Dynamic environmental management (09-Mar-2010)

PRIORITY

This application is a continuation-in-part of the following U.S. patent Application Publication and patent Applications: U.S. Patent Application Publication No. 20060206291, entitled “COMMISIONING OF SENSORS,” filed on Mar. 11, 2005 and published on Sep. 14, 2006, now U.S. Pat. No. 7,117,129; U.S. Patent Application Publication No. 20060214014, entitled “TEMPERATURE CONTROL USING A SENSOR NETWORK,” filed on Mar. 25, 2005 and published on Sep. 28, 2006; and U.S. patent application Ser. No. 11/699,402, entitled “COMPUTERIZED TOOL FOR ASSESSIGN CONDITIONS IN A ROOM,” filed on Jan. 29, 2007; all of which are herein incorporated by reference in their entireties.

BACKGROUND

A data center may be defined as a physical location, for example, a room that houses one or more components, such as computer systems, that are capable of generating heat. The computer systems may be arranged in a number of racks. These racks are configured to house a number of computer systems which typically include a number of printed circuit boards (PCBs), mass storage devices, power supplies, processors, micro-controllers, and semi-conductor devices, that dissipate relatively significant amounts of heat during their operation.

Increases in system-level compaction of data centers have resulted in increases of server-level and rack-level power densities and dissipations that place significant pressure on traditional data center thermal management systems. Conventional data center thermal management involves traditional systems that use computer room air conditioning (CRAC) units to pressurize a raised floor plenum of a data center with cool air that is passed to equipment racks via ventilation tiles distributed throughout the raised floor. Temperature is typically monitored and controlled based on a single sensory feedback signal, which acts as a global indication of the heat being dissipated in the data center, at the hot air return of the CRAC units away from the equipment racks. This conventional mode of operation allows no local flexibility in how the cooling is delivered to the servers or computers in the data center, and there is no local state feedback information from different areas of the data center. Consequently, due primarily to a lack of distributed environmental sensing, traditional thermal management systems often operate conservatively with reduced computational density, added operational expense, and unnecessary redundancy due to poor utilization.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments are illustrated by way of example and not limited in the following figure(s), in which like numerals indicate like elements, in which:

FIG. 1A shows a simplified perspective view of a data center, according to an embodiment of the invention;

FIG. 1B is a simplified plan view of the data center depicted in FIG. 1A;

FIG. 2 is a block diagram of a sensor commissioning system according to an embodiment of the invention;

FIG. 3 illustrates a flow diagram of an operational mode of a method for commissioning sensors, according to an embodiment of the invention;

FIG. 4A illustrates a flow diagram of an operational mode of a method for commissioning temperature sensors, according to an embodiment of the invention;

FIG. 4B illustrates a flow diagram of an operational mode for assigning sensors to respective actuator families based upon a neural network, according to an embodiment of the invention;

FIG. 4C illustrates a flow diagram of an operational mode for assigning sensors to respective actuator families based upon a curve fitting algorithm, according to an embodiment of the invention;

FIG. 5 shows a example of a neural network created through implementation of the operational mode depicted in FIG. 4B, according to an embodiment of the invention;

FIG. 6 illustrates a flow diagram of an operational mode for optionally filtering sensors belonging to multiple actuator families, according to an embodiment of the invention;

FIG. 7 illustrates a flow diagram of a method for controlling temperature using a sensor network, according to an embodiment of the invention;

FIG. 8 illustrates a flow diagram of a method for controlling environmental conditions at sensor locations, according to an embodiment of the invention;

FIG. 9 shows a block diagram of a computerized tool configured to assess and visualize thermal performance in a room, according to an embodiment of the invention;

FIG. 10 shows a flow diagram of a method for assessing a thermal profile with respect to equipment housed in a room, according to an embodiment of the invention;

FIG. 11A illustrates an example of a room layout overlaid with a thermal profile, according to an embodiment of the invention;

FIG. 11B illustrates an example of a room layout overlaid with multiple thermal profiles, according to an embodiment of the invention;

FIG. 12 illustrates an example of a room layout overlaid with a thermal profile, according to another embodiment of the invention;

FIGS. 13A-13E, collectively, show a flow diagram of a method for analyzing and visualizing a thermal profile of a room, according to an embodiment of the invention;

FIG. 14 illustrates a block diagram of a controller as an application server operable with software therein to implement an EM architecture for a data center, according to an embodiment of the invention.

FIG. 15 illustrates a computer system, which may be employed to perform the various functions of the controller in FIG. 1.

DETAILED DESCRIPTION

For simplicity and illustrative purposes, the principles of the embodiments are described by referring mainly to examples thereof. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the embodiments. It will be apparent however, to one of ordinary skill in the art, that the embodiments may be practiced without limitation to these specific details. In other instances, well known methods and structures have not been described in detail so as not to unnecessarily obscure the embodiments.

Described herein are systems and methods for an environmental management (EM) architecture for a physical setting or environment, such as a data center. The EM architecture includes a set of one or more management software programs or modules operable to interact with a distributed sensor network and actuators in the data center, to exert control over the actuators based on sensor measurements obtained from the distributed sensor network, and to provide analysis and visualization of thermal profiles in the data center.

In one embodiment, the actuators may be controlled according to a control scheme designed to enable efficient energy utilization by the data center while satisfying one or more user-defined criteria (e.g., thermal management criteria) for environmental management of the data center. An example of a control scheme is disclosed in a co-pending and commonly assigned U.S. patent application as identified from its U.S. Patent Application Publication No. 20060214014, entitled “TEMPERATURE CONTROL USING A SENSOR NETWORK.” Additionally, the control scheme for the actuators may be based on correlations developed between the actuators and the sensors of the distributed sensor network. The correlations may be developed through a process for commissioning the sensors with respect to the actuators. In addition, the correlations generally provide indications of how the sensors of the sensor network may be affected by variations in the outputs of the actuators. An example of a commissioning process suitable for correlating the actuators and the sensors is disclosed in a co-pending and commonly assigned U.S. patent application as identified from its U.S. Patent Application Publication No. 20060206291, entitled “COMMISIONING OF SENSORS”, now U.S. Pat. No. 7,117,129. Also, visualization of the thermal profiles in the data center may be provided to the users, who can then employ such visualization to optimize placement of equipment in the data center for desired or better environmental management. An example of a computerized tool suitable for the aforementioned visualization is disclosed in co-pending and commonly assigned U.S. patent application Ser. No. 11/699,402, entitled “COMPUTERIZED TOOL FOR ASSESSIGN CONDITIONS IN A ROOM.”

In one example, the actuators include computer room air conditioning (CRAC) units capable of varying one or both of volume flow rate and temperature of airflow supplied to sensors in a data center. In this example, the determination of which CRAC units to manipulate, for example, to maintain a particular sensor below a predetermined maximum temperature, is based upon the correlations determined between the CRAC units and the sensors. In addition, the CRAC units may be manipulated according to a selected control scheme as described above.

The systems and methods for a data center EM architecture disclosed herein may also be employed in any reasonably suitable environment containing actuators and sensors, such as, a building containing air conditioning units and sensors. In this regard, although particular reference is made throughout the present disclosure to data centers and CRAC units, it should be understood that the systems and methods disclosed herein may be implemented in other environments. In addition, therefore, the particular references to data centers and CRAC units are for illustrative purposes and are not intended to limit the systems and methods disclosed herein solely to data centers and CRAC units.

With reference first to FIG. 1A, there is shown a simplified perspective view of a section of a data center 100 which may employ various examples of the environmental control system disclosed herein. The terms “data center” are generally meant to denote a room or other space where one or more components capable of generating heat may be situated. In this respect, the terms “data center” are not meant to limit the invention to any specific type of room where data is communicated or processed, nor should it be construed that use of the terms “data center” limits the invention in any respect other than its definition herein above.

The data center 100 is depicted as having a plurality of racks 102a-102n, where “n” is an integer greater than one. The racks 102a-102n may comprise electronics cabinets, aligned in parallel rows and positioned on a raised floor 110. A plurality of wires and communication lines (not shown) may be located in a space 112 beneath the raised floor 110. The space 112 may also function as a plenum for delivery of cooled air from one or more actuators.

Also shown in FIG. 1A are computer room air conditioning (CRAC) units 114a-114n, where “n” is an integer greater than one, which are considered herein as primary actuators 114a-114n. The CRAC units 114a-114n are considered primary actuators 114a-114n because they are configured to manipulate a characteristic of the cooled airflow supplied to the racks 102a-102n through actuation of one or more secondary actuators. The secondary actuators include a device for controlling airflow temperature and a device for controlling the supply flow rates of the cooled air.

