Mitigating power fluctuations using battery energy storage system

Abstract

Disclosed herein is a mechanism to mitigate power fluctuations in a data center by dynamically charging and discharging a battery energy storage system (BESS). According to some examples, a BESS control system may be configured to monitor power demand fluctuations and initiate charging of the BESS during periods of low power demand. One example of periods of low power demand includes active idle where the workload of a server machine is frequently reduced to a small extent. Another example of periods of low power demand includes deep idle where the workload of the server machine is reduced by a greater magnitude than active idle for a longer duration. The system may discharge power from the BESS during peak power demand. The BESS control system can stabilize the data center power system through system-level fast operating voltage and frequency control.

Description

[0001] Large scale synchronous machine learning (ML) workloads can generate power fluctuations in the tens of megawatts range almost instantaneously across various electronic components. Often, such workload changes can require repeated power fluctuations in a cycle where power is fluctuated frequently in short intervals and the cycle is repeated for an extended period of time. Such power fluctuations can impact the power system and / or power system quality and can cause detrimental effects that result in a degradation of power quality and potentially blackout of the entire data center. SUMMARY

[0002] Disclosed herein is a mechanism to mitigate power fluctuations in a data center by dynamically charging and discharging a centralized battery energy storage system (BESS). The BESS can enhance the resilience of a data center during utility disturbances by providing backup power and supporting critical loads when power demand is at a peak.

[0003] One aspect of the disclosure provides a system for mitigating power fluctuations in a data center, the system comprising: a battery energy storage system (BESS); and one or more processors in communication with the BESS, the one or more processors configured to: measure a power demand of the data center; determine whether the power demand decreases below a lower power demand threshold; in response to the power demand decreasing below the lower power demand threshold, initiate charging of the battery energy storage system; and

[0004] Discharging power from the BESS is initiated when a state of charge (SOC) of the BESS reaches an SOC threshold.

[0005] In some examples, the SOC threshold is in a range of 10% to 90%.

[0006] In some examples, the one or more processors are configured to discharge power from the BESS in response to the power demand exceeding an upper power demand threshold.

[0007] In some examples, the discharged power from the BESS is used to supply power to the data center.

[0008] In some examples, the discharged power is used as backup power when the data center stops receiving a supply of power.

[0009] In some examples, discharging the BESS is initiated when an SOC of the BESS is less than the SOC threshold.

[0010] In some examples, the BESS is an energy storage device connected to a server rack included in the data center.

[0011] In some examples, charging of the BESS is initiated during at least one of a deep idle period or an active idle period, wherein the BESS is configured to be continuously charged during both the deep idle period and the active idle period.

[0012] In some examples, deep idle periods and active idle periods are caused by transient workload fluctuations.

[0013] Another aspect of this disclosure provides a method for mitigating power fluctuations in a data center, the method comprising: measuring the power demand of the data center; determining whether the power demand has decreased to below a lower power demand threshold; initiating charging of a battery energy storage system (BESS) in response to the power demand decreasing to below the lower power demand threshold; and deactivating charging of the BESS when the state of charge (SOC) of the BESS reaches the SOC threshold.

[0014] In some examples, the SOC threshold ranges from 10% to 90%.

[0015] In some examples, the method further includes discharging power from the BESS in response to power demand exceeding an upper limit power demand threshold.

[0016] In some examples, the power discharged from BESS is used to supply power to data centers.

[0017] In some examples, the discharged power is used as backup power when the data center stops receiving power.

[0018] In some examples, a discharge of the BESS is initiated when the SOC of the BESS is less than the SOC threshold.

[0019] In some examples, BESS is an energy storage device that connects to server racks included in a data center.

[0020] In some examples, charging of the BESS is initiated during at least one of a deep idle period or an active idle period, wherein the BESS is configured to be continuously charged during both the deep idle period and the active idle period.

[0021] In some examples, deep idle periods and active idle periods are caused by transient workload fluctuations.

[0022] Another aspect of this disclosure provides a non-transitory machine-readable medium including machine-readable instructions encoded on the non-transitory machine-readable medium, the machine-readable instructions being used to perform a method for mitigating power fluctuations in a data center, the method comprising: measuring the power demand of the data center; determining whether the power demand has decreased to below a lower power demand threshold; initiating charging of a battery energy storage system (BESS) in response to the power demand decreasing to below the lower power demand threshold; and deactivating charging of the BESS when the state of charge (SOC) of the BESS reaches the SOC threshold.

[0023] In some examples, the method further includes discharging power from the BESS in response to power demand exceeding an upper limit power demand threshold. Attached Figure Description

[0024] Figure 1 This is a schematic representation of a BESS control system for a data center based on various aspects of this disclosure.

