Effect of Surface Thermal Resistance on the Simulation Accuracy of the Condensation Risk Assessment for a High-Performance Window

Transcription

1 energies Article Effect of Surface Thermal Resistance on Simulation Accuracy of Condensation Risk Assessment for a High-Performance Window So Young Koo, Sihyun Park, Jin-Hee Song Seung-Yeong Song * Department of Architectural Engineering, Ewha Womans University, 52 Ewhayeodae-gil, Seodaemun-gu, Seoul 03760, Korea; soykoo@gmail.com (S.Y.K.); sihyun.park@ewha.ac.kr (S.P.); cco79@naver.com (J.-H.S.) * Correspondence: archssy@ewha.ac.kr; Tel.: Received: 27 December 2017; Accepted: 2 February 2018; Published: 7 February 2018 Abstract: The accuracy of condensation risk assessment depends on accuracy of measured or calculated s. The existing 2D simulation method provides sufficiently accurate results for evaluating average performance values, such as U-values. However, accuracy of predicting s in local areas such as edge-of-glazing frame has been questioned. This study analyzes effect of surface rmal resistance on accuracy of condensation risk assessment for high-performance windows. Experiments three-dimensional simulations were performed for a triple-glazed window. The differences in results between basic experimental test simulations with several different applied boundary conditions were analyzed. The results show that, in simulations, a small change in surface rmal resistance has no significant effect on accuracy of condensation risk assessment of center-of-glazing or frame. However, for edge-of-glazing, accuracy of predicting condensation risk was significantly improved by using increased local surface rmal resistance with simulation. By employing reduced radiation convection at edges or junctions between two surfaces, error between measured calculated factors can be reduced to less than 3%. Keywords: high-performance window; condensation risk; 3D rmal simulation; surface rmal resistance 1. Introduction Recent progress in window technology has enabled development of advanced windows with a U-value of approximately W/m 2 K [1]. However, improvements in insulation performance have not completely solved problem of condensation on internal surface of windows in cold climates. When indoor humidity is not adequately controlled, even a high-performance window is likely to form condensation on its surface. One of main considerations in selection of a window is minimizing condensation on internal surface. There are several different indicators for a condensation risk assessment of a window in different countries. The international organization for stardization (ISO) uses factor (f Rsi ) for a condensation risk assessment [2]. Many European countries, such as France, Germany, Nerls, Austria, UK use factor as an indicator for ir building design criteria to prevent inside surface condensation [3 7]. Korea uses difference ratio (TDR) obtained from a similar equation with factor [8]. The National Fenestration Rating Council (NFRC) has developed condensation resistance (CR). The condensation resistance scale is 1 to 100, with a lower number representing a greater condensation risk [9]. The factor was selected as indicator to perform a condensation risk assessment in this paper. The factor is a construction specific value, which is independent of any difference between interior Energies 2018, 11, 382; doi: /en

2 Energies 2018, 11, of 13 exterior environmental conditions. The internal surface is used as an input to calculate factor. Triple-glazed windows are widely used in cold climate countries, like Sweden Norway. The Passive House stard, which is a building stard for low energy use, requires a window U-value of no more than 0.80 W/m 2 K in cool temperate climate zone [10]. A fenestration energy efficiency rating system has been adopted in Korean building codes since It requires window U-value of no more than 1.00 W/m 2 K for highest rated window, which implies that double glazed windows are no more enough to meet stards. Spacer bars for windows have traditionally been made of aluminum or galvanized steel. These metal spacer bars have high rmal conductivity, thus, create a rmal bridge on interior glazing surface in edge-of-glazing area. Insulating spacers have been introduced to reduce heat loss caused by rmal bridge in edge-of-glazing. Previous research has shown that improving spacers could significantly reduce energy lost through multi-pane windows [11 13]. In high-performance windows, such as triple-glazed windows, insulating spacers can reduce window U-value by 12% compared to U-value with aluminum spacers [11]. Insulating spacers offer stronger rmal performance, but y still represent a rmal bridge that influences rmal performance of a window [14]. Although rmal performance of windows is still evaluated by traditional experimental methods, recent advancements in computer tools allow use of more efficient simulation methods. Window evaluations are typically undertaken using two-dimensional heat transfer tools, such as THERM. Various algorithms have been developed to perform more realistic simulations, many studies have been performed to evaluate accuracy of computer-simulated results. Shin et al. [12] showed that a two-box model could be useful for two-dimensional rmal simulations to investigate impact of spacers on window rmal performance. Arasteh et al. [15] used five different laboratories six different simulators to evaluate rmal transmittance (U w ) of 16 windows. The agreement among simulations was more than three times higher than agreement among experimental results, agreement between tests simulations was reasonable. Gustavsen et al. [16] compared test simulation data of total window rmal transmittance (U w ) glazing of double-glazed argon-filled windows. The windows rmal transmittance showed a quite good agreement between test simulation results. In case of glazing surface, however, poor agreement was found for interior edge-of-glass region. In particular, windows with an insulating spacer showed a larger difference between measured simulated s than those with a conventional aluminum spacer. Gustavsen et al. [17] noted that although algorithms defined in ISO [18] ISO [19] are relatively accurate for windows with U-values of approximately 2.0 W/m 2 K, y are not adequate for accurately evaluating heat transfer through low-conductance frames. Gustavsen et al. [20] used a two-dimensional heat transfer analysis tool based on algorithm in ISO to calculate rmal transmittance of high-performance frames. In every case, measured value of rmal transmittance was higher than calculated value. In particular, PVC frame had more than twice error percentage for measurement relative to aluminum frame. When calculating window s using THERM, Kohler et al. [21] showed that heat transfer coefficient convective heat transfer inside glazing cavity have a strong effect on accuracy of calculated s. Generally, experimental method for a condensation risk assessment requires use of stardized environmental conditions such as interior, exterior, relative humidity, wind speed. For example, condensation resistance (CR) of NFRC is determined based on outside conditions of 18 C with a 5.5 m/s wind inside conditions of 21 C with indoor relative humidity of 30%, 50%, 70% [9]. The rmal surface resistances are not specified in environmental conditions for experiment. For simulation method, surface rmal