The cooled air may be delivered from the space 112 to the racks 102a-102n through vent tiles 118 located between some or all of the racks 102a-102n. The vent tiles 118 are shown as being located between rows 102 and 104 and 106 and 108. The cooled air contained in the space 112 may include cooled air supplied by one or more primary actuators 114a-114n. Thus, characteristics of the cooled air, such as, temperature, pressure, humidity, flow rate, etc., may substantially be affected by the operations of one or more of the primary actuators 114a-114n. In this regard, characteristics of the cooled air at various areas in the space 112 and the cooled air supplied to the racks 102a-102n may vary, for example, due to mixing of the cooled air. In other words, the characteristics of the cooled air supplied to a particular location in the data center 100 may differ from that of the cooled air supplied by a single primary actuator 114a.

At least one condition, for example, temperature, pressure, or humidity, of the cooled air supplied to various areas of the data center 100 may be detected by a distributed sensor network having sensors 120a-120n, where “n” is an integer greater than one. As shown, the sensors 120a-120n are represented as diamonds to distinguish them from other elements depicted in FIG. 1A. In addition, the sensors 120a-120n are depicted as being positioned to detect the at least one condition at the inlets of the racks 102a-102n. In this example, the sensors 120a-120n may environmental sensors such as temperature sensors or humidity sensors. In another example, the sensors 120a-120n may be positioned within the space 112 near respective vent tiles 118 to detect the temperature, pressure, or humidity of the cooled air supplied through the respective vent tiles 118. Thus, although the sensors 120a-120n are depicted as being located on the raised floor 110, the sensors 120a-102n may be positioned at various other reasonably suitable locations, including, for example, near or within some or all of the components 116.

In any regard, the sensors 120a-120n may be employed to detect the at least one condition at various primary actuator 114a-114n settings. In addition, the sensors 120a-120n may be assigned to the families of one or more primary actuators 114a-114n. A primary actuator 114a-114n “family” may be defined as a grouping of sensors 120a-120n that respond to the various primary actuator 114a-114n settings to levels greater than a predefined threshold level. In other words, the sensor 120a may be considered as being in the primary actuator 114a family if the response of the sensor 120a exceeds a predefined threshold level at various primary actuator 114a-114n settings. Various manners in which the sensors 120a-120n may be assigned to the one or more primary actuator 114a-114n families is described in greater detail below and also in the co-pending and commonly assigned U.S. Patent Application Publication No. 20060206291, entitled “COMMISIONING OF SENSORS.”

The vent tiles 118 may comprise manually or remotely adjustable vent tiles. In this regard, the vent tiles 118 may be manipulated to vary, for example, the mass flow rates of cooled air supplied to the racks 102a-102n.

The racks 102a-102n are generally configured to house a plurality of components 116 capable of generating/dissipating heat, for example, processors, micro-controllers, high-speed video cards, memories, semi-conductor devices, and the like. The components 116 may be elements of a plurality of subsystems (not shown), for example, computers, servers, bladed servers, etc. The subsystems and the components may be operated to perform various electronic functions, for example, computing, switching, routing, displaying, and the like.

The areas between the rows 102 and 104 and between the rows 106 and 108 may comprise cool aisles 122. These aisles are considered “cool aisles” because they are configured to receive cooled airflow from the vent tiles 118, as generally indicated by the arrows 124. In addition, and as shown, the racks 102a-102n generally receive cooled air from the cool aisles 122. The aisles between the rows 104 and 106, and on the rear sides of rows 102 and 108, are considered hot aisles 126. These aisles are considered “hot aisles” because they are positioned to receive air that has been heated by the components 116 in the racks 102a-102n, as indicated by the arrows 128. It should be understood that the designations “cool aisles” and “hot aisles” are merely descriptive of the particular arrangement of the data center 100, and alternative embodiments are contemplated wherein the data center 100 may be arranged such that there are no separate “cool aisles” or “hot aisles” without deviating from the scope of the present disclosure.

The sides of the racks 102a-102n that face the cool aisles 122 may be considered as the fronts of the racks 102a-102n and the sides of the racks 102a-102n that face away from the cool aisles 122 may be considered as the rears of the racks 102a-102n. For purposes of simplicity and not of limitation, this nomenclature will be relied upon throughout the present disclosure to describe the various sides of the racks 102a-102n. Although not shown, the racks 102a-102n may be positioned with their rear sides adjacent to one another (wherein there is no longer any hot aisle). In this example, the vent tiles 118 may be provided in each aisle 122 and 126. In addition, the racks 102a-102n may comprise outlets on top panels thereof to enable heated air to flow out of the racks 102a-102n.

As described herein above, the primary actuators 114a-114n generally operate to cool received heated air as indicated by the arrows 128. In addition, the primary actuators 114a-114n may supply the racks 102a-102n with airflow that has been cooled and humidity-controlled (at rack inlets), through any reasonably suitable known manners and may thus comprise widely available, conventional CRAC units. For example, the primary actuators 114a-114n may comprise vapor-compression type air conditioning units, chilled water air conditioning units, etc.

Also shown in FIG. 1A is a controller 130 configured to perform various functions in the data center 100. The controller 130 may receive data from the primary actuators 114a-114n and the sensors 120a-120n and may perform various computations on the data. In one regard, the controller 130 may operate to assign the sensors 120a-120n into one or more primary actuator 114a-114n families. According to an example, the controller 130 may implement the commissioning procedures noted earlier to assign the sensors 120a-120n into the one or more primary actuator 114a-114n families.

The controller 130 may also operate the primary actuators 114a-114n based upon the correlations between the primary actuators 114a-114n and the sensors 120a-120n. In operating the primary actuators 114a-114n, the controller 130 may select and implement one or more control schemes as described in greater detail herein below.

Although the controller 130 is illustrated in FIG. 1A as comprising a component separate from the components 116 housed in the racks 102-108, the controller 130 may comprise one or more of the components 116 without departing from a scope of the data center 100 disclosed herein. In addition, or alternatively, the controller 130 may comprise software configured to operate on a computing device, for example, one of the components 116.

With reference now to FIG. 1B, there is shown a simplified plan view of the data center 100 depicted in FIG. 1A. The data center 100 is shown as including primary actuators 114a-114n positioned at various locations throughout the data center 100. A plurality of vent tiles 118 are also illustrated in FIG. 1B and are configured to deliver cooled airflow to racks 102a-102n as described above. It should be appreciated that the data center 100 may include any reasonably suitable number of racks 102a-102n and primary actuators 114a-114n without departing from the data center 100 illustrated in FIG. 1B.

As described herein above, the vent tiles 118 and the racks 102a-102n are positioned on a raised floor 110, beneath which lies a space 112 (FIG. 1A). The space 112 is in fluid communication with the primary actuators 114a-114n and generally operates, in one respect, as a plenum for supplying cooling airflow from the primary actuators 114a-114n to be delivered through the vent tiles 118. In most instances, the space 112 may comprise a relatively open space that is accessible by cooling airflow supplied by a plurality of the primary actuators 114a-114n. In this regard, the cooling airflow supplied by the primary actuators 114a-114n may mix in the space 112. Therefore, the cooling airflow supplied to the racks 102a-102n by the vent tiles 118 may have originated from more than one of the primary actuators 114a-114n.

Although particular reference is made throughout the present disclosure to a raised floor 110, it should be understood that various other types of cooling arrangements may be employed without departing from the systems and methods disclosed herein. For example, the data center 100 may include a lowered ceiling which may also include a space that is operable as a plenum. In addition, or alternatively, the data center 100 may include ceiling mounted heat exchangers, in-row coolers, rack-mounted coolers, or any other cooling units that are operable to provide cool air to a shared environment.

Also shown in FIG. 1B are the sensors 120a-120n, which are illustrated as being positioned with respect to each of the racks 102a-102n. As also stated above, the sensors 120a-120n may also, or in the alternative, be positioned to detect the at least one condition within the space 112. In addition, the sensors 120a-120n may comprise sensors contained in some or all of the components 116. As a further example, the sensors 120a-120n may be positioned near or within the primary actuators 114a-114n. In any regard, the sensors 120a-120n may be grouped in various primary actuator 114a-114n families based upon various criteria. The various primary actuator 114a-114n families 132a-132n corresponding to respective primary actuators 114a-114n are illustrated in FIG. 1B. As shown, the sensors 120a-120n are considered as being within the families 132a-132n of those primary actuators 114a-114n.

