KR102351681B1 - System and method for active disturbance rejection based thermal control - Google Patents

Abstract

An active Disturbance Rejection thermal control (ADRC) system and method according to an embodiment of the present invention includes, at a first ADRC controller, receiving a first temperature measurement from a first thermal zone, and , to generate a first output control signal for controlling the first cooling element. The first output control signal includes a first estimated temperature for the first thermal zone and a first estimated temperature for the first thermal zone computed by a first extended state observer (ESO) of the first ADRC controller. generated by interference.

Description

SYSTEM AND METHOD FOR ACTIVE DISTURBANCE REJECTION BASED THERMAL CONTROL

Embodiments of the present invention relate to controlling thermal-cooling in servers and data centers.

A rapid increase in data causes a corresponding increase in data centers. As part of data center growth, the density of computing infrastructure has also increased. For example, 1U storage servers are popular and feature increasingly powerful processors and high-density storage (eg, flash storage). It is becoming increasingly difficult to provide adequate cooling for these high-density data centers, and the cost of providing cooling is increasing in the overall power consumption of the data center.

Since the above-described information is only for helping understanding of the background of the embodiments of the present invention, information that does not constitute the prior art may be included.

The present invention is to solve the above technical problem, the present invention can provide a system and method for active interference cancellation-based thermal control with reduced power consumption.

Some embodiments of the present invention provide systems and methods for active jamming thermal control. In various embodiments, the system receives, at a first active Disturbance Rejection thermal control (ADRC) controller, a first temperature measurement from the first thermal zone. The constant system may generate a first output signal for controlling the first cooling element by using the first ARDC controller. In various embodiments, the first output control signal includes a first estimated temperature for the first thermal zone and a first estimate calculated by a first extended state observer (ESO) of the first ADRC controller. generated according to the interference.

In various embodiments, the system receives, at a second ADRC controller, a second temperature measurement from a second thermal zone. The second ADRC controller may generate a second output control signal for controlling the second cooling element. In various embodiments, the second output control signal is generated according to a second estimated temperature and a second estimated disturbance computed by a second ESO of the second ADRC controller.

In various embodiments, the system calculates, in the first ADRC controller, a first temperature tracking error by comparing the first temperature measurement to a first target temperature, and in the second ADRC controller, the second temperature measurement is compared with the second target temperature to calculate a second temperature tracking error. In various embodiments, the system may modify at least one of the first output control signal and the second output control signal based on the first and second temperature tracking errors.

In various embodiments, the first thermal zone includes at least one of an SSD, a processor, or a graphics processing unit.

In various embodiments, the first ESO is a maximum temperature read from the first thermal zone, an estimate of the maximum temperature read from the first thermal zone, an estimate of total disturbance, the first output control signal, a supervisor gain ( observer gain), and second order ESO (second order ESO) that operates according to at least one adjustable parameter.

In various embodiments, the supervisor gain is inversely proportional to the values of adjustable supervisor bandwidth and sampling time.

In various embodiments, the first output control signal (u _i ) includes the equations of the first target temperature (r), the first estimated temperature (z ₁ ), and the first estimated disturbance (z _{2 )} contains the same function,

The ω _c is an adjustable controller bandwidth, and the b ₀ is an adjustable parameter.

In various embodiments, the storage server includes a first device and the first thermal sensor and the first cooling element, wherein the first thermal zone is included in a first thermal zone including a first temperature sensor and a first cooling element; and a first active disturbance rejection thermal control (ADRC) controller coupled thereto. In various embodiments, the first ADRC controller receives a first temperature measurement of the first thermal zone from the first temperature sensor, a first estimated temperature for the first thermal zone, a first output control signal to the first cooling element according to a first estimated disturbance computed by a first extended state observer (ESO) for is configured to provide

In various embodiments, the storage server includes a second device included in a second thermal zone including a second temperature sensor and a second cooling element, and a second ADRC controller connected to the second temperature sensor and the second cooling element. include more In various embodiments, the second ADRC controller receives a second temperature measurement from the second temperature sensor and generates a second output control signal for controlling the second cooling element, and the second output control signal is a second estimated temperature of the second thermal zone, a second estimated disturbance computed by the second ESO for the second thermal zone, and a second target temperature for the second thermal zone.

