Real-Time Occupant Based Plug-in Device Control Using ICT ... - MDPI

energies Article

Real-Time Occupant Based Plug-in Device Control Using ICT in Office Buildings Woo-Bin Bae, Sun-Hye Mun and Jung-Ho Huh * Department of Architecture Engineering, University of Seoul, Seoul 130-743, Korea; [email protected] (W.-B.B.); [email protected] (S.-H.M.) * Correspondence: [email protected]; Tel.: +82-02-6490-2757 Academic Editor: Joseph H. M. Tah Received: 2 January 2016; Accepted: 18 February 2016; Published: 1 March 2016

Abstract: The purpose of this study is to reduce the unnecessary plug loads used by computers, monitors, and computer peripheral devices, all of which account for more than 95% of the entire plug loads of an office building. To this end, an occupant-based plug-in device control (OBC-P) software was developed. The OBC-P software collects real-time information about the presence or absence of occupants who are connected to the access point through the Wifi and controls the power of monitors or computers, while a standby power off device controls computer peripheral devices. To measure the plug load saving of the occupant-based plug-in device control, an experiment was conducted, targeting 10 occupants of three research labs of the graduate school, for two weeks. The experiment results showed that it could save the plug loads of monitors and computer peripheral devices by 15% in the Awake mode, and by 26% in the Sleep mode. Keywords: occupancy detection; occupant based control; plug-in device control; real-time occupant information collection; Wifi

1. Introduction According to the U.S. Department of Energy (DOE), about 33% of the entire energy consumption of an office building is consumed by plug loads, which are expected to surge to 49% by 2030 [1,2]. In case of an office building, 95% of plug loads are consumed by computers (69%), monitors (9%), and computer peripheral devices (17%), except for severs and air handling units [3]. In the past, the reduction of plug loads in office buildings was heavily dependent on the improvement of the behaviors of occupants, by raising their awareness about the importance of energy saving [1,3–5]. However, as the dependence on occupants’ behaviors inevitably caused inconvenience to occupants, it was difficult to expect continuous plug load saving effects. In recent years, the standby power off device has been widely used to reduce the unnecessary plug loads of computer peripheral devices of an office building [1,3]. However, the standby power off devices can save the plug loads of computer peripheral devices only when occupants turn off the computer; otherwise, it cannot reduce the plug loads. Therefore, the plug load saving effect using the existing standby power off devices was still dependent on occupants’ behaviors, for instance, their awareness to turn off computers. However, office building occupants usually do not pay for energy consumption, and the act of pulling off plugs is often perceived as inconvenience. As a result, unnecessary plug loads are wasted by the current office buildings. Against this backdrop, it is necessary to develop a way of reducing unnecessary plug loads of computers, monitors and computer peripheral devices without being dependent on behavioral tendencies.

Energies 2016, 9, 143; doi:10.3390/en9030143

www.mdpi.com/journal/energies

Energies 2016, 9, 143

2 of 15

2. Limitation of the Conventional Plug Load Saving Strategies The existing plug load saving measures can be divided into the education-based method (namely, the improvement of the energy saving behavior of occupants), the software-based measures, and the hardware-based method [3]. The education-based method aims at improving the energy saving behavior of the occupants, and has its own advantages; it does not incur any additional costs and can be applied to most of the plug-in devices. It involves turning off the unnecessary devices or pulling out the plugs whenever an occupant moves to another place. However, this inevitably causes inconvenience to the occupants, and energy saving is too much dependent on the behaviors of occupants to expect a continuous energy saving effect. Although the software-based method that can be remotely controlled by a building superintendent does not incur additional costs, it cannot control computer peripheral devices, and a remote-control software needs to be installed to enable users to remotely control computers and monitors from a personal computer. Also, as it can terminate the operation of a computer or a monitor, it is likely to cause inconvenience to occupants. The hardware-based method is to use a standby power off device or to replace the existing plug-in device with a highly efficient one. The mechanisms of standby power off devices can be divided into the scheduling-based method and the standby power detection and cut-off method. The scheduling-based method is to connect plug-in devices with schedulers to a standby power off device, and to adjust the scheduler setting of the standby power off device to allow the plug-in devices to automatically go off. The method is effective in reducing the plug loads of the plug-in devices with built-in schedulers. However, it cannot respond to the irregular use of plug-in devices by the occupants. The power detection & standby power-off method is to connect the power cable of a computer to the master outlet of the standby power-off device and computer peripheral devices to the controlled outlets. After the standby power-off device is installed, if an occupant turns off the computer, the standby power of the computer peripheral devices is automatically discontinued, or when an occupant turns on the computer, the power is supplied to computer peripheral devices. This method is effective in cutting off standby power. However, the reduction of plug loads is dependent on the occupant’s act of turning off the power of the computer. The replacement method is to preplace a low efficient device with a high efficient device to save plug loads. This method is expected to achieve a higher energy saving efficiency than any other method, but the replacement method is not cost-efficient in most cases. To make up for the shortcoming of the existing plug load reduction methods which are overly dependent on behavioral tendencies, the development of occupant-based plug-in controls is in a great demand in recent years [1,4,6]. The occupant-based plugin device control is to collect real-time information about the presence or absence of occupants and, based on the occupant information, to control the plug-in devices. Harris et al. [7] proposed a method of controlling computers and monitors through the bluetooth of smartphones after installing an occupant-based plugin device control software in a computer. However, as bluetooth consumes a greater amount of battery than Wifi, it is not widely used. Also, the study by Harris did not include the controlling of computer peripheral devices. In a study by Ken Christensen et al. [8], computers, monitors, and computer peripheral devices were connected to the standby power-off device, a control software was installed in a computer, while a control application was installed in a smartphone. Subsequently, the computers, monitors, and computer peripheral devices were controlled through the Wifi of the smartphone. When an occupant left the office room for a short period of time, the monitor would go off and the standby power of computer peripheral devices would be cut off. Also, the monitors and computer peripheral devices are turned off right before the occupant enters the office room. Therefore, a plug load saving effect can be achieved while minimizing the inconvenience of the occupant. However, in the study by Ken Christensen et al., the computer was always turned on even when the occupant was absent for a short period, and an analysis on plug in saving quantities was not conducted. Also, another

Energies 2016, 9, 143

3 of 15

Energies 2016, 9, 143

3 of 15

Energies 2016, 9, 143

3 of 15

disadvantage is that the occupant, who was not obliged to pay for energy consumption, had to is that the occupant, who was not obliged to pay for energy consumption, had to experience the experience the inconvenience of installing an application in the smartphone. is that the occupant, whoan was not obliged for energy consumption, had to experience the inconvenience of installing application in to thepay smartphone. The purpose of the studyanisapplication to developinathe new mechanism to control computers, monitors, and inconvenience smartphone. The purposeofofinstalling the study is to develop a new mechanism to control computers, monitors, and computer peripheral according to information about the presence or absence ofmonitors, occupants, The purposedevices ofdevices the study is to develop a new mechanism to control computers, andusing computer peripheral according to information about the presence or absence of occupants, theusing Wifi and a standby power cutoff device, and installing a control software in a computer, but without computer peripheral devices power according to information the presence orsoftware absence of the Wifi and a standby cutoff device, andabout installing a control inoccupants, a computer, using the Wifi and a standby power cutoff device, and end, installing a this control software in aoccupant-based computer, an thebut installation of an application aapplication smartphone. this this developed an without the installation of in an in To a smartphone. Tostudy end, this study developed but without the installation of an application in a smartphone. To this end, this study developed an plug-in control software and analyzed plug load saving rates through an experiment. occupant-based plug-in control software and analyzed plug load saving rates through an experiment. occupant-based plug-in control software and analyzed plug load saving rates through an experiment.