Some of the sensors 120a-120n, for example, the sensors 120a-120n in a first section 134a may be included in the family 132a of a single primary actuator 114a. Some of the other sensors 120a-120n, for example, the sensors 120a-120n in a second section 134b may be included in the families 132a and 132b of two primary actuators 114a and 114b. In addition, some of the sensors 120a-120n, for example, the sensors 120a-120n in a third section 134c may be included in the families 132a-132c of three primary actuators 114a-114c. As such, for example, one or more of the sensors 120a-120n may belong to more than one primary actuator 114a-114n family.

It should, in any regard, be understood that the families 132a-132n depicted in FIG. 1B are for purposes of illustration and are not intended to limit the data center 100 and its components in any respect. It should also be understood that the depiction of the families 132a-132n in FIG. 1B are for illustrative purposes only and are not meant to limit the data center 100 in any respect.

FIG. 2 illustrates an environmental control system 202 operable to provide environmental management of the data center 100. It should be understood that the following description of the environmental control system 202 is but one manner of a variety of different manners in which such a system may be configured. In addition, it should be understood that the environmental control system 202 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 environmental control system 202. For example, the environmental control system 202 may include any number of sensors, memories, processors, CRAC units, etc., as well as other components, which may be implemented in the operations of the environmental control system 202.

Generally speaking, the environmental control system 202 is employed to control the primary actuators 114a-114n, to thereby control at least one environmental condition at the sensor locations 120a-120n. A determination of primary actuator 114a-114n families may be used to determine which of the primary actuators 114a-114n are to be manipulated in response to conditions detected by the sensors 120a-120n. The control over the primary actuators 114a-114n may be effectuated through manipulation of one or more secondary actuators 222 and 224. One of the secondary actuators 222 may comprise a variable frequency drive (VFD) for controlling an airflow volume varying device, such as a blower or fan. The other secondary actuator 224 may comprise a device for controlling the temperature of the cooled air supplied by the primary actuators 114a-114n. such as water chillers, compressors, valves, etc. Thus, the secondary actuator 224 may depend upon the type of primary actuator 114a-114n in which the secondary actuator 224 is located. For example, if a primary actuator 114 comprises a vapor-compression type CRAC unit, a secondary actuator 224 therein may comprise a variable speed compressor configured to vary the temperature of the airflow supplied by the CRAC unit. Likewise, if a primary actuator 114 comprises a chilled-water (or any other refrigerant type) CRAC unit, a secondary actuator 224 therein may comprise a two or three-way valve configured to control the temperature of a coolant configured to receive heat from the airflow.

The secondary actuators 224 may also comprise devices for varying other characteristics of the airflow supplied by the primary actuators 114a-114n. The secondary actuators 224 may comprise, for example, humidifiers/dehumidifiers configured to vary the humidity of the airflow supplied by the primary actuators 114a-114n. In this example, the sensors 120a-120n may comprise humidity sensors. In addition, therefore, the primary actuators 114a-114n in this example may operate to maintain the humidity levels at the sensor 120a-120n locations within predefined thresholds.

In any respect, the controller 130 is configured to control the primary actuators 114a-114n and thus the secondary actuators 222 and 224. Instructions from the controller 130 may be transmitted over a network 221 that operates to couple the various components of the environmental control system 202. Although not shown, the controller 130 may be equipped with or have access to software and/or hardware to enable the controller 130 to transmit and receive data over the network 221. The network 221 generally represents a wired or wireless structure in the data center 100 for the transmission of data between the various components of the environmental control system 202. The network 221 may comprise an existing network infrastructure or it may comprise a separate network configuration installed for the purpose of environmental control by the controller 130.

The sensors 120a-120n may be configured to transmit collected data over the network 221 for storage and processing. As stated above, the sensors 120a-120n may comprise sensors configured to detect at least one environmental condition at various locations in the data center 100. The at least one environmental condition may comprise temperature, humidity, or pressure and the sensors 120a-120n may be configured to detect at least one of these conditions. The sensors 120a-120n may also be configured to compare detected environmental conditions with predefined environmental conditions to determine differences between the detected environmental conditions and the predefined environmental conditions. The sensors 120a-120n may transmit these differences as signals to the controller 130, where the strengths of the signals correspond to the difference levels. In addition, the controller 130 may vary operations of the actuator control module 218 according to the types of environmental condition detected and the magnitudes of the signals received from the sensors 120a-120n.

Commissioning of Sensors

The commissioning of the sensors 120a-n with respect to the primary actuators 114a-n, as disclosed in a co-pending and commonly assigned U.S. patent application as identified from its U.S. Patent Application Publication No. 20060206291, entitled “COMMISIONING OF SENSORS”, now U.S. Pat. No. 7,117,129, is described below with reference to FIGS. 3-6.

FIG. 3 illustrates a flow diagram of an operational mode 300 of a method for commissioning sensors, according to an example. For illustrative purposes only and not to be limiting thereof, the operation mode 300 is discussed in the context of the block diagram 200 illustrated in FIG. 2.

The operational mode 300 may be implemented to commission the sensors 120a-120n with respect to a plurality of actuators, for example, CRAC units 114a-114n. More particularly, the operational mode 300 may be implemented to relate the sensors 120a-120n to the actuators. In addition, those sensors 120a-120n that are influenced to a predefined level by a particular actuator are considered to be within that actuator's family.

In the operational mode 300, the controller 130 may determine correlations between the sensors 120a-120n and a plurality of actuators at step 302. Manners in which these correlations may be determined are described in greater detail herein below with respect to the operational modes 400 (FIG. 4A), 600 (FIG. 6), and 700 (FIG. 7). The controller 130 may also calculate correlation indexes of the sensors 120a-120n, which are functions of the plurality of actuator settings and a particular actuator, from the correlations at step 304. Examples of how the correlation indexes of the sensors 120a-120n may be calculated are described in greater detail herein below with respect to the operational modes 450 (FIG. 4B) and 470 (FIG. 4C). In addition, the controller 130 may assign each of the sensors 120a-120n to at least one actuator family at step 306. Again, a more detailed description of this step is provided below with respect to the operational modes 450 (FIG. 4B) and 470 (FIG. 4C).

With particular reference now to FIG. 4A, there is shown a flow diagram of an operational mode 400 of a method for commissioning sensors, according to an example. It is to be understood that the following description of the operational mode 400 is but one manner of a variety of different manners in which an embodiment of the invention may be practiced. It should also be apparent to those of ordinary skill in the art that the operational mode 400 represents a generalized illustration and that other steps may be added or existing steps may be removed, modified or rearranged without departing from a scope of the operational mode 400.

The operational mode 400 may be implemented to commission the sensors 120a-120n in a data center 100. More particularly, the operational mode 400 may be implemented to relate the sensors 120a-120n to the CRAC units 114a-114n. In addition, those sensors 120a-120n that are influenced to a predefined level by a particular CRAC unit 114a-114n are considered to be within that CRAC unit's 114a-114n family.

The operational mode 400 may be initiated at step 402 in response to any of a number of stimuli or conditions. For example, the operational mode 400 may be initiated with activation of the components in the data center 100, such as, the CRAC units 114a-114n. In addition, or alternatively, the operational mode 400 may be manually initiated or the controller 130 may be programmed to initiate the operational mode 400 at various times, for a set duration of time, substantially continuously, etc.

Once initiated, the controller 130 may set the CRAC units 114a-114n to a first distribution level at step 404. The first distribution level may comprise a first flow rate (VFD setting) and temperature of the airflow supplied by the CRAC units 114a-114n, which are common for the CRAC units 114a-114n. In addition, the controller 130 may wait for a period of time at step 406, prior to recording temperature information received from the sensors 120a-120n at step 408. The controller 130 may allow this time period to elapse in order to enable a relatively steady-state of operation to be reached. The time period may be based upon, for example, the loading on the CRAC units 114a-114n. In addition, during the time period at step 406, the controller 130 may determine a median temperature reading for one or more of the sensors 120a-120n in the event that the temperatures detected by one or more of the sensors 120a-120n oscillate during the time period. In this case, the temperature measurements recorded at step 408 may comprise time-averaged values.

