lifecycle — even late in construction or after commissioning. Because the fan units / modules and coolant distribution units / modules share a modular form factor, they can be added, removed, or rearranged as customer requirements evolve. As rack densities rise and 25MF-366724995Attorney Docket No.: 355532000440more customers adopt liquid-cooled equipment, traditional fixed-infrastructure data centers will struggle to adjust because they are built around predetermined ratios of air to liquid cooling. Customers who do not yet know the mix of air-cooled versus liquid-cooled hardware they will deploy often force traditional operators to install both systems in full on day one, driving unnecessary capital expense and higher operating costs.

[0093] The data center systems disclosed herein can avoid this limitation through its modular cooling platform, which includes interchangeable units / modules / components forming a complete cooling assembly. Each unit / module / component can have a defined function but can also be designed for independent installation and removal without requiring changes to the overall system. This modularity can be central to the cooling architecture.

[0094] When the data hall demands different cooling temperatures in different areas, multiple assemblies can be dedicated to different cooling modes simultaneously. For example, some assemblies may operate at elevated coolant temperatures to serve dense liquid-cooled racks, while others operate at lower discharge temperatures to support air-cooled racks. This can allow the data centers to tune energy efficiency and distribute capacity without compromising resiliency.

[0095] The designs and systems disclosed herein can also allow operators to install only the cooling equipment needed at any moment. If a customer requires only air cooling, fan-based units / modules can be deployed and coolant distribution units / modules can be omitted. If liquid cooling will be added later, coolant distribution units / modules can be installed only when required, including during live operation. If both systems are desired on day one, the correct ratio can be deployed initially and adjusted later. Furthermore, different areas of the data hall can be configured differently: network racks may require only air cooling, while Al clusters may require substantial liquid cooling. The modular nature of the fan units / modules and coolant distribution units / modules can allow each zone to be equipped and reconfigured independently over time.

[0096] For example, FIG. 6 illustrates three different flexible configurations (A, B, and C) to meet different specifications for cooling. The configurations (A, B, C) illustrate data hall 101, computer devices 203 (darker / filled in computer devices can represent those that are air-cooled only, whereas white computer devices can represent liquid-cooled computer devices (e.g., racks with direct to chip cooling)), fan units / modules 300, coolant distribution unit / module 400, a wall / barrier 605 (e.g., a wire mesh barrier) separating the mechanical areas / region 609 from the data hall 101, data center walls 608, optional supply / return flowline for primary coolant fluid 606, and potential CDU / M 607. In some embodiments, the supply / return flowline 26MF-366724995Attorney Docket No.: 355532000440for primary coolant fluid can be used when a coolant distribution unit and / or fan units / modules can be separated from a control & pump module on the other side of the data hall. Not shown is the fluid lines of the secondary coolant fluid to the data hall and corresponding racks / computer devices. In some embodiments, potential CDU / M 607 can represent being able to add liquid cooling (i.e., coolant distribution units / modules) in the future (i.e, the possibility of retrofitting a 100% air-cooled facility to have liquid cooling).

[0097] In some embodiments, the mechanical areas / regions can contain a cooling system (e.g., the fan units / modules, coolant distribution units / modules, and / or control & pump units / modules). As shown in configurations A and B, the cooling subsystems (a) and (b) can be separated from the data hall 101 in the mechanical areas / regions 609. In contrast, configuration C illustrates cooling subsystems (a) and (b) within the data hall 101. In some embodiments, configuration A illustrates a flexible cooling configuration / system to meet a liquid 40°C specification for cooling. For example, configuration A can illustrate a situation when a customer provides an IT rack specification requiring 40°C fluid temperature circulated to the rack in the secondary fluid network. In some embodiments, configuration B illustrates a flexible cooling configuration / system to meet a liquid 30 / 26°C specification for cooling. For example, configuration B can illustrate a situation when a customer provides an IT rack specification requiring 30°C or 26°C fluid temperature circulated to the rack in the secondary fluid network showing the flexibility of the cooling configuration. In some embodiments, configuration C illustrates a flexible cooling configuration / system to meet a 100% air cooled specification for cooling.

[0098] In some embodiments, a cooling system and / or subsystem can be in the form of one or more arrays or walls. For example, one or more fan units / modules, one or more coolant distribution units / modules, and / or one or more control & pump units / modules can be combined to form one or more arrays or walls. In some embodiments, the one or more fan units / modules, one or more coolant distribution units / modules, and / or one or more control & pump units / modules can be physically connected to form one or more arrays or walls. In some embodiments, the one or more fan units / modules, one or more coolant distribution units / modules, and / or one or more control & pump units / modules can be stacked together to form one or more arrays or walls. In some embodiments, the units / modules can be connected to one another on any side, top, or bottom.

[0099] For example, the cooling system or architecture can be configured by selecting numeric combinations of fan units / modules, coolant distribution units / modules, and control & pump units / modules. Since these units / modules can be standardized in physical form factor,27MF-366724995Attorney Docket No.: 355532000440interfaces, and / or control, the units / modules can be interoperable so that any combination of fan units / modules, coolant distribution units / modules, and control & pump units / modules can be used to meet a specific air / liquid cooling ratio for a given data hall requirement.