In various embodiments, the first ADRC controller is further configured to calculate a first temperature tracking error by comparing the first temperature measurement to the first target temperature, and wherein the second ADRC controller calculates the second temperature measurement as the and calculate a second temperature tracking error compared to the second target temperature. In various embodiments, at least one of the first output control signal or the second output control signal is modified based on the first and second temperature tracking errors.

In various embodiments, the first device includes at least one of an SSD, a processor, or a graphics processing unit.

In various embodiments, the first ESO is a maximum temperature read from a first thermal zone, an estimate of the maximum temperature read from the first thermal zone, an estimate of total disturbance, the first output control signal, an observer gain gain), and a second order ESO (second order ESO) operating according to at least one adjustable parameter.

In various embodiments, the monitor gain is inversely proportional to the values of adjustable monitor bandwidth and sampling time.

In various embodiments, the first output control signal ui is the first target temperature r, the first estimated temperature z1, and the first estimated disturbance z2, contains functions.

The ωc is an adjustable controller bandwidth, and the b ₀ is an adjustable parameter.

In various embodiments, a data center includes a plurality of thermal zones comprising an ADRC. In various embodiments, the data center is connected to a first thermal zone including a first temperature sensor and a first cooling element, the first temperature sensor and the first cooling element, and from the first temperature sensor Receive a first temperature of a thermal zone, and the first cooling according to a first estimated temperature, a first estimated disturbance computed by a first estimated state observer (ESO), and a first target temperature. a second thermal zone comprising a first active disturbance rejection thermal control (ADRC) controller configured to provide a first output control signal to an element, a second temperature sensor, and a second cooling element, the second temperature coupled to a sensor and the second cooling element, configured to receive a second temperature measurement from the second temperature sensor, and to generate a second estimated temperature, a second output signal for controlling the second cooling element, The second output control signal is generated according to a second estimated disturbance computed by the second ESO, and a second target temperature.

In various embodiments, the first ADRC controller is further configured to compare the first temperature measurement to the first target temperature to generate a first temperature tracking error, and wherein the second ADRC controller compares the second temperature measurement to the and generate a second temperature tracking error compared to a second target temperature, wherein at least one of the first output control signal and the second output control signal is modified based on the first and second temperature tracking errors. do.

In various embodiments, the first thermal zone includes at least one server and the second thermal zone includes at least one server.

In various embodiments, the first thermal zone comprises at least a portion of the second thermal zone.

In various embodiments, the first ESO is a maximum temperature read from a first thermal zone, an estimate of the maximum temperature read from the first thermal zone, an estimate of total disturbance, the first output control signal, an observer gain gain), and a second order ESO (second order ESO) operating according to at least one adjustable parameter.

In various embodiments, the supervisor gain is inversely proportional to the values of controllable supervisor bandwidth and sampling time.

According to an embodiment of the present invention, a system and method for active interference cancellation-based thermal control with reduced power consumption are provided.

Some embodiments may be understood in more detail from the following descriptions provided in conjunction with the accompanying drawings.
1 shows an exemplary server using ADRC thermal management in accordance with various embodiments of the present invention.
2 shows a method of performing ADRC-based thermal control for a thermal zone according to various embodiments of the present disclosure.
3 shows an exemplary ADRC controller according to various embodiments of the present invention.
4 shows an exemplary ADRC controller including a plurality of ADRC controllers for a plurality of thermal zones in accordance with various embodiments of the present invention.
5 shows a method of controlling cooling elements for a plurality of thermal zones according to various embodiments of the present invention.