Following elapse of the time period at step 406, the controller 130 may record the temperature measurements obtained by the sensors 120a-120n, as indicated at step 408. As stated above, the temperature information may be stored through implementation of the data storage module 216. The controller 130 may instruct a CRAC unit 114a to change the temperature of the airflow by a specified amount (N°) at step 410. The specified amount (N°) may comprise an amount that differs from the first distribution level temperature by a relatively discernable amount. Thus, for example, the specified amount (N°) may range from, for example, ±1° C. to 20° C. or more. The CRAC unit 114a may change the temperature of the airflow by the specified amount (N°) through, for example, varying operations of the actuator B 224, which may comprise a compressor, a chiller, a valve, etc.

The controller 130 may again wait for a period of time at step 412, prior to recording temperature information received from the sensors 120a-120n at step 414. The controller 130 may allow this time period to elapse in order to enable a relatively steady-state of operation to be reached following the supply air temperature change in the CRAC unit 114a. Following elapse of the time period at step 412, the controller 130 may again record the temperature measurements obtained by the sensors 120a-120n, as indicated at step 414. In addition, during the time period at step 412, the controller 130 may determine a median temperature reading for one or more of the sensors 120a-120n in the event that the temperatures detected by one or more of the sensors 120a-120n oscillate during the time period. In this case, the temperature measurements recorded at step 412 may comprise time-averaged values.

At step 416, the controller 130 may calculate a sensor-to-actuator correlation coefficient (C_i) for the sensors 120a-120n (i). The actuators are the CRAC units 114a-114n. As such, the correlation coefficient (C_i) is a function of the relative level of influence the CRAC units 114a-114n have over the sensors 120a-120n. Thus, for example, the higher the correlation coefficient (C_i) value for a sensor 120a-120n, the greater the influence a CRAC unit 114a-114n has over that sensor 120a-120n. In addition, the calculated correlation coefficients (C_i) for the CRAC units 114a-114n and the sensors 120a-120n may be stored in the memory 204.

Although the correlation coefficients (C_i) may be determined through any number of suitable correlation algorithms, the following algorithm may be employed to calculate the correlation coefficients (C_i) of the sensors 120a-120n.

C i = ( T 1 - T 2 ) N . Equation ⁢ ⁢ ( 1 )
In this equation, T₁ is the temperature measurement recorded at step 408 and T₂ is the temperature measurement recorded at step 414. In addition, N is the specified amount of supply air temperature change for the CRAC unit 114a at step 410.

By way of example, if the temperature of the sensor 120a recorded at step 408 (T₁) is 20° C., the temperature of the sensor 120a recorded at step 414 (T₂) is 25° C., and the change in temperature of the supply air (N) is +10° C., the correlation coefficient (C_i) between the sensor 120a and the CRAC unit 114a is 0.5. As another example, if the temperature of the sensor 120b recorded at step 408 (T₁) is 20° C., the temperature of the sensor 120b recorded at step 414 (T₂) is 21° C., and the change in temperature of the supply air (N) is +10° C., the correlation coefficient (C_i) between the sensor 120b and the CRAC unit 114a is 0.10. As such, the sensor 120a has a greater correlation to the CRAC unit 114a. Thus, changes to the supplied airflow from the CRAC unit 114a are likely to have a greater impact on conditions at the sensor 120a as compared with the sensor 120b.

At step 418, the temperature of the CRAC unit 114a may be reset to the temperature at the first distribution level set at step 404. In addition, it may be determined whether correlations between other CRAC units 114b-114n and the sensors 120a-120b are to be made at step 420. If it is determined that additional correlations are to be determined, the temperature of the airflow supplied by a next CRAC unit 114b may be varied by the specified amount (N°) at step 422. The temperature of the airflow supplied by the CRAC unit 114b may be varied in manners as described herein above with respect to step 410.

Following step 422, the controller 130 may again wait for a period of time at step 412, prior to recording temperature information received from the sensors 120a-120n at step 414, as described above. In addition, the controller 130 may calculate a sensor-to-actuator correlation coefficient (C_i) for the sensors 120a-120n (i) and the CRAC unit 114b at step 416, as also described above. Moreover, the temperature of the CRAC unit 114b may be reset to the temperature at the first distribution level set at step 404.

Steps 412-422 may be repeated for the remaining CRAC units 114c-114n. In this regard, the correlations between all of the CRAC units 114a-114n and the sensors 120a-120n may be determined and recorded. If there are no further CRAC units 114a-114n for which correlations to the sensors 120a-120n are to be determined, it may be determined as to whether correlations are to be determined at an additional distribution level at step 424. If “yes”, the CRAC units 114a-114n may be set to a next distribution level at step 426. The next distribution level may comprise characteristics that differ from the first distribution level. As such, either or both of the flow rate and the temperature of the cooled airflow supplied by the CRAC units 114a-114n may differ from their settings in the first distribution level.

As shown, following step 426, steps 406-424 may be repeated to calculate and record the correlations between the sensors 120a-120n and the CRAC units 114a-114n at the next distribution level. In addition, step 426, and steps 406-424, may be repeated for a number of different distribution levels. For example, these steps may be repeated for a predetermined number of iterations, where the predetermined number of iterations may be chosen according to the desired size of the sensor-to-actuator correlations. In addition, or alternatively, these steps may be repeated for a predetermined period of time. In any regard, once the correlation data has been recorded and no further data is to be collected at different distribution levels, the collected data may be processed in either of two examples, as indicated by the identifier “A”.

The first example is illustrated in the flow diagram of an operational mode 450 illustrated in FIG. 4B. As shown in FIG. 4B, following a “no” condition at step 424, the correlation data collected at step 416 for the various distribution levels and CRAC unit 114a-114n settings are fed into a neural network teaching algorithm, as indicated at step 452. The neural network teaching algorithm may, for example, comprise the correlation determination module 214 depicted in FIG. 2. In addition, the correlation coefficients (C_i) may be used to teach the neural network of the initial relationships between the CRAC unit 114a-114n settings and the conditions detected by the sensors 120a-120n. The neural network may use the initial relationships to estimate correlation coefficients (C_i) relating to various CRAC unit 114a-114n settings as described in greater detail herein below.

The neural network teaching algorithm may be implemented to generate a neural network as indicated at step 454. A diagram of a neural network 500 generated at step 454, according to an example, is also illustrated in FIG. 5. As shown in FIG. 5, the neural network 500 includes an input layer 502, a hidden layer 504, and an output layer 506. The input layer 502 includes input neurons A-N 510a-510n, the hidden layer 504 includes hidden neurons A-N 512a-512n, and the output layer 506 includes output neurons A-N 514a-514n, where “n” is an integer greater than one. The ellipses “ . . . ” positioned between various neurons in the neural network 500 generally indicate that the neural network 500 may include any reasonably suitable number of additional neurons.

The input neurons 510a-510n may represent CRAC unit 114a-114n setpoints of the neural network 500, and may comprise, for example, volume flow rates of the airflow supplied by the CRAC units 114a-114n (VFD speeds), CRAC unit 114a-114n supply temperatures, etc. The output neurons 514a-514n may comprise correlation coefficients (C_i) of the sensors 120a-120n. The correlation coefficients (C_i) may be fed into the neural network teaching algorithm to teach the neural network 500 of the initial relationships between the CRAC unit 114a-114n setpoints and the correlation coefficients (C_i) the sensors 120a-120n. In this regard, the output neurons 514a-514n may comprise the correlation coefficients (C_i) determined at step 416, which the teaching algorithm may implement to generate neural network 500.

In addition, based upon the initial relationships between the CRAC unit (i) setpoints (X_i) and the correlation indexes (Y_j) for the sensors (j), the neural network 500 may determine weights (W_ij) between the CRAC unit (i) setpoints (X_i) and the correlation indexes (Y_j). The weights, which are assigned to each interaction, may be randomly selected and modified to reduce the mean square error as the learning epoch proceeds. In addition, the relationships between various CRAC unit (i) setpoints (X_i) and correlation coefficients (Y_j) for the sensors (j) may be defined by the following equation:
Y_j=Sum(W_ij*X_i+B_j), for all the CRAC units (i).  Equation (2)
In this equation, B_j are offsets of the correlation coefficients (Y_j). Equation (2) may be employed in the neural network 500 to determine the correlation coefficients (Y_j) that correspond to various CRAC unit (i) setpoints, which were not fed into the neural network 500 at step 452.

