Distributed Power Units for Data Centers

Abstract

A technique for providing auxiliary power to a data center typically served by an electric utility is disclosed. The system includes a controller, a first stamp, and a second stamp. A common DC bus couples the first stamp to the second stamp on the DC side. A common AC bus couples the first stamp to the second stamp on the AC side (e.g., the output of the stamp). An AC breaker is disposed on the AC bus between the first stamp and the second stamp. The controller controls whether the AC breaker is open or closed. The default state of the AC breaker is open. Control of the breaker can be for various reasons, including whether the utility is available or whether the utility is unavailable. The utility may be unavailable due to a utility failure or other reason.

Description

[Background technology]

[0001] background A data center is a facility that contains a variety of different computing systems and architectures. Data centers are generally designed to store and share data, applications, or even provide infrastructure to clients. Examples of various systems included in a data center include servers, switches, routers, security systems, etc.

[0002] Maintaining and operating a data center can be a huge task. For example, customers typically expect their data to be available upon request at any time of the day. This therefore means that data centers must be designed to be operational at all times. As a result, the work associated with providing power to data centers is often considered a high priority.

[0003] Most data centers are powered through a local power grid as their primary or main power source. This power grid provides a continuous flow of energy to the data center in a reliable manner. Data centers often actually include a variety of different transformers and other power components to convert incoming power into a power signal that can be used by the computing elements within the data center.

[0004] In addition to the power grid, data centers often actually include a backup power source, such as a generator, an uninterruptible power supply (UPS), a battery system, or a capacitor system, for each line-up of the data center, where line-up refers to a set of computing systems that can optionally be powered directly by the backup power system. This secondary power system is designed to provide power to the data center in the event of a power grid failure.

[0005] Solid oxide fuel cell (SOFC) technology is an excellent energy source for baseload power and can operate independently or in parallel with a utility (i.e., the power grid). A variety of different fuel cell inverters can operate at a fixed power setpoint and can be used to generate a fixed amount of power (grid-following mode). In grid-following mode, the utility can optionally make up for any difference, delta, or shortfall that may exist between the fuel cell's power generation and load demand. In traditional SOFC grid-forming applications (e.g., stand-alone) mode, the fuel cell's step load capability relies on pre-packaged ultracapacitors for the loads, and the behavior of these loads is predictable or designed to enable specific load-shedding schemes in microgrid (MG) operation. However, clearly, the behavior of loads in data centers is often unpredictable. The use of SOFC technology in data centers therefore requires improved techniques for managing how SOFCs are used to provide power to data centers.

[0006] The subject matter claimed herein is not limited to embodiments that solve any disadvantages or that operate only in environments such as those described above. Rather, this background is provided to illustrate one example technology area where some embodiments described herein may be practiced. Summary of the Invention [Means for solving the problem]

[0007] Quick Overview SUMMARY OF THE INVENTION The embodiments disclosed herein generally relate to systems, apparatus, and methods for providing auxiliary power to a data center serviced by an electric utility.

[0008] Some embodiments include a controller, a first stamp, and a second stamp. Of course, any number of stamps can be used, with each stamp made up of several fuel cell-inverter pairs connected to one step-up transformer. Some embodiments further include a common direct current (DC) bus coupling the first stamp to the second stamp and a common alternating current (AC) bus coupling the first stamp to the second stamp. An AC breaker is disposed on the AC bus between the first stamp and the second stamp. The controller controls whether the AC breaker is open or closed. The default state of the AC breaker is open.

[0009] According to some embodiments, a set of stamps operates in a grid-following mode, and the stamps are coupled together using a common AC bus. For example, some embodiments determine that a power providing utility for a data center is currently available to provide power to the data center. According to this embodiment, a set of breakers located on a common AC line remains open. This set of breakers couples multiple stamps on the AC line that can provide auxiliary power to the data center. According to this embodiment, the stamps operate in a grid-following mode while the utility is providing power to the data center.

[0010] According to some embodiments, a set of stamps provides power to a data center, and the stamps are coupled together using a common AC bus. For example, some embodiments determine that the power providing utility for the data center is currently unavailable to provide power to the data center. This embodiment determines that the utility is currently unavailable due to a utility fault. This embodiment leaves open a set of breakers located on a common AC line. This set of breakers couples multiple stamps on the AC line that can provide auxiliary power to the data center. This embodiment activates a set of inverters associated with the stamps. The stamps provide power to the data center.

[0011] This Summary is provided to introduce selected concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of the claimed subject matter.

[0012] Additional features and advantages will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the teachings herein. The features and advantages of the invention may be realized and obtained by means of the instruments and combinations particularly pointed out in the appended claims. Features of the invention will become more fully apparent from the following description and the appended claims, or may be learned by the practice of the invention as set forth hereinafter.

[0013] BRIEF DESCRIPTION OF THE DRAWINGS To describe the manner in which the foregoing and other advantages and features can be obtained, the subject matter briefly described above will be more particularly described by reference to specific embodiments illustrated in the accompanying drawings, in which the embodiments will be described and explained with additional specificity and detail, with the understanding that these drawings illustrate exemplary embodiments only and are not therefore to be considered limiting in scope. [Brief explanation of the drawings]

[0014] [Figure 1] 1 shows an example of an electricity provider. [Figure 2] 1 illustrates an example of a data center. [Figure 3] 1 illustrates an improved architecture that can provide supplemental or secondary power to a data center. [Figure 4] Various examples of stamps are shown. [Figure 5] 1 shows an example of a power supply unit. [Figure 6] 1 illustrates an architecture designed to facilitate control of specific stamps. [Figure 7]1 shows a process flow detailing how the stamp can be controlled. [Figure 8] 1 illustrates an exemplary computer system capable of performing any of the disclosed operations. DETAILED DESCRIPTION OF THE INVENTION

[0015] Detailed Description The disclosed embodiments generally relate to a system for providing auxiliary power to a data center typically served by an electric utility. Embodiments also relate to control of the system.

[0016] Some embodiments include a controller, a first stamp, and a second stamp. A common DC bus couples the first stamp to the second stamp, and a common AC bus couples the first stamp to the second stamp. An AC breaker is disposed on the AC bus between the first stamp and the second stamp. The controller controls whether the AC breaker is open or closed. The default state of the AC breaker is open.

