Resolving Stack Effect Problems in a High-Rise Office Building By

Home , ASHRAE Handbook, Lobby (room), Stack effect

sustainability

Article Resolving Stack Effect Problems in a High-Rise Ofﬁce Building by Mechanical Pressurization

Jung-yeon Yu 1, Kyoo-dong Song 2,* and Dong-woo Cho 3

1 Department of Architectural Engineering, Graduate School, Hanyang University, Seoul 04763, Korea; [email protected] 2 Department of Architecture and Architectural Engineering, Hanyang University, ERICA Campus, Ansan, Gyeonggi-Do 15588, Korea 3 Korea Institute of Civil Engineering & Building Technology, 283 Goyangdae-Ro, Ilsanseo-gu, Goyang-Si, Gyeonggi-Do 10223, Korea; [email protected] * Correspondence: [email protected]; Tel.: +82-10-3630-5135; Fax: +82-31-408-6335

Received: 5 September 2017; Accepted: 22 September 2017; Published: 26 September 2017

Abstract: In high-rise buildings, the stack effect causes various problems, especially problems related to excessive pressure differences across main entrance doors and elevator doors, particularly in heating seasons. To reduce the stack effect, this study aims to find effective operation schemes for the HVAC systems in a 60-story commercial building, located in Seoul, Korea. Field measurements were conducted to identify the problems related to the stack effect in the building. Computer simulations were conducted to examine the effectiveness of various HVAC operation schemes in reducing the stack effect. Then, an optimum and effective operation scheme was adopted from the computer simulation results and applied in the field. The adopted scheme was used to pressurize the upper zone of the building. Through field application and an adjustment process, a proper amount of air volume was found to effectively pressurize the upper zone of this building, solving the problems related to the stack effect. The required air volume for pressurization was maintained in the building by reducing the volume of the exhaust air (EA) while maintaining a constant volume of outdoor air (OA).

Keywords: stack effect; high-rise buildings; mechanical pressurization; HVAC operation schemes

1. Introduction Stack effect takes place in buildings due to buoyancy of heated air moving upward. The stack effect in buildings plays a positive role in intermediate and cooling seasons by increasing the natural ventilation through the buildings [1,2]. Buildings having an atrium space, large scale factory buildings having a high ceiling combined with operable openings, and solar chimneys are some good examples. However, the stack effect in high-rise buildings in heating seasons causes various problems [3] such as malfunctioning elevator doors, annoying whistling noises in elevator halls, increased infiltration through the main entrance, increased heating loads, and the rapid spread of fire and smoke if a fire breaks out [4,5]. Therefore, many studies have been conducted to develop architectural and mechanical methods to reduce the stack effect in high-rise buildings [3–24]. In Korea, many high-rise buildings have been built for residential and commercial uses since the early 1980s. Up until the early 2000s, many of these high-rise buildings were designed without considering proper architectural and mechanical methods to reduce the stack effect. As a result, many of these buildings have been remodeled or are being investigated in an attempt to solve problems related to the stack effect, and the stack effect has emerged as an important issue in designing high-rise buildings.

Sustainability 2017, 9, 1731; doi:10.3390/su9101731 www.mdpi.com/journal/sustainability Sustainability 2017, 9, 1731 2 of 17

According to previous studies [3,4,6–10], one effective architectural measure to reduce the stack effect is to increase the number of walls between the elevator shaft and the building envelope. However, many commercial buildings require more openness on typical floors for office spaces consisting of multiple work stations divided by low-height interior partitions. For these types of buildings, mechanical methods may be considered to reduce infiltration at floors below the neutral pressure level, such as pressurization of the building interior by HVAC systems [3–5,11–16]. The goal of this study was to find effective operation schemes for the HVAC systems in a 60-story commercial building located in Seoul, Korea to reduce the stack effect during the heating season. The building had suffered a severe stack effect problem related to large pressure differences across the elevator doors on the first floor. HVAC operation schemes were developed to reduce this problem by pressurizing floors in a vertical range; various floor ranges were considered and compared, as well as various quantities of air for pressurization. This study was conducted by taking the following steps.

(1) Field measurements were conducted to identify the magnitude of the pressure difference causing the malfunction of elevator doors in this particular building. (2) Computer simulations were conducted to identify the proper vertical grouping of ﬂoors and the required volume of air for pressurization by the HVAC systems considering the winter outdoor design dry bulb temperature condition in the Seoul area. (3) The results from the computer simulations were then applied to the actual building by pressurizing the upper part of the building, including ﬂoors 40–60. During this phase, the actual volume of air for pressurization was adjusted according to measured pressure differences across the elevator doors.

2. Literature Review

2.1. Stack Effect in High-Rise Buildings When the stack effect occurs in a high-rise open-plan office building, the greatest pressure differences across the elevator doors are usually measured at the first floor and the top floor (Figure1a). However, when the upper floors are pressurized by the HVAC systems, the air flow rate from the elevator shaft to the office spaces on the upper floors is reduced, and the air flow rate from the lobby space on the first floor to the elevator shaft is also reduced (Figure1b). This method can greatly alleviate problems such as malfunctioning elevator doors and noise by reducing the pressure differences across the elevator shaft. Pressurizing every floor uniformly has little effect on the pressure differences across the floors and vertical shafts. If stack pressure across the entrance door is a concern, the pressurization of the first floor is often adopted [17]. In addition, when the inside of an elevator or stairwell shaft is pressurized, the pressure difference across the elevator doors on the first floor decreases while one on each floor of the neutral pressure level or above increases [18,19]. Sustainability 2017, 9, 1731 3 of 17 Sustainability 2017, 9, 1731 3 of 15

Warm Cold

Elevator Office Room Shaft

Main Lobby

(a) (b)

