WO2015114580A1

WO2015114580A1 - Method and system for monitoring and control of cooling of data centers based on it load

Info

Publication number: WO2015114580A1
Application number: PCT/IB2015/050711
Authority: WO
Inventors: Shrikant Bhat; Niket KAISARE
Original assignee: Abb Technology Ltd.
Priority date: 2014-01-31
Filing date: 2015-01-30
Publication date: 2015-08-06

Abstract

A method and system for monitoring and control of cooling of data centers is disclosed. The method includes recording an IT load distribution for a rack in the data centre and defining a temperature reference value (T_ref), a cooling setting (U_min) and a temperature band (T_band) for the rack based on the IT load distribution for the rack. A rack inlet temperature (T_rack) is measured at periodic intervals; and cooling setting (U_min) is adjusted based on the IT load distribution, when the rack inlet temperature is below a difference between (T_ref) and (T_band).

Description

METHOD AND SYSTEM FOR MONITORING AND CONTROL OF COOLING OF DATA CENTERS BASED ON IT LOAD

FIELD OF THE INVENTION AND USE OF INVENTION

[0001] The invention generally relates to the field of cooling of data centers and more specifically relates to a method and system for monitoring and control of cooling of data centres based on IT load distribution and a temperature band approach.

PRIOR ART AND PROBLEM TO BE SOLVED

[0002] A data centre is typically a physical space that contains computer systems and associated telecommunication and storage systems. Large data centres can be industrial scale operations and use large amount of electricity for running of the computer systems. Due to large power usage there is associated heat generation which needs to be effectively removed to maintain the health of the computer systems and associated electronics. This results in deploying cooling equipment and cooling control strategies for effective cooling of data centres.

[0003] There are approaches focusing on localized cooling (such as using automated ventilation tiles) and model predictive control to reduce the cooling energy requirements for data centres. There is also prior art that focuses on the integrated control of data center using inverse gain matrix approach to address interactions between various cooling components. These approaches focus on manipulating the supply air flow rate and temperature to restrict the rack inlet air temperature below a certain safe threshold limit. This is expected to reduce the blower power consumption by unto 36%. The bounds on the decision variables in such cases are chosen considering the worst case scenario (for reliability) and are uniform across all the sections in data center. There is also provision of changing the temperature set point based on the information on IT load.

[0004] FIG. 1 is a prior art schematic of a raised floor design, wherein the cool air flows are directed from Control Room Air Conditioning (CRAC) through floor plenum and automated vent tiles to the equipment racks. There are multiple such vent tiles across data center floor. The flow through these tiles can be independently controlled by manipulating the tile openings. The controlled cool air flow is then introduced from cold aisles, gets heated while it passes through the server to hot aisles, and is sent back to CRACs where it gets cooled by chiller water.

[0005] FIG. 2 is a prior art schematic of the control structure for the CRAC as described in. In this case, the supply air temperature ( T_suppiy or SAT), blower speed- Variable Frequency Drive (VFD) and tile openings (Uj) are controlled to maintain the rack inlet temperature below certain threshold value, Tref.

[0006] However, this approach does not address the tradeoff between reliability and energy efficiency, as explained below.

[0007] The objective function at the lower-level MPC (model predictive controller) in the prior art system is:

VFDwjx < VFDj < VFD_MiiXt U_mtn < U < 3_mml, Ύ ^ ίη— ^ref

subject to: where, Rv and RSAT are cost factors associated with blower power and cost of cooling respectively. Thus, the above MPC formulation minimizes the energy cost (blower power using VFD (Variable Frequency Drive) and cost of cooling the supply air to T_suppiy ), while maintaining rack inlet temperature below a threshold, Tref. As per the prior art, the threshold value Tref can be varied across different sections of a data center based on the IT load to reduce energy consumption. This approach has following shortcomings:

• Even if T_ref is changed based on the IT load (higher T_ref for lower IT load and vice versa), Track,in might not be able to reach closer to T_ref unless Umin (tile opening in this case) is also reduced.

• Adapting T_ref and Umin with IT load might lead to loss of reliability. OBJECTS OF THE INVENTION

[0008] The invention described herein aims to increase energy efficiency of air-cooled data centres while ensuring reliability by defining a temperature band (instead of temperature set point) for cooling control and using the information on IT load distribution within a rack to optimally determine the temperature band and adapt the bounds on the constraints. This method also facilitates more hours of economizer (free cooling) integration.