[0017] According to some embodiments, a set of stamps operates in grid-following mode, and the stamps are coupled together using a common AC bus. The utility providing power for the data center is currently available to provide power to the data center. A set of breakers located on a common AC line is open. This set of breakers couples multiple stamps on the AC line that can provide auxiliary power to the data center. According to this embodiment, the stamps operate in grid-following mode while the utility is providing power to the data center.

[0018] According to some embodiments, a set of stamps provides power to a data center, and the stamps are coupled together using a common AC bus. Due to a utility failure, the power providing utility for the data center is currently unavailable to provide power to the data center. According to this embodiment, a set of breakers located on a common AC line remains open. This set of breakers couples multiple stamps on the AC line that can provide auxiliary power to the data center. This embodiment activates a set of inverters associated with the stamps. The stamps provide power to the data center.

[0019] Examples of technical benefits, improvements and practical applications The following section outlines some exemplary refinements and practical applications brought about by the disclosed embodiments, but it will be understood that these are merely examples and that the embodiments are not limited to only these refinements.

[0020] The disclosed embodiments provide many benefits, advantages, and practical applications in the field of power system management, particularly for data centers. Specifically, the disclosed embodiments advantageously provide modular and flexible data center power control applications. The embodiments also provide for the integration of power sources (e.g., SOFC or any other equivalent technology) into a data center's lineup, where (as noted above) the lineup refers to the set of computing systems that can optionally be powered directly by a SOFC system.

[0021] Embodiments also advantageously provide a full DC bus connection between all stamps, where a stamp refers to a set of one or more energy sources (e.g., SOFCs) and one or more ultracapacitors connected to a step-up transformer. Embodiments also allow for new connections through normally open circuit breakers present on the AC side of the stamps. Yet another benefit relates to the control process used to manage these various different components.

[0022] By implementing the disclosed principles, embodiments allow for unlimited expansion of power system planning, such as by a step-by-step process or by including several modules (also known as power units) at a time and adding more as lineup or demand increases. The disclosed embodiments can also be fully deployed in a short period of time (e.g., one day or less).

[0023] Unlike conventional SOFC configurations that work with a central unit installed to supply fixed loads, the disclosed novel design allows for flexibility and customized, scalable solutions at the lineup level. This includes flexibility in sizing each Energy Server (ES) (i.e., a group of specific energy sources such as SOFCs) and the stamps (i.e., several Energy Servers connected to one step-up transformer) on which these ESs operate. Optionally, embodiments allow for equal or unequal stamp sizes (e.g., stamp sizes can be, but need not be, equal due to data center design considerations).

[0024] Over the next few years, the demands on data centers are likely to increase significantly. It is often desirable to be able to disconnect the data center from the traditional power grid and control the data center's microgrid (i.e., a local power generation system that provides baseload power coverage for the data center). The baseload power for the microgrid comes from local energy resources, which may be renewable and / or environmentally friendly.

[0025] When it comes to baseload applications, there are several options available, and it has been shown that SOFCs can significantly outperform other existing technologies (e.g., natural gas (NG) or H2 conversion to electricity) in terms of efficiency and emissions. The disclosed embodiments can build SOFC-based microgrids that provide reliable, scalable, on-demand power.

[0026] As mentioned above, the data center industry is experiencing exponentially increasing demand, and this demand must be met by building data centers locally. However, in some cases, local utilities / grids may not be able to provide the capacity or consistency required to power these data centers. Therefore, data center builders must bring their own microgrids, which include distributed energy resources. These distributed energy resources are often desirably (i) designed to reduce emissions, (ii) highly reliable, (iii) redundant, and (iv) scalable. The disclosed solution can be scaled up or down to meet any design constraints.

[0027] Advantageously, the disclosed embodiments also provide natural gas-to-electricity conversion efficiencies of the high 50% to low 60% range, and when combined with some heat recovery technology, this efficiency can be increased by 20-30%. Note how SOFCs can also operate with hydrogen instead of natural gas as fuel, if the infrastructure becomes available. Utilizing H2 can also make the entire power production an emissions-free process (assuming environmentally friendly hydrogen as the fuel). Accordingly, the above benefits, and many others, will now be described in more detail in the remaining sections of this disclosure.

[0028] Business entities and data centers Attention now turns to Figure 1, which illustrates an exemplary utility 100 that can provide power to a power grid. Utilities 100 can be any type of power generating utility without limitation. Examples of utilities can include any type of fossil fuel utility, hydroelectric utility, solar utility, wind utility, etc. As shown in Figure 1, utility 100 provides power to remotely located components via any number of transmission lines.

[0029] Figure 2 shows power lines providing power from utility 100 of Figure 1. In this exemplary situation, these power lines are connected to data center 200. Data center 200 includes any number of lineups, such as lineup 205. As previously mentioned, a lineup refers to a set of computing elements that can be collectively powered by a utility and a secondary energy source, such as the SOFCs mentioned above. A lineup can include any number of servers, routers, switches, etc. Optionally, power can be separated between the various lineups.

[0030] Although not labeled, data center 200 may also include, without limitation, any number of cooling or temperature regulation systems, ventilation systems, computing elements, etc. Providing a consistent and reliable power signal to data center 200 is highly desirable. Typically, that power signal is provided by a utility. However, occasionally, the utility may go offline or experience a voltage drop or blackout. To account for such potential situations, data centers typically include or have access to a second power source, such as a battery bank, capacitor bank, SOFC, or any other power generation element. These elements may operate in parallel with the utility. If the utility goes offline, these other power generation elements can then be used to provide power to the power grid for at least some period of time. The disclosed embodiments are directed to improved configurations and uses of secondary or auxiliary power systems.

[0031] Auxiliary Power Design System The disclosed embodiments provide a distributed energy solution for meeting the load demands of a data center using an energy resource or power supply unit (e.g., SOFC). The power supply unit provides power to the data center either in parallel with a utility (e.g., in so-called grid-following mode) or as an independent voltage source (e.g., in so-called grid-forming mode). The power supply unit functions as a current source when in parallel with a utility, and as a voltage source when operating in a microgrid or so-called islanded mode.

