U.S. patent application number 14/350376 was filed with the patent office on 2014-08-28 for managing airflow provisioning.
The applicant listed for this patent is Cullen E. Basch, Alan A. Mcreynolds, Zhikui Wang, Rongliang Zhou. Invention is credited to Cullen E. Basch, Alan A. Mcreynolds, Zhikui Wang, Rongliang Zhou.
Application Number | 20140244046 14/350376 |
Document ID | / |
Family ID | 48290428 |
Filed Date | 2014-08-28 |
United States Patent
Application |
20140244046 |
Kind Code |
A1 |
Zhou; Rongliang ; et
al. |
August 28, 2014 |
MANAGING AIRFLOW PROVISIONING
Abstract
In an implementation, a method for managing airflow provisioning
in an area comprising a plurality of racks, wherein a plurality of
fluid moving devices are to supply airflow to the plurality of
racks through a plurality of adjustable vent tiles, includes
accessing a model that describes airflow transport and distribution
within the area, said model comprising a plurality of parameters,
determining values for the plurality of parameters, and
implementing the model to partition the area into a plurality of
fluid moving device zones of influence with a desired level of
overlapping among the plurality of fluid moving device zones of
influence.
Inventors: |
Zhou; Rongliang; (Mountain
View, CA) ; Wang; Zhikui; (Fremont, CA) ;
Basch; Cullen E.; (Los Gatos, CA) ; Mcreynolds; Alan
A.; (Los Altos, CA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Zhou; Rongliang
Wang; Zhikui
Basch; Cullen E.
Mcreynolds; Alan A. |
Mountain View
Fremont
Los Gatos
Los Altos |
CA
CA
CA
CA |
US
US
US
US |
|
|
Family ID: |
48290428 |
Appl. No.: |
14/350376 |
Filed: |
November 11, 2011 |
PCT Filed: |
November 11, 2011 |
PCT NO: |
PCT/US11/60434 |
371 Date: |
April 8, 2014 |
Current U.S.
Class: |
700/276 |
Current CPC
Class: |
H05K 7/20745 20130101;
F24F 11/30 20180101; H05K 7/20836 20130101 |
Class at
Publication: |
700/276 |
International
Class: |
F24F 11/00 20060101
F24F011/00 |
Claims
1. A method for managing airflow provisioning in an area comprising
a plurality of racks, wherein a plurality of fluid moving devices
are to supply airflow to the plurality of racks through a plurality
of adjustable vent tiles, said method comprising: accessing a model
that describes airflow transport and distribution within the area,
said model comprising a plurality of parameters; determining, by a
processor, values for the plurality of parameters; and implementing
the model to partition the area into a plurality of fluid moving
device zones of influence with a desired level of overlapping among
the plurality of fluid moving device zones of influence.
2. The method according to claim 1, wherein the model further
describes effects of actuations on the plurality of fluid moving
devices on the airflow transport and distribution within the
area.
3. The method according to claim 1, wherein the model further
describes effects of actuations on the plurality of fluid moving
devices and the adjustable vent tiles on the transport and
distribution of airflow supplied into the plurality of racks.
4. The method according to claim 3, wherein the model is described
by the following equation: T ( k + 1 ) = T ( k ) + { i = 1 N CRAC g
i [ SAT i ( k ) - T ( k ) ] VFD i ( k ) } { j = 1 N tile b j U j (
k ) ] } + C , ##EQU00002## wherein T represents a rack inlet
temperature, k and k+1 represent discrete time steps, SAT.sub.i and
VFD.sub.i are a supply air temperature and a blower speed of the
ith fluid moving device, U.sub.j is the opening of the jth
adjustable vent tile, N.sub.CRAC and N.sub.tile are the number of
fluid moving devices and adjustable vent tiles, respectively, and
wherein g.sub.i and b.sub.j are the parameters that capture
influences of each fluid moving device i and adjustable vent tile
j, respectively, and C denotes a temperature change.
5. The method according to claim 1, wherein implementing the model
to partition the area into a plurality of fluid moving device zones
of influence with a level of overlapping further comprises:
determining influence levels of the plurality of fluid moving
devices on a plurality of rack inlet temperatures; for each of the
plurality of rack inlet temperatures, calculating ratios between
each of the determined influence levels and a largest influence
level of the determined influence levels for that rack inlet
temperature; and partitioning the data center into the plurality of
fluid moving device zones of influence based upon the calculated
ratios.
6. The method according to claim 5, further comprising: setting an
overlapping threshold value for the ratios, wherein overlapping
threshold value is to substantially control the level of
overlapping among the plurality of fluid moving device zones of
influence.
7. The method according to claim 1, wherein implementing the model
further comprises implementing the model to simultaneously control
the plurality of fluid moving devices and adjustable vent tiles to
manage airflow provisioning in the area.