[0100] In some embodiments, a cooling wall or cooling array can be made up of at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, or at least 8 units / modules. In some embodiments, a cooling wall or cooling array can include units / modules stacked on top of one another. In some embodiments, any combinations of 0-7 fan units / modules and 0-1 coolant distribution units / modules are possible. In some embodiments, a cooling wall or cooling array may include a blank unit / module or panel. The blank unit / module or panel can have the same form factor or size as a fan unit / module, coolant distribution unit / module, or a control & pump unit / module disclosed herein including the same length, height, and / or depth as any fan unit / module, coolant distribution unit / module, or a control & pump unit / module disclosed herein. In some embodiments, the blank unit / module or panel 703 can be a standard block off as shown in FIG. 7.

[0101] In some embodiments, the units / modules are easy to install and configure with most solutions centered around the control & pump unit / module, which can provide power and control to the other units / modules within the cooling wall or cooling array (that can be an 8 unit / module cooling wall or array).

[0102] In some embodiments, for a given project, a target liquid: air ratio (e.g., 80 / 80 vs. 100 / 10 LC:AC) can be selected. In some embodiments, the target liquid:air ratio can be based on customer defined operating parameters or industry standard parameters. Based on that ratio and megawatt (MW) target (i.e., IT load), a unit / module configuration can be chosen. For example, FIG. 7 illustrates a cooling wall / array 700 that is for 100% liquid cooling. This cooling wall / array includes 1 coolant distribution unit / module 400 and one control & pump unit / module 500. In other words, such a unit / module configuration can be referred to as a 0-1-1 configuration (0 fan units / modules - 1 coolant distribution unit / module - 1 control & pump unit / module). FIG. 7 also illustrates blank units / modules or panels 703. In some embodiments, the blank units / modules or panels can keep air from the data hall from bypassing the fan units / modules (heat exchanger(s) of the fan units / modules) and / or can create an airflow block from exhaust air or supply air bypassing the fan units / modules. In other words, the blank units / modules or panels can be used to guide exhaust air to the fan units / modules.

[0103] FIG. 8 illustrates a cooling wall / array 700 that is for 100% air cooling. This cooling wall / array includes 7 fan units / modules 300 and one control & pump unit / module 500. In other28MF-366724995Attorney Docket No.: 355532000440words, such a unit / module configuration can be referred to as a 7-0-1 configuration (7 fan units / modules - 0 coolant distribution unit / module - 1 control & pump unit / module).

[0104] FIG. 9 illustrates a cooling wall / array 700 that can be up to a 100:80 liquid to air cooled ratio. This cooling wall / array includes 5 fan units / modules 300, 1 coolant distribution unit / module 400, and 1 control & pump unit / module 500. In other words, such a unit / module configuration can be referred to as a 5-1-1 configuration (5 fan units / modules - 1 coolant distribution unit / module - 1 control & pump unit / module). Such a configuration can offer both liquid and air cooling up to a 100% liquid-cooled: 80% air-cooled ratio.

[0105] FIG. 10 illustrates a cooling wall / array 700 that can be predominantly for air cooling. This cooling wall / array includes 3 fan units / modules 300, 0 coolant distribution unit / module, and 1 control & pump unit / module 500. In other words, such a unit / module configuration can be referred to as a 3-0-1 configuration (3 fan units / modules - 0 coolant distribution unit / module - 1 control & pump unit / module). In some embodiments, the 3-0-1 configuration can be paired with the 0-1-1 configuration to achieve 100% liquid-cooled: 10% air cooled ratio. In some embodiments, a 3-0-1 configuration only has the air-based heat rejection from a data hall (e.g., IT racks) achieving a 10% total data hall / IT rack heat rejection via air cooled components in the rack. A 0-1-1 configuration can circulate the heat rejection in the secondary fluid network between the direct to chip cold plates in the racks and the at least one heat exchanger (e.g., liquid-to-liquid heat exchanger) of a coolant distribution unit / module, thereby capable of achieving a 100% liquid cooled heat rejection capacity.

[0106] A person of ordinary skill in the art will recognize that these configurations can be scaled or combined to support larger facilities or different redundancy requirements. As shown above, reconfiguration of the air vs. liquid ratio per data hall can be possible by changing only the counts of fan, cooling distribution, and / or control & pump units / modules, with minimal changes to plant piping and headers. In addition, this cooling architecture / system can allow for late-binding of configuration up to shortly before handover. Thus, data center sites may be planned for one ratio (e.g., 80% air-cooling (AC) / 80% liquid-cooling (LC)) and can be modified later (e.g., 100% LC / 10% AC) by adjusting unit / module counts, without redesigning the entire mechanical plant. In addition, the systems, designs, and architectures disclosed herein can support fungibility between cloud and Al workloads (i.e., the same mechanical plant can be rebalanced as demands change). Furthermore, the systems, designs, and architectures disclosed herein can provide a productized, repeatable module family for multiple campuses and regions, rather than one-off mechanical designs per project or data center.29MF-366724995Attorney Docket No.: 355532000440