In addition, the number of neurons 510a-510n, 512a-512n, and 514a-514n per layer 502-506 may be modified to increase the accuracy of the neural network model depicted in FIG. 5. By way of example, the number of hidden neurons 512a-512n may be increased to thereby increase the complexity in the relationship between the input neurons 510a-510n and the output neurons 514a-514n. The final model may comprise layers 502-506 of neurons 510a-510n, 512a-512n, and 514a-514n with weights and connections with associated biases that link up the input neurons 510a-510n to the output neurons 514a-514n. In one respect, therefore, the neural network 500 may be employed to determine the correlation coefficients (C_i) of the sensors 120a-120n that were not determined through implementation of the method 400. In this regard, for example, the neural network 500 may be capable of interpolating correlation coefficients (C_i) for various CRAC unit 114a-114n setpoints.

Referring back to FIG. 4B, at step 456, correlation indexes (C_i,j,k) of the sensors 120a-120n may be determined from the neural network 500 generated at step 454. The correlation indexes (C_i,j,k) for the sensors 120a-120n may broadly be defined as functions of the VFD speeds for a plurality of CRAC units 114a-114n and a particular CRAC unit 114a. The function may be in the form of:
C_i,j,k=F(VFD₁, VFD₂, . . . , VFD_n, CRAC_k), where C_i,j,k is the correlation index for the sensor (i), VFD₁ . . . VFD_n are various VFD setpoints (j) for the CRAC units 114a-114n, and the CRAC_k refers to a particular CRAC unit (k).  Equation (3)

At step 458, the correlation indexes (C_i,j,k) of the sensors 120a-120n may be compared to a predefined threshold value. The predefined threshold may be based upon an infrastructure efficiency as determined, for example, by an average value for the correlation coefficients. A high average value of correlation coefficients generally represents an efficient infrastructure, whereas a low average value of correlation coefficients represents a less efficient infrastructure. In any event, it may be determined as to which of the correlation indexes (C_i,j,k) exceed the predefined threshold value at step 458. For those correlation indexes (C_i,j,k) that exceed the predefined threshold value, the associated sensor (i) may be assigned to the particular CRAC unit (k) for the VFD setpoints (j) contained in the function of Equation (3) at step 460. In addition, the correlation indexes (C_i,j,k) for a particular sensor (i) may vary with varying thermal management requirements or varying VFD setpoints (j). In other words, the correlations between any sensor (i) and any CRAC unit (k) are functions of the VFD setpoints (j). Thus, although a sensor (i) may be strongly correlated with a CRAC unit (k) for a particular set of VFD setpoints (j), the same sensor (i) may not be strongly correlated with the same CRAC unit (k) at a different set of VFD setpoints (j).

The remaining sensors 120a-120n may either be assigned to respective CRAC unit 114a-114n families based upon these criteria. In addition, the sensors 120a-120n may belong to multiple families for any given set of VFD setpoints (j). The correlating information regarding the CRAC unit 114a-114n families and their associated sensors 120a-120n may be stored, for example, in a look-up table, in a map, etc.

The sensors 120a-120n may be assigned to respective CRAC units 114a-114n through implementation of the operational modes 400 and 450, thereby commissioning the sensors 120a-120 with respect to the CRAC units 114a-114n in a data center 100. The commissioning process depicted in the operational modes 400 and 450 may include additional steps. For example, data may be historically logged and periodically fed to the neural network teaching algorithm to update the CRAC unit 114a-114n families. This data may also be used to refine the estimates determined by the neural network 500, such as, for example, in the event that an initial commissioning process utilized a relatively limited number of VFD settings to, for example, reduce the time required to perform the initial commissioning process.

Following either or both of steps 458 and 460, it may be determined as to whether the operational modes 400 and 450 are to continue at step 462. The determination of whether to continue the operation modes 400 and 450 may be based upon whether it is desired to, for example, commission the sensors 120a-120n on an ongoing basis. Thus, for example, the operational modes 400 and 450 may be continued at step 462 to substantially continuously update the CRAC unit 114a-114n families. If a “yes” condition is reached at step 462, the operational modes 400 and 450 may be repeated beginning at step 404. If, however, a “no” condition is reached at step 462, the operational modes 400 and 450 may end as indicated at step 464.

The second example is illustrated in the flow diagram of an operational mode 470 illustrated in FIG. 4C. As shown in FIG. 4C, following a “no” condition at step 424, the correlation data collected at step 416 for the various distribution levels and CRAC unit 114a-114n settings are fed into a curve fitting algorithm, as indicated at step 472. The curve fitting algorithm may, for example, comprise the correlation determination module 214 depicted in FIG. 2. In addition, the curve fitting algorithm may comprise any reasonably suitable, traditional curve fitting algorithm used to fit a multi-variant, polynomial function to the data set that defines the correlation index (C_i,j,k).

The curve fitting algorithm may be implemented to determine the correlation indexes (C_i,j,k) for the sensors 120a-120n at step 474. Although a number of various equations may be employed, an example of a polynomial function for two CRAC units 114a-114b may be represented as follows:

C i , j , k = ∑ m = 0 M ⁢ ∑ n = 0 M ⁢ a l ⁢ VFD m ⁢ VFD n , Equation ⁢ ⁢ ( 4 )
where i is the sensor 120a-120n number, j is the CRAC unit 114a-114n distribution set, k is the CRAC unit 114a-114n number, and a_l is a coefficient.

The summation in Equation (4) may be expanded to determine the correlation indexes (C_i,j,k) with additional CRAC units 114a-114n. In any regard, the data from the CRAC unit 114a-114n distribution set (j) may be used to define the coefficients as of Equation (4). Although not shown, a filtering process may be performed following step 416 to reduce the number of CRAC units 114a-114n to consider for various sensors 120a-120n. More particularly, for example, those CRAC units 114a-114n having a relatively limited effect on a sensor 120a may be removed from the calculation of the coefficients (a), as described, for example, with respect to step 814 in FIG. 8 below.

In addition, an equation representing the multi-variant, polynomial function may be implemented for each of the sensors 120a-120n. In addition, the equation may be used to calculate the coefficients (a) for each of the sensors 120a-120n. A standard statistical regression method, for example, through software such as, MATLAB, MS EXCEL, MATHEMATICA, and the like, may be employed to calculate the coefficients (a), with the remaining inputs of the equation having been determined through implementation of the operational mode 400.

At step 476, the correlation indexes (C_i,j,k) of the sensors 120a-120n may be compared to a predefined threshold value. The predefined threshold value may be based upon an infrastructure efficiency as determined, for example, by an average value for the correlation coefficients. A high average value generally represents an efficient infrastructure, whereas a low average value represents a less efficient infrastructure. In any event, it may be determined as to which of the correlation indexes (C_i,j,k) exceed the predefined threshold value at step 476. For those correlation indexes (C_i,j,k) that exceed the predefined threshold value, the associated sensor (i) may be assigned to the particular CRAC unit (k) for the VFD setpoints (j) contained in the function of Equation (3) at step 478. In addition, the correlation indexes (C_i,j,k) for a particular sensor (i) may vary with varying thermal management requirements or varying VFD setpoints (j). In other words, the correlations between any sensor (i) and any CRAC unit (k) are functions of the VFD setpoints (j). Thus, although a sensor (i) may be strongly correlated with a CRAC unit (k) for a particular set of VFD setpoints (j), the same sensor (i) may not be strongly correlated with the same CRAC unit (k) at a different set of VFD setpoints (j).

The remaining sensors 120a-120n may either be assigned to respective CRAC unit 114a-114n families based upon these criteria. In addition, the sensors 120a-120n may belong to multiple families for any given set of VFD setpoints (j). The correlating information regarding the CRAC unit 114a-114n families and their associated sensors 120a-120n may be stored, for example, in a look-up table, in a map, etc.

The sensors 120a-120n may be assigned to respective CRAC units 114a-114n through implementation of the operational modes 400 and 470, thereby commissioning the sensors 120a-120 with respect to the CRAC units 114a-114n. Following either or both of steps 476 and 478, it may be determined as to whether the operational modes 400 and 470 are to continue at step 480. The determination of whether to continue the operation modes 400 and 470 may be based upon whether it is desired to, for example, commission the sensors 120a-120n on an ongoing basis. Thus, for example, the operational modes 400 and 470 may be continued at step 480 to substantially continuously update the CRAC unit 114a-114n families. If a “yes” condition is reached at step 480, the operational modes 400 and 470 may be repeated beginning at step 404. If, however, a “no” condition is reached at step 480, the operational modes 400 and 470 may end as indicated at step 484.

With reference now to FIG. 6, there is shown a flow diagram of an operational mode 600 for optionally filtering sensors 120a-120n belonging to multiple actuator or CRAC unit 114a-114n families. The operational mode 600 may comprise a filtering algorithm stored in the memory 204 that may be implemented by the controller 130. The controller 130 may also update the information stored in the data storage module 216 based upon the results of the filtering algorithm.