3. 3. Occupant OccupantBased BasedPlug-in Plug-inDevice Device Control Control

3. Occupant Based Plug-in Device Control

3.1.3.1. Development of Software Development of Software 3.1. Development of Software Figure 1 shows the role of OBC-P software (version 1.0) and and standby standbypower powercut-off cut-off device for Figure 1 shows (version 1.0) 1.0) device Figure 1 showsthe therole roleofofOBC-P OBC-P software software (version and standby power cut-off device forfor thethe plug-in device control. The Occupant Based Control-Plugged in Device (OBC-P) software was plug-in device Control-PluggedininDevice Device(OBC-P) (OBC-P) software was the plug-in devicecontrol. control.The TheOccupant Occupant Based Based Control-Plugged software was developed forforthe control of occupant-based plug-indevices. devices. The OBC-P software was developed developed the control of occupant-based plug-in The OBC-P software was developed developed for the control of occupant-based plug-in devices. The OBC-P software was developed by by by Matlab Matlab and has graphical user interface (GUI) as seen in Figure The software OBC-P software collects and has aa graphical user interface (GUI) as seen inFigure Figure The2. OBC-P software collects realMatlab and has a graphical user interface (GUI) as seen in 2.2.The OBC-P collects realreal-time occupant information through Wifi, and plays a role to control the main monitor or computers. time occupant information through Wifi, and plays a role to control the main monitor or computers. time occupant information through Wifi, and a role to control the main monitor or computers.

OBC-P OBC-P software software

Plug-in Plug-in device device control control

Awakemode mode Awake

Sleep Sleepmode mode

Control offoff Control

Occupantpresence presence Occupant check(see (seeFig. Fig. 6) 6) check

Occupant Occupantpresence presence check check(see (seeFig. Fig.6)6)

Monitorcontrol control Monitor (see Fig. 11)

Monitor Monitorand andcomputer computer control (see Fig. 13)

Computer peripheral

Computer peripheral Computer control (seeperipheral Fig. 13)

Uncontrolled Uncontrolled

(see Fig. 11)

Cut-off

Cut-off device device

Computer peripheral control (see Fig. 11) control (see Fig. 11)

control (see Fig. 13)

control (see Fig. 13)

Computer peripheral Computer peripheral is controlled by is controlled by monitor on/off mode

monitor on/off mode

Figure 1. Role of OBC-Psoftware software and cutoff cutoff device device control. Figure 1.1.Role devicefor forplug-in plug-in device control. Figure RoleofofOBC-P OBC-P software and and cutoff device for plug-in device control.

MAC address

MAC address Control mode

Control mode

Control status

Control status

Figure 2. OBC-P software.

Figure2. 2.of OBC-P The GUI of the OBC-P softwareFigure consists threesoftware. sections. The first section is where the Media OBC-P software. Access Control (MAC) address is entered. Every smartphone has its unique MAC address, and when GUI of the OBC-Pthrough software three sections. The first section is whereonthe anThe occupant is connected theconsists Wifi of aof smartphone, the MAC address is displayed theMedia AP The GUI of the OBC-P software consists of three sections. The first section is where thewhen Media Access Control address is entered. Every smartphone has its unique MAC of address, and (Access Point)(MAC) administration page, as seen in Figure 3. The absence and presence an occupant is Access Control (MAC) address is entered. Every smartphone has its unique MAC address, and when an determined occupant isby connected through the Wifi of athat smartphone, addresssoftware is displayed on the comparing the MAC address is enteredthe intoMAC the OBC-P with the oneAP an (Access occupant connected through the Wifi ofsecond a smartphone, MAC address is displayed onThe theisAP administration page, as seen in Figure 3. The absence and presence ofselected. an occupant storedPoint) inisthe AP administration page. The section is the where a control model is

(Access Point) by administration page, asaddress seen in that Figure 3. The absence presence of with an occupant determined comparing the MAC is entered into the and OBC-P software the one is determined the MACpage. address is entered the OBC-P software the one stored stored in by thecomparing AP administration Thethat second sectioninto is where a control modelwith is selected. The

Energies 2016, 9, 143 Energies 2016, 9, 143

4 of 15 4 of 15

plug-in device has beenThe developed into two types: aAwake monitors Energies 2016, 9,control 143 4 of plug-in 15 and in the AP administration page. second section is where controlmode modeltoiscontrol selected. The computer peripheral devices, and Sleep mode to control monitor, computer peripheral devices, and device control has been developed into two types: Awake mode to control monitors and computer plug-in device control has been developed into two types: Awake mode to control monitors and computers. The third indicates the controlling status, which showsdevices, the ON/Off status of peripheral devices, andsection Sleep mode to control peripheral and computers. computer peripheral devices, and Sleep mode monitor, to controlcomputer monitor, computer peripheral devices, and plug-in devices in a real-time basis; “ON” is expressed in “1”, while “OFF” is expressed in “0”. The third section indicates the controlling status, which shows the ON/Off status of plug-in devices in computers. The third section indicates the controlling status, which shows the ON/Off status of

a real-time “ON” is expressed “1”,iswhile “OFF” is expressed in “0”. plug-inbasis; devices in a real-time basis; in “ON” expressed in “1”, while “OFF” is expressed in “0”. MAC address MAC address

Network name

Signal strength

Room3 name Network

Connect hours Connect hours 04:04:54

Signal strength

Room3 Room3

04:04:54 03:57:20

Room3 Room3

03:57:20

01:20:50

Room3

01:20:50

Figure 3. Access point administration page. Figure pointadministration administration page. Figure3.3.Access Access point page.

3.1.1. Real-Time Occupant Information Acquisition Real-Time Occupant InformationAcquisition Acquisition 3.1.1.3.1.1. Real-Time Occupant Information Figure 4 shows how the occupant collectedin in a real-time basis by OBC-P the OBC-P Figure 4 shows how theoccupant occupant information information isis a real-time basis by the Figure 4 shows how the information iscollected collected in a real-time basis by the OBC-P software through Wifi. (1) When an occupant is connected to the access point through the software through Wifi. (1) When an occupant is connected to the access point through the Wifi Wifi of a of a software through Wifi. (1) When an occupant is connected to the access point through the Wifi smartphone (the(the connection totothe induced synchronizing the ON/OFF smartphone connection theWifi Wifiof of the the smartphone smartphone isisinduced byby synchronizing the ON/OFF of a smartphone (the connection to the Wifi of the smartphone is induced by synchronizing the status of computers,monitors, monitors, and and computer computer peripheral the the connection and and status of computers, peripheraldevices deviceswith with connection ON/OFF status oftocomputers, monitors, and computer peripheral devices with the connection disconnection the Wifi); (2) the OBC-P software gains access to the AP administration phase to and disconnection to the Wifi); (2) the OBC-P software gains access to the AP administration phase to disconnection to the Wifi); (2) the OBC-P software gains access to the AP administration phase to collect the MAC address, Wifinetwork network name, name, and information; (3) Based on theon the collect the MAC address, thethe Wifi andsignal signalintensity intensity information; (3) Based collect the MAC address,the theOBC-P Wifi network signal intensitythe information; (3) Based on the collected information, softwarename, gathersand information location of an collected information, the OBC-P software gathers informationabout about the location ofoccupant an occupant (Wifi information, network name,the signal intensity), the presence absence of occupants addresses), the collected OBC-P software gathers or information about the(MAC location of an occupant (Wifi network name, signal intensity), the presence or absence of occupants (MAC addresses), the number of occupants (number of MAC addresses), and the identities of occupants (MAC addresses); (Wifi network name, signal intensity), the presence or absence of occupants (MAC addresses), the number of occupants (number ofto MAC addresses), and the identities of occupants (MAC addresses); (4) The OBC-P software logs in the AP administration on a regular to collect occupants’ number of occupants (number of MAC addresses), and page the identities of basis occupants (MAC addresses); (4) The OBC-P software logs in to the AP administration page on a regular basis to collect occupants’ information on a real-time (4) The OBC-P software logs basis. in to the AP administration page on a regular basis to collect occupants’ information on a real-time basis. information on a real-time basis. AP administration page Room1

Room1

Room2

Room2

Room3

Room3

(1)

AP administration page

(2)

(2) Identity

(1)

Location

(3)

Presence, Count

Identity

Location

Building

(3)

Presence, Count Room1

(4)

Room2

Room3

Building Room1

Room2

Room3

(4)

Figure 4. Real-time occupant information collection.