[0107] In some embodiments, because the system can be designed so that cooling components can be added or removed without disrupting IT operations, it may be preferred to locate the cooling equipment outside the sensitive IT environment of the data hall such as in separate mechanical areas / regions of the data center. Placing cooling equipment outside the data hall can avoid disruption during construction, expansion, and / or ongoing maintenance while preserving the security and operational integrity of the IT space. The modularity of the cooling system can also enable straightforward retrofits: many existing data centers were designed exclusively for air-cooled racks, and much of their available space outside the data hall is occupied by air-moving equipment. As operators introduce liquid-cooling infrastructure into these legacy environments, traditional designs often place liquid-cooling components directly inside the IT hall near the racks, which can introduce risks related to access, security, and / or equipment servicing. By allowing air-cooled and / or liquid-cooled IT equipment to coexist within a common thermal environment while the mechanical equipment (e.g., cooling systems and / or cooling subsystems) can remain outside the data hall (for example, in mechanical area(s) / region(s) of the data center), the system can maintain service continuity and minimize operational disruption. This layout can also allow sections of the data center or data hall to be temporarily isolated or staged during retrofits without affecting ongoing IT operations.

[0108] FIG. 11 illustrates a cross section of data center system 1100. In some embodiments, the data center can have multiple stories or floors. For example, data center 1100 has two stories or floors as shown in FIG. 11. In some embodiments, the data hall 101 can be separated by a wall / barrier 605 from mechanical areas / regions 609. The mechanical areas / regions can include the cooling systems and / or cooling subsystems 102. The cooling systems or subsystems can include any combination of the one or more fan units / modules, one or more coolant distribution units / modules, and one or more control & pump units / modules disclosed herein. In some embodiments, the cooling systems or subsystems can each be a cooling wall / array made up of any combination of the one or more fan units / modules, one or more coolant distribution units / modules, and one or more control & pump units units / modules disclosed herein. As shown in FIG. 11, a data center can also include external heat rejection systems 1303 (e.g., dry coolers, air cooled chillers, roof chillers, and / or adiabatic dry chillers). FIG. 11 also shows exemplary cooling and heat flow through data center 1100. For example, cool air 1310 (shown as solid arrows) from fan units / modules of a first or front side 1102 of a cooling system or subsystem 102 can be supplied to the data hall 101 for cooling the one or more computer devices 203 with this air. The computer devices can draw in this cold air. The heat from the one or more computer devices can be transferred to the cool air forming a heated 30MF-366724995Attorney Docket No.: 355532000440fluid 1309 (i.e., hot exhaust air from the computer devices (IT racks)) shown as dotted arrows. This heated fluid can flow above the computer devices through a ceiling 1101 and towards a second or back side 1103 of the cooling system or subsystem. Heat from this heated fluid (i.e., hot exhaust air from data hall) can be transferred to a primary coolant fluid 1307 in the cooling system or subsystem, thereby forming cool air to be reused for cooling the computer devices in the data hall. The heated primary coolant fluid 1308 (from heat from the heated fluid) can be sent to heat rejection system 1303 where the heat from the heated primary coolant fluid can be transferred to ambient or outside air forming heated ambient or outside air 1312. In some embodiments, the data center can have a flow channel or system for air to be recirculated by the one or more fan units / modules into and out of the data hall as shown in FIG. 11.

[0109] In some embodiments, the cooling system or subsystem 102 can include one more coolant distribution units / modules. Secondary coolant fluid 1412 from a coolant distribution unit / module can be supplied to the data hall 101 for cooling the one or more computer devices 203 with liquid. Heat from the one or more computer devices can be transferred to the secondary coolant fluid forming a heated secondary coolant fluid 1411. This heated secondary coolant fluid can return back to a coolant distribution unit / module of a cooling system or subsystem. Heat from the heated secondary coolant fluid can be transferred to primary coolant fluid 1307 in the cooling system or subsystem, thereby forming cooled secondary coolant fluid to be reused for cooling computer devices in the data hall. As previously explained, the heated primary coolant fluid 1308 (from heat from the secondary coolant fluid) can be sent to heat rejection system 1303 where the heat from the heated primary coolant fluid can be transferred to ambient or outside air forming heated ambient or outside air 1312.

[0110] During operation, the various units / modules can be monitored and / or controlled via a supervisory control and data acquisition (SCAD A) or building management system. Additional units / modules may be brought online, taken offline, and / or rebalanced to accommodate shifts in workload mix, such as increased deployment of liquid-cooled Al racks. The systems disclosed herein provide several advantages over conventional designs, including lifecycle flexibility, reducing stranded capacity; improved economic efficiency, enabling higher IT load per unit of utility power; operational fungibility, supporting diverse and evolving tenant workloads; and / or repeatability and scalability, enabling standardized deployment across multiple sites. Furthermore, because the one or more units / modules can share standardized interfaces and can be designed to operate as a coordinated family, the selected configuration may be modified late in the project lifecycle. For example, up to several months before handover, the number of fan units / modules and coolant distribution units / modules may be 31MF-366724995Attorney Docket No.: 355532000440adjusted to reflect updated tenant requirements or hardware plans, while leaving primary piping headers, mechanical rooms, and / or external heat-rejection systems substantially unchanged.