8. The method according to claim 7, wherein implementing the model
further to simultaneously control the plurality of fluid moving
devices and adjustable vent tiles further comprises: accessing a
cost function; and determining a coordinated actuation of the
plurality of fluid moving devices and adjustable vent tiles through
use of the model to minimize the cost function while substantially
maintaining rack inlet temperatures within predetermined
ranges.
9. An apparatus for managing airflow provisioning in an area
comprising a plurality of fluid moving devices and a plurality of
adjustable vent tiles, said apparatus comprising: a memory storing
at least one module comprising machine readable instructions to:
access a model that describes airflow transport and distribution
within the area, said model comprising a plurality of parameters;
determine values for the plurality of parameters; and implement the
model to partition the area in to a plurality of fluid moving
device zones of influence with a level of overlapping among the
plurality of fluid moving device zones of influence; and a
processor to implement the at least one module.
10. The apparatus according to claim 9, wherein the at least one
module further comprises machine readable instructions to:
determine influence levels of the plurality of fluid moving devices
on a plurality of rack inlet temperatures; for each of the rack
inlet temperatures, calculate ratios between each of the determined
influence levels and a largest influence level of the determined
influence levels for that rack inlet temperature; and partition the
data center into the plurality of fluid moving device zones of
influence based upon the calculated ratios.
11. The apparatus according to claim 9, wherein the at least one
module further comprises machine readable instructions to: access a
cost function; determine a coordinated actuation of the plurality
of fluid moving devices and adjustable vent tiles through use of
the model to minimize the cost function while substantially
maintaining rack inlet temperatures within predetermined ranges;
and output the determined coordinated actuation.
12. The apparatus according to claim 19, wherein the model is
described by the following equation: T ( k + 1 ) = T ( k ) + { i =
1 N CRAC g i [ SAT i ( k ) - T ( k ) ] VFD i ( k ) } { j = 1 N tile
b j U j ( k ) ] } + C , ##EQU00003## wherein T represents a rack
inlet temperature, k and k+1 represent discrete time steps,
SAT.sub.i and VFD.sub.i are a supply air temperature and a blower
speed of the ith fluid moving device, U.sub.j is the opening of the
jth adjustable vent tile, N.sub.CRAC and N.sub.tile are the number
of fluid moving devices and adjustable vent tiles, respectively,
and wherein g.sub.i and b.sub.j are the parameters that capture
influences of each fluid moving device i and adjustable vent tile
j, respectively, and C denotes a temperature change.
13. A non-transitory computer readable storage medium on which is
embedded at least one computer program, said at least one computer
program implementing a method for managing airflow provisioning in
an area comprising a plurality of fluid moving devices and a
plurality of adjustable vent tiles, said at least one computer
program comprising computer readable code to: access a model that
describes airflow transport and distribution within the area, said
model comprising a plurality of parameters; determine values for
the plurality of parameters; and implement the model to
simultaneously control the plurality of fluid moving devices and
the plurality of adjustable vent tiles, said at least one computer
program further comprising computer readable code to: access to a
cost function; and determine a coordinated actuation of the
plurality of fluid moving devices and adjustable vent tiles through
use of the model to minimize the cost function while substantially
maintaining rack inlet temperatures within predetermined
ranges.
14. The non-transitory computer readable storage medium according
to claim 14, wherein the model is described by the following
equation: T ( k + 1 ) = T ( k ) + { i = 1 N CRAC g i [ SAT i ( k )
- T ( k ) ] VFD i ( k ) } { j = 1 N tile b j U j ( k ) ] } + C ,
##EQU00004## wherein T represents a rack inlet temperature, k and
k+1 represent discrete time steps, SAT.sub.i and VFD.sub.i are a
supply air temperature and a blower speed of the ith fluid moving
device, U.sub.j is the opening of the jth adjustable vent tile,
N.sub.CRAC and N.sub.tile are the number of fluid moving devices
and adjustable vent tiles, respectively, and wherein g.sub.i and
b.sub.j are the parameters that capture influences of each fluid
moving device i and adjustable vent tile j, respectively, and C
denotes a temperature change.
15. The non-transitory computer readable storage medium according
to claim 14, said at least one computer program further comprising
computer readable code to: determine influence levels of the
plurality of fluid moving devices on a plurality of rack inlet
temperatures; for each of the plurality of rack inlet temperatures,
calculate ratios between each of the determined influence levels
and a largest influence level of the determined influence levels
for that rack inlet temperature; and partition the data center into
the plurality of fluid moving device zones of influence with a
desired level of overlapping among the plurality of fluid moving
device zones of influence based upon the calculated ratios.
Description
BACKGROUND
[0001] Data centers typically include multiple cooling units, such
as, computer room air conditioning (CRAG) units, arranged to supply
cooling airflow to a plurality of servers arranged in a rows of
racks. The cooling airflow is often supplied through vent tiles
distributed at multiple locations on a raised floor. More
particularly, the fluid moving devices supply cooling airflow into
a plenum formed beneath the raised floor and the cooling airflow is
supplied to the servers through the vent tiles.