In order to control computers, monitors and computer peripheral devices, it is necessary to Figure Figure 4. 4. Real-time Real-time occupant occupant information information collection. collection. compare real-time information to find out whether the occupants are present or absent. Figure 5 shows how to confirm the presence and absence of occupants. (1) The OBC-P software logs in to the In order to control computers, monitors and computer peripheral devices, it is necessary to In order to control computers, and computer peripheral it isintensity necessary to AP administration page to collect monitors MAC addresses, Wifi network names,devices, and signal compare real-time information to find out whether the occupants are present or absent. Figure 5 information; (2)information Figure 6 displays flowchart of checking the absence and presence of occupants. compare real-time to finda out whether the occupants are present or absent. Figure 5 shows shows how to confirm the presence and absence of checks occupants. (1) The software logs in to the of all, thethe MAC address of theabsence OBC-P software whether APOBC-P administration page howFirst to confirm presence and of occupants. (1) The an OBC-P software logs inexists to the AP AP administration to collect MACin the addresses, Wifi network and signal intensity or not. If thepage MACpage address notaddresses, exist AP administration page,and itnames, issignal judgedintensity that an occupant administration to collectdoes MAC Wifi network names, information; information; (2) accessible Figure 6 range displays a flowchart the and presence of occupants. is out of the of the Wifi, and itofis checking determined as absence an unoccupied room. If the MAC (2) Figure 6 displays a flowchart of checking the absence and presence of occupants. First of all, the exists in the AP administration page, it will check whether thean signal intensity is higher thanexists First address of all, the MAC address of the OBC-P software checks whether AP administration page MAC address of the OBC-P software checks whether an AP administration page exists or not. If the 80% or not. If the signal intensity is higher than 80%, it will be judged as an occupied room. The or not. If the MAC address does not exist in the AP administration page, it is judged that an time occupant MAC address does existininorder the APprevent administration page, it is judged that an occupant is out of the is set at onenot second overload of computers. is outdelay of the accessible range of thetoWifi, andthe it is determined as an unoccupied room. If the MAC

accessible range of the Wifi, and it is determined as an unoccupied room. If the MAC address exists in address exists in the AP administration page, it will check whether the signal intensity is higher than the AP administration page, it will check whether the signal intensity is higher than 80% or not. If the 80% or not. If the signal intensity is higher than 80%, it will be judged as an occupied room. The time signal intensity is higher than 80%, it will be judged as an occupied room. The time delay is set at one delay is set at one second in order to prevent the overload of computers. second in order to prevent the overload of computers.

Energies 2016, 9, 143

5 of 15

Energies 2016, 9, 143 Energies 2016, 9, 143

5 of 15 5 of 15 Room1 Room1

Room2 Room2

Room3 Room3

(1) (1)

(2) (2)

Collect MAC address, network name and signal strength Collect MAC address, network name and signal strength

Occupant presence check Occupant presence check

Figure5.5. Occupant Occupant presence check Figure checkmethod. method. Figure 5. Occupant presence presence check method. Start Start

Connect AP administration page Connect AP administration page

Collect MAC address, Collect address, network nameMAC and signal strength network name and signal strength

Yes Yes

Is there MAC Is address? there MAC address?

No No

No No

Signal Signal strength ≥ 80 ? strength ≥ 80 ? Yes Yes

Occupied Occupied

Unoccupied Unoccupied

Time delay Time delay

Figure 6. Occupant presence check algorithm. Figure6.6.Occupant Occupant presence presence check Figure checkalgorithm. algorithm.

Figures 7 and 8 show an experiment on the plug-in load saving and explains why the occupancy Figures 7 and 8 show an experiment on the plug-in load saving and explains why the occupancy signal intensity set atan 80%. Two experimental locations the second floor of why the experimental Figures 7 andwas 8 show experiment on the plug-in loadon saving and explains the occupancy signal intensity was set at 80%. Two experimental locations on the second floor of the experimental place used the same AP. In cases of AP1 and AP2, the signal intensity was reduced to 80%, when signal intensity was set at 80%. Two experimental locations on the second floor of the experimental place place used the same AP. In cases of AP1 and AP2, the signal intensity was reduced to 80%, when occupants moved 11cases m away from theAP2, AP with the smartphone in the hand. to In 80%, case of AP 3occupants on the used the same AP. In of AP1 and the signal intensity was reduced when occupants moved 11 m away from the AP with the smartphone in the hand. In case of AP 3 on the fourth11floor, the signal intensity wasthe reduced to 80%,inwhen occupants moved 153m from the moved m away from the AP with smartphone the hand. In case of AP onaway the fourth floor, fourth floor, the signal intensity was reduced to 80%, when occupants moved 15 m away from the AP with the smartphone. In the case of AP3, the signal intensity of 90% was considered appropriate, theAP signal intensity was reduced to 80%, when occupants moved 15 m away from the AP with with the smartphone. In the case of AP3, the signal intensity of 90% was considered appropriate,the but this study setcase up theAP3, signal at 80% of in order to considered simplify the plug-in device control smartphone. In the theintensity signal intensity was appropriate, but this study but this study set up of the signal intensity at 80% in90% order to simplify the plug-in device control software. Therefore, after the Wifi was activated by occupants, when they moved away by about setsoftware. up the signal intensity at 80% in order simplify by theoccupants, plug-in device control software. Therefore, after Therefore, after the Wifi wastoactivated when they moved away by about 11–15 m from the experimental locations, they were determined to be absent in the rooms, but when m from the experimental locations, determined to be absent rooms, but when the11–15 Wifi was activated by occupants, when they they were moved away by about 11–15 in m the from the experimental the occupants moved within 11–15 m from the assess point of the experimental locations, they were the occupants moved within 11–15 m from theinassess point of thewhen experimental locations, they were locations, they were determined to be absent the rooms, but the occupants moved within determined to be present in the rooms. When applying the methodology of this study to other determined to be present When applying of this to other 11–15 m from the assess pointinofthe therooms. experimental locations, the theymethodology were determined to study be present in the buildings, signal intensity criteria can vary according to the number and placement of obstacles, the buildings, intensity criteria can vary according number and placement of obstacles, thecan rooms. Whensignal applying the methodology of this study to tothe other buildings, signal intensity criteria performance of Wifi access point and etc. performance of Wifi access point and etc. vary according to the number and placement of obstacles, the performance of Wifi access point and etc.

Energies 2016, 9, 143

6 of 15

Energies 2016, 9, 143 Energies 2016, 9, 143

6 of 15 6 of 15 60% 60% 65% 65% 70% 70% 75%

Signal strength Signal strength

75% 80% 80% 85% 60% 85%

AP2

AP1

AP2

AP1

60% 65% 65% 70% 70% 75% 75% 80% 80% 85%

Signal strength Signal strength

85%

Figure 7. Experimental location (second floor).