Various features of the present invention and methods for achieving the present invention may be understood in more detail with reference to the accompanying drawings and the following detailed description of embodiments. Hereinafter, embodiments will be described in more detail with reference to the accompanying drawings. Like reference numbers refer to like elements throughout. However, the present invention may be embodied in various other forms and should not be construed as being limited to the described embodiments. Rather, these embodiments are provided as examples to help the general understanding of the present invention, and will fully convey the technical features and aspects of the present invention to those skilled in the art. Accordingly, processes, elements, and techniques that are unnecessary for one of ordinary skill in the art to fully understand the features and aspects of the present invention may not be described. Unless otherwise noted, throughout the appended drawings and the detailed description, like reference numerals refer to like elements, and their description will not be repeated. In the drawings, the relative sizes of elements, layers, and regions may be exaggerated for clarity.

In the following detailed description, a number of specific descriptions are provided for convenience of description and understanding of various embodiments. However, various embodiments may be implemented without detailed description or with one or more equivalent substitutes. In other instances, well-known structures and devices are shown in block diagram form in order not to unnecessarily obscure various embodiments.

When an element, layer, region, or component is referred to as being “on, connected to, or coupled to” another element, layer, region, or component, the other element, layer, region, or component is directly ( directly) or there may be one or more intermediate elements, layers, regions, or configurations. However, the term “directly connected” refers to one component being directly connected to another component without an intermediate component. On the other hand, other expressions describing a relationship between constructs, such as "between, immediately between" or "adjacent to or directly adjacent to," may be interpreted similarly. Additionally, an element or hierarchy is two elements. When referring to being between elements or layers, it may be understood that only an element or layer may exist between elements or components, or one or more intermediate elements or layers may further exist.

Terms used in the text are illustrative only for describing specific embodiments, and the present invention is not limited thereto. As used herein, singular terms are intended to include plural forms, unless the context clearly dictates otherwise. When the term "comprises" is used in the specification, it defines the presence of open features, integers, steps, operations, elements, and/or configurations, but one or more other features, integers, etc. It does not exclude the presence or addition of groups, steps, acts, elements, configurations, and/or groups thereof. As used herein, the term "and/or" includes any combination or portion of one or more of the associated listed lists.

As used herein, the terms "substantial", "about, approximately" and similar terms are used as terms of approximation, not as terms of degree; It is intended to account for inherent deviations in measured or computed values that would be recognized by one of ordinary skill in the art. As used herein, the term "about, approximately" is inclusive of the recited value, and is considered by one of ordinary skill in the art to account for errors and questionable measurements (eg, limitations of the measurement system) associated with measurement of a particular quantity. It means that it is within an acceptable deviation of the determined specific value. For example, “about” can mean within one or more standard deviations or within ± 30%, 20%, 10%, 5% of a stated value. As used herein, the terms "use, using, and used" may be considered synonymous with "utilize, utilizing, and utilized." Also, the term "exemplary" is intended to refer to an "example or illustration."

If particular embodiments are implemented differently, the particular process order may result in a different order than the order described. For example, two successively described processes may be performed substantially simultaneously or may be performed in an order opposite to that described.

Various embodiments are described herein with reference to cross-sectional views that are schematic views of embodiments and/or intermediate structures. As such, deviations from the forms shown should be expected, for example, as a result of manufacturing techniques and/or tolerances. Moreover, specific structural or functional descriptions described in the text are merely examples for describing embodiments according to the spirit of the present invention. That is, the embodiments described in the text should not be construed as being limited to the specifically illustrated shapes of the regions, but should include, for example, deviations in shape resulting from improvement. For example, implant regions depicted as rectangular will typically have rounded or curved features at their edges and/or slopes of implant concentration rather than a binary-change from implant to non-implant region. . Likewise, the buried area formed by the implant may result in some implants in the area between the buried area and the surface from which the implant is born. That is, the regions shown in the drawings are schematic in nature, and their shapes are not meant to depict the actual shape of the region of the device, and are not intended to be limiting.