[0025] Figure 2 The example BESS control system is based on various aspects of this disclosure.

[0026] Figure 3 Vertically aligned graphs are provided to represent the workload and BESS charge rate for various aspects according to this disclosure.

[0027] Figure 4 Vertically aligned graphs are provided to represent the charge / discharge cycles of the BESS and rack-mounted batteries according to various aspects of this disclosure.

[0028] Figure 5 This is an example flowchart of a BESS control system based on various aspects of this disclosure. Detailed Implementation

[0029] This disclosure relates to mitigating power fluctuations in data centers by dynamically charging and discharging a battery energy storage system (BESS). According to some examples, a BESS control system (“System”) can be configured to monitor power demand fluctuations and initiate charging of the BESS during periods of low power demand. One example of a period of low power demand includes active idle, where the workload of server machines frequently decreases by small amounts. Another example of a period of low power demand includes deep idle, where the workload of server machines decreases by a larger amount over a longer duration than active idle. The System can discharge power from the BESS during periods of peak power demand. The BESS control system can stabilize the data center power system through system-level fast-acting voltage and frequency control.

[0030] When power demand decreases, the system can continuously charge the BESS, and when power demand is at its peak, the system can continuously discharge the BESS, thereby effectively managing power fluctuations and maintaining grid stability. The system can reduce the number of charge or discharge cycles of the BESS by avoiding discharging the BESS until its state of charge (SOC) reaches a predetermined threshold SOC. The threshold can be a percentage of SOC, such as 80%, 90%, 95%, etc. In other examples, the threshold can be a value such as a unit of measurement.

[0031] Once the BESS reaches a predetermined threshold SOC, the power discharged from the BESS can be used to smooth peak loads to prevent extreme spikes. In some examples, BESS discharge can begin even when the power supply stops (e.g., a data center power outage), even if the BESS's SOC is below the threshold.

[0032] By implementing the BESS described herein, server racks can be more densely packed with hardware by eliminating the space in the rack used for in-rack batteries and instead using BESSs mounted separately outside the server rack. The power stored in the BESS can also provide backup power for extended periods, from minutes to hours, when no power is being transferred from the load. Furthermore, by actively managing the charge and discharge cycles of the BESS as described herein, the system can extend the lifespan of the BESS to be relatively longer than that of other types of power backup systems.

[0033] Figure 1 This is a schematic representation of a BESS control system 100 for a data center. The BESS control system 100 includes a controller 102, a power supply 120, and server racks 110A-C. Each server rack is connected to sensors 104A-C and BESS 108A-C, respectively. Each sensor can be configured to measure changes in the workload of each server rack. Each sensor can also be configured to monitor the power consumption of each server rack and the power received from the power supply 120. Each sensor can transmit the measured data to the controller 102 at pre-configured intervals.

[0034] The BESS108A-C can be configured to store the power supplied to each server rack. A BESS can refer to any type of energy storage system capable of flexibly accumulating, storing, and discharging electrical energy. The BESS108A-C can include, but is not limited to, lithium-based systems, flow batteries, lead-acid batteries, etc. A BESS can replace in-rack batteries. A BESS can be a large-scale battery energy storage system (BESS) installed in a data center at a medium voltage level with a large aggregate capacity. For example, a BESS can include lithium-ion batteries, a DC / AC inverter, a BESS main control system, and a human-machine interface (HMI). In some examples, a BESS can be installed in a data center at a low voltage level with distributed capacity. In such applications, the BESS can be installed close to the server rack. The server rack 110A-C includes one or more computing devices for various purposes such as internet hosting, cloud computing, storage, computing, networking, machine learning, etc. The computing device may include a processor comprising one or more individual circuits, transistors, and / or other components. Each operation performed at the circuit may require at least a small amount of power, and therefore, each operation generates a certain amount of heat as a byproduct. Although each circuit can generate relatively little heat, each server rack can generate a considerable amount of heat in total.

[0035] Each sensor can send power consumption measurements in watts and / or volts to the controller 102 via bus 116. Bus 116 may include a physical layer that implements the communication protocol between the power consumption sensors attached to each server rack or computing device and the controller 102. The BESS108A-C can be configured to communicate with the controller 102 via bus 116.

[0036] Controller 102 can be configured to control the charging and discharging of the BESS108A-C. For example, the BESS108A-C can be configured to continuously charge and store power when power demand is below a predefined threshold or when power demand begins to decrease as the workload of the corresponding server rack decreases. The BESS108A-C can be configured to discharge the stored power when power demand is at its peak, so that the power supplied from the BESS108A-C can help manage power fluctuations and maintain the stability of power supply 120. Power discharged from the BESS108A-C can also help smooth peak loads to prevent extreme spikes in power demand.