[0111] In some embodiments, the units / modules (e.g., fan units / modules, coolant distribution units / modules, etc.) that make up the liquid-based server cooling systems and air-based server cooling systems can be products that can be readily replaced, removed, and / or added over the lifespan of a data center. In some embodiments, during the planning and permitting stage of a data center, the data center can obtain permits for both 100% air-based cooling system and 100% liquid-based cooling system. For example, the permit can say the data center will have 200% capacity (100% air, 100% liquid) then within a set time period (e.g., 6-8 weeks) before the data center is finished, a certain amount of air-based cooling system units / modules and / or liquid-based cooling system units / modules can be omitted from the permit set such that the exact configuration desired is obtained.

[0112] In some embodiments, the air-based cooling system units / modules and / or liquid-based cooling system units / modules can be in a form factor and / or size that one can remove, replace, and / or add to the data center without disrupting the data center. In some embodiments, the fan units / modules making up the air-based cooling system and / or coolant distribution units / modules making up the liquid-based cooling system can be in form factor that allows for addition and / or removal of the units / modules in the data hall / center without significant (or any) modification of other cooling units / modules that are not being removed / replaced.

[0113] For example, on the building inside there may be at least one fan unit / module in a cooling wall or array. If liquid cooling is required, then liquid-to-liquid coolant distribution units / modules can be added, and segments / units of the wall or array can be removed. Initially buildings can be permitted with both fan units / modules and coolant distribution units such that both 100% air and liquid cooling can be provided. When more information is available about the exact racks mix (for example), excess fan units / modules and coolant distribution units / modules can be omitted from the design by submitting an updated mechanical system permit set. The removed fan units / modules can be replaced with blanking panels (for example 703) such that air is blocked (i.e., no air escape routes). In addition, these fan units / modules can be replaced / removed while the data center is in operation. This workflow can be beneficial as it can reduce the permitting time to down to a few weeks as compared to a few months.

[0114] In some embodiments, the electrical architecture can mirror the mechanical modularity. For example, each major subsystem can be designed with N+l redundancy or better. In some embodiments, fan units / modules, control & pump units / modules, coolant distribution units / modules, and / or heat rejection system / equipment can be sized and arranged so the loss of 32MF-366724995Attorney Docket No.: 355532000440any single component does not compromise the cooling capacity required to support the data hall (i.e., committed IT load). In critical areas, reserve modules can be tied into switchover cabinets, enabling rapid transfer of load and / or continuous operation during concurrent maintenance or equipment fault scenarios. In combination, the electrically independent lineups and the component-level redundancy within both the mechanical and / or electrical subsystems can localize most maintenance and fault events, avoid taking large portions of the data center offline, and / or preserve thermal compliance for the IT environment under a wide range of contingencies.

[0115] In some embodiments, the redundancy architectures for cooling systems disclosed herein can include multi-layer redundancy applied to the fan units / modules, coolant distribution units / modules, and / or control & pump units / modules disclosed herein, including architectures that can provide both primary and reserve cooling paths for improved resilience and capacity utilization. As shown in FIG. 12, conventional mechanical cooling systems in data centers are commonly designed using conventional redundancy schemes such as N+l, in which sufficient capacity is provided to meet a design load (N) plus one additional unit; and N+2, in which two additional units or equivalent excess capacity are provided. These schemes are typically applied at the component level, such as chillers, pumps, or air handling units, and assume relatively static loads and homogeneous cooling paths.

[0116] The primary difference of the convention design in FIG. 12 and the disclosed design (i.e., (N+l)+l) can be the primary cooling loop that is a loop within a loop. In other words, the conventional design can include a loop consisting of supply / retum piping between the chillers and pumps which can be a very complex system. This can allow any chiller to fail connected to this second loop and the system can still operate. In the (N+l)+l system, there can be a reserve heat rejection capacity (N+l) on the primary coolant fluid network that can kick on and there can be a simplified piping arrangement. In addition, there can also be +1 redundancy to the N+l by utilizing reserve components (i.e., one or more coolant distribution units / modules, one or more fan units / modules, and / or one or more control & pump units / modules) to get additional redundancy.

[0117] In the data centers disclosed herein that employ modular cooling systems and subsystems, cooling capacity can be delivered via multiple parallel paths; air and liquid cooling systems can be integrated rather than independent; and IT loads, particularly Al and ML workloads, may fluctuate rapidly and unpredictably. Conventional N+l or N+2 models may not map cleanly onto these modular, multi-path systems. Applying traditional redundancy approaches can result in: overbuilding of mechanical infrastructure; complex and brittle 33MF-366724995Attorney Docket No.: 355532000440switchover logic; inefficient utilization of installed capacity; and / or difficulty aligning redundancy planning with modular cooling units and / or multi- site deployment.