Figure location(second (secondfloor). floor). Figure7.7.Experimental Experimental location 70% 70% 75% 75% 80% 80% 85% 85%

Signal strength Signal strength

90% 90% 95% 95%

AP3 AP3

Figure 8. Experimental location (fourth floor). Figure8.8.Experimental Experimental location Figure location(fourth (fourthfloor). floor). 3.1.2. Plug-in Device Control 3.1.2. Plug-in Device Control 3.1.2. Plug-in Device Control plug-in device control, the OBC-P software and the standby power For the occupant-based For the occupant-based device control,power the OBC-P software and the power cut-off device seen in Figure 9 plug-in are used. The standby cut-off device consists of standby the Always For the occupant-based plug-in device control, the OBC-P software and the standby power ON cut-off cut-off device seen in and Figure 9 are used. The standby power cut-off device consists of the Always ON outlet, Master outlet, Controlled outlets. device seen in Figure 9 are used. The standby power cut-off device consists of the Always ON outlet, outlet, Master outlet, Controlled The Always ONand outlet does notoutlets. have the power cut-off function as ordinary socket outlets. Master outlet, and Controlled outlets. The Always ON outlet does not have it theconsumes power cut-off function Therefore, if a plug-in device is activated electricity, but as if ordinary a plug-insocket deviceoutlets. is not The Always ON outlet doesis not have the power cut-off function asa ordinary socketis outlets. Therefore, if a plug-in device activated it consumes electricity, but if plug-in device not activated, it consumes standby power. Therefore, if aitplug-in device is activated it consumes electricity, but if a plug-in device is not activated, activated, consumes standby power. The Master outlet also does not have the power cut-off function as ordinary socket outlets. it consumes standby power. The Master outlet does not have depending the power on cut-off function as ordinary socket outlets. However, it regulates thealso controlled outlets, whether a plug-in device connected to the However, it regulates thedoes controlled outlets, depending onfunction whetheras a plug-in device connected to the The Master outlet also not have the power cut-off ordinary socket outlets. However,

it regulates the controlled outlets, depending on whether a plug-in device connected to the Master outlet

Master outlet is activated or not. In other words, if a plug-in device connected to the Master outlet is activated, it supplies electricity to the Controlled outlets, but if a plug-in device connected to the Master outlet is not activated, it cuts off electricity supplied to the Controlled outlets (standby power cut-off). Energies 2016, 9, 143 7 of 15 The mechanism of ordinary standby power cut-off devices is that a computer is plugged into the Master outlet, while monitors and computer peripheral devices, including personal lighting, are is activatedinto or not. other words, if a Therefore, plug-in device Master is activated, it connected the In Controlled outlets. if theconnected computertoisthe turned off,outlet the standby power supplies electricity to the Controlled outlets, but if a plug-in device connected to the Master outlet is supplied to computer peripheral devices will be cut off. However, this method cannot reduce plug not activated, it cuts off electricity supplied to the Controlled outlets (standby power cut-off). loads, when the computer is turned on. Controlled outlets

Master outlet

Always ON outlet

Figure 9. 9. Cutoff Cutoff device. device. Figure

In order to control unnecessary plug-in devices when occupants are not present, this study The mechanism of ordinary standby power cut-off devices is that a computer is plugged into connected the computer to the Always ON outlet, the main monitor to the Master outlet, and the Master outlet, while monitors and computer peripheral devices, including personal lighting, are computer peripheral devices including the sub-monitor, personal lighting, and speakers to the connected into the Controlled outlets. Therefore, if the computer is turned off, the standby power Controlled outlets. Therefore, if the power of the main monitor is turned off, it cuts off the standby supplied to computer peripheral devices will be cut off. However, this method cannot reduce plug power supplied to the computer peripheral devices connected to the Controlled outlets. loads, when the computer is turned on. In the plug-in control methods of the OBC-P software, this study developed the Awake mode, In order to control unnecessary plug-in devices when occupants are not present, this study Sleep mode, and control off. In the Awake mode (as seen in Figure 10), if an occupant leaves the office connected the computer to the Always ON outlet, the main monitor to the Master outlet, and computer room for a short period of time, the computer consumes the normal electricity, and the main monitor peripheral devices including the sub-monitor, personal lighting, and speakers to the Controlled outlets. consumes standby power, while the supply of standby power to computer peripheral devices is cut Therefore, if the power of the main monitor is turned off, it cuts off the standby power supplied to the off. Also, before the occupant enters the office room, the main monitor and computer peripheral computer peripheral devices connected to the Controlled outlets. devices are activated. Therefore, it can reduce plug loads of the main monitor and computer In the plug-in control methods of the OBC-P software, this study developed the Awake mode, Sleep peripheral devices while reducing the inconvenience of the occupant. Figure 11 shows the algorithm mode, and control off. In the Awake mode (as seen in Figure 10), if an occupant leaves the office room for of the Awake mode. When the OBC-P software is set at the Awake mode, it collects real-time a short period of time, the computer consumes the normal electricity, and the main monitor consumes occupant information and checks whether an occupant is present or absent in the room; if the standby power, while the supply of standby power to computer peripheral devices is cut off. Also, before occupant is present, the OBC-P software activates the main monitor and the computer peripheral the occupant enters the office room, the main monitor and computer peripheral devices are activated. devices by disabling the standby power cut-off device. If the occupant is absent, the OBC-P software Therefore, it can reduce plug loads of the main monitor and computer peripheral devices while reducing turns off the power of the main monitor and cut off the standby power to the computer peripheral the inconvenience of the occupant. Figure 11 shows the algorithm of the Awake mode. When the OBC-P devices by enabling the standby power cut-off device. software is set at the Awake mode, it collects real-time occupant information and checks whether an In the Sleep mode (as seen in Figure 12), when the occupant leaves the office room for a short occupant is present or absent in the room; if the occupant is present, the OBC-P software activates the period of time, the computer switches to the saving mode, and the main monitor consumes standby main monitor and the computer peripheral devices by disabling the standby power cut-off device. If the power, while the electricity supply to computer peripheral devices is cut off. Therefore, it can save occupant is absent, the OBC-P software turns off the power of the main monitor and cut off the standby more plug loads than the Awake mode. But when the occupant returns to the office room, the power to the computer peripheral devices by enabling the standby power cut-off device. occupant must click the mouse or press the keyboard once to reactivate the plug-in devices connected In the Sleep mode (as seen in Figure 12), when the occupant leaves the office room for a short period to the independent socket outlet and the dependent socket outlets. Figure 13 shows the algorithm of of time, the computer switches to the saving mode, and the main monitor consumes standby power, the Sleep mode. while the electricity supply to computer peripheral devices is cut off. Therefore, it can save more plug If it is set in the Sleep mode, the OBC-P software collects real-time occupant information, and if loads than the Awake mode. But when the occupant returns to the office room, the occupant must click it determines the absence of the occupant, it switches the power of the computer to the saving mode. the mouse or press the keyboard once to reactivate the plug-in devices connected to the independent After that, the standby power cut-off device discontinues the supply of standby power to computer socket outlet and the dependent socket outlets. Figure 13 shows the algorithm of the Sleep mode. peripheral devices. If an occupant is present, the OBC-P software detects a movement of the mouse If it is set in the Sleep mode, the OBC-P software collects real-time occupant information, and if it or a button click of the keyboard. Only when the mouse moves or the keyboard button is clicked, the determines the absence of the occupant, it switches the power of the computer to the saving mode. OBC-P software activates the monitor and computer. After that, the standby power cut-off device discontinues the supply of standby power to computer peripheral devices. If an occupant is present, the OBC-P software detects a movement of the mouse or a button click of the keyboard. Only when the mouse moves or the keyboard button is clicked, the OBC-P software activates the monitor and computer.

Energies 2016, 9, 143 8 of 15 Energies 2016, 9, 143 8 of 15 Energies 2016, 9, 143 8 of 15 The Control-off mode is a control model that does not regulate plug-in devices regardless the Energies 2016, 9, 143 8of ofthe 15 The Control-off mode is a control model that does not regulate plug-in devices regardless of

absenceThe or presence presence of ofmode an occupant. occupant. Control-off is a control model that does not regulate plug-in devices regardless of the absence or an absence or presence of an occupant. Before arrival Before arrival Before arrival

After leave After leave

O : Power dissipation O : Power dissipation

△: Standby power X : Standby power cutoff △: Standby power X : Standby power cutoff

After leave

△: Standby power X : Standby power cutoff

O : Power dissipation Computer peripheral Monitor Computer Computer peripheral Monitor Computer

Computer peripheral Monitor Computer Computer peripheral Monitor Computer

Computer peripheral Monitor Computer

O O

O

O O

O

O O

O

O O

Computer peripheral Monitor Computer

O O O

O

X

X X

X

X X

X

X X

△ O △ O O

△

Figure 10. 10. Awake Awake mode. mode. Figure Figure mode. Figure 10. 10. Awake mode. awake mode awakemode mode awake

Occupant OccupantInformation Information Collection Collection Occupant Information Collection

Yes Yes Yes

Occupied? Occupied? Occupied?