The operational mode 600 may be initiated at step 602 in response to any of a number of stimuli or conditions. For example, the operational mode 600 may be initiated with activation of the components in the data center 100, such as, the CRAC units 114a-114n. In addition, or alternatively, the operational mode 600 may be manually initiated or the controller 130 may be programmed to initiate the operational mode 600 at various times, for a set duration of time, substantially continuously, etc.

At step 604, the CRAC unit 114a-114n families may be determined. More particularly, the operational mode 400, 600, or 700 and either of operational 450 or 470 may be performed at step 604 to determine which sensors 120a-120n belong to which CRAC unit 114a-114n families. This information may be stored, for example, in the memory 204. In addition, the controller 130 may access this information to determine whether any of the sensors 120a-120n belong to multiple families at step 606. If none of the sensors 120a-120n belongs to multiple CRAC unit 114a-114n families, it may be determined as to whether the operational mode-600 is to continue at step 607. If it is determined that the operational mode 600 is to continue, the operational mode 600 may be repeated beginning at step 604. In this regard, the operational mode 600 may run in a substantially continuous manner, for example, with each control cycle, to filter sensors 120a-120n belonging to multiple CRAC unit 114a-114n families. In addition, or alternatively, the operational mode 600 may be continued for a predetermined period of time, a predetermined number of iterations, substantially indefinitely, etc. If it is determined that the operational mode 600 is to be discontinued at step 607, the operational mode 600 may end as indicated at step 608. The operational mode 600 may be re-initiated under any of the conditions described with respect to step 604.

However, if at least one sensor 120a-120n is determined to belong to multiple CRAC unit 114a-114n families, the controller 130 may identify those sensors 120a-120n at step 610. In addition, the controller 130 may compare the correlation levels between the sensors 120a-120n and the CRAC units 114a-114n. More particularly, for each of the identified sensors 120a-120n, the controller 130 may determine whether a correlation difference (CD) among the CRAC unit 114a-114n families is significant at step 612. The correlation difference (CD) may be considered as being significant if it exceeds a correlation difference threshold. The correlation difference threshold may be based upon infrastructure efficiency as determined by the average magnitude of correlation coefficients. Thus, for example, the higher the average magnitude of correlation coefficients, the greater the correlation difference threshold.

By way of example, if the sensor 120a has a correlation index of 0.7 to the CRAC unit 114a and a correlation index of 0.1 to the CRAC unit 114b, the correlation difference (CD) between the CRAC unit 114a and the CRAC unit 114b is 0.6. If the correlation difference threshold is 0.5, then the correlation difference between the CRAC unit 114a and the CRAC unit 114b is considered to be significant and a “yes” condition is reached at step 612. However, if the correlation difference threshold is 0.7, then the correlation difference between the CRAC unit 114a and the CRAC unit 114b is considered to be insignificant and a “no” condition is reached at step 612.

As shown in the example above, if the identified sensors 120a-120n are more closely related to a particular CRAC unit 114a-114n, the “yes” condition is reached at step 612. In this regard, the sensor 120a-120n is considered to be significantly affected by that particular CRAC unit 114a-114n. As such, at step 614, those sensors 120a-120n that meet the “yes” condition at step 612 may be filtered out. In other words, those sensors 120a-120n may be removed from further filtering operations described in the operational mode 600.

At step 616, following either or both of the “no” condition at step 612 and the filtering operation at step 614, the remaining identified sensors 120a-120n may be sorted according to whether they require heating or cooling. More particularly, those sensors 120a-120n that require heating may be sorted into one group and those sensors 120a-120n that require cooling may be sorted into another group. Sensors 120a-120n that are considered as requiring “heating” may be defined as sensors 120a-120n whose temperatures may be raised to reach a predefined operating level. Sensors 120a-120n that are considered as requiring “cooling” may be defined as sensors 120a-120n whose temperatures may be lowered to reach the predefined operating level. In addition, or alternatively, the designation of whether sensors 120a-120n require cooling or heating may be based upon a comparison between the temperatures measure by the sensors 120a-120n and a temperature reference matrix.

The sensors 120a-120n may be grouped as indicated at step 616 to substantially prevent sensors 120a-120n assigned to multiple CRAC unit families 114a-114n from causing at least one of the CRAC units 114a-114n to cool and at least another one of the CRAC units 114a-114n to heat. In other words, step 616 may be performed to distinguish the sensors 120a-120n and to substantially prevent CRAC units 114a-114n from competing against each other in attempting to maintain the sensors 120a-120n in their families at their desired temperatures.

At step 618, the identity of a control sensor 120a-120n in each CRAC unit 114a-114n family may be determined. In addition, the magnitude of a temperature difference detected by the control sensor 120a-120n may be determined at step 618. The control sensor 120a-120n may be considered as the sensor 120a-120b having the highest difference between the sensed temperature and a reference temperature for each CRAC unit 114a-114n family. The reference temperature for the sensors 120a-120n may comprise a desired temperature for the sensors 120a-120n and may differ for the sensors 120a-120n. In addition, the magnitude of the temperature difference is the sign and amount of the difference in sensed and reference temperatures. In situations where a CRAC unit 114a family includes a sensor 120a requiring cooling and a sensor 120b requiring heating, the sensor 120a requiring cooling is considered as the control sensor for that CRAC unit 114a family, even if the magnitude of the temperature difference is greater for the sensor 120b requiring heating is greater than the magnitude of the temperature difference for the sensor 120a requiring cooling. In addition, the control sensors requiring cooling may be designated as requiring cooling and the control sensors requiring heating may be designated as requiring heating at step 618.

By way of example, in the CRAC unit 114a family, if the sensor 120a has a sensed temperature of 30° C. and a reference temperature of 25° C. and the sensor 120b has a sensed temperature of 32° C. and a reference temperature of 25° C., the sensor 120b is considered as the control sensor for the CRAC unit 114a family. In addition, the magnitude of the temperature difference of the control sensor is +7° C. The sensor 120b is considered the control sensor because the sensor 120b may have the greatest level of control over the CRAC unit 114a since its sensed temperature has the greatest deviation from the reference temperature.

At step 620, it is determined as to whether the control sensors identified at step 618 are the control sensors of multiple CRAC unit 114a-114n families. For those control sensors that are the control sensors of respective single CRAC unit 114a-114n families, the heating or cooling designation given to them and the other sensors 120a-120n in their respective CRAC unit 114a-114n families at step 618 may be removed at step 622. In addition, the sensors 120a-120n may be checked at step 624, to determine whether they are all in at least one CRAC unit 114a-114n family. Should it be determined that at least one of the sensors 120a-120n is not in a CRAC unit 114a-114n family, the operational mode 400 and either of operational mode 450 and 470 may be performed to determine the at least one CRAC unit 114a-114n family of the at least one sensor 120a-120n. In any regard, at step 607, it may be determined as to whether the operational mode 600 is to continue for those control sensors and other sensors 120a-102n in the CRAC unit 114a-114n families 114a-114n of those control sensors. If it is determined that the operational mode 600 is to continue for those control sensors and other sensors 120a-120n, the operational mode 600 may be repeated for those control sensors and other sensors 120a-120n beginning at step 604. If it is determined that the operational mode 600 is to be discontinued at step 607, the operational mode 600 for those control sensors and other sensors 120a-120n may end as indicated at step 608. The operational mode 600 may also be re-initiated under any of the conditions described with respect to step 604.

For those control sensors identified as the control sensors of multiple CRAC unit 114a-114n families, it may be determined whether the CRAC units 114a-114n of the multiple CRAC unit 114a-114n families are at different setpoints at step 626. In other words, for those CRAC units 114a-114n, it may be determined whether the differences in supply air temperatures between the CRAC units 114a-114n exceed a CRAC temperature difference threshold. For those multiple CRAC unit 114a-114n families having CRAC units 114a-114n that do not exceed the CRAC temperature difference threshold, steps 622, 624, 607 and 608 may be performed as described above. In one regard, since the CRAC units 114a-114n are within the CRAC temperature difference threshold, multiple CRAC units 114a-114n may both be used to control the control sensor. However, if it is determined that the CRAC units 114a-114n are at different setpoints or outside of the CRAC temperature difference threshold, it may be determined whether the CRAC units 114a-114n are required to heat or cool the control sensor at step 628.