[0118] Disclosed herein is a two-layer redundancy architecture, referred to herein as an (N+l)+l redundancy model (and shown in FIG. 12), for the modular data center cooling systems disclosed herein. In some embodiments, a primary cooling system or subsystem can include a plurality of the units / modules disclosed herein (e.g., one or more fan units / modules, one or more coolant distribution units / modules, one or more control & pump units / modules, and / or heat rejection equipment / systems) arranged to provide N+l redundancy, such that N units / modules are sufficient to meet a design cooling load (e.g., data hall cooling load) and one additional units / modules provides redundancy within the primary system at the units / modules level. In some embodiments, the primary cooling system, subsystem, or path (i.e., the primary fan units / modules, primary coolant distribution units / modules, primary control & pump units / modules, and primary heat rejection equipment / systems) can be designed for N+l redundancy (N units plus one redundant unit) to meet full design load. For example, the one or more fan units / modules, one or more coolant distribution units / modules, and one or more control & pump units / modules of the cooling system can be combined into standardized units (i.e., integrated cooling units) sized to serve a known MW of IT load (e.g., per row or group of rows). For a given design load, N such units can be used with one additional unit providing the “+1” redundancy in the primary system. In some embodiments, cooling capacity can be provided by standardized, integrated cooling units that may include air cooling functionality corresponding to one or more fan units / modules, liquid cooling functionality corresponding to one or more coolant distribution units / modules, and / or pumping and control functionality corresponding to one or more control & pump units / modules. Each integrated cooling unit can be rated to serve a known amount of IT load and / or can be configured to operate as a discrete, replaceable block within the overall cooling system. For a given design IT load (or data hall), a primary cooling system can include N integrated cooling units sufficient to meet the full design cooling demand and one additional integrated cooling unit, providing redundancy within the primary system. Under normal operation, the N units may carry the majority of the cooling load; the additional unit may be lightly loaded, load-sharing, or idle; and / or the primary system can tolerate failure or maintenance of a single unit without loss of cooling capacity.

[0119] In some embodiments, a reserve cooling system (e.g., reserve fan units / modules, reserve coolant distribution units / modules, reserve one control & pump units / modules, and / or reserve heat rejection equipment / systems) can provide a separate additional redundancy layer, effectively constituting a second “+1,” that is independent of the primary system and 34MF-366724995Attorney Docket No.: 355532000440configured to backstop the primary system in the event of systemic failures, curtailment events, and / or large load excursions. This distinct reserve system path can provide a second “+1” layer, creating effectively two independent surplus paths: one within the primary system, and one as a reserve system. This separation can allow redundancy to be modeled and deployed at the level of complete cooling paths rather than individual components. Accordingly, such a system can separate redundancy within a primary cooling system from a distinct system-level reserve; apply redundancy models to integrated fan units / modules, coolant distribution units / modules, and / or control & pump units / modules treated as a single product family; and / or provide a repeatable redundancy pattern aligned with standardized capacity blocks (e.g., per 200 MW data center segment). In some embodiments, the reserve fan units / modules, reserve coolant distribution units / modules, and / or reserve control & pump units / modules can be any one of the fan units / modules, coolant distribution units / modules, and / or one control & pump unit / module disclosed herein. In some embodiments, the reserve cooling system can include an additional or reserve integrated cooling unit. As explained above, cooling capacity can be provided by standardized, integrated cooling units that may include air cooling functionality corresponding to one or more fan units / modules, liquid cooling functionality corresponding to one or more coolant distribution units / modules, and / or pumping and control functionality corresponding to one or more control & pump units / modules. In some embodiments, cooling capacity can be provided by standardized, integrated cooling units that may include air cooling functionality corresponding to one or more reserve fan units / modules, liquid cooling functionality corresponding to one or more reserve coolant distribution units / modules, and / or pumping and control functionality corresponding to one or more reserve control & pump units / modules. In some embodiments, the reserve cooling system can include a separate cooling system employing different cooling technology (e.g., a distinct cooling system (e.g., chiller vs. indirect evaporative) sized to backstop N units in case of systemic issues). In some embodiments, the reserve system can be configured to backstop the primary system in events such as: simultaneous or cascading failures within the primary system; utility curtailment or “use it or lose it” events; rapid IT load swings associated with ML or Al workloads; and / or maintenance or commissioning activities affecting multiple primary units. In some embodiments, the reserve system may be normally idle, lightly loaded, and / or operated in a standby mode.

[0120] During normal operation, N integrated cooling units can carry the bulk of the cooling load, with the +1 integrated cooling unit sharing or absorbing contingency events (e.g., maintenance, localized failures, etc.) In larger events (e.g., simultaneous failures, curtailment,35MF-366724995Attorney Docket No.: 355532000440or major ML power swings), the reserve system can engage, acting as the second redundant layer. In some embodiments, control logic can coordinate load sharing and / or automatic switchover between primary and reserve layers.

[0121] In some embodiments, the (N+l)+l redundancy architecture can be applied per standardized capacity block, such as a 200 MW data center segment (or more or less than 200 MW). For each block: the number of integrated cooling units required to satisfy N can be determined; one additional unit can provide primary system redundancy; and reserve system capacity can be sized to provide system-level protection. This approach can support repeatable deployment across multiple sites and phases. In some embodiments, the (N+l)+l model can be applied per amount (e.g., 200, more than 200, less than 200) MW data center with detailed forecasts of how many integrated cooling units can be required and what reserve capacity may be needed to satisfy service level agreements and / or regulatory constraints. This can support lumpy IT demand and ML load swings while maintaining grid code compliance and “use it or lose it” obligations.

[0122] In some embodiments, control logic implemented for example via a supervisory control system, can coordinate operation of the primary and reserve systems. Such control logic may: monitor operating states of integrated cooling units; balance load across primary units; detect fault or curtailment conditions; and / or automatically engage the reserve system when required. A person of ordinary skill in the art will recognize that such control logic may be implemented using known building management or SC AD A systems.