No No No

byby OBC-P software OBC-P software

by OBC-P software

by cutoff device Monitor ON Monitor ON (Power Consumption) Monitor ON

by cutoff device by cutoff device

Monitor OFF Monitor OFF (Standby MonitorPower) OFF

(Power Consumption) (Power Consumption)

(Standby Power) (Standby Power)

Computer Peripheral ON Computer Peripheral ON (PowerPeripheral Consumption) Computer ON (Power Consumption) (Power Consumption)

Computer Peripheral OFF Computer Peripheral OFF (Cutoff Standby Power) Computer Peripheral OFF (Cutoff Standby Power) (Cutoff Standby Power)

Time delay Time delay Time delay

Figure11. 11. Awake Awake mode mode algorithm. Figure algorithm. Figure 11. Awake mode algorithm. Figure 11. Awake mode algorithm.After leave

Before arrival

△: Standby power X : Standby power cutoff

O : Power dissipation

Before arrival Before arrival

After leave After leave

O : Power dissipation O : Power dissipation Computer peripheral Monitor Computer

Computer peripheral Monitor Computer Computer peripheral Monitor Computer

O

O O

O

O O

O

O O

O

O O

△: Standby power X : Standby power cutoff △: Standby power X : Standby power cutoff

Computer peripheral Monitor Computer

Computer peripheral Monitor Computer Computer peripheral Monitor Computer

O

X

O O

X

X X

Figure 12. Sleep mode.

Figure 12. Sleep mode. Figure 12. 12. Sleep Figure Sleep mode. mode.

X

X X

△ △ X X

△ △ △ △

Energies 2016, 9, 143 Energies 2016, 9, 143

9 of 15 9 of 15

sleep mode

Occupant Information Collection

Time delay

No

Yes

Occupied?

Any Move? (Mouse or Keyboard)

No

Monitor OFF (Standby Power)

Yes

by OBC-P software by cutoff device

Monitor ON (Power Consumption) Computer Standby Mode (Standby Power) Computer ON (Power Consumption)

Computer Peripheral ON (Power Consumption)

Computer Peripheral OFF (Cutoff Standby Power)

Time delay

Figure 13. 13. Sleep Sleep mode mode algorithm. algorithm. Figure

3.2. Selection of Representative Value The Control-off mode is a control model that does not regulate plug-in devices regardless of the In order to estimate load saving quantities of the occupant-based plug-in device control in absence or presence of anplug occupant. each control mode, the same occupancy schedule was applied for comparative analysis. Therefore, 3.2. Selection ofconsumption Representativeof Value the electricity each occupant’s plug-in devices in four control modes (when all the plug-in devices activated, an occupant leaves room for a short period of control time after In order to are estimate plugwhen load saving quantities of the theoffice occupant-based plug-in device in setting the control to Awake mode, an occupant leaves the office room after setting the control to each control mode, the same occupancy schedule was applied for comparative analysis. Therefore, Sleep mode, or when all plug-in are plug-in deactivated) were measured times before the the electricity consumption of eachdevices occupant’s devices in four control500 modes (when all the experiment, and these were selected as the representative values of plug loads. plug-in devices are activated, when an occupant leaves the office room for a short period of time after Figures 14 and showmode, the representative valuesthe of plug were measured setting the control to15 Awake an occupant leaves officeloads roomwhich after setting the control500 to times Sleep at an interval seconds, when all the plugwere devices were activated to the presence of mode, or whenofallthree plug-in devices are deactivated) measured 500 timesdue before the experiment, Occupant A andselected Occupant B, and when theyvalues were absent short leave after setting the control to and these were as the representative of plugfor loads. the Awake mode. In case of Occupant A, the mean value of plug loads is similar the median Figures 14 and 15 show the representative values of plug loads which were to measured 500value. times Therefore, if either the mean value or the value is setwere as the representative there was at an interval of three seconds, when allmedian the plug devices activated due to value, the presence of not much difference in the total electricity consumption. However, when Occupant B was present Occupant A and Occupant B, and when they were absent for short leave after setting the control and theAwake plug-in devices activated, there a difference by plug aboutloads 10W is between mean value to the mode. In were case of Occupant A, was the mean value of similarthe to the median and theTherefore, median value. In this either the value orvalue the median has to be selected as value. if either thecase, mean value ormean the median is set asvalue the representative value, the representative value. If the mean value is used as the representative value of plug loads, it is likely there was not much difference in the total electricity consumption. However, when Occupant B was to overestimate electricity consumption. The there meanwas value tends to show a greater present and the the plug-in devices were activated, a difference by about 10 Wresponse between to thea more distant value from the center [9]. Therefore, this study used the median value as mean value and the median value. In this case, either the mean value or the median value has tothe be representative selected as the value. representative value. If the mean value is used as the representative value of plug loads, it is likely to overestimate the electricity consumption. The mean value tends to show a greater response to a more distant value from the center [9]. Therefore, this study used the median value as the representative value.

Energies 2016, 9, 143 Energies 2016, 9, 143 Energies 2016, 9, 143

10 of 15 10 of 15 10 of 15 Occupant A, occupied Occupant A, occupied

60 60

Histogram Histogram Normal distribution Normal distribution mean mean median median

50 50

Histogram Histogram Normal distribution Normal distribution mean mean median median

60 60 50 50

Frequency Frequency

40 40

Frequency Frequency

Occupant A, unoccupied(awake mode) Occupant A, unoccupied(awake mode)

70 70

30 30

20 20

40 40 30 30 20 20

10 10

10 10

00 175 175

180 180

185 185

190 190

195 195 Watt Watt

200 200

205 205

210 210

00 86 86

215 215

88 88

90 90

92 92

94 94

Watt Watt

96 96

98 98

100 100

102 102

104 104

Figure plug load. Figure 14. 14. Occupant A, representative representative value of Figure 14. Occupant A, representative value value of of plug plugload. load. Occupant OccupantB, B,occupied occupied

300 300

250 250

250 250

200 200

200 200 Histogram Histogram Normal Normal distribution distribution mean mean median median

150 150

Frequency Frequency

Frequency Frequency

Occupant B, unoccupied(awake unoccupied(awake mode) mode) Occupant B,

300 300

100 100

100 100

50 50

50 50

00 140 140

160 160

180 180

200 200 Watt Watt

220 220

240 240

260 260

Histogram Histogram Normal distribution distribution Normal mean mean median median

150 150

0 73

74

75

76

77

78 Watt

79 79

80 80

81 81

82 82

83 83

Figure Figure 15. 15. Occupant Occupant B, B, representative Figure 15. Occupant B, representative representative value value of of plug plugload. load.

4.4.Plug PlugLoad Load Saving Experiment 4. Plug LoadSaving SavingExperiment Experiment Thisstudy studyconducted conducted an an experiment experiment to to measure This study conducted due to plug-in This an experiment to measure the the plug plug load load saving savingquantities quantitiesdue dueto toplug-in plug-in device control. Table 1 shows the occupants experiment and the types device control. Table 1 shows the occupants who participated in the types of the device control. Table 1 shows the occupants who participated in the experiment and the typesof ofthe the used plug-in devices. Figures 7 and 8 display The experiment was conducted used plug-in devices. Figures 7 and 8 display the experimental places. was conducted used plug-in devices. Figures 7 and 8 display the experimental places. The experiment was conducted bytargeting targetingaaatotal totalof of10 10occupants occupants in in three three research research labs by targeting total of 10 occupants in three research labs at at University University of weeks. by of Seoul Seoul for for two two weeks. weeks. Table 1. 1. Occupants Occupants and Table 1. Table Occupants and plug-in plug-in devices. devices.