[0032] In grid following mode, the system is provided with a set point and then produces power at a desired level (i.e., the set point) depending on voltage and frequency.

[0033] As an example, assume the system is instructed to operate at 1 MW, in which case the system operates to provide 1 MW of power.

[0034] When the system is in grid-forming mode, the system follows the load as it increases or decreases. If a utility is available, the utility makes up any shortfall between the power generated by the system and the load demand. That is, in grid-forming mode, the system serves the load. If there is a utility failure, a seamless transition can occur between grid-forming and grid-following modes. Optionally, a combination of batteries, ultracapacitors, and solid-state fuel cells can be used to cover the load for long periods of time.

[0035] The voltage source operation of the power supply unit with droop capability allows for the scalable addition of other distributed energy resources to the same microgrid system, where the second or subsequent distributed energy resources may operate in grid-forming or grid-following modes. The proposed distributed energy resource configuration provides a reliable, scalable, and redundant 24 / 7 power source to data centers.

[0036] The disclosed solution and architecture consists of multiple stamps, each of which can operate at different voltage levels depending on the DC load block. The power rating of the stamps can be increased or decreased depending on the demand load and design. Each stamp consists of one or more energy servers. Both the size of the power supply units / modules within the energy servers and the number of energy servers within a stamp can be changed, thus the disclosed modular solution is 100% flexible.

[0037] The system can initially perform a black start (i.e., the process of starting the system for the very first time, such as after a power outage event) with or without a utility (i.e., power grid). After black start, the modules in the stamp are continuously monitored and simultaneously maintainable. The design can include a customizable number of ultracapacitors to provide instantaneous power for rapid load changes. This solution can also optionally utilize natural gas infrastructure or biogas. The fuel cell system does not necessarily rely on batteries as an internal resource, although batteries can be used. Note that natural gas infrastructure is much more reliable than electric infrastructure, which greatly increases the availability of this scheme compared to utility availability.

[0038] Exemplary Architecture Attention is now directed to Figure 3, which illustrates an exemplary architecture 300 that may be used to provide the disclosed benefits and advantages. Architecture 300 is shown as including or connected to utility 305, which represents utility 100 of Figure 1. Utility 305 may be connected to the system in several different ways. For example, Figure 3 illustrates a left-hand path in which utility 305 is connected to a first transformer 310 and corresponding breaker (not labeled and shown as a square below transformer 310). Additionally, Figure 3 illustrates a right-hand path in which utility 305 is connected to a second transformer 315 and corresponding breaker (not labeled and shown as a square below transformer 315).

[0039] Architecture 300 includes service 300A. As used herein, the term service refers to an automated program responsible for performing various actions based on inputs. In some cases, service 300A may be a deterministic service that operates completely without randomization factors given a set of inputs. In other cases, service 300A may be or include a machine learning (ML) or artificial intelligence engine.

[0040] As used herein, reference to any type of machine learning or artificial intelligence may include any type of machine learning algorithm or device, convolutional neural networks, multi-layer neural networks, recurrent neural networks, deep neural networks, decision tree models (e.g., decision trees, random forests, and gradient boosted trees), linear regression models, logistic regression models, support vector machines (SVMs), artificial intelligence devices, or any other type of intelligent computing system. Any amount of training data may be used to train (and possibly subsequently refine) a machine learning algorithm to dynamically perform the disclosed operations.

[0041] In some implementations, service 300A is a cloud service running in a cloud environment. In some implementations, service 300A is a local service running on a local device, such as a device in a data center. In some implementations, service 300A is a hybrid service including a cloud component in communication with a local component. Optionally, a machine learning engine can be used to adjust the thresholds, probability values, increment values, and confidence values.

[0042] Service 300A can be used to dynamically control the various stamps, switches, breakers, transformers, and other components shown in Figure 3. Service 300A can operate architecture 300 in a grid-following mode or a grid-forming mode.

[0043] Transformers 310 and 315 are step-down transformers that step down the high voltage coming from utility 305 to a medium voltage (e.g., voltages above about 1000V and below about 69KV). Transformers 310 and 315 may be considered part of their respective medium voltage distribution systems. Another set of transformers, described below, step down the medium voltage to a lower voltage usable by the data center.

[0044] The left and right paths are also connected to a series of switches, one of which is labeled switch 320A. This switch may be used in some embodiments to determine which path is used to provide power to data center 325, which represents data center 200 in FIG. 2. When the left path is providing power to the data center, the left portion of each switch is closed and the right portion of each switch is open. When the right path is providing power to the data center, the right portion of each switch is closed and the left portion of each switch is open.

[0045] Architecture 300 also shows several breakers (e.g., empty rectangular boxes), one of which is labeled breaker 330. These breakers can be used as cut-off switches to cut off power to data center 325.

[0046] Additionally, any number of other transformers (e.g., transformer 335) may be provided to further modify the voltage level of the power signal sent to data center 325. For example, transformer 335 converts a medium voltage to a low voltage supplied to data center 325.

[0047] In accordance with the disclosed principles, architecture 300 may be further constructed to include any number of so-called stamps, such as stamp 340, stamp 345, stamp 350, stamp 355, etc. Stamps may also be associated with switches, as illustrated by switch 320B. These stamps include the scalable modules (also known as power supply units) previously described. Architecture 300 is designed so that any number of stamps can be easily incorporated into architecture 300 to provide secondary or auxiliary power to data center 325. In this exemplary situation, there are four stamps, but it will be understood how the number of stamps may vary based on the design needs of data center 325. Furthermore, additional stamps may be added to architecture 300 as additional computing elements are added to data center 325 or as the load on an existing data center 325 increases.

[0048] 3 is built with scalable distributed stamps connected to the medium-voltage side of step-down transformers (e.g., transformers 310 and 315), and is a unique configuration relative to conventional power systems. That is, this particular configuration (i.e., stamps sized to meet lineup demands with tunable built-in redundancy and stamps connected at the medium-voltage level) is novel.