[0123] The (N+l )+l redundancy architecture can be applied to modular cooling architectures disclosed herein that include fan units / modules, coolant distribution units / modules, and / or control & pump units / modules; integrated fan wall and coolant distribution units; and / or end-to-end cooling paths serving defined IT capacity loads / blocks. Advantageously, the disclosed redundancy architecture can provide higher resilience than conventional N+l designs while, in certain topologies, requiring fewer resources than native N+2 designs. In some embodiments, the (N+l)+l redundancy architecture can provide improved resilience compared to simple N+l designs; reduced overbuild relative to certain N+2 implementations; alignment with modular cooling product families; simplified capacity planning and procurement; and / or consistent redundancy behavior across air-cooled and liquid-cooled IT loads.

[0124] In some embodiments, the cooling designs and architectures disclosed herein can minimize reliance on variables that carry the highest uncertainty, such as local water supply, and can maximize performance through high temperature cooling loops and / or low approach36MF-366724995Attorney Docket No.: 355532000440temperature heat exchangers. The result can be a system that can support mixed workloads and evolving IT requirements without rearchitecting the entire mechanical plant.

[0125] In some embodiments, facility cooling systems may be based on indirect evaporative cooling and may include dry coolers with evaporative media pads, for example. Additionally or alternatively, facility cooling systems may include chillers (e.g. trim chillers and / or aircooled chillers) and / or adiabatic cooling systems such as adiabatic chillers. An exemplary facility cooling system may include one or more fan units / modules, for example one or more sets of fan units / modules disclosed herein, and / or one or more evaporative cooling blocks, for example one or more cooling units based on indirect evaporative cooling. Facility cooling system redundancy may be designed to meet an “N+2” standard per fan unit / module and / or per evaporative cooling block, where “N” can represent a minimum number of units for normal operation and “+2” represents two backup units. In some embodiments, additional data center redundancies may include an “N+l” configuration for medium-voltage room split-system cooling units, an “N+l” configuration for high-voltage to medium-voltage client control room split-system cooling units, and / or an “N+l” configuration for critical fuel oil pumps. An exemplary data center may also include makeup water storage tank redundancy set at “N” and / or an “N+l” redundancy configuration for industrial water purification skid systems.

[0126] The occupancy classifications and / or construction types of data centers disclosed herein may align with applicable building codes. For example, the main occupancy group may be classified as an electronic data processing facility, with accessory storage areas, and / or may be classified as required by the local Authority Having Jurisdiction. The construction type of an exemplary data center may be a fully sprinkler-protected structure (e.g., Type II-B). In some implementations, the construction type may vary based on regional construction materials, practices, and / or regulations.

[0127] Office space within an exemplary data center may include one or more private offices and / or open office space that may accommodate a plurality of data center operators. In some implementations, if a data center is the first facility on a campus, the office space of the data center may be expanded to include additional open office seating and / or one or more additional private offices. In some implementations, an exemplary data center may include additional rooms based on operational requirements. A data center may include restrooms in accordance with local regulations, for example restrooms may be included within each security zone of an exemplary data center. Restrooms may include one or more shower rooms, one or more toilets, one or more sinks, and / or one or more dressing areas. A data center may include a quiet room that may include a one or more countertops, one or more sinks, one or more mini-fridges, one 37MF-366724995Attorney Docket No.: 355532000440or more electrical outlets, and / or furniture. A data center may additionally or alternatively include storage areas, for example accounting for at least 1% of the total footprint of the data center. An exemplary data center may include, within the one or more data halls, a designated disk destruction area and / or a wet room for liquid cooling rack preparation.

[0128] To facilitate equipment transportation, the doorways, hallways, and / or corridors of a data center may be designed to accommodate objects with approximate dimensions of 20 feet in length, 10 feet in height, and / or 6 feet in depth. A data center may include a dedicated battery storage area that may be protected by an automatic sprinkler system designed to meet Extra Hazard Group 1 design density specifications including, for example, 0.30 GPM per square foot over a 2,500 square foot area with a 500 GPM hose allowance. Dedicated battery backup storage areas may include air-sampling smoke detection systems in accordance with fire safety regulations. Server rack dimensions may be standardized. For example, a server rack may be 2 feet in width, 3 feet 6 inches in depth, and / or 8 feet in height.

[0129] The security and access control layout of an exemplary data center may include a security entry area, for example a main building entrance, a waiting area, and / or a security control room. A data center may include a “red zone,” for example a high-security area that may include server racks and / or other sensitive equipment. The red zone may include enhanced security measures, for example one or more controlled entry points, access authentication, and / or monitoring systems.

[0130] A data center may include one or more conference areas that may be designed to support staff activities and may include at least one conference room accessible from an entry lobby. A data center may additionally or alternatively include break rooms and / or food preparation areas. The mechanical areas of an exemplary data center may account for 10% to 30% of the total footprint which may ensure sufficient clearance for maintenance. The electrical areas of an exemplary data center may account for 10% to 20% of the total footprint, while building circulation may account for 3% to 10% of the total footprint.