[0002] The cooling units are typically operated to substantially
ensure that the temperatures in the servers are maintained within
predetermined temperature ranges. That is, to largely prevent the
servers from reaching temperature levels at which the servers
operate inefficiently or are harmful to the servers, the cooling
units are typically operated to supply cooling resources at lower
temperatures and/or at higher volume flow rates than are necessary
to maintain the servers within the predetermined temperature
ranges. This over-provisioning of cooling resources is inefficient,
increases operational costs of the data center, and shortens the
life span of the cooling units.
BRIEF DESCRIPTION OF DRAWINGS
[0003] Features of the present disclosure are illustrated by way of
example and not limited in the following figure(s), in which like
numerals indicate like elements, in which:
[0004] FIG. 1 illustrates a simplified block diagram of a section
of a data center, according to an example of the present
disclosure;
[0005] FIG. 2 shows a block diagram of a system for managing
airflow provisioning in the data center depicted in FIG. 1,
according to an example of the present disclosure.
[0006] FIG. 3 illustrates a flow diagram of a method for managing
airflow provisioning in the data center depicted in FIG. 1,
according to an example of the present disclosure;
[0007] FIGS. 4 and 5, respectively, depict flow diagrams of methods
of implementing the model disclosed herein in managing airflow
provisioning depicted in FIG. 3, according to two examples of the
present disclosure;
[0008] FIG. 6 depicts a control diagram that includes an MPC that
implements the model disclosed herein, according to an example of
the present disclosure; and
[0009] FIG. 7 illustrates a block diagram of a computing device to
implement the methods depicted in FIGS. 3-5, according to example
of the present disclosure.
DETAILED DESCRIPTION
[0010] For simplicity and illustrative purposes, the present
disclosure is described by referring mainly to an example thereof.
In the following description, numerous specific details are set
forth in order to provide a thorough understanding of the present
disclosure. It will be readily apparent however, that the present
disclosure may be practiced without limitation to these specific
details. In other instances, some methods and structures have not
been described in detail so as not to unnecessarily obscure the
present disclosure. As used herein, the term "includes" means
includes but not limited to, the term "including" means including
but not limited to. The term "based on" means based at least in
part on. In addition, the variables "l", "m", and "n" are intended
to denote integers equal to or greater than one and may denote
different values with respect to each other.
[0011] Disclosed herein are a method and apparatus for managing
airflow provisioning in an area, such as, a data center. More
particularly, the airflow provisioning is managed through
implementation of a model that describes airflow transport and
distribution within the area. According to an example, the model
comprises a physics based state-space model. In addition,
parameters for the model are determined and the model is
implemented in managing the airflow provisioning. The airflow
provisioning includes the determination of the temperatures and
volume flow rates of airflow supplied by a plurality of fluid
moving devices as well as the volume flow rates of airflow supplied
through a plurality of adjustable vent tiles.
[0012] The model disclosed herein is a holistic model, in that,
zonal and local level actuations are coordinated. More
particularly, the model disclosed herein captures the airflow
resources provisioning, transport, and distribution, and
incorporates both zonal airflow actuation, including the fluid
moving device supply air temperature and blower speed, and local
airflow provisioning actuation (for instance, from adaptive vent
tiles). One result of this coordination is that fighting among
various airflow actuations is substantially eliminated, while
substantially optimal airflow provisioning efficiency is attained.
In another regard, implementation of the model disclosed herein
enables data centers to be partitioned into fluid moving device
zones of influence with adjustable levels of overlapping among the
fluid moving device zones of influence.
[0013] Implementation of the model disclosed herein also enables
dynamic prediction of transient trajectories of the rack inlet
temperatures based upon their current thermal statuses for any
given zonal and local airflow actuations. In other words, the model
disclosed herein may be implemented to dynamically predict how the
rack inlet temperatures evolve over time. Implementation of the
model disclosed herein further enables the determination of future
rack inlet temperatures to be determined once current rack inlet
temperatures and airflow actuations to be applied are given. In
other words, future rack inlet temperatures may be determined
without performing iterative equation solving. Moreover, the model
disclosed herein enables all of the above features to be attained
in a computationally efficient manner because, according to an
example, the model is explicit and only involves relatively simple
calculations.
[0014] In a further regard, implementation of the model disclosed
herein enables real-time airflow actuation optimization at both the
zonal and local levels, for instance, through minimization of a
cost function. As such, the airflow optimization techniques
disclosed herein are able to detect thermal anomalies or
inefficient airflow statuses and are able to correct those issues
in a timely manner. Moreover, through use of a properly defined
cost function(s), the apparatus disclosed herein actively seeks the
optimal settings for all the fluid moving devices and local airflow
provisioning actuation mechanisms to satisfy the target thermal
status, while minimizing the cost function(s) of interest.