Figure 1. Typical air pressure distributions in high-rise buildings when the stack effect exists with Figure 1. Typical air pressure distributions in high-rise buildings when the stack effect exists with HVAC pressurization on floors above neutral pressure level: (a) without pressurization; and (b) with HVAC pressurization on floors above neutral pressure level: (a) without pressurization; and (b) with pressurization on upper floors. pressurization on upper floors. 2.2. Previous Studies 2.2. PreviousTamura Studies [4] used field measurements and computer simulations to identify various pressure profilesTamura under [4 ]which used fieldthe stack measurements effect occurs and according computer to simulationsthe outdoor temperature, to identify various partitioning pressure profilesschemes, under and whichoperation thestack conditions effect occursof the accordingventilation to systems. the outdoor Tamura temperature, developed partitioning a stack effect schemes, andevaluation operation index conditions called the of thethermal ventilation draft systems.coefficient Tamura (TDC). developed The TDC aindicates stack effect the evaluationlevel of air- index tightness of the building envelope relative to the interior divisions. In that study, various architectural called the thermal draft coefficient (TDC). The TDC indicates the level of air-tightness of the building and mechanical measures were suggested to reduce the stack effect. While Tamura and Wilson envelope relative to the interior divisions. In that study, various architectural and mechanical measures originally applied the TDC concept to a whole building, Hayakaya et al. [6] applied the TDC to were suggested to reduce the stack effect. While Tamura and Wilson originally applied the TDC individual floors. Jo et al. [7] evaluated the characteristics of pressure distributions in high-rise concept to a whole building, Hayakaya et al. [6] applied the TDC to individual floors. Jo et al. [7] residential buildings according to interior partitioning and elevator zoning schemes, and suggested evaluateda separation the method characteristics by installing of pressure so-called distributions “air lock doors” in high-risebetween the residential elevator doors buildings and entrances according to interiorto the residential partitioning units and to reduce elevator the zoning pressure schemes, differences and across suggested those a doors. separation Lstiburek method [8] suggested by installing so-calledthe inclusion “air lockof air doors” barriers between to control the elevator infiltration doors and and exfiltration, entrances tothereby the residential maintaining units the to air- reduce thetightness pressure of the differences interior spaces across and those the doors.building Lstiburek as a whole. [8 ]Based suggested on the theinformation inclusion and of airthe barriersdesign to controlguidelines infiltration suggested and by exfiltration, these studies thereby to reduce maintaining the stack the effect, air-tightness most high-rise of the interiorbuildings spaces recently and the buildingbuilt in Korea as a whole. have been Based designed on the information to include vari andous the architectural design guidelines measures suggested such as by air-tightened these studies to reducebuilding the envelopes, stack effect, revolving most high-rise doors for buildings the main entrance, recently builtair-lock in Koreadoors for have elevator been designedhalls, dedicated to include variouselevators architectural for underground measures parking such lots, as air-tightened and vertical zoning building of elevator envelopes, shafts. revolving doors for the main entrance,These air-lock architectural doors formeasures elevator have halls, practically dedicated solved elevators stack for effect underground problems parking in high-rise lots, and verticalresidential zoning buildings, of elevator in which shafts. numerous walls separate the housing units. However, because commercialThese architectural buildings have measures many open-plan have practically office spaces, solved the stack stack effect effect problems in these in buildings high-rise has residential not buildings,been effectively in which reduced numerous by architectural walls separate measures the alone. housing Therefore, units. other However, measures because to reduce commercial the buildingsstack effect have in high-rise many open-plan commercial office buildings spaces, have the stack been effectdeveloped in these and buildings applied to has several not been buildings. effectively reducedThese measures by architectural include measuresnatural or alone. mechanical Therefore, cooling other of measures elevator toshafts reduce and the mechanical stack effect in pressurization of elevator shafts and office spaces. Yu et al. [9] suggested that elevator shafts should high-rise commercial buildings have been developed and applied to several buildings. These measures be located in the perimeter zone to lower the air temperature within elevator shafts. In the study, include natural or mechanical cooling of elevator shafts and mechanical pressurization of elevator stack effects were measured and compared between an elevator shaft located in the core zone and shafts and office spaces. Yu et al. [9] suggested that elevator shafts should be located in the perimeter Sustainability 2017, 9, 1731 4 of 17 zone to lower the air temperature within elevator shafts. In the study, stack effects were measured and compared between an elevator shaft located in the core zone and another one located in the perimeter zone with a glass wall exposed to the outdoors. The study found that the perimeter shaft had a much weaker stack effect owing to the lower air temperature within the shaft, resulting in a pressure difference across the elevator door in the perimeter shaft that is about half that of the core shaft. Lee et al. [20] conducted computer simulations and measurements of a building with multiple elevator shafts, located in Seoul, to investigate the effectiveness of mechanically cooling the elevator shafts with cold OA to reduce the pressure difference across elevator doors. In the computer simulations, the pressure difference was reduced by about 27% by cooling the elevator shafts from 22 ◦C to 12 ◦C; field measurements showed a reduction of about 25% in air velocity through the elevator door on the first floor and by about 10% on the upper floors when the elevator shafts were cooled. However, this method requires extra building space for fan systems and ductwork, which should be prepared from the design stages, and also incurs extra costs for the mechanical systems. Therefore, this method may not be a practical solution for existing high-rise buildings if there is not enough space to accommodate the added mechanical systems and ductwork. Some notable studies [3–5,11–16] mostly related to mechanical measures to reduce stack effect were conducted with various operation schemes using the HVAC systems to pressurize building spaces. Tamura and Wilson [11] used the ventilation system to reduce the air pressure in the upper zone and increase the pressure in the lower zone to reduce infiltration through the building skins. As a result, the operation could reduce infiltrated air volume by reducing the pressure differences across the building skins, but the pressure differences across the vertical shafts, including elevator doors and stairwell doors, increased. On the other hand, Tamblyn [12] suggested increasing air pressure in the upper zone and decreasing pressure in the lower zone in order to reduce the pressure difference across the elevator shaft. Tamblyn emphasized the importance of airtightness of the building envelope to suppress infiltration, which might increase due to the lowered interior air pressure at the lower zone. In the research project conducted by ASHRAE [13] for an existing high-rise building, the upper zone and mid zone were pressurized and the lower zone excluding the first floor for the lobby was depressurized. The results showed that the pressure differences across the elevator doors for the upper zone, mid zone, and the lobby floor were reduced. However, it was found that the pressure differences across the elevator doors for the lower zone exceeded 25 Pa, which was known as the maximum pressure difference that does not cause the malfunction of the elevator doors. The research suggested that the mechanical balancing of air pressures for a limited number of floors at the top and bottom would be effective for buildings with loose airtight skins and buildings of more than 180 m in height. From these previous studies, it was revealed that the depressurization of the lower zone caused high pressure differences across the elevator doors and building envelope for the lower zone. Therefore, pressurization of the upper zone only was adopted and tested in this study. Then, this study focused on finding the optimum vertical floor ranges and volumes of air for pressurization.