Electrical or electronic devices and/or other associated devices or components according to the embodiments of the invention described herein may include appropriate hardware, firmware (eg, application-specific integrated circuits (ASICs); It may be implemented using software or a combination of software, firmware, and hardware For example, various components of these devices may be formed in one integrated circuit (IC) chip or separate IC chips Moreover, the various configurations of these devices may be implemented on a flexible printed circuit film, a tape carrier package (TCP), a printed circuit board (PCB), or on a single substrate. Furthermore, various components of such devices may be in communication with other system components and executing computer program instructions to perform various functions described herein, one or more computing devices. may be a process or thread running on one or more processors Computer program instructions are stored in a memory that may be implemented in a computing device using a standard memory device such as random access memory (RAM). The program instructions may be stored in other non-transitory computer readable media, such as, for example, a CD-ROM, a flash drive, etc. Those skilled in the art will also do so without departing from the spirit and aspect of the exemplary embodiments of the present invention. It will be appreciated that the functions of various computing devices may be combined or integrated into a single computing device, or the functionality of a particular computing device may be distributed among one or more other computing devices.

Unless defined otherwise, all terms including technical/scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in the public dictionary are to be interpreted as having a meaning consistent with their meaning in the context of the related art and/or detailed description of the present invention, and unless specifically defined in the text, ideal or It should not be interpreted in an overly formal sense.

Embodiments of the present invention include systems and methods for thermal control based on active disturbance rejection. In various embodiments, the system includes one or more controllers that control and monitor cooling of one or more drive zones and processor zones. The system can be configured such that active disturbance rejection control (ADRC)-based thermal management 1 improves thermal control in data centers, servers, and storage systems. The system improves thermal control efficiency by compensating for thermal reading as well as workload fluctuations. The improved thermal control therefore reduces cooling cost, thereby reducing the overall cost of servers and users of data centers. Moreover, the system improves performance by reducing thermal throttling of the processor, GPUs, and storage drives.

In the related art, a number of thermal management systems are used. For example, some systems have native cooling (eg, native server fan speed control), dynamic cooling (eg, dynamic fan speed control based on temperature), and proportional (PID) -integral-derivative) control. PID controllers calculate a difference between a target value and a current value, and apply a correction based on proportional, integral, and derivative items. While each of these methods has various advantages, each also has various disadvantages. For example, PID controllers require a good balance between the three controller gains, and compromise between transient response, robustness, or disturbance rejection ability. can do it

1 shows an exemplary server using ADRC thermal management in accordance with various embodiments of the present invention.

Referring to FIG. 1 , in various embodiments, server 100 includes an ADRC controller 110 that manages cooling of a plurality of thermal (eg, temperature) zones. For example, in various embodiments, server 100 may include a plurality of processors, storage drives, and other components such as memory (eg, DRAM), GPUs, power supplies, and other components. It may be a 1U storage server including In various embodiments, the server 100 may be divided into a plurality of thermal zones, each of which includes various components. In some embodiments, each of the thermal zones may primarily include a single type of components (eg, processors or storage drives), while in other embodiments, the thermal zones may include multiple components. have. In some embodiments, the temperature zones may be physically separated, while in other embodiments there may be no physical boundaries between the zones. In various embodiments, each temperature zone may include one or more active cooling elements. For example, each zone may contain one or more fans. In some embodiments, cooling elements may be associated with one or more thermal zones.

In various embodiments, the ADRC controller 110 may include multiple inputs and multiple outputs (MIMO) for receiving temperature measurements and controlling the cooling elements. For example, in various embodiments, the ADRC controller 110 includes a temperature input for each device (eg, processor, GPU, or storage drive) in each of the zones, and sends a temperature input to each of the corresponding cooling elements. control outputs for

In various embodiments, server 100 includes one or more processor zones 120 , 130 , 140 , one or more drive zones 150 , 160 , 170 , and one or more other zones 180 . , 190) may be included. In various embodiments, each of the processor zones 120 , 130 , 140 may include one or more CPUs. In various embodiments, each of drive zones 150 , 160 , 170 may include one or more storage drives (eg, SSDs, HDDs, etc.). In various embodiments, each of the other zones 180 , 190 may include one or more other computing components, such as power supplies, GPUs, memory, or other server components. In various embodiments, each of the zones may include one or more cooling elements 125 , 135 , 145 , 155 , 165 , 175 , 185 , 195 . For example, each of the cooling elements may include one or more adjustable speed fans.