[0037] Figure 2This is a block diagram illustrating an example computing device 200 according to various aspects of this disclosure. The computing device may take various configurations, such as, for example, a controller or microcontroller, or a processor such as a CPU, GPU, or ASIC including a Tensor Processing Unit (TPU). The computing device may further include a BESS control system 202. The BESS control system 202 may be configured to control the charging and discharging of the BESS to mitigate power demand fluctuations in the server rack.

[0038] The BESS control system 202 may include a processor 210, a memory 204 including data 206 and instructions 208, a BESS charging module 212 and a BESS discharging module 214, and other components typically found in server computing devices. In other examples, such operation may be performed by one or more computing devices in a data center or elsewhere.

[0039] Memory 204 may store information accessible by processor 210, including instructions 208 executable by processor 210. Memory may also include data 206 that can be retrieved, manipulated, or stored by processor 210. Memory 204 may be a type of non-transitory computer-readable medium capable of storing information accessible by processor 210, such as a hard disk drive, solid-state drive, magnetic tape drive, optical storage, memory card, ROM, RAM, DVD, CD-ROM, writable memory, and read-only memory. Processor 210 may be a well-known processor or another lesser-known type of processor. Alternatively, processor 210 may be a dedicated controller, such as an ASIC.

[0040] Instructions 208 may be a set of instructions (such as machine code) that are directly executed by processor 210, or a set of instructions (such as scripts) that are indirectly executed. In this regard, the terms "instruction," "step," and "program" are used interchangeably herein. Instructions 208 may be stored in object code format for direct processing by processor 210, or stored in other types of computer languages, including sets of scripts or stand-alone source code modules that are interpreted on demand or compiled in advance. For example, instructions 208 may include instructions for... Figure 1 The BESS108A-C depicted in the figure performs charging or discharging commands based on inputs from the BESS charging module 212 and the BESS discharging module 214.

[0041] Data 206 can be retrieved, stored, or modified by processor 210 according to instructions 208. For example, although the system and method are not limited to a specific data structure, data 206 can be stored in a computer register, stored as a table with multiple different fields and records in a relational database, or stored in an XML document. Data 206 can also be formatted in a computer-readable format, such as, but not limited to, binary values, ASCII, or Unicode. Furthermore, data 206 may include information sufficient to identify relevant information, such as numbers, descriptive text, proprietary codes, pointers, references to data stored in other memory (including other network locations), or information used by functions to calculate relevant data. Data 206 may include historical data relating to the correlation between workload and power demand. Additionally or alternatively, data 206 may include test data obtained based on experiments performed using the BESS108A-C.

[0042] BESS charging module 212 can command BESS108A-C to charge during active idle or deep idle states. In both states, power demand decreases as the workload of each server rack decreases. Active idle state can refer to frequent fluctuations in power demand and workload. The amplitude of these fluctuations is smaller than that of deep idle state. Deep idle state can refer to a state where power demand and workload fluctuate at a greater amplitude. When sensor 104A-C sends fluctuation information to controller 102, BESS charging module 212 can be configured to begin charging BESS108A-C.

[0043] BESS discharge module 214 can command BESS108A-C to discharge the charged power when server racks 110A-C are in an active idle or deep idle state. When sensor 104A-C determines that power demand will peak due to a sudden increase in workload on each server rack exceeding a threshold level, BESS discharge module 214 can command BESS108A-C to discharge the power. The discharged power from BESS108A-C can be used by server racks 110A-C along with the power supplied from power supply 120, preventing the power supplied from power supply 120 from experiencing extreme spikes.

[0044] Figure 2Processor 210 and memory 204 are shown functionally within the same box, but processor 210 and memory 203 may alternatively include multiple processors and memories, which may or may not be stored in the same physical housing. For example, some of the instructions 208 and data 206 may be stored on a removable CD-ROM, while others may be stored within a read-only computer chip. Some or all of the instructions and data may be stored in a location physically remote from processor 210 but still accessible to that processor. Similarly, processor 210 may include a collection of processors that may or may not operate in parallel.

[0045] It should be understood that in this example, the BESS charging module 212 and the BESS discharging module 214 are shown as part of the BESS control system 202. In other examples, the BESS charging module 212 and the BESS discharging module 214 may be implemented in one or more other systems or computing devices.