[0015] With reference first to FIG. 1, there is shown a simplified
perspective view of a section of an area 100, in this instance, a
data center, in which a method and apparatus for managing airflow
provisioning may be implemented, according to an example. The data
center 100 is depicted as having a plurality of racks 102a-102n, a
plurality of fluid moving devices 114a-114l (only fluid moving
devices 114a-114b are depicted in FIG. 1), and a plurality of
sensors 120a-120n. The racks 102a-102n are depicted as being
positioned on a raised floor 110 and as housing electronic devices
116. The electronic devices 116 comprise, for instance, computers,
servers, bladed servers, disk drives, displays, etc. As shown in
FIG. 1, airflow, such as cool airflow, is delivered through
adjustable vent tiles 118a-118m in the floor 110 to the racks
102a-102n. The fluid moving devices 114a-114b generally operate to
supply airflow into a space 112 beneath the raised floor 110, and
in certain instances to cool heated airflow (indicated by the
arrows 124). The fluid moving devices 114a-114b may comprise, for
instance, air conditioning (AC) units that have actuators for
controlling the temperature and the volume flow rate of the cooled
airflow supplied by the fluid moving devices 114-114b. In other
examples, the fluid moving devices 114a-114b comprise heaters
having actuators to control the temperature and volume flow rate of
heated airflow supplied by the fluid moving devices 114a-114b.
[0016] The adjustable vent tiles (AVTs) 118a-118m comprise manually
and/or automatically adjustable vent tiles. In any regard, the AVTs
118a-118m may be adjusted to thereby vary the volume flow rate of
airflow supplied through the AVTs 118a-118m. When the AVTs
118a-118m comprise automatically adjustable vent tiles, actuators
(not shown) are provided to vary the operational settings of the
AVTs 118a-118m. In addition, each of the AVTs 118a-118m may also
include an interface through which the AVTs 118a-118m may receive
instruction signals from a controller 130. The operational settings
of the AVTs 118a-118m may include the opening levels of the AVTs
118 that may be used to vary the volume flow rate of the airflow
and, in some instances, a speed level of local fans used to vary
the flow rates of the airflow through the AVTs 118a-118m. The AVTs
118a-118m may have many different suitable configurations and are
thus not to be limited to any particular type of adjustable vent
tile.
[0017] In any regard, the airflow contained in the space 112 may
include airflow supplied by more than one of the fluid moving
devices 114a-114n, and in certain instances, airflow recirculated
into the space 112 from above the floor 110. Thus, characteristics
of the airflow, such as, temperature, pressure, humidity, flow
rate, etc., delivered to various locations in the data center 100
may substantially be affected by the operations of multiples ones
of the fluid moving devices 114a-114n. As such, conditions at
various locations in the data center 100 may substantially be
affected by the operations of more than one of the fluid moving
devices 114a-114n.
[0018] The sensors 120a-120n may be networked, in a wired and/or
wireless manner, with the controller 130 to convey detected
condition information to the controller 130. The detected
conditions may include, for instance, temperatures at the inlets of
the racks 102a-102n, temperatures at the outlets of the adjustable
vent tiles 118, etc. The detected conditions may, in addition or
alternatively, include other environmental conditions, such as,
pressure, humidity, airflow velocity, etc. In this regard, the
sensors 120a-120n comprise any suitable types of sensors to detect
the conditions.
[0019] As discussed in greater detail herein below, environmental
condition information collected by the sensors 120a-120n is used to
determine various parameters of a model that describes airflow
transport and distribution within the data center 100. In one
example, the model comprises a physics based state-space model. As
also discussed in greater detail herein below, the model further
describes effects of actuations on the fluid moving devices
114a-114n as well as the settings of the adjustable vent tiles 118
on the airflow transport and distribution within the data center
100. In this regard, the model disclosed herein is a holistic
model. Moreover, the model is implemented to manage airflow
provisioning in the data center 100.
[0020] In one example, values obtained through implementation of
the model are used to partition the data center 100 into a
plurality of fluid moving device 114a-114n zones of influence with
varying levels of overlapping among the plurality of fluid moving
device 114a-114n zones. In another example, the obtained values are
used to simultaneously control the plurality of fluid moving
devices 114a-114n and the adjustable vent tiles 118 to manage
airflow provisioning in the data center 100. In a further example,
the obtained values are used in the minimization of a cost function
to simultaneously control the fluid moving devices 114a-114n and
the adjustable vent tiles 118, substantially in real time.
[0021] It should be understood that the data center 100 may include
additional elements and that some of the elements described herein
may be removed and/or modified without departing from a scope of
the data center 100. In addition, the data center 100 may comprise
a data center that is in a fixed location, such as a building,
and/or a data center that is in a movable structure, such as a
shipping container or other relatively large movable structure.
Moreover, although particular reference has been made in the
description of the area 100 as comprising a data center, it should
be understood that the area 100 may comprise other types of
structures, such as, a conventional room in building, an entire
building, etc.