[0049] Another unique aspect of the disclosed embodiments is that the DC connections of stamps can be connected not only between stamps but also within them, so that a DC connection from one stamp is coupled to a DC connection on another stamp, so that the connected DC sides form a large pool of DC energy that grows as more modules (i.e., power supply units) are added.

[0050] Yet another unique aspect of the disclosed embodiments is that the stamps' AC connections can also be connected between stamps. The AC sides are connected through various circuit breakers (e.g., breakers 370, 375, 380) that are normally open. As described in more detail below, the breakers (e.g., breakers 370, 375, and 380) are closed during power startup when no utility power is available (e.g., so-called island operation) and open once the island startup process is complete, limiting fault current on the AC side. This allows each fuel cell module (i.e., power unit) to supply its own load (i.e., lineup), while all stamps can use a common pool of DC energy.

[0051] Normally open circuit breakers on the AC side (e.g., breakers 370, 375, and 380) eliminate the need for synchronization between every AC power supply unit, even as size increases, such as by adding new stamps. This design provides the flexibility to customize each module / power supply unit and add them together to fit the size of the data center. This concept provides a unique and novel auxiliary power scheme to allow for phased builds or full-size one-time deployment.

[0052] As previously mentioned, each respective stamp may be responsible for serving a particular lineup within data center 325. For example, stamp 340 may serve a first lineup within data center 325, stamp 345 may serve a second lineup within data center 325, stamp 350 may serve a third lineup within data center 325, and stamp 355 may serve a fourth lineup within data center 325.

[0053] In accordance with the disclosed principles, architecture 300 can be designed so that the stamps have a common direct current (DC) bus, as will be shown below. Additionally, architecture 300 can also be designed so that the stamps have a common alternating current (AC) bus, as shown by common AC line 360. Note that common AC line 360 ​​can include a logical disconnect in the form of a normally open breaker, as shown by breakers 370, 375, and 380. Figure 3 also shows stamp output 365 and how output 365 connects to the utility's lines that supply power to data center 325.

[0054] That is, architecture 300 is constructed to provide various different breakers between the various stamps. For example, breaker 370 is an AC breaker on common AC line 360, with breaker 370 logically located between stamp 340 and the other stamps. Breakers 375 and 380 are also AC breakers, providing additional disconnection between the various stamps. As will be described in more detail below, these breakers can be controlled during the black start and ramp-up process. In some cases, breakers 370, 375, and 380 are normally open breakers, and opening and closing these breakers can be used to control inrush currents that may occur during black start. When operating in grid-forming mode, the demands on the fuel cell / power unit are even more stringent. However, connecting the fuel cell on the AC side helps reduce the burden on the fuel cell to regulate inrush currents, especially during startup.

[0055] Optionally, architecture 300 may include a flywheel 385 for providing power to data center 325. More details on this aspect are provided below. In general, flywheel 385 may serve as a substitute for an ultracapacitor.

[0056] FIG. 4 shows an example stamp 400 that may represent any of the stamps shown in FIG. 3. Stamp 400 is shown as including a first power supply unit 405 and a second power supply unit 410. Ellipsis 415 indicates how any number of power supply units may be included in a single stamp. The power supply units are connected to each other at the point where they share a common DC bus, as indicated by common DC 420. Stamp 400 is connected to or optionally includes a transformer 425 that is different from transformer 335 of FIG. 3.

[0057] FIG. 4 shows a second stamp 430. Stamp 430 includes a first power supply unit 435 and a second power supply unit 440. Ellipsis 445 indicates how this stamp may include any number of power supply units. Optionally, the number of power supply units in stamp 400 may differ from the number of power supply units in stamp 430. As indicated by common DC 450, the power supply units in stamp 430 also share a common DC bus. Furthermore, stamp 400 may be connected to stamp 430 via a common DC bus, as indicated by common DC 455. Thus, the power supply units within a single stamp have a common DC bus. Furthermore, the common DC bus may be used to connect the power supply units in the first stamp to the power supply units of the second stamp.

[0058] FIG. 4 also shows how stamps 400 and 430 can share a common AC bus, as indicated by common AC 460. Common AC 460 represents common AC line 360 ​​in FIG. 3. Breaker 465 is also shown. Breaker 465 represents any of breakers 370, 375, or 380 in FIG. 3. Breaker 465 can be considered a normally open breaker. When breaker 465 (or breakers 370, 375, or 380) is open, each stamp serves only its own respective lineup, and synchronization between the various stamps is not performed or required. When the breaker is open, each stamp operates independently and supplies its own load on the AC side. One main benefit of breakers 465, 370, 375, and 380 occurs when black start is required and when the utility is unavailable.

[0059] Figure 5 provides further details regarding power supply units 405, 410, 435 and 440 of Figure 4. Specifically, Figure 5 shows a unit 500 that may represent any of the power supply units described above.

[0060] Unit 500 is shown as including an energy server 505 that includes a first SOFC 510 and a second SOFC 515. Ellipsis 520 indicates how any number of SOFCs may be included in energy server 505. While Figure 5 shows an example involving the use of SOFCs, one skilled in the art will recognize how any other type of energy source may also be used or may alternatively be used in place of SOFCs.

[0061] SOFCs output DC power, often (but not always) at about 50 KW to about 75 KW. Energy servers can output about 300 KW to about 325 KW.

[0062] FIG. 5 also shows how unit 500 includes an ultracapacitor, as indicated by capacitor 525. Capacitor 525 is provided to facilitate ramp-up events, bringing unit 500 online quickly to provide power. Capacitor 525 may subsequently operate to immediately discharge power while the SOFC ramps up. The SOFC may also be used to charge capacitor 525 as it operates. Optionally, capacitor 525 may be a 100 KW or 1500 KJ type capacitor. Of course, other sizes may be used as well. An alternative form of capacitor 525 may be the flywheel discussed earlier.

[0063] Capacitors 525 and energy servers 505 are shown as sharing a common DC bus, as indicated by common DC 530. This common DC 530 represents common DC 420 in Figure 4. As a result of this common DC bus, if any particular data center is heavily loaded, any number of capacitors and / or power supply units may be used to facilitate servicing that load.