3. Methods

3.1. Identifying Stack Effect Problems by Field Measurements A series of field measurements were conducted in a high-rise office building located in Seoul, Korea to identify problems related to the stack effect. As shown in Figure2, the building has a total height of 245 m, including the lobby height of 9 m and a typical floor height of 4 m. The area of the lobby on the first floor is 2250 m2, and the areas of most floors vary from 1800 to 2250 m2, varying due to the curved shape of the building. The total volume of the building is about 500,000 m3. The passenger elevator shafts are designed for three separate vertical zones: the low-rise zone elevators serve Floors 1–20, mid-rise zone elevators depart from the first floor and stop at Floors 20–37, and high-rise zone elevators depart from the first floor and stop at the 37th floor and Floors 40–60. The 38th, 39th, 61st Sustainability 2017, 9, 1731 5 of 17 and 62nd floors are used as mechanical rooms, so the passenger elevators do not stop at these floors. InSustainability addition, 2017 two, 9 emergency, 1731 staircases serve all 60 floors. 5 of 15

FigureFigure 2. 2.Section Section and and elevator elevator shaft shaft zones zones of of the the test test building. building.

MeasurementsMeasurements of of pressure pressure differences differences (∆Ps) (Δ acrossPs) across the elevator the elevator doors weredoors conducted were conducted on selected on floorsselected between floors thebetween first and the 60thfirst floors:and 60th Floors floors: 1, 4,Floors 8, 12, 1, 16, 4, 8, 20, 12, 25, 16, 29, 20, 33, 25, 37, 29, 41, 33, 45, 37, 48, 41, 52 45, and 48, 56. 52 ◦ Duringand 56. the During measurements, the measurements, the outdoor the andoutdoor indoor and air indoor temperatures air temperatures were, respectively, were, respectively,−5.2 C and −5.2 ◦ 22°C C,and and 22 the°C, outdoorand the outdoor wind velocity wind velocity was less was than less 1 m/s than at 1 am/s height at a ofheight 16 m of from 16 m the from ground the ground level, whichlevel, meanswhich thatmeans external that external wind pressure wind pressure on the buildingon the building was negligible. was negligible. Figure3 Figure shows 3 floor shows plans floor ofplans the first of the floor first and floor a typical and a floor typical (50th floor floor) (50th between floor) thebetween 40th andthe 60th40th floorsand 60th of the floors test of building, the test includingbuilding, theincluding measurement the measurement points. points. FigureFigure4 4shows shows the the measuredmeasured Δ∆PsPs across across the the elevator elevator doors doors at atdifferent different floors floors and and shafts. shafts. The Thegreatest greatest ΔP∆ wasP was 121 121 Pa Paacross across one one of the of the first-floor first-floor doors doors of ofthe the high-rise high-rise elevator elevator shaft. shaft. When When the the ΔP ∆wasP was recorded recorded at at121 121 Pa, Pa, the the elevator elevator doors doors for for the the high-rise high-rise elevator shaft on thethe firstfirst floorfloor were were not not closed.closed. These These malfunctions malfunctions completely completely ceased ceased when when the the∆ ΔPP was was reduced reduced to to about about 100 100 Pa. Pa. This This value value waswas much much greater greater than than the the maximum maximum value value of of 25 25 Pa Pa recommended recommended by by ASHRAE ASHRAE [ 13[13]] because because the the elevatorelevator doors doors installed installed on on the the first first floor floor of of this this building building were were manufactured manufactured to to properly properly operate operate at at largelarge∆ ΔPP across across the the doors doors that that exceeded exceeded ASHARE ASHARE standards. standards. The The∆ ΔPP for for the the low-rise low-rise and and mid-rise mid-rise elevator doors on the first floor were measured as −25 Pa and 52 Pa, respectively. Therefore, in this study, the evaluation criterion for pressure difference was set to 100 Pa. Sustainability 2017, 9, 1731 6 of 17

elevator doors on the ﬁrst ﬂoor were measured as −25 Pa and 52 Pa, respectively. Therefore, in this

Sustainabilitystudy, the 2017 evaluation, 9, 1731 criterion for pressure difference was set to 100 Pa. 6 of 15

FigureFigure 3. Floor 3. Floor plans plans and and measurement measurement points points on onthe the 1st 1st floor floor and and a typical a typical floor floor (50th (50th floor) floor) serviced serviced byby a high-rise a high-rise elevator elevator shaft. shaft. Sustainability 2017, 9, 1731 7 of 17 Sustainability 2017, 9, 1731 7 of 15

FigureFigure 4.4.Pressure Pressure differences differences across across the the elevator elevator doors doors (outdoor (outdoor air air temperature: temperature:− −5.25.2◦ °C,C, indoorindoor airair temperature:temperature: 22 22◦ °C).C).