2 shows a method of performing ADRC-based thermal control for thermal zones according to various embodiments of the present disclosure.

Referring to FIG. 2 , an ADRC controller (eg, ADRC controller 110 of FIG. 1 ) may be configured to monitor a plurality of temperature zones of a server (eg, server 100 of FIG. 1 ). In various embodiments, each of the temperature zones may include a separate ADRC control loop that provides an output for monitoring temperature, estimating disturbance, and managing cooling elements (eg, controlling fan speed). For example, a server may include i row zones. In various embodiments, the ADRC controller receives one or more temperature measurements from one or more temperature sensors of the ith thermal zone. (S200) In various embodiments, temperature sensors may include temperature sensors embedded in the devices (eg, a temperature sensor included in each of SSD, processor, GPU), while in other embodiments, The temperature sensor may be for a zone containing a plurality of devices (eg, located in a server chassis). In various embodiments, the ADRC controller uses the temperature reading to determine a temperature tracking error (eg, the difference between a set temperature and an estimated maximum temperature), and determines the estimated temperature and the estimated maximum temperature. Interference can be calculated. (S210) This will be described in more detail below. Based on the temperature tracking error, the estimated temperature, and the estimated disturbance, the ADRC controller may provide an output control signal to the cooling element(s) to control the output of the cooling element(s). (For example, adjust the fan speed) (S220)

In various embodiments, the ADRC controller is a MIMO system that is a combination of multiple single-input-single-output (SISO) or MIMO systems. In various embodiments, the ADRC controller may include a plurality of ADRC controllers. For example, in various embodiments, there may be a separate ADRC controller for each thermal zone. Thus, each thermal zone is an individual ADRC control that includes an extended state observer (ESO) that estimates the real-time temperature of the thermal zone along with external disturbances (eg, workload) affecting the thermal zone. It can contain loops. For example, external disturbances can arise from unknown workloads sent to devices operating in the thermal zone (eg, I/O sent to storage devices, workloads sent to processors, etc.) may include the resulting lumped uncertainties. In various embodiments, ESO may be a second order ESO (second order ESO) defined according to Equations 1 to 5 below.

y _i is the temperature read from the zone (eg, the highest temperature read from one SSD in a zone containing a plurality of SSDs), and z ₁ is the maximum temperature read from the zone (max temperature). Estimation, z ₂ is the total disturbance into the loop, u _i is the output control signal controlling the cooling elements I ₁ and I ₂ are the observer gains δ, b ₀ , and values for α are adjustable parameters that can be selected based on overall system tuning.

Equations 4 and 5 can be rewritten to show the output control signal as a function of the ESO outputs. For example, ESO Equations 4 and 5 may be converted into a linear form as in Equation 6 .

If the zero order hold method is used, the discrete-time form of ESO is expressed as Equation (7).

T _s is the sampling time.

In various embodiments, the supervisor gain β should be selected to position all poles (eg, four) of Equation 8 as shown below.

β is defined by Equation (9).

In equation (9), ω ₀ is the adjustable monitor bandwidth. Thus, in various embodiments, the supervisor gain β may be a single tuning parameter for ESO. With properly tuned ESO, the watcher states will track the state of the system. That is, the output control signal u _i may be expressed by Equation (10).

R is the temperature set point (eg, target or desired operating temperature), ω _c is the adjustable controller bandwidth, and b ₀ is the adjustable parameter set by the user. That is, ESO estimates the maximum temperature and workload disturbance, and ADRC can actively compensate for the workload disturbance.