[0064] By providing a common DC 530, all stamps have access to a common pool of DC energy, including the power stored in the capacitors. Obviously, any number of capacitors can be included in this architecture.

[0065] This common connection on the DC side also handles sudden load changes that may occur on each line-up or temporary uneven loads on the AC side: if one line-up is overloaded, it can draw more energy from the DC side as long as energy is available without causing a voltage drop on the DC bus, which acts like the heartbeat of the whole system.

[0066] The inclusion of more modules / power units results in more step load capability being offered for each lineup. Additionally, since SOFC technology is designed for base load applications, sharing a large DC energy bus is advantageous as the ultracapacitors provide step load energy for a short period of time before the SOFCs increase their power and take over the load from the ultracapacitors.

[0067] The same power sharing feature applies to situations of overload on one or more lineups. This is also true when the load on the AC side of the inverter suddenly drops on one lineup, because the extra energy on the DC side can be compensated for by other medium-voltage lineups that may be experiencing increased load. Thus, by connecting every DC bus, this design provides an energy reservoir with a relatively flat energy-versus-time graph compared to when the individual DC buses are not coupled together. Because it is difficult to predict changes in maximum load on a single lineup (e.g., due to having to handle customer demand rather than the characteristics of the data center lineup), the disclosed embodiments are highly advantageous because they provide a high level of flexibility in dealing with unpredictable loads.

[0068] Capacitor 525 and energy server 505 are also shown as coupled or connected to a power inverter 535. Inverter 535 converts DC power to AC power. Optionally, a breaker 540 can be located downstream of inverter 535. Inverter 535 is then connected to a transformer, such as transformer 425 in FIG. 4. Often, power inverter 535 is a 355 KVA inverter, although other sizes can be used.

[0069] Inverter 535 may be considered a main or central inverter that serves multiple SOFCs. However, in some implementations, each respective SOFC may be provided with or associated with its own corresponding inverter. Thus, in some cases, the number of inverters in a system may match the number of power supply units or SOFCs. In other cases, each power supply unit may be provided with a single inverter, and the power supply unit may include multiple different SOFCs.

[0070] In island operation, the first stamp self-starts the other stamps and the entire system by closing a normally open breaker (e.g., breaker 370 in FIG. 3). Once the system has started and the voltage and frequency have stabilized, those other breakers (e.g., breakers 375 and 380) can be opened to avoid many fault currents on the AC side and to isolate the modules / power supply units, thereby minimizing the risk of various interferences. This is beneficial because if the outputs of the power supply units are combined, they can send power to the other side, which can cause problems for the inverter.

[0071] If the utility is available, these breakers (e.g., breakers 370, 375, and 380) may be used as normally open. If the utility is present, black start can be performed by this utility. If the utility fails and the power supply units (e.g., SOFC) take over the load after the power outage, these breakers do not need to be closed because the power supply units (e.g., the fuel cells in those units) go into standby and turn back on as soon as the system is fully isolated from the grid. This time interval (i.e., the time it takes the inverter to go into standby and back) is also used for the inverter to switch from current source or grid following mode to voltage source or grid forming mode.

[0072] That is, while the power grid is operating, the SOFCs can operate in parallel with the grid. In the event of a utility failure, the system will isolate the system from the grid for a period of time (e.g., 2 seconds or whatever time is set based on local codes). The SOFCs can go into standby mode or seamlessly switch to grid-forming mode. This process can be deployed in stages (e.g., possibly 10 MW, 20 MW, etc.) or all at once.

[0073] A control mechanism, service, or controller is used to control normally open breakers (e.g., breakers 370, 375, and 380) depending on operating conditions. For example, service 300A in Figure 3 may act as this control mechanism.

[0074] The controller uses a control algorithm that observes the utility voltage and frequency and the status of each breaker. This algorithm is unique due to the novel configuration disclosed herein. In general, the control mechanism commands breakers to open or close depending on whether the system is operating with or without the utility. These breakers can also be used to supply any combination of modules / power supply units to any combination of the lineup. This feature is especially advantageous when a downstream transformer or data center breaker fails. These breakers can be used to connect two adjacent modules / power supply units together, and power can bypass the failed utility component.

[0075] In this regard, embodiments utilize a central controller (e.g., service 300A) that coordinates all activity of the stamps. Each stamp may optionally also have its own controller, which can communicate with the central controller.

[0076] The utility status (e.g., available / unavailable) affects the status of the normally open breaker. If the utility is available, the controller keeps the breaker open unless there is a need to combine modules. If the utility is not available, the breaker is closed only during start-up. Once start-up is fully successful, the breaker can be opened. If the utility fails and the power supply unit (e.g., SOFC) takes over the load, the breaker remains open once the system starts up with the utility. This is because when the utility fails, the inverter goes into standby and waits for the utility isolation signal to return to voltage source (grid-forming) mode before taking over the load.

[0077] Control Process Broadly speaking, the disclosed embodiments are directed to a distributed, reliable, potentially zero-emission, redundant, scalable, 24 / 7 available fuel cell solution, the operation and black-start details of which are described herein. The disclosed approach allows the system to black-start with or without an enterprise. The design, block connections, integration into data center infrastructure, and control logic are unique compared to traditional architectures and methodologies. Generally, when an enterprise performs black-start, normally open breakers remain open during startup. Otherwise, the breakers are closed, so that the first stamp powers the remaining stamps. Normally open breakers can also close to couple modules / power units when any equipment downstream of one module fails, allowing power to pass through that path.

[0078] Any DC bus on the stamp can be connected, providing a pool of DC energy, including ultracapacitors, for each lineup. Inverter-based technology has low fault current, so flywheel-based technology can also be applied to the system. Additionally or alternatively, rotary UPS, motors, or generator sets can be used instead of flywheels or capacitors. The use of flywheels can further improve stability, fault current, and fast injection / absorption of power. Embodiments can also detect low fault currents by utilizing zone selective interlock (ZSI) or voltage control (restraint) protection.