3.2. Computer Simulations of HVAC Operations 3.2. Computer Simulations of HVACOperations An initial computer simulation was conducted to validate the accuracy of the computer model An initial computer simulation was conducted to validate the accuracy of the computer model by by comparing the calculated results with the results obtained from the initial measurements comparing the calculated results with the results obtained from the initial measurements mentioned mentioned above for the existing conditions without pressurization when the outdoor and indoor air above for the existing conditions without pressurization when the outdoor and indoor air temperatures temperatures were −5.2 °C and 22 °C, respectively. were −5.2 ◦C and 22 ◦C, respectively.

3.2.1.3.2.1. ValidationValidation of of Computer Computer Model Model TheThe computer computer program program used used in thisin this study study was CONTAMwas CONTAM [25], developed [25], developed by the National by the InstituteNational ofInstitute Standards of Standards and Technology and Technology (NIST) in (NIST) the United in the States. United Figure States.5 showsFigure CONTAM5 shows CONTAM models formodels the firstfor the floor first and floor a typical and a floortypical (50th floor floor) (50th serviced floor) serv by theiced high-rise by the high-rise elevator elevator shaft. The shaft. computer The computer model includesmodel includes major elevator major shafts,elevator stairwell shafts, shafts,stairwell ceiling-mounted shafts, ceiling-moun air diffusers,ted air doors diffusers, for the doors office for zones the andoffice the zones building and envelope.the building envelope. TableTable1 1shows shows the the air air leakage leakage data data for for the the computer computer simulation simulation which which were were measured measured values values or or valuesvalues recommended recommended by by Tamura Tamura [ 4[4]] and and ASHRAE ASHRAE [ 26[26]].. Figure Figure6 6shows shows the the measured measured and and calculated calculated ∆ΔPs.Ps. AsAs shownshown inin thethe figure,figure, thethe∆ ΔPP profilesprofiles ofof thethe twotwo valuesvalues showedshowed aa closeclose patternpattern andand thethe relativerelative errorerror betweenbetween the the greatest greatest measured measured and and calculated calculated∆ ΔPsPs was was 7%. 7%. Sustainability 2017, 9, 1731 8 of 17 Sustainability 2017, 9, 1731 8 of 15

Figure 5. The CONTAM model for the 1st floor and a typical floor (50th floor) serviced by a high-rise Figure 5. The CONTAM model for the 1st floor and a typical floor (50th floor) serviced by a high-rise elevator shaft. elevator shaft. Table 1. Air leakage data used for initial computer simulations.

Location Air Leakage DataSource Elevator door EqLA75 1 240 cm2/item(closed) Tamura [4] Stairwell door EqLA75 130 cm2/item(closed) Tamura [4] Revolving door EqLA4 2 1020 cm2/item Measured data Swing door EqLA4 53 cm2/item(closed) ASHRAE [26] Sliding door EqLA4 100 cm2/item(closed) ASHRAE [26]

Door for office room EqLA4 2.1 m2/item(open) Measured data Exterior wall of lobby EqLA75 4.88 cm2/m2 (AL/Awall) Tamura [4] Exterior wall of typical floors EqLA75 3.60 cm2/m2 (AL/Awall) Tamura [4]

1 EqLA75: Equivalent leakage area at 75 Pa, 2 EqLA4: Equivalent leakage area at 4 Pa. Sustainability 2017, 9, 1731 9 of 17

Table 1. Air leakage data used for initial computer simulations.

Location Air Leakage Data Source 1 2 Elevator door EqLA75 240 cm /item(closed) Tamura [4] 2 Stairwell door EqLA75 130 cm /item(closed) Tamura [4] 2 2 Revolving door EqLA4 1020 cm /item Measured data 2 Swing door EqLA4 53 cm /item(closed) ASHRAE [26] 2 Sliding door EqLA4 100 cm /item(closed) ASHRAE [26] 2 Door for ofﬁce room EqLA4 2.1 m /item(open) Measured data 2 2 Exterior wall of lobby EqLA75 4.88 cm /m (AL/Awall) Tamura [4] 2 2 Exterior wall of typical ﬂoors EqLA75 3.60 cm /m (AL/Awall) Tamura [4] 1 2 EqLA75: Equivalent leakage area at 75 Pa, EqLA4: Equivalent leakage area at 4 Pa. Sustainability 2017, 9, 1731 9 of 15

◦ FigureFigure 6. 6.Measured Measured and and calculated calculated∆ ΔPsPs without without pressurization pressurization (outdoor (outdoor air air temperature: temperature:− −5.25.2 °C,C, ◦ indoorindoor air air temperature: temperature: 22 22 °C).C).