3 shows an exemplary ADRC controller according to various embodiments of the present invention.

Referring to FIG. 3 , in various embodiments, the ADRC controller 300 receives temperature information from one or more temperature sensors 310 and outputs the output of one or more cooling elements 320 to a thermal zone. can be controlled The thermal zone may have various sizes. For example, in various embodiments, one or more SSDs may be operating in a thermal zone inside a storage server, but in other embodiments, a thermal zone may include one or more servers operating inside a data center. have. Various embodiments may be expanded or contracted and may include various levels of ADRC controllers (eg, ADRC-based control for cooling within a server and ADRC-based control for cooling within a data center). .

In various embodiments, the ADRC controller 300 receives the target temperature r. The target temperature may be a beneficial operating temperature for one or more devices operating within the thermal zone. For example, in various embodiments, the thermal zone may include a plurality of SSDs, and the target temperature may be 60 degrees Celsius. In various embodiments, the ADRC controller 300 also receives the zone temperature y from the temperature sensor(s). In various embodiments, the zone temperature r may include the highest temperature read from the thermal zones. In other embodiments, the zone temperature y may include an average temperature reading from temperature sensors in the thermal zone.

In various embodiments, the ADRC includes a comparison component 330 that compares the zone temperature and the target temperature to generate a temperature tracking error. In various embodiments, the zone temperature (y) to generate an estimate of the maximum temperature of a read out of the estimation and zone (z ₁₎ of the total interference for the loop (z2) in accordance with Equation 1 to Equation (5) is ESO ( 340) is provided. In various embodiments, the ADRC controller 300 includes an output controller 350 that provides an output to the cooling element(s) 320 . In various embodiments, the output controller 350 receives the tracking error, the target temperature r, and the outputs z ₁ , z ₂ of the ESO 340 , and an output control signal u _i (eg, mathematical according to Equation 10). For further calculation for calculating the estimated maximum temperature z1 as shown in Equation 4, the output control signal u _i may be fed back to the ESO 340 .

In various embodiments, the ADRC controller 300 may also receive input from other ADRC controllers. For example, in various embodiments, cooling elements may be shared across thermal zones, or not all of the cooling elements may be active at the same time. In various embodiments, one or more other ADRC controllers 300 may provide their tracking error for comparison, and the cooling elements for the zone with the highest tracking error may be activated. have. This is explained in more detail with reference to FIGS. 4 and 5 .

4 shows an exemplary ADCR controller including a plurality of ADRC controllers for a plurality of thermal zones in accordance with various embodiments of the present invention.

Referring to FIG. 4 , in various embodiments, the ADRC controller 400 is configured to manage cooling of the first to eighth thermal zones Zone1 to Zone8_. For example, in various embodiments, ADRC controller 400 may be a MIMO system that is a combination of a plurality of single-input-single-output (SISO) or MIMO systems, and separate ADRC controllers for each of the thermal zones ( 410-480). For example, in various embodiments, the first ADRC controller 410 may be configured to monitor the first thermal zone Zone1 and control the cooling elements FAN1 and FAN2 . The second ADRC controller 420 may be configured to monitor the second thermal zone Zone2 and control the cooling elements FAN1 and FAN3 . The third ADRC controller 430 may be configured to monitor the third thermal zone Zone3 and control the cooling element FAN4 . The fourth ADRC controller 440 may be configured to monitor the fourth thermal zone Zone4 and control the cooling element FAN1 . The fifth ADRC controller 450 may be configured to monitor the fifth thermal zone Zone5 and control the cooling elements FAN1 and FAN6 . The sixth ADRC controller 460 may be configured to monitor the sixth thermal zone Zone6 and control the cooling element FAN7 . The seventh ADRC controller 470 may be configured to monitor the seventh thermal zone Zone7 and control the cooling elements FAN1 and FAN8 . The eighth ADRC controller 480 may be configured to monitor the eighth thermal zone Zone8 and control the cooling element FAN1 .