SUMMARY OF THE INVENTION

[0009] In one aspect, a method for monitoring and control of cooling of data centers is disclosed. The method includes recording an IT load distribution for a rack in the data centre and defining a temperature reference value (Tref), a cooling setting - associated with Umin), and a temperature band (Tband) for the rack based on the IT load distribution for the rack. The method further includes measuring a rack inlet temperature (Track) at periodic intervals; and adjusting cooling setting (Umin) based on the IT load distribution, when the rack inlet temperature is below a difference between (Tref) and (Tband).

[0010] In another aspect a system for monitoring and control of cooling of data centers is provided. The system comprises a rack with a given IT load distribution; a measurement module for measuring rack inlet temperature value at periodic intervals; an energy management module to control a cooling setting for the rack; and a controller comprising predefined temperature reference value, pre-defined cooling setting, and pre-defined temperature band, based on the IT load distribution, where the controller is configured to receive rack inlet temperature value at the periodic intervals, and use a difference between the a difference between (T_ref) and (Tband) to compare with the rack inlet temperature value, and is further configured to adjust a cooling setting (associated with effective power usage -Umin) when the rack inlet temperature is below the difference between (T_ref) and (Tband).

DRAWINGS

[0011] These and other features, aspects, and advantages of the present invention will become better understood when the following detailed description is read with reference to the accompanying drawings in which like reference numerals represent corresponding parts throughout the drawings, wherein:

[0012] FIG. 1 is a prior art schematic of a raised floor design, wherein the cool air flows are directed from Control Room Air Conditioning (CRAC);

[0013] FIG. 2 is a prior art schematic of the control structure for the CRAC; [0014] FIG. 3 is a schematic illustration of reference temperature, temperature band and change in control setting possible through the method and system of the invention;

[0015] FIG. 4 is a control schematic for the method and system of the invention;

[0016] FIG. 5 is a flowchart showing exemplary steps for the method for monitoring and control of cooling of data centers according to one aspect of the invention; and

[0017] FIG. 6 is a diagrammatic representation of the system for monitoring and control of cooling of data centers according to one aspect of the invention.

DETAILED DESCRIPTION OF THE INVENTION

[0018] As used herein and in the claims, the singular forms "a," "an," and "the" include the plural reference unless the context clearly indicates otherwise.

[0019] A method and system for defining a temperature band for monitoring and control of cooling in data center is provided where for each section or rack of the data centre, the upper limit of the temperature band is determined by desired energy efficiency value (cooling requirement) and width of the band is determined by the reliability requirements for the data centre considering the distribution of IT load in that section.

[0020] Thus the method and system of invention allows for adapting the bounds on the decision variables when the temperature in a given section is outside the previously defined temperature band and resetting/readapting the bounds when the temperature is brought back within the temperature band. It may be noted here that the method and system of the invention can be adapted for other data centre cooling designs (other than air cooling) such as liquid cooling and hybrid air-liquid cooling.

[0021] In one specific embodiment, adapting the limit for temperature will also be accompanied by incorporation of the economizer mode in the cooling management, where the IT load is low and therefore temperature TRef can be higher and cooling can be achieved through ambient air flow itself.

[0022] The different embodiments described herein effective address the issue of optimality i.e. maintaining the energy efficiency vs. reliability tradeoff for the effective data centre cooling. [0023] The method and system described herein involve defining a temperature band, Tband, based on the IT load distribution in a given rack as shown in FIG. 3. The control loop defined by Umin (achieved through airflow speed/control or cooling setting) is adapted only when Trackjn goes below T_ref -Tband. Also change in Umin (airflow speed/control or cooling setting) is dependent on IT load distribution as shown in the control schematic of FIG. 4.

[0024] FIG. 5 illustrates a flow chart 10 for a more detailed method for monitoring and control of cooling of data centers. As explained herein, the method includes a step 12 for recording an IT load distribution for a rack in the data centre; a step 14 for defining a temperature reference value (Tref) for the rack based on the IT load distribution for the rack; a step 16 for defining a cooling setting (Umin) for the rack based on the IT load distribution; a step 18 for defining a temperature band (Tband) based on the IT load distribution; a step 20 for measuring a rack inlet temperature (Track) at periodic intervals; and a step 22 for adjusting cooling setting (Umin) based on the IT load distribution, when the rack inlet temperature is below a difference between (Tref) and (Tband).