Laboratory Laboratory

Occupant Computer Computer Notebook Monitor Personal Lighting Occupant Lighting Occupant Computer Notebook Monitor Personal Lighting A 1 2 1 A 1 A 1 2 1 1 BB B 111 2 2 1 C 1 2 C C 11 2 1 1 D 1 1 1 D 1 D 11 1 E 1 2 1 E 1 2 1 E 1 F 1 2 1 1 FF 11 2 G 1 2 1 GH 2 1 G 111 2 2 1 1 H 1 2 1 H I 11 12 2 1 1 1 2 II J 1111 - 2 1 11 J 1 J 10 -9 21 17 10 11 10 17 10 10 99 22 17 10

Laboratory

11

1

2

22

33

3 Total

Total Total

The equation to calculate the plug load saving quantity was established in reference to Option A Theequation equation to to calculate calculate the the plug plug load load saving saving quantity quantity was was established established in to Option A in reference reference of theThe IPMVP (International Performance Measurement and Verification Protocol) [10]. to Option A ofthe theIPMVP IPMVP(International (International Performance Performance Measurement Measurement and and Verification Verification Protocol) of Protocol) [10]. [10].

The plug load saving quantity was calculated by multiplying the Baseline Representative Value ( BRV𝑖 ) with the Baseline Period ( BP𝑖 ) and then by subtracting the multiplied value of the Representative Value (𝐶𝑅𝑉𝑖 ) and the Control Period (𝐶𝑃𝑖 ) from it. Herein, “i” is the control mode of plug-in devices (i = 1: when all the plug-in devices are activated, i = 2: when an occupant leaves the Energies 2016, 9,for 143 a short period of time after setting the control to Awake mode, i = 3: an occupant 11 of 15 office room leaves the office room for a short period of time after setting the control to Sleep mode, and i = 4: when all plug-in devices of an occupant are deactivated). 5. Results and Discussion Plug Load Saving “

4 ÿ

pBRVi ˆ BPi ´ CRVi ˆ CPi q

(1)

i “1

To confirm the accuracy of the occupancy OBC-P software, experiments of the case of leaving The plug was the calculated by multiplying the Baseline Representative Value the room andload the saving case ofquantity occupying room with a smart phone were performed 20 times, (BRV ) with the Baseline Period (BP ) and then by subtracting the multiplied value of the Representative i i respectively. As a result of 100% accuracy, plug-in devices have been activated before the occupants Value (CRV ) and the Control Period (CP ) from it. Herein, “i” is the control mode of plug-in devices i i occupy the room and deactivated after the occupants are out. (i = 1:Figures when all the plug-in devices are activated, i = 2: when an occupant leaves the office for 16–21 show the results of the experiment on the occupant-based plug-in deviceroom control, a short period of time after setting the control to Awake mode, i = 3: an occupant leaves the office room which was conducted by targeting a total of 10 occupants for two weeks. The baseline system was set for follows: a short period time aftercame setting control to turned Sleep mode, and i = devices, 4: when all plug-in devices of as after of occupants to the work, they on plug-in such as computers, an occupant are deactivated). monitors and computer peripheral devices; during the short leave period, they did not control the plug-in devices; when they went to home, they turned off the plug-in devices. 5. Results and Discussion Figure 16 shows the plug load of Occupant A on Wednesday. Occupant A came to work at To confirm the accuracy the occupancy of theleaves case ofbetween leaving the the around 10 AM and went to of home at around OBC-P 2 PM, software, and thereexperiments were five short room and the case of occupying the room with a smart phone were performed 20 times, respectively. clock-in time and the clock-out time. When Occupant A used the occupant-based plug in device As a result 100% accuracy, plug-in devices have been activated theleave occupants occupy the control, theof plug-in devices were automatically controlled during before the short period, and as a room and deactivated after the occupants are out. results, unnecessary plug loads were reduced. When the occupant set the control to the Awake mode, Figures 16–21 the results of the experiment on baseline the occupant-based plug-in device it could reduce the show plug loads by 10.2%, compared to the system. In case of the Sleepcontrol, mode, which was the conducted by by targeting it can save plug loads 18.2%. a total of 10 occupants for two weeks. The baseline system was set asFigure follows: to work, turned onAplug-in as computers, 17 after showsoccupants the total came occupied time they of Occupant for twodevices, weeks such and the plug load monitors and computer peripheral devices; during the short leave period, they did not control quantities. In case of Occupant A, the short leave period and the period from the clock-out timethe to plug-in they went to home, they turned devices. the nextdevices; clock-inwhen time varied widely depending on the off daythe of plug-in the week, and the plug load quantities Figure 16 shows the of aOccupant A on Wednesday. Occupant A came to work around changed accordingly. In plug mostload cases, longer unoccupied time means a smaller plug load at quantity. 10 A.M. and went to home at around 2 P.M., and there were five short leaves between the clock-in time However, in cases where the unoccupied time was similar (Tuesday, Wednesday and Thursday of andsecond the clock-out When Occupant A in used occupant-based in device control, the plug-in the week),time. the plug load quantity thethe Awake mode was plug decreased when the period from devices weretime automatically the increased. short leaveIn period, and as a results, plug the clock-in to the nextcontrolled clock-outduring time was the Sleep mode, whenunnecessary the unoccupied loadswas wereincreased, reduced. When theload occupant set the control to the Awake mode, show it could reduce the plug time the plug quantity was decreased. These results the differences inloads plug by 10.2%, compared to the baseline system. mode In caseduring of the Sleep mode, it can save the plug loads by 18.2%. load quantities depending on the control the short leave period, as seen in Figure 16. Occupied Unoccupied 200 180 160 140

Watt

120 100 70 60 40 20 0 0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

Time of day

Figure Figure 16. 16. Occupant Occupant A, A, plug plug load load (Wednesday). (Wednesday).

Figure 17 shows the total occupied time of Occupant A for two weeks and the plug load quantities. In case of Occupant A, the short leave period and the period from the clock-out time to the next clock-in time varied widely depending on the day of the week, and the plug load quantities changed accordingly. In most cases, a longer unoccupied time means a smaller plug load quantity. However, in cases where the unoccupied time was similar (Tuesday, Wednesday and Thursday of the second week), the plug load quantity in the Awake mode was decreased when the period from the clock-in

Energies 2016, 9, 143

12 of 15

time to the next clock-out time was increased. In the Sleep mode, when the unoccupied time was increased, the plug load quantity was decreased. These results show the differences in plug load quantities depending on the control mode during the short leave period, as seen in Figure 12 16. Energies 12 Energies2016, 2016,9,9,143 143 of of 1515 3.03.0

24.0 24.0

12 of 15 20.0 24.0 20.0

2.02.02.5

16.0 20.0 16.0

24.0 24.0

24.0 24.0

24.0

24.0

13.3 13.3

13.4 13.4

13.9 13.9

1.01.5 1.0

12.6 12.6

11.9 11.9

14.6 14.6

14.2 14.2

13.6 13.6

24.0

16.0 12.0

12.0

24.0

12.0 8.0

8.0

13.3 11.9

13.9

13.6 14.6

14.2

13.3

12.6

0.51.0 0.5

13.3 13.3

24.0 24.0

16.8 16.8

13.4

8.0 4.0

16.8

3.3

3.2

4.0

3.4

3.3

3.2

0.00.5 0.0

24.0 24.0

3.4

1.7

3.0 1.5

1.7

3.0

1.5

1.5

1.5

Mon Tue Wed Thu Fri Sat Sun Mon Tue Mon Tue Wed Thu Fri Sat Sun Mon Tue short leave 3.3 period the period time 3.4 from the clock-out time to the next clock-in 3.2 3.0 1.7 the period from the short leave period 1.5clock-out time to the next clock-in time 1.5

0.0 Mon

Time [hour]

1.52.0 1.5

Time [hour] Time [hour]

2.52.53.0

Plug load [kWh] Plug load [kWh]

Plug load [kWh]

Energies 2016, 9, 143

1.8

1.8

2.7

2.7

Wed Thu Wed Thu existing system 2.7 1.8 existing system

2.0

2.0

Fri Sat Fri Sat awake mode 2.0 awake mode

Tue Wed Thu Fri Sun unoccupied Mon Tue Wed Thu plug Fri Sat Figure 17. Occupant A,Sat daily time and load.