In addition, at step 628, the control sensor may be assigned to the CRAC unit 114a family having the CRAC unit 114a, whose supply air temperature is the farthest away from the temperature of the control sensor. That CRAC unit 114a family may be considered as the primary family for that control sensor. The other CRAC unit 114b-114n families to which the control sensor is assigned may be considered as secondary families for that control sensor. More particularly, if the control sensor requires cooling, the CRAC unit 114a family having the CRAC unit 114a with the highest supply air temperature is selected as the primary family for that control sensor. Alternatively, if the control sensor requires heating, the CRAC unit 114a family having the CRAC unit 114a with the lowest supply air temperature is selected as the primary family for that control sensor.

If the control sensor requires cooling, all overlapping points having the same sign as the control sensor are removed from the secondary family, as indicated at step 630. More particularly, the sensors 120a-120n that are in both the primary and secondary families and that have the same sign (for example, requiring either heating or cooling) as the control sensor are removed from the secondary family. In this regard, control over the temperature of the control sensor is given to the CRAC unit 114a that has the highest supply air temperature, if that supply air temperature is greater than the temperature of the control sensor. In this regard, the CRAC unit 114a having the highest supply air temperature may be able to reduce its supply air temperature. Otherwise, if all of the potential control CRAC units 114a-114n have supply air temperatures below the control sensor temperature, control is given to the CRAC unit 114 that is closest in temperature to the control sensor temperature. As such, the CRAC unit 114a that has the greatest supply temperature may be controlled, when the control sensor is in the cooling mode.

If the control sensor requires heating, which means that no other sensors 120a-120n in the primary family require cooling, all overlapping points having the opposite sign as the control sensor are removed from the primary family, as indicated at step 632. More particularly, the sensors 120a-120n that are in both the primary and secondary families and that have the opposite sign as the control sensor are removed from the primary family. In this regard, control over the temperature of the control sensor is given to the CRAC unit 114a that has the lowest supply air temperature, if that supply air temperature is less than the temperature of the control sensor. In this regard, the CRAC unit 114a may be able to increase its supply air temperature. Otherwise, if all of the potential control CRAC units 114a-114n have supply temperatures above the control sensor temperature, control is given to the CRAC unit 114a that is the closest in temperature to the control sensor. Thus, the CRAC unit 114a that has the lowest supply temperature may be controlled, when the control sensor is in the heating mode.

Following step 632, the heating or cooling designation given to the sensors 120a-120n at step 616 may be removed at step 622. In addition, the sensors 120a-120n may be checked at step 624, to determine whether they are all in at least one CRAC unit 114a-114n family. Should it be determined that at least one of the sensors 120a-120n is not in a CRAC unit 114a-114n family, the operational mode 400 and either of operational mode 450 and 470 may be performed to determine the at least one CRAC unit 114a-114n family of the at least one sensor 120a-120n. In addition, or alternatively, sensors 120a-120n that are not in any of the CRAC unit 114a-114n families may be removed from consideration from the operational mode 400 or the cooling infrastructure may be re-arranged, such as by moving vent tiles 118, to bring the sensors 120a-120n within at least one of the CRAC unit 114a-114n families. In any regard, steps 607 and 608 may be performed as described above.

Control of Actuators

The control of the primary actuators 114a-n based on the commissioning of the sensors 120a-n, as disclosed in a co-pending and commonly assigned U.S. patent application as identified from its U.S. Patent Application Publication No. 20060214014, entitled “TEMPERATURE CONTROL USING A SENSOR NETWORK,” is described below with reference to FIGS. 7-8.

FIG. 7 illustrates a flow diagram of a method 700 for controlling an environmental condition using a sensor network, according to one embodiment. For illustrative purposes only and not to be limiting thereof, the operation mode 300 is discussed in the context of the block diagram 200 illustrated in FIG. 2.

The method 700 may be implemented to control the temperature of the airflow supplied to the sensors 120a-120n by the primary actuators 114a-114n. In the method 700, the controller 130 may commission the sensors 120a-120n of a network of sensors at step 702. The controller 130 may also select one of a plurality of control schemes for operating a primary actuator 114a-114n configured to vary temperatures of the sensors 120a-120n at step 704. The selection of the control scheme may be based upon energy utilization requirements of the plurality of control schemes. In addition, the controller 130 may implement the selected one of the plurality of control schemes to vary the temperatures detected by the sensors 120a-120n at step 706. A more detailed description of the steps outlined in the method 700 is provided below with respect to the method 800.

With particular reference now to FIG. 8, there is shown a flow diagram of a method 800 for controlling environmental conditions at the sensor 120a-120n locations, according to one embodiment. For illustrative purposes only and not to be limiting thereof, the operation mode 300 is discussed in the context of the block diagram 200 illustrated in FIG. 2.

The method 800 may be implemented to control primary actuators (CRAC units) 114a-114n to thereby control environmental conditions at the sensor 120a-120n locations. For example, the primary actuators 114a-114n may be controlled as disclosed below to control at least one of temperature, humidity, and pressure at the sensor 120a-120n locations.

The method 800 may be initiated at step 802 in response to any of a number of stimuli or conditions. For example, the method 800 may be initiated with activation of components, such as, the CRAC units 114a-114n. In addition, or alternatively, the method 800 may be manually initiated or the controller 130 may be programmed to initiate the method 800 at various times, for a set duration of time, substantially continuously, etc.

Once initiated, the controller 130 may implement the sensor commissioning module 214 to commission the sensors 120a-120n at step 804. More particularly, the sensors 120a-120n may be assigned to primary actuator 114a-114n families according to, for example, the level of influence the primary actuators 114a-114n have over the sensors 120a-120n. In addition, a sensor 120a may be considered as being in the family of a primary actuator 114a if the level of influence of that actuator 114a over the sensor 120a exceeds a predefined threshold. In addition, if a sensor 120a is assigned to multiple primary actuator 114a-114n families, a filtering process may be implemented to assign the sensor 120a to the primary actuator 114a that has the greatest level of influence over the sensor 120a. The filtering process may also be implemented to keep the sensor in multiple primary actuator 114a-114n families if the influence levels of the primary actuators 114a-114n are within predefined thresholds. An examples of a suitable commissioning process and a suitable filtering process is as described earlier.

At step 806, for each primary actuator 114a-114n family, the controller 130 may implement the control sensor selection module 216 to choose the control sensor 120n. As described above, the control sensor 120n may be defined as the sensor 120a-120n with the largest positive temperature difference from a setpoint temperature in each primary actuator 114a-114n family. If all of the sensors 120a-120n have negative temperature differences from the at least one setpoint, then the control sensor 120n may be defined as the sensor 120a-120n with the largest negative difference from the setpoint temperature. The setpoint used to determine the control sensor 120n may also vary between the sensors 120a-120n and is thus not required to be identical for all of the sensors 120a-120n.

In general, the temperatures of the control sensors 120a-120n are used to control the primary actuators 114a-114n of the families to which the control sensors 1201a-120n belong. More particularly, the following steps may be performed to vary the setpoints of the secondary actuators 222 and 224 to thereby bring the temperatures of the control sensors 120a-120n to within desired temperature ranges. The desired temperature ranges, in this case, may generally represent acceptable levels of error in controlling the temperatures of the sensors 120a-120n. Thus, for example, the desired temperature ranges may be larger if a higher level of error is acceptable.

At step 808, the controller 130 may receive signals from the control sensors 120a-120n of the primary actuator 114a-114n families. The signals are generally control signals for operating respective primary actuators 114a-114n. The signals are generated according to the levels of error between measured temperatures and predefined temperatures. More particularly, if the temperature difference is relatively small, the signal sent to the controller 130 is also relatively small. Alternatively, if the temperature difference is relatively large, the signal sent to the controller 130 is also relatively large. The controller 130, in this case, may include a proportional, integral (PI) controller or a proportional, integral, derivative (PID) controller, and may control operations of the primary actuators 114a-114n (to output desired supply air temperature or SAT setpoints) based upon the magnitudes of the control signals received from the control sensors 120a-120n as described in greater detail herein below with respect to step 812. In addition, each of the primary actuators 114a-114n may include respective PI, PID, other suitable controllers.

For each primary actuator 114a-114n family having a control sensor 120a-120n that transmits a control signal indicating an error, the controller 130 may implement one of many possible control schemes to manipulate the secondary actuators 222 and 224, as indicated at step 810. As described above with respect to the control system 202, one of the secondary actuators 222 may comprise a variable frequency drive (VFD) for controlling an airflow volume varying device, such as a blower or fan. The other secondary actuator 224 may comprise a device for controlling the temperature of the cooled air supplied by the primary actuators 114a-114n, and may depend upon the type of primary actuator 114a-114n in which the secondary actuator 224 is located. Additionally, each primary actuator 114a-114n may include another secondary actuator (not shown) for humidity control, such as a humidifier, a de-humidifier, or the like.