3.2.2.3.2.2. Computer Computer Simulation Simulation AsAs mentioned mentioned above,above, thisthis studystudy suggestedsuggested pressurizationpressurization ofof thethe upperupper partpartof ofthe the building buildingby by thethe HVAC HVAC systems systems to to reduce reduce the the great great∆ ΔPsPs across across the the elevator elevator doors doors that that were were confirmed confirmed byby field field measurements.measurements. InIn thisthis study,study, the the group group of of floors floors on on the the upper upper part part of of the the building building to to be be pressurized pressurized waswas determined determined through through computer computer simulations. simulations. In addition, In addition, the required the air required volume forair pressurization volume for (VpressurizationPA) was also determined. (VPA) was also determined. TheThe design design outdoor outdoor and and indoor indoor air air temperatures temperatures for for heating heating in in the the computer computer simulations simulations were were −−11.3 ◦°CC andand 2020 ◦°CC[ [27],27], respectively.respectively. TheThe outdooroutdoor windwind velocityvelocity waswas setset to to 0 0 m/s m/s toto eliminateeliminate thethe effectseffects of of external external wind.wind. TableTable2 2shows shows the the pressurization pressurization conditions conditions for for different different groups groups of of floors. floors. TheThe basebase casecase refers refers to to the the existing existing condition condition without without pressurization. pressurization. CasesCases 1–31–3 involveinvolve variousvarious pressurizationpressurization conditions. conditions. Case Case 1 1 was was where where the the whole whole building building was was pressurized. pressurized. Case Case 2 2 was was where where FloorsFloors 23–60 23–60 were were pressurized, pressurized, where where the the mid-rise mid-rise and and high-rise high-rise elevators elevatorswere were operated. operated. InIn CaseCase 2,2, FloorsFloors 38 38 and and 39 39 were were excluded excluded because because they they are are mechanical mechanical floors. floors. Case Case 3 3 was was where where Floors Floors 40–60 40–60 were pressurized, where only the high-rise elevators were operated. For each case, HVAC operations were simulated with three different VPA volumes; these are denoted as A, B, and C in the case names. For Cases 1–3, the respective volumes of VPA of 3000, 6000 and 9000 m3/h were used for each floor.

Table 2. Pressurization conditions used in the computer simulations.

Cases Floors for Pressurization VPA Per Floor [m3/h] Total VPA [m3/h] Base case None None None Case 1A 3000 180,000 Case 1B 1–60 6000 360,000 Case 1C 9000 540,000 Case 2A 3000 108,000 Case 2B 23–37 and 40–60 6000 216,000 Case 2C 9000 324,000 Case 3A 3000 63,000 Case 3B 40–60 6000 126,000 Case 3C 9000 189,000 Sustainability 2017, 9, 1731 10 of 17 were pressurized, where only the high-rise elevators were operated. For each case, HVAC operations were simulated with three different VPA volumes; these are denoted as A, B, and C in the case names. 3 For Cases 1–3, the respective volumes of VPA of 3000, 6000 and 9000 m /h were used for each ﬂoor. Table 2. Pressurization conditions used in the computer simulations.

3 3 Cases Floors for Pressurization VPA Per Floor [m /h] Total VPA [m /h] Base case None None None Case 1A 3000 180,000 Case 1B1–60 6000 360,000 Case 1C 9000 540,000 Case 2A 3000 108,000 Case 2B23–37 and 40–60 6000 216,000 Case 2C 9000 324,000 Case 3A 3000 63,000 Case 3B40–60 6000 126,000 Case 3C 9000 189,000

3.2.3. Simulation Results Figures7–9 show the simulation results of ∆P across the elevator doors obtained from Cases 1, 2, and 3 compared to the base case. As shown in the figures, most of the ∆Ps across the elevator doors except for the high-rise elevator doors on the first floor were smaller than 100 Pa. Therefore, only the ∆Ps across the high-rise elevator doors on the first floor were summarized in Table3 with the required air volume for each case. As shown in the table, ∆P for the base case was 141.5 Pa, exceeding the threshold of 100 Pa. When the whole building was pressurized, ∆P was smaller than 100 Pa (96.9 Pa) for Case 1B and the required total air volume was calculated as 360,000 m3/h. When the mid zone (Floors 23–37) and upper zone (Floors 40–60) were pressurized, ∆P was 87.6 Pa for Case 2B and the required total air volume was 216,000 m3/h. Finally, ∆P and the total air volume were further reduced to 89.3 Pa and 126,000 m3/h, respectively for Case 3B, in which only the upper zone was pressurized. From this comparative computer simulation process, Case 3B was selected as a candidate pressurization scheme, because this scheme required a comparatively smaller volume of air than the other schemes did. However, Case 3B could not be considered an optimum scheme because ∆P (89.3 Pa) was noticeably smaller than 100 Pa, suggesting the possibility of further reducing the total air volume. Therefore, further computer simulations were performed to find an optimum scheme by reducing the air volume for pressurization until ∆P became smaller than but closer to 100 Pa. Finally, it was found that when the total air volume was reduced to 105,000 m3/h, the ∆P was increased to 99.8 Pa. Figure 10 shows the floor-by-floor distributions of ∆P across the elevator doors due to this final pressurization scheme. Sustainability 2017, 9, 1731 10 of 15

3.2.3. Simulation Results Figures 7–9 show the simulation results of ΔP across the elevator doors obtained from Cases 1, 2, and 3 compared to the base case. As shown in the figures, most of the ΔPs across the elevator doors except for the high-rise elevator doors on the first floor were smaller than 100 Pa. Therefore, only the ΔPs across the high-rise elevator doors on the first floor were summarized in Table 3 with the required air volume for each case. As shown in the table, ΔP for the base case was 141.5 Pa, exceeding the threshold of 100 Pa. When the whole building was pressurized, ΔP was smaller than 100 Pa (96.9 Pa) for Case 1B and the required total air volume was calculated as 360,000 m3/h. When the mid zone (Floors 23–37) and upper zone (Floors 40–60) were pressurized, ΔP was 87.6 Pa for Case 2B and the required total air volume was 216,000 m3/h. Finally, ΔP and the total air volume were further reduced to 89.3 Pa and 126,000 m3/h, respectively for Case 3B, in which only the upper zone was pressurized. From this comparative computer simulation process, Case 3B was selected as a candidate pressurization scheme, because this scheme required a comparatively smaller volume of air than the other schemes did. However, Case 3B could not be considered an optimum scheme because ΔP (89.3 Pa) was noticeably smaller than 100 Pa, suggesting the possibility of further reducing the total air volume. Therefore, further computer simulations were performed to find an optimum scheme by reducing the air volume for pressurization until ΔP became smaller than but closer to 100 Pa. Finally, it was found that when the total air volume was reduced to 105,000 m3/h, the ΔP was increased to 99.8Sustainability Pa. Figure2017 ,109, 1731shows the floor-by-floor distributions of ΔP across the elevator doors due to11 this of 17 final pressurization scheme.