[0022] Although the controller 130 is illustrated in FIG. 1 as
comprising an element separate from the electronic devices 116, the
controller 130 may comprise or be integrated with an electronic
device 116 without departing from a scope of the data center 100
disclosed herein. In addition, or alternatively, the controller 130
may comprise a set of machine readable instructions to operate on a
computing device, for instance, one of the electronic devices 116
or a different computing device. Moreover, although a single
controller 130 has been depicted in FIG. 1, a plurality of
controllers 130 may be implemented to respectively control
individual ones or groups of fluid moving devices 114a-114b and, in
further examples, individual ones or groups of AVTs 118a-118m.
[0023] Turning now to FIG. 2, there is shown a block diagram of a
system 200 for managing airflow provisioning in an area 100, such
as the data center depicted in FIG. 1, according to an example. It
should be understood that the system 200 may include additional
components and that some of the components described herein may be
removed and/or modified without departing from the scope of the
system 200. For instance, the system 200 may include any number of
sensors 120a-120n, memories, processors, fluid moving devices
114a-114l, AVTs 118a-118m, as well as other components, which may
be implemented in the operations of the system 200.
[0024] As shown, the system 200 includes the fluid moving devices
114a-114l, the AVTs 118a-118m, the sensors 120a-120n, the
controller 130, a data store 220, a processor 230, and a network
240. The controller 130 is further depicted as including an
input/output module 202, a data collection module 204, a model
accessing module 206, a parameter determining module 208, a
managing module 210, and an actuation module 212. According to an
example, the controller 130 comprises machine readable instructions
stored, for instance, in a volatile or non-volatile memory, such as
DRAM, EEPROM, MRAM, flash memory, floppy disk, a CD-ROM, a DVD-ROM,
or other optical or magnetic media, and the like. In this example,
the modules 202-212 comprise modules of machine readable
instructions stored in the memory, which are executable by the
processor 230. According to another example, the controller 130
comprises a hardware device, such as, a circuit or multiple
circuits arranged on a board. In this example, the modules 202-212
comprise circuit components or individual circuits, which the
processor 230 is to control. According to a further example, the
controller 130 comprises a combination of modules with machine
readable instructions and hardware modules.
[0025] In any regard, the processor 230 receives detected condition
information from the sensors 120a-120n over the network 240, which
operates to couple the various components of the system 200. The
network 240 generally represents a wired or wireless structure in
the data center 100 for the transmission of data and/or signals
between the various components of the system 200. In addition, the
processor 230 stores the detected condition information received
from the sensors 120a-120n in the data store 220, which may
comprise any suitable memory upon which the processor 230 may store
data and from which the processor 230 may retrieve data. The data
store 220 may comprise DRAM, EEPROM, MRAM, flash memory, floppy
disk, a CD-ROM, a DVD-ROM, or other optical or magnetic media, and
the like. Although the data store 220 has been depicted as forming
a separate component from the controller 130, it should be
understood that the data store 220 may be integrated with the
controller 130 without departing from a scope of the system
200.
[0026] According to an example, the controller 130 outputs the
determined operational settings of the fluid moving devices
114a-114l and, in some instances, the AVTs 118a-118m, such as but
not limited to volume flow rate set point(s), instructions
pertaining to the determined volume flow rate set point(s),
determined supply temperature set point(s), instructions pertaining
to the determined supply temperature set point(s), determined
operational settings and/or instructions pertaining to the
determined operational settings through the input/output module
202. Thus, for instance, the determined volume flow rate set
points, determined supply temperature set points, and the
determined operational settings may be outputted to a display upon
which the outputted information may be displayed, a printer upon
which the outputted information may be printed, a network
connection over which the outputted information may be conveyed to
another computing device, a data storage device upon which the
outputted information may be stored, etc. According to another
example, the controller 130 communicates instruction signals over
the network 240 to the fluid moving devices 114a-114l and/or the
AVTs 118a-118m. In this example, the fluid moving devices 114a-114l
may vary the volume flow rates and/or supply air temperatures of
the fluid moving devices 114a-114l to reach the determined set
points as instructed by the controller 130. According to another
example, the operational settings of the AVTs 118a-118m are varied
to cause the AVTs 118a-118m to have the operational settings as
instructed by the controller 130.
[0027] Various manners in which the modules 202-212 of the
controller 130 may operate are discussed with respect to the
methods 300-500 depicted in FIGS. 3-5. It should be readily
apparent that the methods 300-500 respectively depicted in FIGS.
3-5 represent generalized illustrations and that other elements may
be added or existing elements may be removed, modified or
rearranged without departing from the scopes of the methods
300-500.
[0028] With reference first to FIG. 3, there is shown a flow
diagram of a method 300 for managing airflow provisioning in an
area, such as, a data center 100, according to an example. At block
302, a model that describes airflow transport and distribution
within the area is accessed, for instance, by the model accessing
module 206. The model may be stored in the data store 220 and the
model accessing module 206 may access the model from the data store
220. The model comprises a plurality of parameters and describes
the effects of actuations on the plurality of fluid moving devices
114a-114l on the airflow transport and distribution within the
area. More particularly, the model describes the effects of
actuations on the plurality of fluid moving devices 114a-114l and
the AVTs 118a-118m on the transport and distribution of airflow
supplied into the electronic devices 116. In this regard, the model
disclosed herein is a holistic and efficient model because the
model takes as input both the zonal airflow provisioning actuation
of the fluid moving devices 114a-114l and the local airflow
provisioning actuation of the AVTs 118a-118m.