[0046] Figure 3 A vertically aligned graph showing the correlation between workload and BESS charging rate in an active idle state is shown. Curve 304 represents the SOC of the rack-mount battery. Curve 302 represents the workload, and curve 306 represents the SOC of the BESS. At T1, the workload decreases, and the BESS begins charging. At T1, curve 306 shows a step increase in SOC at T1 and T2 as the workload decreases again. The BESS can be configured to continue charging and accumulating power until it reaches a predefined SOC threshold (not shown). Rack-mount batteries with smaller capacities can only be repeatedly charged and discharged because they typically need to maintain their SOC at approximately 50%. To maintain the SOC of the rack-mount battery at approximately 50%, it can charge when the workload decreases below a lower threshold (not shown) and discharge when the workload increases above an upper threshold (not shown).

[0047] Figure 4A vertically aligned graph showing the charge / discharge cycle of the BESS and in-rack battery during deep idle states is shown. Graph line 402 represents the SOC of the in-rack battery. Graph line 404 represents the power consumption level of the server rack. Graph line 406 represents the workload, and graph line 408 represents the SOC of the BESS. The in-rack battery begins to charge as the workload decreases. However, because the workload decreases by a greater amount during deep idle states than during active idle states, the SOC of the in-rack battery can easily reach 100% and drop below 50% as workload and power demand increase. Conversely, even during deep idle states, the SOC of the BESS can gradually increase without discharging the stored power until the SOC reaches an upper threshold. A charged BESS can be configured to discharge the stored power when it reaches a lower threshold level or when regular power operation ceases.

[0048] Figure 5 An example flowchart of the BESS control system is shown. According to block 502, the system can be configured to measure the power demand of a data center. The power demand can change as the workload of the server racks increases or decreases. When the workload decreases, such as during active idle or deep idle states, the power demand decreases.

[0049] According to box 504, the system can be configured to determine whether power demand has decreased below a lower power demand threshold. According to some examples, if power demand decreases below the lower power demand threshold, the power supply can reduce the amount of power delivered to each server rack. The BESS control system can maintain a stable power supply from the power source by charging the BESS with unwanted power during idle periods.

[0050] According to box 506, the system can be configured to initiate charging of the battery energy storage system (BESS) in response to a decrease in power demand below a first power demand threshold. The BESS begins charging, and the power supply does not need to reduce the amount of power delivered to each server rack. The BESS can continue charging until the State of Charge (SOC) reaches a predetermined SOC threshold, such as a percentage. Example thresholds may range from 10% to 90%, but it should be understood that other configurations are possible.

[0051] According to box 508, the system can be configured to release charging of the BESS when the BESS's State of Charge (SOC) reaches a SOC threshold. The system can also be configured to discharge power stored in the BESS when the BESS reaches the SOC threshold or when the system detects that power is no longer being supplied by the power source—such as in the event of a power outage. In some examples, the system can be configured to discharge power stored in the BESS when power demand exceeds a threshold level. In such examples, the power discharged from the BESS can be used by the server rack, so that the power supply does not need to suddenly increase the power supplied to the server rack.

[0052] The advantage of the BESS control system described in this paper lies at least in its ability to efficiently charge and discharge power using BESS when a certain amount of power is no longer needed due to a sudden decrease in workload and power demand. When workload and power demand increase sharply over a short period, stored power can be discharged to alleviate the burden on the power supply, thereby supplying a significantly increased amount of power over a short period. Therefore, the BESS control system promotes the stability of power supply and avoids hardware degradation of power supply units. The BESS control system can also mitigate hardware degradation of server components because each server rack can receive a near-constant amount of power supply.

[0053] By implementing the BESS described herein, server racks can be more densely packed with hardware by eliminating the space in the rack used for in-rack batteries and instead using BESSs mounted separately outside the server rack. The power stored in the BESS can also provide backup power for extended periods, from minutes to hours, when no power is being transferred from the load. Furthermore, by actively managing the charge and discharge cycles of the BESS as described herein, the system can extend the lifespan of the BESS to be relatively longer than that of other types of power backup systems.

[0054] In this specification, the phrase "configured to" is used in various contexts relating to a computer system, hardware, or part of a computer program, engine, or module. When a system is said to be configured to perform one or more operations, this means that appropriate software, firmware, and / or hardware is installed on the system that, when in operation, causes the system to perform one or more operations. When hardware is said to be configured to perform one or more operations, this means that the hardware includes one or more circuits that, when in operation, receive input and generate output based on the input and corresponding to one or more operations. When a computer program, engine, or module is said to be configured to perform one or more operations, this means that the computer program includes one or more program instructions that, when executed by one or more computers, cause those computers to perform one or more operations.