Figure 7. Comparison of pressure differences across the elevator doors: Case 1 vs. the base case. Figure 7. Comparison of pressure differences across the elevator doors: Case 1 vs. the base case. Sustainability 2017, 9, 1731 11 of 15

FigureFigure 8. 8. ComparisonComparison of of pressure pressure differences differences across across the the elevator elevator doors: doors: Case Case 22 vs. vs. the the base base case. case.

Figure 9. Comparison of pressure differences across the elevator doors: Case 3 vs. the base case.

Table 3. Pressure differences across high-rise elevator doors on the first floor and total air volumes for pressurizations.

Cases ΔP[Pa] Total VPA [m3/h] Base case 141.5 None Case 1A 121.4 180,000 Case 1B 96.9 360,000 Case 1C 70.7 540,000 Case 2A 116.4 108,000 Case 2B 87.6 216,000 Case 2C 58.3 324,000 Case 3A 117.1 63,000 Case 3B 89.3 126,000 Case 3C 61 189,000 Sustainability 2017, 9, 1731 11 of 15

Sustainability 2017, 9, 1731 12 of 17 Figure 8. Comparison of pressure differences across the elevator doors: Case 2 vs. the base case.

Figure 9. Comparison of pressure differences across the elevator doors: Case 3 vs. the base case. Figure 9. Comparison of pressure differences across the elevator doors: Case 3 vs. the base case.

Table 3. Pressure differences across high-rise elevator doors on the first floor and total air volumes Table 3. Pressure differences across high-rise elevator doors on the ﬁrst ﬂoor and total air volumes for pressurizations. for pressurizations. Cases ΔP[Pa] Total VPA [m3/h] 3 CasesBase case ∆ 141.5P [Pa] NoneTotal VPA [m /h] Case 1A 121.4 180,000 Base case 141.5 None Case 1ACase 1B 121.496.9 360,000 180,000 Case 1BCase 1C 70.7 96.9 540,000 360,000 Case 1CCase 2A 116.4 70.7 108,000 540,000 Case 2ACase 2B 116.487.6 216,000 108,000 Case 2BCase 2C 58.3 87.6 324,000 216,000 Case 2CCase 3A 117.1 58.3 63,000 324,000 Case 3ACase 3B 117.189.3 126,000 63,000 Case 3BCase 3C 89.3 61 189,000 126,000 Case 3C 61 189,000 Sustainability 2017, 9, 1731 13 of 17 Sustainability 2017, 9, 1731 12 of 15

FigureFigure 10. 10. PressurePressure differences differences across across elevator doors due to the optimum pressurization scheme.

3.3. Field Application and Adjustment 3.3. Field Application and Adjustment

3.3.1. Implementing Implementing the Simulation Result in an Actual Building The optimum pressurizationpressurization schemescheme identifiedidentified by by the the computer computer simulation simulation was was implemented implemented in inan an actual actual building building by controllingby controlling the HVAC the HVAC system, system, shown shown schematically schematically in Figure in 11Figure. The diagram11. The diagramrepresents represents the groups the of groups fans and of fans dampers and dampers for the pressurized for the pressurized zone (Floors zone 40–60). (Floors It40–60). shows It fans shows for fansthe supply for the airsupply (SA) air and (SA) return and air return (RA) air and (RA) various and dampersvarious dampers for OA, for recirculated OA, recirculated air (CA) air and (CA) EA. andIt also EA. shows It also the shows dedicated the dedicated exhaust fansexhaust for thefans rest for rooms. the rest rooms. At first, first, a a simple simple operation operation was was tried tried by byadmitting admitting outdoor outdoor air but air not but allowing not allowing exhaust exhaust air from air thefrom HVAC the HVAC system system and the and restrooms the restrooms in order in order to incr toease increase the air the pressure air pressure in the in office the office rooms. rooms. As depictedAs depicted in Figure in Figure 11, 11only, only the the SA SA fans fans were were operated operated at ata aconstant constant speed speed and and the the OA OA and and CA dampers were operated to bring in in outdoor outdoor air air and and circulate circulate the the air air through through the the HVAC HVAC system, system, while while thethe EA EA dampers dampers were were closed. At At the the same same time, time, the the RA RA fans fans and and EA EA fans fans for for the the rest rest rooms rooms were were not operated. Then, Then, measurements measurements were were conducted conducted to to ex examineamine if if the the evaluation evaluation criterion criterion of of 100 100 Pa Pa or smaller for Δ∆PP across across high-rise high-rise elevator elevator doors doors on on the the first first floor floor was was achieved. achieved. The The measurements measurements were were conducted from 10:00 p.m. to to 12:00 12:00 a.m. a.m. on on 2 2 Febr Februaryuary 2012 2012 when when the the outdoor outdoor air air temperatures temperatures were were between −10.610.6 and and −−11.211.2 °C,◦C, which which were were typical typical design design outd outdooroor temperatures temperatures in in heating heating seasons seasons for for thethe Seoul Seoul area. The The indoor indoor temperature temperature was maintained at 20 °C.◦C. From this simple HVAC system operation, a VPA close to the simulation result was obtained as From this simple HVAC system operation, a VPA close to the simulation result was obtained shownas shown in inTable Table 4.4 .By By controlling controlling the the HVAC HVAC system system with with the various combinationscombinations ofof damperdamper operations, a VPA of 109,000 m33/h was obtained by opening the OA dampers by 40% and CA dampers operations, a VPA of 109,000 m /h was obtained by opening the OA dampers by 40% and CA dampers by 60%, while the EA dampers were closed. The The measured measured Δ∆PP across across the high-rise elevator doors on thethe first first floor floor was 95 Pa by this operation, while it was 135 Pa and the elevator doors could not not be closed when the target zone (Floors 40–60) was no nott pressurized. Figure Figure 1212 showsshows the the floor-by-floor floor-by-floor Δ∆Ps before and after pressurization. Sustainability 2017, 9, 1731 14 of 17 Sustainability 2017, 9, 1731 13 of 15