5 shows a method of controlling cooling elements for a plurality of thermal zones according to various embodiments of the present invention.

4 and 5 , in various embodiments, the ADRC controller 400 is configured to receive temperature measurements from a plurality of thermal zones and to activate one or more cooling elements. For example, in various embodiments, the first ADRC controller 410 may read the temperature sensor(s) from the first thermal zone Zone1. (S500) At the same time, the second ADRC controller 420 may read the temperature sensor(s) from the second thermal zone Zone2. (S505) In various embodiments, the first and second ADRC controllers 410 and 420 may generate the first and second temperature tracking errors. ( S510 , S520 ) In various embodiments, the ADRC 400 may not simultaneously activate all of the cooling elements for the first thermal zone Zone1 and the second thermal zone Zone2 . For example, the first thermal zone Zone1 and the second thermal zone Zone2 may share a cooling element (eg FAN1 ) or all of the zones due to some limitation (eg power limitation) It may not be possible to activate the cooling element for For example, in a server, a first thermal zone Zone1 includes one or more processors, a second thermal zone Zone2 includes one or more SSDs, and a first fan FAN1 includes a server chassis The fan may be a fan, the second fan FAN2 may be a CPU fan, and the third fan FAN3 may be an SSD fan. That is, when controlling the first and second thermal zones, the first fan FAN1 may be controlled by only one ADRC controller. That is, in various embodiments, the ADRC may compare the first and second tracking errors. (S520) In various embodiments, when the first tracking error is greater than the second tracking error, the first ADRC activates the cooling elements according to the temperature readings of the first thermal zone. (S530) Conversely, if the first tracking error is smaller than the second tracking error, the second ADRC activates the cooling elements according to the temperature readings of the second thermal zone. (S535)

Accordingly, the above-described embodiments of the present invention provide a system and method for active disturbance rejection based thermal control. ADRC-based systems can better control configuration temperature and improve fan power consumption efficiency, while avoiding device degradation (eg, device throttling) caused by variations in workload. For example, in various embodiments, an ADRC-based system may better regulate the temperature of a configuration when compared to conventional systems (eg, intrinsic cooling, dynamic cooling, and PID-based cooling). For example, ADRC-based thermal control systems are more robust to ambient temperature than PID-based systems. ADRC-based systems are more resistant to fast changes and/or unknown workloads than native, dynamic, and PID-based systems. ADRC-based systems are more resistant to changes in CPU utilization ratio than native and PID-based systems. ADRC-based systems can further minimize temperature overshot and undershot than native, dynamic, and PID-based systems. Furthermore, ADRC-based systems can better avoid device throttling compared to dynamic and PID-based systems. These features allow ADRC-based systems to provide higher component performance than conventional systems.

In various embodiments, an ADRC-based thermal control system may improve fan power control efficiency. For example, unlike native and dynamic systems, ADRC-based systems provide a control signal that can vary continuously. Similarly, ADRC-based systems are capable of temperature estimation error correction, whereas intrinsic and dynamic-based systems are not. ADRC-based systems are more efficient in power consumption and have a low control effort compared to native and dynamic systems. ADRC-based systems allow for smoother control effort when compared to dynamic and PID-based systems, thereby reducing fan wear and thereby increasing the life expectancy of the fan.

The detailed description shows exemplary embodiments of the present invention, and the present invention is not limited thereto. Although some exemplary embodiments have been described, it will be fully understood by those skilled in the art that various changes are possible in the exemplary embodiments without departing from the novel teachings and advantages of the exemplary embodiments. Accordingly, all such modifications are intended to be included in the exemplary embodiments as defined in the appended claims. In the claims, functional claims are intended to cover structures as described in the text that perform the referenced function, as well as structural equivalents thereof, as well as equivalent structures. Therefore, the foregoing descriptions are exemplary embodiments, and should not be construed as being limited to the specific embodiments described. Modifications to the described exemplary embodiments as well as other embodiments are intended to be included within the scope of the appended claims. The technical spirit of the present invention will be defined by the following claims and their equivalents.