[0079] Note how any equivalent fuel cell technology can be used for this design. The scope covers seamless transition from current source to voltage source and vice versa or break-before-make, depending on the design and load requirements. So-called green hydrogen can also be used as fuel. This solution overcomes any barriers to building data centers in areas with limited or no power capacity. Power generation for data centers can be local and emission-free, which is an important milestone to achieve. Since both the number of fuel cells in an energy server and the number of energy servers in a stamp are scalable, the reliability and redundancy of this solution can cater to any design. Furthermore, redundancy increases with the number of fuel cells in a fixed load.

[0080] Attention is now directed to Figures 6 and 7. Figure 6 illustrates an example architecture 600 that can facilitate the operations disclosed herein, including the processes illustrated in the flowchart of Figure 7. That is, the architecture 600 of Figure 6 can be used to facilitate each process in this flowchart. Accordingly, frequent references are made between Figures 6 and 7. The architecture 600 of Figure 6 includes a controller 605, which is an example implementation of the service 300A of Figure 3. The controller 605 is an example of the central controller mentioned above. The controller 605 can be a service of the type described above.

[0081] In some cases, the controller 605 can be implemented in a cloud environment and thus is a cloud-based service. In some cases, the controller 605 can be implemented locally on the device. In some cases, the controller 605 can be a hybrid service that includes cloud-based and local components.

[0082] Controller 605 may control or manage any number of local stamp-based controllers, such as stamp controller 610 and stamp controller 615. For example, stamp controller 610 may control the operation of stamp 340 of FIG. 3. Similarly, stamp controller 615 may control the operation of stamp 345 of FIG. 3. Such operations include, but are not limited to, controlling various different breakers, inverters, capacitors (e.g., discharging, charging, etc.), and power units (e.g., SOFC) of the stamp. Controller 605 may also control various other resources 620 (e.g., BESS or PV).

[0083] The controller 605 can facilitate the operations shown in the process flow 700 of Figure 7. With reference to the discussion of Figure 7, reference is made to several methods and method operations that may be performed. Although the method operations may be discussed in a particular order or may be depicted in a flowchart as occurring in a particular order, no particular order is required unless specifically stated, or because an operation depends on another operation being completed before one operation is performed.

[0084] Exemplary Methods The following discussion now refers to several methods / processes and method actions that may be performed. While the method actions may be discussed in a particular order or may be depicted in a flowchart as occurring in a particular order, unless specifically stated, no particular order is required or an action depends on another action being completed before it can be performed. Process flow 700 may be implemented by service 300A of FIG. 3. Because controller 605 of FIG. 6 is an example implementation of service 300A, process flow 700 may also be implemented by controller 605.

[0085] Process flow 700 may be implemented within architecture 300 of FIG. 3 . For example, this architecture is typically designed to provide auxiliary power to a data center serviced by an electric utility. The architecture or system may include a controller, a first stamp, and a second stamp. The system / architecture may further include a common DC bus coupling the first stamp to the second stamp. The system may further include a common AC bus coupling the first stamp to the second stamp. Still further, the system may include an AC breaker disposed on the AC bus between the first stamp and the second stamp. The controller controls whether the AC breaker is open or closed. Furthermore, the default state of the AC breaker is open.

[0086] Process flow 700 (i.e., method) initially includes operation 705 of determining whether a utility is available. As an example, controller 605 may have a voltage or current sensing mechanism attached to a line from the utility. If a signal is observable on the line, controller 605 may determine that the utility is currently providing power to the data center. If no signal is observable, controller 605 may determine that the utility is not currently providing power to the data center.

[0087] If the utility is available (i.e., the yes branch), the controller 605 can facilitate operation 705A, which includes leaving open the normally open breakers. For example, breakers 370, 375, and 380 are left open on common AC line 360 ​​(or breaker 465 in FIG. 4 ). The breakers may include Wireless Fidelity (Wi-Fi) connections, Bluetooth connections, or some other wireless or wired connection. The controller 605 can communicate with these various different breakers to control their state. In some cases, the breakers may be considered devices in the Internet of Things (IoT) and, optionally, may be controlled from the cloud.

[0088] Operation 705B then involves powering up the system. This power up can be performed using either the utility or the power supply unit. Powering up generally involves powering up the entire system, which means first starting up all resources, then connecting loads, and monitoring the system for a period of time until everything is fully stabilized.

[0089] Act 705C includes operating the power supply unit and the utility in parallel with each other, where the power supply unit is operated in a grid-following mode.

[0090] Operation 705D includes determining whether a power grid fault has occurred, such as by analyzing voltage and frequency characteristics. If a power grid fault has occurred, the controller 605 isolates the data center from the power grid (e.g., using switch 320A in FIG. 3 or an upstream breaker near transformers 310 and 315) within a threshold number of seconds (e.g., possibly 2 seconds). Operation 705E includes transitioning the power supply units to a standby mode. After confirming the power grid isolation, the power supply units can recover. Alternatively, the power supply units may seamlessly transition to a grid-forming mode.

[0091] On the other hand, if the utility is not available (i.e., the No branch), operation 710 is performed. As previously mentioned, the controller may determine that the utility is not available by attempting to observe a power signal on the line from the utility. If no signal is observable, the utility is likely not online or at least not providing power to the data center.

[0092] Operation 710 includes determining whether the unavailability of the entity is the result of an entity failure. If the unavailability is due to an entity failure, operation 710A is performed. Any type of entity failure may occur. As various examples, an entity failure may include a power line cut. An entity failure may also include a situation where the power source for the entity is currently offline. In fact, any type of failure may occur. The controller 605 may communicate with the entity to determine whether the entity is offline. In some cases, the controller 605 may trigger a query or alert to be sent to an entity administrator to determine whether an entity failure has occurred.

[0093] Action 710A involves leaving open the normally open breakers, for example, breakers 370, 375, and 380 (or breaker 465 in FIG. 4).

[0094] Then, operation 710B includes isolating the data center from the utility. Then, operation 710C includes switching an inverter (e.g., inverter 535 in FIG. 5) from standby mode to active mode. Then, operation 710D includes causing the power supply units to provide power in a reliable, redundant, and continuous manner. On the other hand, if the unavailability is not due to a utility fault (i.e., no branch), operation 715A is performed. The unavailability may be due to maintenance. For example, a technician may be currently managing or maintaining the line to the utility or possibly the utility itself. The breakers, in this case, are for addressing black start issues when the system starts up in grid-forming mode. Operation 715A includes closing breakers that are normally open. For example, breakers 370, 375, and 380 (or breaker 465 in FIG. 4) are closed.