[0131] An exemplary data hall may be designed for high floor loads, supporting up to 405 pounds per square foot for steel frame structures and 390 pounds per square foot for concrete frame structures. Server racks may weigh up to 5000 pounds when moved across a data hall floor. Certain structures associated with a data center, including data halls, centralized water treatment buildings, and / or security structures, may be designated as higher-risk facilities, corresponding to risk category IV, and may be held to stricter structural, mechanical, and electrical standards. A Structural Engineer of Record may be responsible for verifying dead,38MF-366724995Attorney Docket No.: 355532000440live, wind, and / or snow loads in compliance with applicable regulations. The full live load of server racks and / or cable racks may be accounted for in determining the total seismic mass.

[0132] Data center security may be structured using a zone-based approach, dividing an exemplary data center into green, yellow, and / or red zones. Security may be categorized into security zones and security elements. Security zones may physically divide spaces into highland low-security regions using walls and / or electrically locking doors connected to an access control system. Security elements may include additional features such as access control points, metal detectors, and / or disk destruction stations as mentioned above. Security elements may include a yellow / red security station, a security turnstile (yellow zone entry), metal detection (red zone security checkpoint), disk destruction stations, disk destruction station rooms, an exterior fence, a pedestrian turnstile, a pedestrian gate, a guard booth, a crash-rated barrier arm, and / or a pin-badge talk-back access control stanchion. A “zone defense” security arrangement may be implemented, where each zone may allow for a different level of access, corresponding to a green, yellow, and red zone scheme. A security fence may be installed around the equipment yards, designating that space as a yellow zone.

[0133] Exemplary control systems may automate and / or monitor environmental conditions, electrical power distribution, and / or energy-efficiency measures. These systems may provide automation and / or monitoring of data center critical infrastructure, for example server racks and / or associated equipment. Control systems may not include ownership of fire or life- safety systems but may provide monitoring and / or alarming functionality where applicable.

[0134] An exemplary mechanical or cooling system may monitor temperature, pressure, humidity, and / or outside air conditions to maintain acceptable operating ranges for servers and / or associated IT equipment. A mechanical or cooling system may also evaluate local conditions and / or automate energy-saving measures by adjusting mechanical HVAC equipment. An exemplary electrical system may monitor power distribution from an incoming utility supply including the server-critical load, ensuring continuous and / or stable power delivery. An exemplary mechanical and / or electrical system may include a graphical user interface for visualizing real-time conditions, performing equipment overrides, and / or adjusting system setpoints. If any local conditions exceed acceptable thresholds, an exemplary control system may generate user notifications and / or alarms.

[0135] Historical telemetry and / or trending data for an exemplary data center may be collected and / or stored locally. An exemplary control system may include a built-in filtering and / or reporting functionality to enable customized data analysis. Control systems may be designed with redundancies to mitigate power loss, network issues, and / or equipment failures. Critical 39MF-366724995Attorney Docket No.: 355532000440control components may rely on hard-wired connections for reliable automation interactions, while other data integrations may be performed through accepted network protocols.

[0136] The design of an exemplary data center, including doors, elevators, loading docks, and / or security passageways, may enable transport of large IT equipment skids and / or pallets into data halls. For example, an IT equipment skid may be 20 feet in length, 8 feet in height, and / or 4 feet in depth. Conventional data centers may be designed to accommodate only individual rack movement, whereas disclosed data centers may enable larger equipment skids to be maneuvered directly into a data hall. This approach may provide flexibility for future technological advancements that require larger infrastructure, ensuring compatibility with evolving IT equipment form factors. By accommodating large IT equipment skids, an exemplary data center may enable efficient equipment deployment while maintaining security and operational efficiency.

[0137] Those skilled in the art will recognize that the systems and methods described herein are representative, and deviations from the explicitly disclosed embodiments are within the scope of the disclosure.

[0138] In the description of the various embodiments, it is to be understood that the singular forms “a,” “an,” and “the” used in the following description are intended to include the plural forms as well, unless the context clearly indicates otherwise. It is also to be understood that the term “and / or” as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed terms. It is further to be understood that the terms “includes,” “including,” “comprises,” and / or “comprising,” when used herein, specify the presence of stated features, integers, steps, operations, elements, components, and / or units but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, units, and / or groups thereof.

[0139] It is understood that embodiments described herein can include “consisting” and / or “consisting essentially of’ embodiments. For all methods, systems, products, data centers, assemblies, etc. described herein, the methods, systems, products, data centers, assemblies, etc. can either comprise the listed components or steps, or can “consist of’ or “consist essentially of’ the listed components or steps. When a system, product, data center, assembly, etc. is described as “consisting essentially of’ the listed components, the system, product, data center, assembly, etc. contain the components listed, and may contain other components which do not substantially affect the performance of the system, product, data center, assembly, etc., but either do not contain any other components which substantially affect the performance of the system, product, data center, assembly, etc. other than those components expressly listed; or 40MF-366724995Attorney Docket No.: 355532000440do not contain a sufficient concentration or amount of the extra components to substantially affect the performance of the system, product, data center, assembly, etc.. When a method is described as “consisting essentially of’ the listed steps, the method contains the steps listed, and may contain other steps that do not substantially affect the outcome of the method, but the method does not contain any other steps which substantially affect the outcome of the method other than those steps expressly listed.

[0140] This application discloses several numerical ranges in the text and figure. The numerical ranges disclosed inherently support any range or value within the disclosed numerical ranges, including the endpoints, even though a precise range limitation is not stated verbatim in the specification because this disclosure can be practiced throughout the disclosed numerical ranges.