[0055] While this document has described the technology with reference to specific examples, it should be understood that these examples are merely illustrative of the principles and applications of the technology. Therefore, it should be understood that various modifications can be made, and other arrangements can be designed without departing from the spirit and scope of the technology as defined by the appended claims.

[0056] Unless otherwise stated, the foregoing alternative examples are not mutually exclusive, but can be implemented in various combinations to achieve unique advantages. Since these and other variations and combinations of the features discussed above can be utilized without departing from the subject matter defined by the claims, the foregoing description should be presented in an illustrative rather than restrictive manner. Furthermore, the examples described herein and the provision of terms such as “such as,” “comprising,” etc., should not be construed as limiting the subject matter of the claims to specific examples; rather, the examples are intended to illustrate only one of many possible implementations. Additionally, the same reference numerals in different figures may identify the same or similar elements.

Claims

1. A system for mitigating power fluctuations in data centers, characterized in that, The system includes: Battery Energy Storage System (BESS); and One or more processors, which communicate with the BESS, are configured to: Measure the power demand of the data center; Determine whether the electricity demand has decreased to below the lower limit electricity demand threshold; In response to the power demand decreasing below the lower power demand threshold, charging of the battery energy storage system is initiated; and When the State of Charge (SOC) of the BESS reaches the SOC threshold, the charging of the BESS is released.

2. The system as described in claim 1, characterized in that, in, The SOC threshold is in the range of 10% to 90%.

3. The system as described in claim 1, characterized in that, in, The one or more processors are configured to discharge power from the BESS in response to the power demand exceeding an upper limit power demand threshold.

4. The system as described in claim 3, characterized in that, in, The power discharged from the BESS is used to supply power to the data center.

5. The system as described in claim 4, characterized in that, in, When the data center stops receiving power, the discharged power is used as backup power.

6. The system as described in claim 5, characterized in that, in, When the SOC of the BESS is less than the SOC threshold, the discharge of the BESS is initiated.

7. The system as described in claim 5, characterized in that, in, The BESS is an energy storage device connected to a server rack included in the data center.

8. The system as described in claim 1, characterized in that, in, The charging of the BESS is initiated during at least one of a deep idle period or an active idle period, wherein the BESS is configured to charge continuously during both the deep idle period and the active idle period.

9. The system as described in claim 8, characterized in that, in, The deep idle period and the active idle period are caused by instantaneous workload fluctuations.

10. A method for mitigating power fluctuations in a data center, characterized in that, The method includes: Measure the power demand of the data center; Determine whether the electricity demand has decreased to below the lower limit electricity demand threshold; In response to the power demand decreasing below the lower power demand threshold, charging of the battery energy storage system (BESS) is initiated; and When the State of Charge (SOC) of the BESS reaches the SOC threshold, the charging of the BESS is released.

11. The method as described in claim 10, characterized in that, in, The SOC threshold is in the range of 10% to 90%.

12. The method as described in claim 10, characterized in that, Further includes: In response to the power demand exceeding the upper limit power demand threshold, power is discharged from the BESS.

13. The method as described in claim 12, characterized in that, in, The power discharged from the BESS is used to supply power to the data center.

14. The method as described in claim 13, characterized in that, in, When the data center stops receiving power, the discharged power is used as backup power.

15. The method as described in claim 14, characterized in that, in, When the SOC of the BESS is less than the SOC threshold, the discharge of the BESS is initiated.

16. The method as described in claim 14, characterized in that, in, The BESS is an energy storage device connected to a server rack included in the data center.

17. The method as described in claim 10, characterized in that, in, The charging of the BESS is initiated during at least one of a deep idle period or an active idle period, wherein the BESS is configured to charge continuously during both the deep idle period and the active idle period.

18. The method as described in claim 17, characterized in that, in, The deep idle period and the active idle period are caused by instantaneous workload fluctuations.

19. A non-transitory machine-readable medium, characterized in that, Includes machine-readable instructions encoded on the non-transitory machine-readable medium, the instructions being used to perform a method for mitigating power fluctuations in a data center, the method comprising: Measure the power demand of the data center; Determine whether the electricity demand has decreased to below the lower limit electricity demand threshold; In response to the power demand decreasing below the lower power demand threshold, charging of the battery energy storage system (BESS) is initiated; and When the State of Charge (SOC) of the BESS reaches the SOC threshold, the charging of the BESS is released.

20. The non-transitory machine-readable medium as claimed in claim 19, characterized in that, in, The method further includes discharging power from the BESS in response to the power demand exceeding an upper limit power demand threshold.

Images