4.0 0.0 Sun 0.0 Sun sleep mode sleep mode 0.0 Sun

Figure 17. Occupant A, unoccupied time andplug plug load. Occupant A,daily daily unoccupied time and load. shortFigure leave period 17. the period from the clock-out time to the next clock-in time existing system awake mode sleep mode

3.0 3.0 3.0 2.5 2.5

Figure 17. Occupant A, daily unoccupied time and plug load. system Occ_A_existing

Mon Wed Mon Fri Wed Fri

2.5 2.0 2.0

load [kWh] Plug [kWh] Plug load

Plug load [kWh]

Occ_A_existing system Occ_A_awake mode Occ_A_awake mode Occ_A_sleep mode Occ_A_existing system Occ_A_sleep mode Occ_A_awake mode Occ_A_sleep mode

Mon Tue Mon Thu Tue Thu Wed Thu Thu Mon Wed Tue Mon Tue Wed Fri Thu Tue Thu Wed Tue

Fri

Fri

2.0 1.5 1.5

R²= 0.9771 Fri

1.5 1.0 1.0

R²= 0.9771 R²= 0.9934 R²= 0.9771R²= 0.9934 R²= 0.9934

1.0 0.5

R²= 1

R²= 1 R²= 1

0.5

Weekend

Weekend Weekend

0.5 0.0 0.0 14 0.014

15

18 19 20 21 22 23 Unoccupied day 18 19time of20 21 22 23 14 15 16 17 18 19 time of 20day 21 22 23 Unoccupied Figure 18. Occupant A, unoccupied Unoccupied time oftime day and plug load.

15

16

17

16

17

24

24 24

Figure18. 18.Occupant OccupantA, A, unoccupied time and plug load. Figure timeand and plug load. Figure 18. Occupant A,unoccupied unoccupied time plug load.

3.0

Occ_A_existing system Occ_A_awake mode Occ_A_existing system Occ_A_existing system Occ_A_sleep mode Occ_A_awake mode Occ_A_awake mode Occ_B_existing system Occ_A_sleep mode Occ_A_sleep mode Occ_B_awake mode Occ_B_existing system Occ_B_existing system Occ_B_sleep mode Occ_B_awake mode Occ_B_awake mode Occ_B_sleep mode Occ_B_sleep mode

3.0

3.0 2.5

Plug load [kWh] load [kWh] Plug [kWh] Plug load

2.52.5 2.0 2.02.0 1.5

1.51.5 1.0 1.01.0 0.5 0.50.5 0.0 14 0.00.0 14 14

15

16

17

18 19 20 21 22 23 Unoccupied time of day 15 16 17 18 19 20 21 22 23 15 16 17 18 19 20 21 22 23 Unoccupied time of day time and plug load. Figure 19. Occupant A, Occupant B, unoccupied Unoccupied time of day Figure 19. Occupant A, Occupant B, unoccupied time and plug load.

Figure 19.19.Occupant unoccupiedtime timeand andplug plug load. Figure OccupantA, A,Occupant Occupant B, unoccupied load.

24 24

24

Energies 2016, 9, 143 Energies 2016, 9, 143 Energies 2016, 9, 143

13 of 15 13 of 15 13 of 15

25 25

awake mode awake mode sleep mode sleep mode

Frequency Frequency

20 20

15 15

10 10

5 5

0 0

0~5 0~5

5~10 5~10

10~15 10~15

15~20 15~20

20~25 20~25

25~30 30~35 35~40 40~45 25~30 30~35 35~40 40~45 Plug load saving rate (%) Plug load saving rate (%)

45~50 45~50

50~55 50~55

55~60 55~60

60~65 60~65

65~70 65~70

Figure Plug Figure 20. 20. Plug load load saving saving rate. rate. Figure 20. Plug load saving rate. 180 180

15% Reduction 15% Reduction

160 160

26% Reduction 26% Reduction

kWh/(10persons∙2weeks) kWh/(10persons∙2weeks)

140 140 120 120 100 100 80 80

165.02 165.02 140.19 140.19

60 60

121.93 121.93

40 40 20 20 0 0

existing system existing system

awake mode awake mode

sleep mode sleep mode

Figure 21. Comparison of plug load. Figure Figure 21. 21. Comparison Comparison of of plug plug load. load.

Figure 18 shows the plug loads of Occupant A by the unoccupied hours for two weeks. On Figure 18 shows the plug loads of Occupant A by the unoccupied hours for two weeks. On Figureand 18 Sundays, shows the loadsdid of not Occupant by the unoccupied hours for two Saturdays theplug occupant come toAwork, so the plug log quantity was theweeks. same. Saturdays and Sundays, the occupant did not come to work, so the plug log quantity was the same. On Saturdays and Sundays, the occupant did not come to work, the plug log theweek. same. Except for Saturdays and Sundays, the unoccupied time variedsodepending onquantity the daywas of the Except for Saturdays and Sundays, the unoccupied time varied depending on the day of the week. Exceptthe forunoccupied Saturdays and the unoccupied time quantity varied depending on theAlso, day of When timeSundays, was increased, the plug load was decreased. in the caseweek. that When the unoccupied time was increased, the plug load quantity was decreased. Also, in case that When the unoccupied time was increased, themode plug load quantity was decreased. Also,followed in case that the unoccupied time was the same, the Sleep had the greatest plug load saving, by the unoccupied time was the same, the Sleep mode had the greatest plug load saving, followed by unoccupied the same, the Sleep mode hadbaseline the greatest plug load saving, followedwas by the Awake modetime and was the baseline system. In case of the system, the plug load quantity the Awake mode and the baseline system. In case of the baseline system, the2 plug load quantity was the Awakeinmode theeven baseline system. In case of the baseline system, the load quantity was increased someand cases, though the unoccupied time was increased (𝑅 2 plug = 0.9771). increased in some cases, even though the unoccupied time was increased (𝑅2 = 0.9771). increased in somethe cases, evenmode though the unoccupied time was decreasing increased (Rin “ In contrast, Awake showed a more remarkable the0.9771). plug load quantity In contrast, the Awake mode showed a more remarkable decreasing in the plug load quantity Awakewhen mode the showed a more remarkable decreasing in the plug load quantity than In thecontrast, baselinethe system, occupied time increased (𝑅22 = 0.9934). In the Sleep mode,than an than the baseline system, when the occupied time increased (𝑅 = 0.9934). In the Sleep mode, an 2 “ 0.9934). In the Sleep mode, an2increase in the baseline system, when the occupied time increased (R increase in the unoccupied time always led to a decrease in the plug load quantity (𝑅 2 = 1). The increase in the unoccupied time always led to a decrease in the plug load quantity (𝑅 = 1). The the unoccupied time to a decrease in themode plug load quantity (R2to“differences 1). The varying varying results of thealways Awakeled mode and the Sleep can be ascribed in theresults short varying results of the Awake mode and the Sleep mode can be ascribed to differences in the short of theperiod Awakeand mode the Sleepperiod mode from can bethe ascribed to differences theclock-in short leave and leave theand unoccupied clock-out time to the in next time,period as seen in leave period and the unoccupied period from the clock-out time to the next clock-in time, as seen in the unoccupied period fromthe theAwake clock-out time to the next load clock-in time,W as during seen in the Figure 16.leaves In other Figure 16. In other words, mode used a plug of 94.34 short of Figure 16. In other words, the Awake mode used a plug load of 94.34 W during the short leaves of words, the and Awake mode used plug of 94.34 W during the short of occupants and a plug occupants a plug load of a2.39 Wload for an unoccupied period from leaves the clock-out time to the next occupants and a plug load of 2.39 W for an unoccupied period from the clock-out time to the next load of 2.39 W for an unoccupied period from the clock-out time to of the4.05 nextWclock-out time. In contrast, clock-out time. In contrast, the Sleep mode consumed a plug load during the short leave of clock-out time. In contrast, the Sleep mode consumed a plug load of 4.05 W during the short leave of the Sleep mode consumed plugofload 4.05 W duringwhen the short leave of occupants, and deactivated used a plug occupants, and used a plugaload 2.39of during a period all the plug-in devices were occupants, and used a plug load of 2.39 during a period when all the plug-in devices were deactivated by occupants. Therefore, the use of the Sleep mode enables the reduction of plug loads in proportion by occupants. Therefore, the use of the Sleep mode enables the reduction of plug loads in proportion to the unoccupied time. to the unoccupied time.