The possible control schemes for manipulating the secondary actuators 222 and 224 involve a linkage of one of the secondary actuators 222 (VFD's) to the supply air temperature (SAT) setpoints of the primary actuators 114a-114n throughout a range of control. The range of control may be defined as a predefined range of supply air temperatures and VFD settings within which the primary actuators 114a-114n may operate. In this regard, the range of control may be based upon levels that are known to provide adequate levels of cooling airflow at adequate temperatures. In any regard, the linkage between the VFD setpoints (secondary actuator 222) and the SAT setpoints (secondary actuator 224) may be described through the following equation:
VFD=VFD_max+γ(SAT_min−SAT).  Equation (5)
In this equation, VFD_max is the maximum allowable VFD setpoint, SAT_min is the minimum allowable supply air temperature, and γ is (VFD_maxγVFD_min)/(SAT_max−SAT_min). VFD_min is the minimum allowable VFD setpoint and SAT_max is the maximum allowable supply air temperature.

The control schemes may be varied as a function of the SAT setpoints output by a controller, such as a PI or PID controller, in the controller 130 as described above. The specific functionality is dependent on the actuator type of the primary actuators 114a-114n, as further elaborated in U.S. Patent Application Publication No. 20060214014.

Visualization

A visualization tool for analyzing and visualizing thermal profiles in a selected environment, such as the data center 100, as disclosed in a co-pending and commonly assigned U.S. patent application Ser. No. 11/699,402, entitled “COMPUTERIZED TOOL FOR ASSESSIGN CONDITIONS IN A ROOM,” is described below with reference to FIGS. 9-13. This visualization tool provides users with a monitoring tool to instantiate any implemented environmental control for a selected environment. Thus, for example, a system administrator of the data center 100 may employ this visualization tool to instantiate the aforementioned control schemes based on the commissioning of the sensors and optimize placement of equipment in such a data center for environmental management of the data center.

1. A computer readable storage medium on which is embedded one or more computer programs, said one or more computer programs being configured as an environmental management architecture for providing environmental management of a physical location using a sensor network having a plurality of environmental sensors and at least one primary actuator configured to provide an environmental change to the physical location, the one or more computer programs comprising: a communications module that operates to access the plurality of environmental sensors of the sensor network; and an application module that operates to: a) commission the plurality of environmental sensors of the sensor network and receive detected conditions from the plurality of environmental sensors; b) control an operation of the at least one primary actuator to provide environmental management of the physical location based on the commission and detected conditions of the plurality of environmental sensors of the sensor network; and c) provide a graphical layout of an environmental condition of the physical location based on the commissioning and detected conditions of the plurality of environmental sensors and the control of the at least one primary actuator, wherein the graphical layout graphically depicts a mapping of at least one of the detected conditions and values derived from the detected conditions with respect to objects contained in the physical location; wherein the environmental management architecture at least one of resides and operates on top of an operating system of a computing device and network services on which the computing device is connected.

2. The computer readable storage medium of claim 1, said one or more computer programs further comprising: an aggregation module that stores data generated by the sensor network and the application module.

3. The computer readable storage medium of claim 2, wherein the application module includes: a data logger sub-module that operates to log the data generated by the sensor network and the application module in the aggregation module for storage.

4. The computer readable storage medium of claim 1, said one or more computer programs further comprising: an aggregation module that operates to provide an interface for data exchange between the communications module and the application module.

5. The computer readable storage medium of claim 1, wherein the communications module includes at least one of: a communication driver that implements an object-linking-and-embedding-for-process-control (OPC) client to gain access to the plurality of sensors of the sensor network; a communication driver that operates to receive multicasts as transmitted from the plurality of environmental sensors of the sensor network; and a communication driver that operates to directly communicate with each of the plurality of environment sensors of the sensor network.

6. The computer readable storage medium of claim 1, wherein the plurality of environmental sensors of the sensor network are temperature sensors arranged throughout the physical location.

7. The computer readable storage medium of claim 1, said one or more programs further comprising: a graphical user interface (GUI) module that operates to display the graphical layout of the environmental condition of the physical location.

8. The computer readable storage medium of claim 7, wherein the GUI module includes a capability to display a simplified version of the graphical layout on wireless devices.

9. The computer readable storage medium of claim 1, further comprising: a graphical user interface (GUI) module that operates to receive a user input of at least one of a parameter for modifying the operational control of the at least one primary actuator, a parameter for the commission of the plurality of environmental sensors of the sensor network, and an equipment layout at the physical location.

10. The computer readable storage medium of claim 1, wherein the application module further comprises: a notification module that operates to monitor one or more conditions of the system and to generate a notification upon the system achieving a predefined condition.

11. The computer readable storage medium of claim 1, wherein the at least one primary actuator comprises secondary actuators, the secondary actuators include a device for controlling supply airflow volume and a device for controlling supply airflow temperature.

12. A method for implementing an environmental management architecture for providing environmental management of a physical location using a sensor network having a plurality of environmental sensors and at least one primary actuator configured to provide an environmental change to the physical location, wherein the environmental management architecture at least one of resides and operates on top of an operating system of a computing device and network services on which the computing device is connected, the method comprising steps performed by a processor of the computing device of: accessing the plurality of environmental sensors of the sensor network with at least one communication driver to receive detected sensor data; commissioning the plurality of environmental sensors of the sensor network based on the sensor data from the accessing; controlling an operation of the at least one primary actuator to provide environmental management of the physical location based on the commissioning; and generating a display of a graphical layout of an environmental condition of the physical location based on the commissioning, the detected sensor data, and the controlling, wherein the graphical layout graphically depicts a mapping of at least one of the detected sensor data and values derived from the detected sensor data with respect to objects contained in the physical location.

13. The method of claim 12, further comprising: storing data generated by the sensor network and the application module in an aggregation module.

14. The method of claim 12, further comprising: monitoring one or more conditions of the system and to generate a notification upon the system achieving a predefined condition.

15. The method of claim 14, wherein commissioning the plurality of environmental sensors of the sensor network based on the reading of sensor data comprises: calculating a correlation index from the read sensor data, wherein the correlation index provides a relationship between one of the plurality of environmental sensors and the at least one primary actuator; and assigning the one environmental sensor to one of a plurality of actuator families of the at least one primary actuator.

16. The method of claim 15, wherein controlling an operation of the at least one primary actuator comprises: selecting one of a plurality of control schemes for operating the at least one primary actuator based on the commissioning and one or more predefined criteria; implementing the selected one of the plurality of control schemes to operate the at least one primary actuator to vary the environmental condition of the physical location.

17. The method of claim 16, wherein generating a display of a graphical layout of the environmental condition of the physical location comprises: modeling the environmental condition of the physical location based on the sensor data received from the accessing; extracting environmental condition data of the physical location from the modeling; and calculating the graphical layout of the environmental condition of the physical location from the environmental condition data.

18. The method of claim 17, wherein the at least one primary actuator comprises secondary actuators, the secondary actuators include a device for controlling supply airflow volume and a device for controlling supply airflow temperature, and the plurality of control schemes include a scheme for controlling the secondary actuators.

19. A computer-readable storage medium on which is encoded programming code for an environmental management architecture for providing environmental management of a physical location using a sensor network having a plurality of environmental sensors and at least one primary actuator configured to provide an environmental change to the physical location, wherein the environmental management architecture at least one of resides and operates on top of an operating system of a computing device and network services on which the computing device is connected, said programming code including a set of instructions that when executed by a computer operates to: access the plurality of environmental sensors of the sensor network with at least one communication driver to receive detected sensor data; commissioning the plurality of environmental sensors of the sensor network based on the sensor data from the accessing; control an operation of the at least one primary actuator to provide environmental management of the physical location based on the commissioning; and generating a display of a graphical layout of an environmental condition of the physical location based on the commissioning, the detected sensor data, and the controlling, wherein the graphical layout graphically depicts a mapping of at least one of the detected sensor data and values derived from the detected sensor data with respect to objects contained in the physical location.

20. The computer-readable medium of claim 19, wherein the encoded programming code when executed by the computer further operates to: monitor one or more conditions of the system and to generate a notification upon a system achieving a predefined condition.