[0025] The IT load distribution described herein may be uniform or non-uniform. Further the IT load distribution may be low or high based on actual usage.

[0026] It may be noted here that the difference between (T_ref) and (Tband) and adjustment of cooling setting is determined using known techniques in the art.

[0027] A further advantage of the method is ability to operate an economizer mode for allowing ambient cooling and increasing the Umin when IT load distribution is uniform and low.

[0028] FIG. 6 is diagrammatic representation of a system 30 for monitoring and control of cooling of data centers. The system comprises a rack 32 with a given IT load distribution. Data centre racks are well understood to those skilled in the art. A measurement module 34 is used for measuring rack inlet temperature value at periodic intervals. The measurement module may be implemented through a hardware circuitry through known electronic components and may include programmable integrated circuitry chips. An energy management module 38 is provided that is configured to regulate the Umin through a cooling setting. A controller 36 is used for implementing the method of the invention, that includes a hardware or software or a combination of hardware and software module for storing predefined temperature reference value, cooling setting, and temperature band, all of which are set based on the IT load distribution. The controller is configured to receive rack inlet temperature value at the periodic intervals from the measurement module. The controller has a calculator module (a hardware or software or a combination of hardware and software) to calculate a difference between between (Tref) and (Tband) and a comparator module (a hardware or software or a combination of hardware and software) to compare this difference with the rack inlet temperature value. The controller is further configured to trigger a control action to adjust cooling setting (Umin) when the rack inlet temperature is below the difference between (T_ref) and (Tband). It may be appreciated by those skilled in the art that the adjustment may be done automatically to pre-defined settings or alternately the controller may be configured to send an alert signal to an external device to trigger the adjustment action. The external device may be a remote monitor, or a mobile communication device. The system also includes an economizer mode 40 to allow ambient cooling in case the ambient temperature is less than the rack temperature.

[0029] As explained in reference with the method of the invention, the IT load distribution may be uniform or non-uniform, high or low. Also, the system includes an economizer mode to allow ambient cooling and increasing the Umin when IT load distribution is uniform and low.

[0030] Under implementation, the use of method and system of the invention is presented in the following exemplary scenarios: For two cases of data centres with a rack with higher average IT load but with (a) uniform distribution and (b) non-uniform distribution, Tref will be the same for both the cases. However, Tband width will be kept higher for case (b) than case (a) and therefore the change in Umin (airflow speed/control or cooling setting) as per the method and system of the invention will be lower for case (b) than case (a).

[0031] For another two cases with a rack with lower average IT load but with (c) uniform distribution and (d) non-uniform distribution, Tref will be the same for both the cases and higher than scenario presented above. Again, Tband width will be higher for case (d) than case (c), and therefore the change in Umin (airflow speed/control or cooling setting) will be lower for case (d) than case (c). However, Tband for scenarios (c) and (d) will be narrower than scenario (a) and (b), respectively. Also change in Umin for scenarios (c) and (d) will be larger than scenario (a) and (b), respectively. [0032] It may be understood by those skilled in the art that the main objective here to ensure effective cooling while maintaining energy efficiency and reliability of operability of the data centre. Hence wherever there is non-uniform IT load, to take into account the unpredictability of heat being generated due to the non-uniformity, the cooling setting will not be disturbed too much from it' s preset or pre-defined position. Also, when the average IT load is high then again, Umin will not be adjusted to the extent when average IT load is high. In all these scenarios both reliability and energy efficiency objectives are met.

[0033] It would be appreciated by those skilled in the art that the values of Tband and change in Umin can be determined based on empirical correlations, using fuzzy logic or a rule based approach.