[0029] According to an example, the model is a state-space model
based on energy and mass balance principles. In a non-limiting
example, the model is a physics based state-space model. An example
of the physics based state-space model is described by the
following equation:
T ( k + 1 ) = T ( k ) + { i = 1 N CRAC g i [ SAT i ( k ) - T ( k )
] VFD i ( k ) } { j = 1 N tile b j U j ( k ) ] } + C , Eqn ( 1 )
##EQU00001##
[0030] in which T represents a rack inlet temperature, k and k+1
represent discrete time steps, SAT.sub.i and VFD.sub.i are a supply
air temperature and a blower speed of the ith fluid moving device
114a-114l, U.sub.j is the opening of the jth adjustable vent tile
118a-118m, N.sub.CRAC and N.sub.tile are the number of fluid moving
devices 114a-114l and adjustable vent tiles 118a-118m,
respectively, and wherein g.sub.i and b.sub.j are parameters that
capture influences of each fluid moving device i and adjustable
vent tile j, respectively, and C denotes a temperature change due
to reasons such as recirculation and reversed flow.
[0031] At block 304, values for the parameters in the model are
determined, for instance, by the parameter determining module 208.
Generally speaking, the parameter determining module 208 determines
the values for the parameters through an analysis of detected
condition data received from the sensors 120a-120n. More
particularly, the parameter determining module 208 determines
values for the parameters g.sub.i, b.sub.j and C in Eqn (1) through
an optimization process, in which the parameter values that
minimize the difference between the thermal status (rack inlet
temperatures) predicted by the model using the parameters (g.sub.i,
b.sub.j, and C) being evaluated and the detected conditions. The
parameters (g.sub.i, b.sub.j, and C) that result in the least
amount of difference between the thermal status (rack inlet
temperatures) predicted by the model are selected as the values for
the parameters (g.sub.i, b.sub.j, and C). This optimization process
is repeated for each rack inlet temperature since each rack inlet
temperature is characterized by a different set of parameters.
Alternatively, the parameter determining module 208 may implement
the parameter optimization process for a plurality of different
rack inlet temperatures in parallel.
[0032] At block 306, the model is implemented in managing airflow
provisioning in the data center 100, for instance, by the managing
module 210. Various examples of manners in which the model is
implemented at block 306 are described in greater detail with
respect to FIGS. 4 and 5. More particularly, FIGS. 4 and 5,
respectively, depict flow diagrams of methods 400 and 500 of
implementing the model in managing airflow provisioning, according
to two examples. As such, for instance, either or both of the
methods 400 and 500 may be implemented at block 306 in FIG. 3.
[0033] With reference first to FIG. 4, at block 402, influence
levels of the plurality of fluid moving devices 114a-114l on
temperatures at inlets of a plurality of racks 102a-102n are
determined, for instance, by the managing module 210. According to
an example, the influences of each of the fluid moving devices
114a-114l on the temperatures at the inlets of the racks are
captured by the parameter g, in Eqn (1). By way of example, in a
data center 100 having 8 fluid moving devices 114a-114l, each
detected rack inlet temperature will have 8 influence levels, each
representing the influence level of one fluid moving device
114a-114l.
[0034] At block 404, for each of the rack inlet temperatures,
ratios between each of the determined influence levels
corresponding to a particular fluid moving device 114a-114l and a
largest influence level of the determined influence levels for that
particular rack inlet temperature are calculated, for instance, by
the managing module 210. Thus, in the example above, a respective
ratio between each of the 8 influence levels and the largest
influence level for a particular rack inlet temperature are
calculated. As such, one of the ratios will be 1 because one of the
ratios will be between the largest influence level and itself and
the remaining ratios will be less than 1. In addition, block 404
may be repeated for each of the rack inlet temperatures to
determine the respective ratios of the influence levels
corresponding to each of the fluid moving devices 114a-114l.