Energies 2016, 9, 143

14 of 15

load of 2.39 during a period when all the plug-in devices were deactivated by occupants. Therefore, the use of the Sleep mode enables the reduction of plug loads in proportion to the unoccupied time. Figure 19 shows the plug load by hour during the unoccupied period of Occupant A and Occupant B. Even though Occupant A and Occupant B had similar unoccupied periods of time, their normal use power consumption and standby power consumption of computers, monitors and computer peripheral devices were different, and their short leave period and their unoccupied period from the clock-out time to the next click-in time were also different, leading to differences in the plug load quantities among occupants. Figure 20 shows the plug load saving rates of 10 occupants during the weekday depending on the varying control modes. The Awake mode had the daily minimum plug load saving rate of 1.39%, the daily maximum plug load saving rate of 39.80%, and the average plug load saving rate of 18%. The Sleep mode saves a daily minimum plug load of 1.55%, the daily maximum plug load of 65.76%, and the average plug load of 26.93%. The reason why the Sleep mode had a higher average plug load saving rate than the Awake mode was that occupants additionally controlled the computer to the power save mode when they left the rooms for short leave. Figure 21 shows the comparison of plug load quantities when 10 occupants used the occupant-based plug-in device control for two weeks. The Awake mode had a higher plug load saving rate by about 15% than the baseline system, while the Sleep mode could reduce the plug loads by about 26% compared to the baseline system. 6. Conclusions and Future Directions This study was conducted to develop a way of reducing unnecessary plug loads while minimizing the inconvenience of occupants by using an occupant-based automatic control of computers, monitors and computer peripheral devices, all of which account for approximately 95% of plug loads of an office building. To this end, the study developed the OBC-P software which uses the Wifi of smartphones and standby power cutoff devices. The OBC-P software collects information about occupants who are connected to the AP administration page through the Wifi, and based on this information, it controls the power on/off of computers and monitors, while a standby power cutoff device is used to control the power of computer peripheral devices. In order to measure the plug load saving rate in case of the use of the occupant-based plug-in device, the experiment was conducted targeting 10 occupants in three research labs of the graduate school for two weeks. The results of the study can be summarized as follows: (1)

(2)

When an occupant went out for a short break, the Awake mode automatically switched the main monitor to the standby mode and interrupted the supply of standby power to computer peripheral devices. Also, it reactivated the main monitor and computer peripheral devices right before the occupant returned to the room. The Awake mode could reduce the plug loads of computers, monitors, and computer peripheral devices by 15% compared to the baseline system. When an occupant left the room for a short period of time, the Sleep mode automatically switched the main monitor to the standby mode and interrupted the supply of standby power to computer peripheral devices. Thus, since it switched the computer to the saving mode, more plug loads could be saved than with the Awake mode. However, the computer, main monitor and computer peripheral devices were reactivated only when the occupant moved the mouse or touched the keyboard button. The Sleep mode could reduce the plug loads of computers, monitors, and computer peripheral devices by 26% compared to the baseline system.

This study has its limitations. Some test subjects likely felt that the computer gets slow when they run OBC-P software for a long time (10 h or more). It is estimated to be associated with the performance of the computer. It was difficult to collect correct information about the presence or absence of occupants, with increasing distance between the occupants and their smartphones. Also, when an occupant was within the range to the wireless access point but had left his or her workspace, the plug-in devices were not controlled. Further studies need to be conducted using a sensor which

Energies 2016, 9, 143

15 of 15

can locate the position of those occupants who are not connected to the access point through the Wifi. In addition, when the occupant's smart phone is connected to the other Wifi access point in the adjacent area, it will have to be added, the algorithm for determining the occupied or not. Acknowledgments: This research was supported by a grant (No. 2014R1A2A2A01006494) of Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT (Information and Communication Technology) and Future Planning, Republic of Korea. Author Contributions: The author Woo-Bin Bae developed the methodology, performed the experiment and wrote the full manuscript. Sun-Hye Mun discussed the results and implications at all stages. The author Jung-Ho Huh advised all tasks, double-checked the results and the whole manuscript. All authors proof read the paper. Conflicts of Interest: The authors declare no conflict of interest.

References 1.

2. 3.

4. 5. 6.

7.

8.

9. 10.

Lobato, C.; Pless, S.; Sheppy, M.; Torcellini, P. Reducing plug and process loads for a large scale, low energy office building: Nrel’s research support facility. In Proceedings of the ASHRAE Winter Conference, Las Vegas, NV, USA, 29 January–2 February 2011; p. 29. Office of Energy Efficiency and Renewable Energy in U.S. Department of Energy. Buildings Energy Data Book. Available online: http://buildingsdatabook.eere.energy.gov/ (accessed on 29 February 2016). Mercier, C.; Moorefield, L. Commercial Office Plug Load Savings and Assessment: Final Report; Produced by ECOVA and Supported through the California Energy Commission’s Public Interest Energy Research Program; Ecos: Durango, CO, USA, 2011. Lobato, C.; Sheppy, M.; Brackney, L.; Pless, S.; Torcellini, P. Selecting a control strategy for plug and process loads. Contract 2012, 303, 275–3000. Sheppy, M.; Lobato, C. Assessing and Reducing Plug and Process Loads in Commercial Office and Retail Buildings; National Renewable Energy Laboratory: Golden, CO, USA, 2011. Kaneda, D.; Jacobson, B.; Rumsey, P.; Engineers, R. Plug load reduction: The next big hurdle for net zero energy building design. In Proceedings of the ACEEE Summer Study on Energy Efficiency in Buildings, Pacific Grove, CA, USA, 15–20 August 2010; pp. 120–130. Harris, C.; Cahill, V. Power management for stationary machines in a pervasive computing environment. In Proceedings of the 38th Annual Hawaii International Conference on System Sciences, Big Island, HI, USA, 3–6 January 2005. Christensen, K.; Melfi, R.; Nordman, B.; Rosenblum, B.; Viera, R. Using existing network infrastructure to estimate building occupancy and control plugged-in devices in user workspaces. Int. J. Commun. Netw. Distrib. Syst. 2014, 12, 4–29. [CrossRef] Williams, R.H. Electrical Engineering Probability; West Publishing Company: St. Paul, MN, USA, 1991. International Performance Measurement and Verification Protocol. Available online: http://www.nrel.gov/ docs/fy02osti/31505.pdf (accessed on 29 February 2016). © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons by Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).