Sustainability 2017, 9, 1731 13 of 15

Figure 11. HVAC operation scheme for the pressurized zone (Floors 40–60). Figure 11. HVAC operation scheme for the pressurized zone (Floors 40–60). Table 4. Damper controls for pressurization. Figure 11. HVACTable operation 4. Damper scheme controls for the for pressurized pressurization. zone (Floors 40–60). Controls OA Damper (%) CA Damper (%) EA Damper (%) VPA (m3/h) ΔP 1 (Pa) 0 100 0 0 135 3 1 Controls OA DamperTable10 (%) 4. Damper CA Damper controls90 (%) for pressurization. EA Damper0 (%) VPA27,333(m /h) ∆110P (Pa) Actual HVAC 020 10080 00 54,667 0 107 135 SystemControls Operations OA Damper (%) CA Damper (%) EA Damper (%) VPA (m3/h) ΔP 1 (Pa) 1030 9070 00 27,33382,000 105 110 Actual HVAC 0 100 0 0 135 2040 8060 00 109,333 54,667 95 107 System Operations 10 90 0 27,333 110 ComputerActual Simulation HVAC 30 - 70 - 0 - 105,00082,000 99.8 105 20 80 0 54,667 107 System Operations 40 60 0 109,333 95 1 Pressure difference30 across high-rise70 elevator doors on0 the first floor.82,000 105 Computer Simulation -40 -60 -0 105,000109,333 95 99.8 Computer Simulation1 Pressure difference - across high-rise - elevator doors on - the ﬁrst ﬂoor. 105,000 99.8 1 Pressure difference across high-rise elevator doors on the first floor.

Figure 12. Pressure differences across elevator doors: before and after pressurization.

3.3.2. Actual HVAC Operation FigureFigure 12.12.Pressure Pressuredifferences differences across across elevator elevator doors: doors: before before and and after after pressurization. pressurization. After deciding the total VPA needed to solve the stack effect problem without considering the ventilation3.3.2. Actual requirement HVAC Operation as described above, actual volumes of OA and EA were adjusted according to the required ventilation. When the test building was ventilated as recommended by ASHRAE [28], After deciding the total VPA needed to solve the stack effect problem without considering the ventilation requirement as described above, actual volumes of OA and EA were adjusted according to the required ventilation. When the test building was ventilated as recommended by ASHRAE [28], Sustainability 2017, 9, 1731 15 of 17

3.3.2. Actual HVAC Operation

After deciding the total VPA needed to solve the stack effect problem without considering the ventilation requirement as described above, actual volumes of OA and EA were adjusted according to the required ventilation. When the test building was ventilated as recommended by ASHRAE [28], in which 29 m3/h/person of fresh OA is required for about 200 people/ﬂoor in the pressurized zone 3 consisting of 21 ﬂoors (Floors 40–60), the actual required OA was about 121,800 m /h. Since the VPA for pressurization was 109,333 m3/h, the dampers for EA and exhaust fans for the restrooms, shown in Figure 12, were operated by the building operators such that the sum of EA from the HVAC system and the restrooms was maintained at about 12,467 m3/h.

4. Conclusions This study was conducted to find an effective HVAC operation scheme to solve the stack effect problem in a 60-story commercial building located in Seoul, Korea. A series of field measurements were conducted in the winter to identify problems related to the stack effect. Then, a series of computer simulations were conducted to find an effective HVAC operation scheme to reduce the stack effect. The results from the computer simulations were implemented in the field and adjusted to find an optimum HVAC system operation scheme. The whole process and results can be summarized as follows:

(1) From the initial field measurements, the evaluation criterion was established for the pressure difference (∆P) across the high-rise elevator doors on the first floor to be below 100 Pa to ensure smooth opening and closing of these elevator doors. (2) The CONTAM computer program was used to find effective HVAC operation schemes. From this procedure, the decided upon scheme was to pressurize the upper zone of the building, from the 40th to 60th floor. Then, further computer simulations were conducted to find a scheme to minimize the total air volume for pressurization (VPA). The scheme selected as the most effective and efficient HVAC operation for this particular building was to pressurize the upper building 3 zone (Floors 40–60) with 105,000 m /h of VPA. (3) This optimized pressurization scheme identified by the computer simulation was implemented in the actual building by controlling the dampers and fans in the HVAC system. At first, a simple operation was attempted to bring in outdoor air (OA) while not allowing exhaust air (EA) in order to increase the air pressure in the office rooms. From this procedure, it was found that 3 when the upper zone of the building (Floors 40–60) was pressurized with 109,333 m /h of VPA under the winter design outdoor temperature condition in heating seasons in the Seoul area, the ∆P across the first-floor elevator door for the high-rise elevator shaft was reduced from 135 Pa to 95 Pa and the elevator door started closing smoothly. (4) Finally, the actual OA was adjusted to 121,800 m3/h by considering ventilation requirements and about 12,467 m3/h of air was exhausted from the zone.