[0035] At block 406, the data center 100 is partitioned into a
plurality of fluid moving device zones of influence, for instance,
by the managing module 210. The partitioning of the data center 100
includes identifying which of the rack inlet temperatures belong to
which of the fluid moving device 114a-114l zones of influence. In
other words, the partitioning of the data center 100 indicates
which of the fluid moving devices 114a-114l have more significant
influences on which of the rack inlet temperatures, and hence are
the fluid moving devices 114a-114l that are to respond to the
thermal status variation of the corresponding rack inlet
temperatures. According to an example, the partitioning is
performed based upon the ratios determined at block 404. More
particularly, for instance, a rack inlet temperature that has the
largest influence level to a first fluid moving device 114a is
considered to be within the zone of influence of the first fluid
moving device 114a. In addition, the rack inlet temperature is also
considered to be within the zone of influence of a second fluid
moving device 114b if the ratio between the influence level of the
second fluid moving device 114b and the influence level of the
first fluid moving device 114a exceeds an overlapping threshold. As
such, a higher overlapping threshold causes a relatively smaller
level of overlapping between the zones of influence to occur
because the rack inlet temperatures are likely to fall within a
fewer number of fluid moving device zones of influence. In
addition, a lower overlapping threshold causes a relatively larger
level of overlapping between the zones of influence to occur
because the rack inlet temperatures are likely to fall within a
larger number of fluid moving device zones of influence. In one
regard, therefore, the level of overlapping between the fluid
moving device 114a-114l zones of influence may substantially be
changed by varying the overlapping threshold.
[0036] Through use of the ratios between the influence levels and
the respective largest influence level for each of the rack inlet
temperatures instead of an absolute influence threshold to
determine the fluid moving device zones of influence, the
possibility of orphaned rack inlet temperatures during the
partitioning process (that is, those rack inlet temperatures that
do not belong to any fluid moving device zone of influence), may
substantially be reduced. In addition, by tuning the overlapping
threshold from 1 to 0, the partitioned zones may have the desired
level of overlapping, ranging from disjoint zones to 100%
overlapping between any two zones. In comparison, prior techniques
for partitioning data centers do not have this flexibility because
the overlapping is dependent on the absolute influence threshold,
which has a relatively narrow range. Moreover, disjoint zone
partitioning of a data center, for example, is impossible with the
prior techniques. This is because the absolute influence threshold
has to be sufficiently low to avoid orphaned rack inlet
temperatures, and this low threshold will inevitably result in
considerable overlapping between neighboring zones. Furthermore,
the zone partition approach disclosed herein may be used for
input-output pairing, which may be crucial for the development of
distributed data center cooling control systems. For centralized
data center cooling control design, the partition approach
disclosed herein may be used to trim weak connections between the
system inputs and outputs, which leads to more efficient controller
design. Moreover, the tuning feature disclosed herein may also be
used to adjust the level of redundancy in a data center based on
operational policies (e.g., by designating varying levels of
redundancy according to service level agreements, etc.).
[0037] In one regard, the fluid moving device zones of influence
may be implemented in determining which of the fluid moving devices
114a-114l is to be manipulated based upon, for instance, changing
conditions in the data center 100.
[0038] Turning now to FIG. 5, at block 502, a cost function is
accessed, for instance, by the managing module 210 from the data
store 220. According to an example, the cost function comprises the
total airflow provisioning power consumption and is defined with
respect to the airflow provisioning actuations available in the
data center 100. The available airflow provisioning actuations in
the data center 100 comprise temperature and volume flow rate of
airflow supplied by the fluid moving devices 114a-114l as well as
the openings in the AVTs 118a-118m.
[0039] At block 504, zonal and local airflow provisioning actuation
are coordinated through use of the model to minimize the cost
function, for instance, by the managing module 210. More
particularly, for instance, a model predictive controller (MPC),
shown in FIG. 6, uses the model to optimize the zonal and local
airflow provisioning by minimizing the cost function. According to
an example, given the current thermal status (rack inlet
temperatures), the MPC may implement the model to predict future
rack inlet temperature trajectories when the trajectories of the
airflow actuations (VFD and SAT of the fluid moving devices
114a-114l and openings of the AVTs 118a-118m) are given. The
prediction of the future rack inlet temperature trajectories may be
used to evaluate all of the possible zonal and local actuations
implemented at discrete time steps with the updated current thermal
status, and thus, thermal anomalies may be handled, and operating
cost may constantly be minimized in response to varying conditions
within the data center 100.
[0040] Additionally at block 504, the model is implemented to
minimize the cost function while substantially maintaining
temperature levels at the rack inlets within predetermined ranges.
Through implementation of the model, which considers both the zonal
and local airflow provisioning actuations for purposes of
minimizing the cost function, fighting among the various zonal and
local airflow provisioning actuations is substantially eliminated.
The following equation describes an example in which the cost
function is the total cooling power:
.SIGMA..sub.i=1.sup.N.sup.CRAC.left
brkt-bot.VFD.sub.i.sup.3(k).sub.R.sub.VFD+(-SAT.sub.i(k)).sub.R.sub.SAT.l-
eft brkt-bot., Eqn (2):
in which the cooling power incurred by all of the fluid moving
devices 114a-114l are summed up, and for fluid moving device i, the
blower power (VFD) increases with the third power of blower speed
(VFD.sub.i) while the chiller power decreases linearly with the
supply air temperature SAT.sub.i.