[0141] The foregoing description, for the purpose of explanation, has been described with reference to specific embodiments and / or examples. However, the illustrative discussions above are not intended to be exhaustive or to limit the invention to the precise forms disclosed. Many modifications and variations are possible in view of the above teachings. The embodiments were chosen and described in order to best explain the principles of the techniques and their practical applications. Others skilled in the art are thereby enabled to best utilize the techniques and various embodiments with various modifications as are suited to the particular use contemplated.41MF-366724995

Claims

Attorney Docket No.: 355532000440CLAIMS1. A data center system comprising:a data hall comprising a plurality of computer devices;a heat rejection system comprising at least one first heat exchanger, wherein the heat rejection system is outside the data hall; anda cooling system comprising:a plurality of fan units, wherein each fan unit comprises at least one fan and at least one second heat exchanger and each fan unit is configured to transfer heat from air from the data hall to a primary coolant fluid; and / or a coolant distribution unit comprising at least one third heat exchanger, wherein the coolant distribution unit is configured to transfer heat from heated secondary coolant fluid from the data hall to primary coolant fluid; anda control & pump unit comprising one or more pumps, wherein the control & pump unit is configured to receive the heated primary coolant fluid from each fan unit and / or the coolant distribution unit and send the heated primary coolant fluid to the heat rejection system.

2. The data center system of claim 1, wherein each fan unit and the coolant distribution unit are in a form factor that allows for addition and / or removal from the data center system.

3. The data center system of any one of claims 1-2, wherein the at least one first heat exchanger is an air-to-liquid heat exchanger, the at least one second heat exchanger is an air-to-liquid heat exchanger, and the at least one third heat exchanger is a liquid-to-liquid heat exchanger.

4. The data center system of any one of claims 1-3, wherein each fan unit has the same length, height, and depth.

5. The data center system of claim 4, wherein the coolant distribution unit has at least one of the same length, height, and depth as each fan unit.

6. The data center system of claim 5, wherein the coolant distribution unit has a length twice as long as the length of each fan unit.42MF-366724995Attorney Docket No.: 3555320004407. The data center system of any one of claims 1-6, wherein the cooling system is in a mechanical area of a data center separated from the data hall.

8. The data center system of claim 7, wherein the mechanical area is separated from the data hall by at least one wire mesh barrier.

9. The data center system of any one of claims 1-8, further comprising a second cooling system comprising:a second plurality of fan units, wherein each of the second plurality of fan units comprises at least one fan and at least one second heat exchanger and each of the second plurality of fan units is configured to transfer heat from air from the data hall to primary coolant fluid; and / ora second coolant distribution unit comprising at least one third heat exchanger, wherein the second coolant distribution unit is configured to transfer heat from heated secondary coolant fluid from the data hall to primary coolant fluid,wherein the first cooling system is on a first side of the data hall and the second cooling system is on a second side of the data hall opposite the first side.

10. The data center system of claim 9, wherein the first cooling system is parallel to the second cooling system.

11. The data center system of any one of claims 9-10, wherein the first cooling system is a first wall or array and the second cooling system is a second wall or array.

12. The data center system of any one of claims 1-11, wherein the cooling system comprises the plurality of fan units and the coolant distribution unit, and the control & pump unit is configured to receive the heated primary coolant fluid from each fan unit and from the coolant distribution unit and send the heated primary coolant fluid to the heat rejection system.

13. The data center system of any one of claims 1-12, wherein the plurality of computer devices are powered by at least one first lineup; and the cooling system is powered by at least one second lineup different from the at least one first lineup.

14. A method comprising:43MF-366724995Attorney Docket No.: 355532000440transferring heat from a data hall comprising a plurality of computer devices to air;flowing the heated air to a plurality of fan units, wherein each fan unit comprises at least one first heat exchanger and at least one fan;transferring heat from the heated air to a primary coolant fluid in each of the fan units;flowing the heated primary coolant fluid from each fan unit to at least one pump of a control & pump unit;transferring the heated primary coolant fluid from the control & pump unit to a heat rejection system comprising at least one second heat exchanger; and transferring heat from the heated primary coolant fluid to air outside the data hall.

15. The method of claim 14, further comprising transferring heat from the data hall to a secondary coolant fluid and flowing the heated secondary coolant fluid to a coolant distribution unit comprising at least one third heat exchanger.

16. The method of claim 15, further comprising transferring heat from the heated secondary coolant fluid to primary coolant fluid in the coolant distribution unit.

17. The method of claim 16, further comprising flowing the heated primary coolant fluid from the coolant distribution unit to the heat rejection system.

18. The method of claim 16, further comprising flowing the heated primary coolant fluid from the coolant distribution unit to the at least one pump of the control & pump unit.

19. The method of claim 15, wherein the at least one first heat exchanger is an air-to-liquid heat exchanger, the at least one second heat exchanger is an air-to-liquid heat exchanger, and the at least one third heat exchanger is a liquid-to-liquid heat exchanger.

20. The method of claim 16, further comprising transferring cooled primary coolant fluid from the heat rejection system to the plurality of fan units, the coolant distribution unit, and / or the control & pump unit.44MF-366724995

Images