[0095] Operation 715B then involves the power supply units performing a black start of the system, which power supply units also run in grid forming mode.

[0096] Optionally, the stamps can be operated incrementally over a defined period of time. For example, the first stamp can be gradually increased over a selected number of seconds. Then, immediately upon detecting a stable voltage generated by the first stamp, one, some, or all of the other stamps can be activated. In this situation, normally open breakers (e.g., breakers 370, 375, and 380) are closed, allowing for fast system startup. This allows for fast system startup regardless of the number of stamps.

[0097] Act 715C then includes calculating the number of power supply units that may be needed. This calculation may optionally be based on design and / or required reliability and redundancy requirements. Act 715D then includes causing the system to provide power to the data center in a reliable, redundant, and uninterrupted manner.

[0098] Advantageously, the disclosed principles may be important for bridging power capacity gaps before utilities can provide the power needed for new data centers (e.g., stand-alone microgrids) or may provide a permanent, scalable, localized power resource solution for data centers. This scheme can work with other distributed energy resources in the system in either grid-forming or grid-following modes. The additional use of flywheel-based technology can also address the low fault current of fuel cell inverters. The proposed scheme can stand alone as a microgrid or can be part of a microgrid system, with the SOFC being a baseload resource and other resources added to the microgrid and working with the SOFC.

[0099] The disclosed principles can be used to accommodate phased expansion or full buildout of a data center. They can also be used to accommodate unpredictable load fluctuations in a data center's load. Traditionally, auxiliary power systems were designed for baseload applications, not for unpredictable applications. This principle can also enable rapid startup of an SOFC unit in standalone mode (without a utility) by connecting all stamps on the AC side through specific breakers (e.g., breakers 370, 375, and 380) (so that the first stamp increases first, and then the remaining stamps can start immediately).

[0100] One key advantage and difference of the disclosed embodiments over conventional techniques relates to the introduction of a distributed and scalable approach for technology designed for baseload demand that can be expanded indefinitely to cover baseload, load steps, and create microgrids with or without other distributed energy resources. This solution can be implemented by any cloud service provider seeking to find local power generation resources for their datacenters and reduce or eliminate carbon emissions from their power generation plans.

[0101] Thus, some of the disclosed embodiments are generally and advantageously directed to the unique configuration of the fuel cell's DC and AC bus connections and their integration with the data center's power infrastructure, operating logic, and black-start process. The embodiments can bridge the power capacity gap before utilities are able to provide the power needed for new data centers (e.g., stand-alone microgrids). The embodiments can also provide a permanent, scalable, localized power resource solution for data centers. This approach can work with other distributed energy resources in the system in either grid-forming or grid-following modes. The additional use of flywheel-based technology can also address the low fault current of the fuel cell inverter.

[0102] Exemplary Computer / Computer System Attention is now directed to Figure 8, which illustrates an exemplary computer system 800 that may include and / or be used to perform any of the operations described herein. For example, computer system 800 may implement controller 605 of Figure 6 or service 300A of Figure 3, or any of the other controllers or services mentioned in this disclosure.

[0103] Computer system 800 can take a variety of different forms. For example, computer system 800 can be implemented as a tablet, desktop, laptop, mobile device, or stand-alone device, such as those described throughout this disclosure. Computer system 800 can also be a distributed system that includes one or more connected computing components / devices that communicate with computer system 800.

[0104] In its most basic configuration, computer system 800 includes a variety of different components. Figure 8 shows that computer system 800 includes one or more processors 805 (also known as hardware processing units) and memory device 810.

[0105] With respect to processor 805, it will be understood that the functions described herein may be performed, at least in part, by one or more hardware logic components, such as processor 805. For example, without limitation, exemplary types of hardware logic components / processors that may be used include a field programmable gate array (FPGA), a programmable or application specific integrated circuit (ASIC), a programmable standard product (ASSP), a system-on-a-chip (SOC), a complex programmable logic device (CPLD), a central processing unit (CPU), a graphics processing unit (GPU), or any other type of programmable hardware.

[0106] As used herein, the terms executable module, executable component, component, module, controller, service, or engine may refer to a hardware processing unit, or to a software object, routine, or method that may be executed on computer system 800. The various components, modules, engines, and services described herein may be implemented as objects or processors (e.g., as separate threads) that execute on computer system 800.

[0107] Storage 810 may be physical system memory, which may be volatile, non-volatile, or some combination of the two. The term memory may also be used herein to refer to non-volatile mass storage devices such as physical storage media. If computer system 800 is distributed, processing, memory, and / or storage functionality may likewise be distributed.

[0108] Storage device 810 is shown as including executable instructions 815. Executable instructions 815 represent instructions that are executable by processor 805 of computer system 800 to perform the disclosed operations, such as those described in various ways.

[0109] As discussed in more detail below, the disclosed embodiments may include or utilize a special-purpose or general-purpose computer, including computer hardware such as, for example, one or more processors (such as processor 805) and system memory (such as storage device 810). The embodiments also include physical and other computer-readable media for carrying or storing computer-executable instructions and / or data structures. Such computer-readable media may be any available media that can be accessed by a general-purpose or special-purpose computer system. Computer-readable media that store computer-executable instructions in the form of data are physical computer storage media or hardware storage devices. Furthermore, computer-readable storage media, including physical computer storage media and hardware storage devices, do not include signals, carrier waves, and propagated signals. In contrast, computer-readable media that carry computer-executable instructions are transmission media, and include signals, carrier waves, and propagated signals. Thus, by way of example and not limitation, the present embodiments may include at least two distinctly different types of computer-readable media: computer storage media and transmission media.