[0034] The described embodiments may be implemented as a system, method, apparatus or article of manufacture using standard programming and engineering techniques related to software, firmware, hardware, or any combination thereof. The described operations may be implemented as code maintained in a "computer readable medium", where a processor may read and execute the code from the computer readable medium. A computer readable medium may comprise media such as magnetic storage medium (e.g., hard disk drives, floppy disks, tape, etc.), optical storage (CD-ROMs, DVDs, optical disks, etc.), volatile and non-volatile memory devices (e.g., EEPROMs, ROMs, PROMs, RAMs, DRAMs, SRAMs, Flash Memory, firmware, programmable logic, etc.), etc. The code implementing the described operations may further be implemented in hardware logic (e.g. an integrated circuit chip, Programmable Gate Array (PGA), Application Specific Integrated Circuit (ASIC), etc.). Still further, the code implementing the described operations may be implemented in "transmission signals", where transmission signals may propagate through space or through a transmission media, such as an optical fibre, copper wire, etc. The transmission signals in which the code or logic is encoded may further comprise a wireless signal, satellite transmission, radio waves, infrared signals, Bluetooth, etc. The transmission signals in which the code or logic is encoded is capable of being transmitted by a transmitting station and received by a receiving station, where the code or logic encoded in the transmission signal may be decoded and stored in hardware or a computer readable medium at the receiving and transmitting stations or devices. An "article of manufacture" comprises computer readable medium, hardware logic, or transmission signals in which code may be implemented. A device in which the code implementing the described embodiments of operations is encoded may comprise a computer readable medium or hardware logic. Of course, those skilled in the art will recognize that many modifications may be made to this configuration without departing from the scope of the present invention, and that the article of manufacture may comprise suitable information bearing medium known in the art.

[0035] The different modules referred herein may use a data storage unit or data storage device that is selected from a set of but not limited to USB flash drive (pen drive), memory card, optical data storage discs, hard disk drive, magnetic disk, magnetic tape data storage device, data server and molecular memory.

[0036] A computer network may be used for allowing interaction between two or more electronic devices or modules, and includes any form of inter/intra enterprise environment such as the world wide web, Local Area Network (LAN), Wide Area Network (WAN), Storage Area Network (SAN) or any form of Intranet.

[0037] While only certain features of the invention have been illustrated and described herein, many modifications and changes will occur to those skilled in the art. It is, therefore, to be understood that the appended claims are intended to cover all such modifications and changes as fall within the true spirit of the invention.

Claims

/ We Claim:

1. A method for monitoring and control of cooling of data centers, the method comprising: recording an IT load distribution for a rack in the data centre; defining a temperature reference value (Tref) for the rack based on the IT load distribution for the rack; defining a cooling setting (Umin) for the rack based on the IT load distribution; defining a temperature band (Tband) based on the IT load distribution; measuring a rack inlet temperature (Track) at periodic intervals; and adjusting cooling setting (Umin) based on the IT load distribution, when the rack inlet temperature is below a difference between (Tref) and (Tband).

2. The method for claim 1 wherein the IT load distribution is one of uniform or nonuniform.

3. The method of claim 1 wherein the IT load distribution high or low with respect to a predefined standard for the rack.

4. The method of claim 1 wherein the difference between (T_ref) and (Tband) and adjustment of cooling setting is determined based on empirical correlations.

5. The method of claim 1 further comprising operating an economizer mode for allowing ambient cooling and increasing the Umin when IT load distribution is uniform and low.

6. A system for monitoring and control of cooling of data centers, the system comprising: a rack with a given IT load distribution; a measurement module for measuring rack inlet temperature value at periodic intervals; an energy management module to control a cooling setting for the rack; a controller comprising predefined temperature reference value, pre-defined cooling setting, and pre-defined temperature band, based on the IT load distribution, wherein the controller is configured to receive rack inlet temperature value at the periodic intervals, and use a difference between the a difference between (Tref) and (Tband) to compare with the rack inlet temperature value, and is further configured to adjust a cooling setting (Umin) when the rack inlet temperature is below the difference between (T_ref) and (Tband).

7. The system of claim 6 wherein the IT load distribution is one of uniform or nonuniform.

8. The system of claim 6 wherein the IT load distribution is one of high or low with respect to a pre-defined standard.

9. The system of claim 6 wherein the difference between (T_ref) and (Tband) and an adjustment of cooling setting is determined based on empirical correlations.

10. The system of claim 6 further comprising an economizer mode to allow ambient cooling and increasing the Umin when IT load distribution is uniform and low.