[0041] At block 506, the coordinated zonal and local airflow
provisioning actuation that minimizes the cost function accessed at
block 502 is outputted, for instance, by the input/output module
202. According to an example, the settings for the VFDs and supply
air temperatures in the fluid moving devices 114a-114l and the
openings of the AVTs 118a-118m that have been determined to
minimize the cost function while meeting temperature requirements
of the electronic devices 116 are outputted. In one example, the
settings are outputted to a display, another computing device, a
data storage, a printer, etc. In another example, the actuation
module 212 outputs control signals to the fluid moving devices
114a-114l and/or the AVTs 118a-118m to cause the fluid moving
devices 114a-114l and/or the AVTs 118a-118m to operate at the
determined settings.
[0042] An example of a control diagram 600 that includes the MPC
602 that implements the model disclosed herein is depicted in FIG.
6. As shown therein, the MPC 602, which comprises the model and an
optimization module (not shown), receives as inputs, a cost
function that the optimization module runs to minimize by selecting
the most appropriate airflow actuations, a threshold temperature (
T.sub.ref) as the constraint of the optimization that future rack
inlet temperatures must stay below, and rack inlet temperatures (
T), for future rack inlet temperature prediction using the model.
In other words, the MPC 602 seeks to determine the optimal zonal
and local airflow provisioning settings in the data center 604
represented by the SAT, VFD, and AVT depicted in FIG. 6, in
response to dynamic IT workload. The airflow resources
provisioning, transport, and distribution are coordinated because
they are considered simultaneously in the same framework to
minimize the airflow provisioning power.
[0043] Some or all of the operations set forth in the methods
300-500 may be contained as utilities, programs, or subprograms, in
any desired computer accessible medium. In addition, the methods
300-500 may be embodied by computer programs, which can exist in a
variety of forms both active and inactive. For example, they may
exist as software program(s) comprised of program instructions in
source code, object code, executable code or other formats. Any of
the above may be embodied on a computer readable storage
medium.
[0044] Example computer readable storage media include conventional
computer system RAM, ROM, EPROM, EEPROM, and magnetic or optical
disks or tapes. Concrete examples of the foregoing include
distribution of the programs on a CD ROM or via Internet download.
It is therefore to be understood that any electronic device capable
of executing the above-described functions may perform those
functions enumerated above.
[0045] Turning now to FIG. 7, there is shown a block diagram of a
computing device 700 to implement the methods depicted in FIGS.
3-5, in accordance with examples of the present disclosure. The
device 700 includes a processor 702, such as a central processing
unit; a display device 704, such as a monitor; a network interface
708, such as a Local Area Network LAN, a wireless 802.11x LAN, a 3G
mobile WAN or a WiMax WAN; and a computer-readable medium 710. Each
of these components is operatively coupled to a bus 712. For
example, the bus 712 may be an EISA, a PCI, a USB, a FireWire, a
NuBus, or a PDS.
[0046] The computer readable medium 710 may be any suitable
non-transitory medium that participates in providing instructions
to the processor 702 for execution. For example, the computer
readable medium 710 may be non-volatile media, such as an optical
or a magnetic disk; volatile media, such as memory; and
transmission media, such as coaxial cables, copper wire, and fiber
optics.
[0047] The computer-readable medium 710 may also store an operating
system 714, such as Mac OS, MS Windows, Unix, or Linux; network
applications 716; and an airflow provisioning management
application 718. The operating system 714 may be multi-user,
multiprocessing, multitasking, multithreading, real-time and the
like. The operating system 714 may also perform basic tasks such as
recognizing input from input devices, such as a keyboard or a
keypad; sending output to the display 704; keeping track of files
and directories on the computer readable medium 710; controlling
peripheral devices, such as disk drives, printers, image capture
device; and managing traffic on the bus 712. The network
applications 716 include various components for establishing and
maintaining network connections, such as machine readable
instructions for implementing communication protocols including
TCP/IP, HTTP, Ethernet, USB, and FireWire.
[0048] The airflow provisioning management application 718 provides
various components for managing airflow provisioning in a data
center 100, as described above. The management application 718 may
thus comprise controller 130. The management application 718 also
includes modules for accessing a model that describes airflow
transport and distribution within the area, the model comprising a
plurality of parameters, determining values for the plurality of
parameters, and implementing the model in managing airflow
provisioning in the data center. In certain examples, some or all
of the processes performed by the application 718 may be integrated
into the operating system 714. In certain examples, the processes
may be at least partially implemented in digital electronic
circuitry, or in computer hardware, machine readable instructions
(including firmware and/or software), or in any combination
thereof.
[0049] Although described specifically throughout the entirety of
the instant disclosure, representative examples of the present
disclosure have utility over a wide range of applications, and the
above discussion is not intended and should not be construed to be
limiting, but is offered as an illustrative discussion of aspects
of the disclosure.
[0050] What has been described and illustrated herein is an example
of the disclosure along with some of its variations. The terms,
descriptions and figures used herein are set forth by way of
illustration only and are not meant as limitations. Many variations
are possible within the spirit and scope of the disclosure, which
is intended to be defined by the following claims--and their
equivalents--in which all terms are meant in their broadest
reasonable sense unless otherwise indicated.
* * * * *