[0110] The computer storage medium (also known as hardware storage) is a computer-readable hardware storage device, such as RAM, ROM, EEPROM, CD-ROM, RAM, flash memory, solid-state drive (SSD) based on phase-change memory (PCM) or other types of memory, or other optical disk storage, magnetic disk storage, or other magnetic storage, or any other medium that can be used to store desired program code means in the form of computer-executable instructions, data, or data structures and that can be accessed by a general-purpose or special-purpose computer. Computer system 800 may also be connected (via wired or wireless connections) to external sensors (e.g., one or more remote cameras) or devices via network 820. For example, computer system 800 may communicate with any number of devices or cloud services to acquire or process data. In some cases, network 820 itself may be a cloud network. Furthermore, computer system 800 may also be connected via one or more wired or wireless networks to remote / separate computer systems configured to perform any of the processes described with respect to computer system 800.

[0111] A network, such as network 820, is defined as one or more data links and / or data switches that enable the transport of electronic data between computer systems, modules, and / or other electronic devices. When information is transferred or provided to a computer over a network (hardwired, wireless, or a combination of hardwired and wireless), the computer properly considers the connection a transmission medium. Computer system 800 includes one or more communication channels used to communicate with network 820. A transmission medium includes a network that can be used to transport data or desired program code means in the form of computer-executable instructions or data structures. Furthermore, these computer-executable instructions can be accessed by a general-purpose computer or a special-purpose computer. Combinations of the above should also be included within the scope of computer-readable media. Upon reaching various computer system components, program code means in the form of computer-executable instructions or data structures can be automatically transferred from a transmission medium to a computer storage medium (or vice versa). For example, computer-executable instructions or data structures received over a network or data link may be buffered in RAM within a network interface module (e.g., a network interface card, or NIC) and then ultimately transferred to the computer system's RAM and / or to a less volatile computer storage medium. Thus, it should be understood that computer system components that also (or even primarily) utilize transmission media may include computer storage media.

[0112] Computer-executable instructions (or computer-interpretable instructions) include, for example, instructions that cause a general-purpose computer, a special-purpose computer, or a special-purpose processing device to perform a certain function or group of functions. Computer-executable instructions may be, for example, binaries, instructions in an intermediate format such as assembly language, or even source code. While we have described subject matter in terms specific to structural features and / or methodological acts, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the described features or acts. Rather, the described features and acts are disclosed as exemplary forms of implementing the claims. Those skilled in the art will appreciate that embodiments may be implemented in network computing environments with many types of computer system configurations, including personal computers, desktop computers, laptop computers, message processors, handheld devices, multiprocessor systems, microprocessor-based or programmable consumer electronics, network PCs, minicomputers, mainframe computers, mobile telephones, PDAs, pagers, routers, switches, and the like. Embodiments may also be practiced in a distributed system environment where local and remote computer systems linked through a network (either by hardwired data links, wireless data links, or a combination of hardwired and wireless data links) each perform tasks (e.g., cloud computing, cloud services, etc.) In a distributed system environment, program modules may be located in both local and remote memory storage devices.

[0113] The present invention may be embodied in other specific forms without departing from its characteristics. The described embodiments are to be considered in all respects only as illustrative and not restrictive. The scope of the invention is, therefore, indicated by the appended claims rather than by the foregoing description. All changes which come within the meaning and range of equivalency of the claims are to be embraced within their scope.

Claims

1. 1. A system for providing auxiliary power to a data center typically served by an electric utility, comprising: a control device; a first stamp; A second stamp; a common direct current (DC) bus coupling the first stamp to the second stamp; a common alternating current (AC) bus coupling the first stamp to the second stamp; an AC breaker disposed on the AC bus between the first stamp and the second stamp; wherein the controller controls whether the AC breaker is open or closed, and a default state of the AC breaker is an open state.

2. The system of claim 1 , wherein the first stamp includes one or more energy sources.

3. The system of claim 2 , wherein the one or more energy sources include a solid oxide fuel cell (SOFC).

4. The system of claim 1 , wherein the second stamp includes one or more energy sources.

5. The system of claim 4 , wherein the one or more energy sources include a solid oxide fuel cell (SOFC).

6. The system of claim 4 , wherein the one or more energy sources include an ultracapacitor.

7. The system of claim 6 , wherein the ultracapacitor is connected to a step-up transformer.

8. The system of claim 1 , wherein a size of the first stamp is equal to a size of the second stamp.

9. The system of claim 1 , wherein the size of the first stamp is different from the size of the second stamp.

10. The system of claim 1 , wherein the first stamp provides power to the data center in parallel with the power providing utility.

11. 1. A method of operating a set of stamps in a grid following mode, the stamps being coupled together using a common alternating current (AC) bus, the method comprising: determining that a power providing entity for a data center is currently available to provide power to said data center; leaving open a set of breakers located on a common AC line, the set of breakers coupling multiple stamps on the AC line that can provide auxiliary power to the data center; operating the plurality of stamps in a grid following mode while the utility is providing the power to the data center; A method comprising:

12. The method of claim 11 , wherein the plurality of stamps provide the power to the data center in parallel with the power providing utility.

13. The method of claim 11 , wherein the plurality of stamps act as current sources.

14. The method of claim 11 , wherein the stamps are configurable to operate at different voltage levels.

15. The method of claim 11 , wherein a first stamp in the plurality of stamps is connected to a medium voltage side of a step-down transformer.

16. The method of claim 11 , wherein direct current (DC) connections of the stamps are interconnected.

17. 1. A method for causing a set of stamps to provide power to a data center, the stamps being coupled together using a common alternating current (AC) bus, the method comprising: determining that a power providing entity for a data center is not currently available to provide power to said data center; determining that the electric power providing entity is currently unavailable due to a failure of the electric power providing entity; leaving open a set of breakers located on a common AC line, the set of breakers coupling multiple stamps on the AC line that can provide auxiliary power to the data center; activating a set of inverters associated with the plurality of stamps; causing the plurality of stamps to provide the power to the data center; and A method comprising:

18. 20. The method of claim 17, wherein direct current (DC) connections of the stamps are interconnected.

19. The method of claim 18 , wherein the DC connections of the stamps are connected across the stamps.

20. 20. The method of claim 19, wherein the plurality of stamps act as a scalable pool of DC energy.

Images