Even though the specific number of floors to be pressurized and the specific air volume for pressurization were applicable to this specific building, the four steps summarized above can be applied to other high-rise buildings to identify and solve the stack effect related problems. This study was conducted under the design outdoor air temperature condition in the Seoul area with negligible effects of external wind. Therefore, further work should be conducted to find more flexible, generalizable, and effective HVAC operation schemes for different boundary layer conditions such as outdoor air temperature, wind speed, and humidity, as discussed in the previous study [29].

Author Contributions: All authors contributed substantially to all aspects of this article. Conﬂicts of Interest: The authors declare no conﬂict of interest. Sustainability 2017, 9, 1731 16 of 17

References

1. Chen, Y.; Tong, Z.; Malkawi, A. Investigating natural ventilation potentials across the globe: Regional and climatic variations. Build. Environ. 2017, 122, 386–396. [CrossRef] 2. Oropeza-Perez, I.; Østergaard, P.A. Energy saving potential of utilizing natural ventilation under warm conditions—A case study of Mexico. Appl. Energy 2014, 130, 20–32. [CrossRef] 3. Lovatt, J.E.; Wilson, A.G. Stack Effect in Tall Buildings; American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.: Atlanta, GA, USA, 1994; pp. 420–431. 4. Tamura, G.T. Smoke Movement and Control in High-Rise Buildings; National Fire Protection Association: Quincy, MA, USA, 1994. 5. Klote, J.H.; Evans, D.H. A Guide to Smoke Control in the 2006 IBC; International Code Council: Washington, DC, USA, 2007. 6. Hayakaya, S.; Togari, S. Study on the stack effect of tall office building (Part1). J. Archit. Inst. Jpn. 1988, 387, 42–52. 7. Jo, J.H.; Lim, J.H.; Song, S.Y.; Yeo, M.S.; Kim, K.W. Characteristics of pressure distribution and solution to the problems caused by stack effect in high-rise residential buildings. Build. Environ. 2007, 42, 263–277. [CrossRef] 8. Lstiburek, J.W. Understanding air barriers. ASHRAE J. 2005, 47, 24–30. 9. Yu, J.Y.; Cho, D.W.; Song, K.D. Case Studies of the Design Alternatives to Minimize Stack Effect Problems in Tall Buildings. In Proceedings of the BUEE 2006, Tokyo, Japan, 10–13 July 2006; pp. 415–420. 10. Yu, J.Y.; Cho, D.W.; Song, K.D. The Design Procedure for Reducing Stack Effect Problems in Tall Complex Building. In Proceedings of the 11th International Conference on Indoor Air Quality and Climate, Copenhagen, Denmark, 17–22 August 2008; p. 773. 11. Tamura, G.T.; Wilson, A.G. Building pressures caused by chimney action and mechanical ventilation. ASHRAE Trans. 1967, 73, 2.2.1–2.2.12. 12. Tamblyn, R.T. Coping with air pressure problems in tall buildings. ASHRAE Trans. 1991, 97, 824–827. 13. ASHRAE. Field Verification of Problems Caused by Stack Effect in Tall Buildings; ASHRAE Research Project Report RP-661; American Society of Heating, Refrigerating, and Air-Conditioning Engineers: Atlanta, GA, USA, 1993. 14. Ross, D.E. HVACDesign Guide for Tall Commercial Buildings; ASHRAE: Atlanta, GA, USA, 2004. 15. Tamblyn, R.T. HVAC system effects for tall buildings. ASHRAE Trans. 1993, 99, 789–792. 16. Klote, J.H. The ASHRAE Design manual for smoke control. Fire Saf. J. 1984, 7, 93–98. [CrossRef] 17. ASHRAE. ASHRAE Handbook 2009 Fundamentals; ASHRAE: Atlanta, GA, USA, 2009. 18. Schmidt, W. Stairwell and elevator shaft pressurization. Fire Saf. J. 1984, 7, 115–116. [CrossRef] 19. Miller, R.S.; Beasley, D. On stairwell and elevator shaft pressurization for smoke control in tall buildings. Build. Environ. 2009, 44, 1306–1317. [CrossRef] 20. Lee, J.H.; Song, D.S.; Park, D.R. A study on the development and application of the E/V shaft cooling system to reduce stack effect in high-rise buildings. Build. Environ. 2010, 45, 311–319. [CrossRef] 21. Klote, J.H.; Milke, J.A. Principle of Smoke Management; ASHRAE: Atlanta, GA, USA, 2002. 22. Klote, J.H. NISTIR 4588: A General Routine for Analysis of Stack Effect; NIST, U.S. Department of Commerce: Gaithersburg, MD, USA, 1991. 23. NRC-CNRC. National Building Code of Canada; NRC-CNRC: Ottawa, ON, Canada, 2005. 24. Lstiburek, J.W. Multifamily Buildings: Controlling Stack Effect-Driven Airflows. ASHRAE J. 2005, 47, 30–39. 25. Dols, W.S.; Walton, G.N. NISTIR 7251-CONTAM User Guide and Program Documentation; National Institute of Standards and Technology: Gaithersburg, MD, USA, 2013. 26. ASHRAE. ASHRAE Handbook 1997 Fundamentals; ASHRAE: Atlanta, GA, USA, 1997. 27. Ministry of Land, Infrastructure and Transport. Energy Saving Design Standard of Building; Ministry of Land, Infrastructure and Transport: Sejong-si, Korea, 2017. Sustainability 2017, 9, 1731 17 of 17

28. ASHRAE. STD 62.1 Ventilation for Acceptable Indoor Air Quality; ASHRAE: Atlanta, GA, USA, 2004. 29. Tong, Z.; Chen, Y.; Malkawi, A. Estimating natural ventilation potential for high-rise buildings considering boundary layer meteorology. Appl. Energy 2017, 193, 276–286. [CrossRef]

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).