Claims (10)

In the method of active disturbance rejection thermal control (ADRC),
receiving, at the first ADRC controller, a first temperature measurement from the first thermal zone; and
generating, by the first ADRC controller, a first output control signal for controlling a first cooling element;
The first output control signal includes a first estimated disturbance for the first thermal zone computed by a first extended state observer (ESO) of the first ADRC controller based on the first temperature measurement and a method generated according to a first estimated temperature for the first thermal zone. The method of claim 1,
receiving, at a second ADRC controller, a second temperature measurement from a second thermal zone while receiving the first temperature measurement; and
generating, by the second ADRC controller, a second output control signal for controlling a second cooling element while generating the first output control signal;
and the second output control signal is generated according to a second estimated disturbance for the second thermal zone and a second estimated temperature for the second thermal zone computed by a second ESO of the second ADRC controller. 3. The method of claim 2,
After the first and second output signals are generated:
calculating, at the first ADRC controller, a first temperature tracking error by comparing the first temperature measurement with a first target temperature;
calculating, at the second ADRC controller, a second temperature tracking error by comparing the second temperature measurement with a second target temperature; and
The method further comprising: modifying at least one of the first output control signal or the second output control signal based on the first temperature tracking error and the second temperature tracking error. The method of claim 1,
wherein the first thermal zone includes at least one of a solid state drive (SSD), a processor, or a graphics processing unit. The method of claim 1,
A first ESO is a maximum temperature read from the first thermal zone, an estimate of the maximum temperature read from the first thermal zone, an estimate of the total room, the first output control signal, an observer gain, at least one A method comprising a second order ESO (second order ESO) operating according to an adjustable parameter of 6. The method of claim 5,
wherein the supervisor gain is inversely proportional to the values of adjustable supervisor bandwidth and sampling times. The method of claim 1,
A first output control signal ui is based on a first target temperature of the first thermal zone, the first estimated temperature, the first estimated disturbance, an adjustable controller bandwidth, and an adjustable parameter. In the storage server,
a first device included in a first thermal zone comprising a first temperature sensor and a first cooling element; and
a first active disturbance rejection thermal control (ADRC) controller connected to the first temperature sensor and the first cooling element;
The first ADRC controller is:
receive a first temperature measurement of the first thermal zone from the first temperature sensor;
a first estimated temperature for the first thermal zone, a first estimated disturbance computed by a first extended state observer (ESO) for the first thermal zone, and a storage server configured to provide a first output control signal to the first cooling element according to a first target temperature for 9. The method of claim 8,
a second device included in a second thermal zone comprising a second temperature sensor and a second cooling element; and
Further comprising a second ADRC controller connected to the second temperature sensor and the second cooling element,
The second ADRC controller is:
receiving a second temperature measurement from the second temperature sensor;
generating a second output control signal for controlling the second cooling element;
The second output control signal includes a second estimated temperature of the second thermal zone, a second estimated disturbance computed by a second ESO for the second thermal zone, and a second target for the second thermal zone. A storage server created by temperature. In a data center,
a first thermal zone comprising a first temperature sensor and a first cooling element;
a first temperature sensor coupled with the first cooling element, the first temperature sensor receiving a first temperature measurement of the first thermal zone from the first temperature sensor, a first estimated temperature, a first extended state monitor (ESO); a first active disturbance rejection thermal control (ADRC) configured to provide a first output control signal to the first cooling element according to a first estimated disturbance calculated by an estimated state observer, and a first target temperature control) controller;
a second thermal zone comprising a second temperature sensor and a second cooling element;
a second active, coupled to the second temperature sensor and the second cooling element, configured to receive a second temperature measurement from the second temperature sensor, and to generate a second output control signal for controlling the second cooling element an active disturbance rejection thermal control (ADRC) controller;
and the second output control signal is generated according to a second estimated temperature and a second estimated disturbance calculated by the second ESO, and a second target temperature.

Images