3 Energies 2018, 11, of 13 resistances are referenced as a fixed combined coefficient or calculated using a detailed algorithm. Thus, surface rmal resistance is one of important factors affecting accuracy of simulated results. From literature review, it is clear that existing simulation method of using two-dimensional rmal analysis tools, such as THERM, can be expected to achieve a satisfactory level of accuracy Energies 2018, 11, x FOR PEER REVIEW 3 of 13 in predicting total window rmal transmittance or of center-of-glazing. However, From accuracy literature of review, simulated it is results clear that is relatively existing lowsimulation for edge-of-glazing method using area, two- vulnerable rmal to condensation, analysis tools, particularly such as THERM, for high-performance can be expected windows. to achieve a satisfactory level which is mostdimensional of The accuracy two-dimensional in predicting heat total transfer window model rmal was transmittance used in previous or studies. of Since center-ofglazing. area However, is characterized accuracy byof multi-dimensional simulated results heat is relatively transfer, low two-dimensional for edge-of-glazing heat area, transfer rmal bridging which is most vulnerable to condensation, particularly for high-performance windows. model might limit accuracy of calculation. Also, effect of surface rmal resistance The two-dimensional heat transfer model was used in previous studies. Since rmal of rmal bridging area was not considered in previous studies. The warm side surface rmal bridging area is characterized by multi-dimensional heat transfer, two-dimensional heat transfer resistance model depends might limit on accuracy product of tested calculation. laboratory Also, hot effect box setup of [22]. surface However, rmal in resistance simulation, of conventional rmal bridging coefficient area was was used not considered in previous in previous method. studies. The warm side surface rmal resistance The aim ofdepends our study on was to product investigate tested effect laboratory of surface hot rmal box setup resistances [22]. However, on in simulation accuracy simulation, for condensation conventional risk coefficient assessment was used of in rmal previous bridging method. area. The study was particularly focused on The high aim of performance our study was triple to investigate glazed windows. effect of The surface experiment rmal resistances was performed on simulation according to Korean accuracy national for stard condensation for risk testassessment method of of dew condensation rmal bridging for windows. area. The The study simulations was wereparticularly performedfocused to findon high surface performance rmal triple resistances glazed windows. reflecting The experiment experimental was performed conditions in according to Korean national stard for test method of dew condensation for windows. The a satisfactory way. Several different surface rmal resistances were applied to simulation simulations were performed to find surface rmal resistances reflecting experimental to calculate internal surface s of a window analyze condensation risk. conditions in a satisfactory way. Several different surface rmal resistances were applied to Three-dimensional simulation to calculate (3D) simulations internal were surface performed, s of a window simulated analyze results were condensation compared with experimental risk. Three-dimensional measurements. (3D) simulations were performed, simulated results were compared with experimental measurements. 2. Methods 2. Methods 2.1. Assessment of Condensation Risk 2.1. Assessment of Condensation Risk A triple-glazed window with a polyvinyl chloride (PVC) frame was evaluated to determine condensation A triple-glazed risk. A rmoplastic window with spacer a polyvinyl (TPS) chloride was applied (PVC) frame to was window, evaluated which to determine was organic polyisobutylene condensation spacer risk. A without rmoplastic metal spacer insert, (TPS) desiccant was applied fillingto window, butyl coating which [23]. was Figure organic 1 shows spacer polyisobutylene bar assemblies spacer of without a conventional metal insert, aluminum desiccant spacer filling butyl a rmoplastic coating [23]. spacer. Figure 1 The shows glazing spacer bar assemblies of a conventional aluminum spacer a rmoplastic spacer. The glazing unit consisted of three 5-mm thick glass panes separated by a 12 mm wide gap filled with 90% argon unit consisted of three 5-mm thick glass panes separated by a 12 mm wide gap filled with 90% argon 10% air. The outermost innermost glass had a low-emissivity coating (emissivity = 0.03) facing 10% air. The outermost innermost glass had a low-emissivity glazing cavity. The window rmal transmittance was 0.98 W/m 2 coating (emissivity = 0.03) facing glazing cavity. The window rmal transmittance was 0.98 K W/m from calculation according 2 K from calculation to ISOaccording using to ISO WINDOW/THERM using WINDOW/THERM 6.3 [24 26]. 6.3 As [24 26]. shown As inshown Figure in 2, Figure window 2, window measured 2 m measured 2 m (W 2 m H) 2 m (W consisted H) consisted of a tilt--turn of a tilt--turn window window combined combined with with a fixed a fixed window. window. (a) (b) Figure 1. Spacer bar assemblies of a conventional aluminum spacer a rmoplastic spacer (TPS): Figure 1. Spacer bar assemblies of a conventional aluminum spacer a rmoplastic spacer (TPS): (a) aluminum spacer; (b) rmoplastic spacer (TPS). (a) aluminum spacer; (b) rmoplastic spacer (TPS). The risk of possible surface condensation was assessed using factor frsi, as defined The riskin ofiso possible [2]. surface The condensation factor wasexpresses assessed using surface in factor dimensionless f Rsi, as defined in ISOform in accordance [2]. The with Equation factor (1). expresses The closer surface factor is to one, in better dimensionless rmal form in accordance performance. with Equation (1). The closer factor is to one, better rmal performance.,, =,, (1)

4 Energies 2018, 11, of 13 f Rsi (x, y, z) = θ si(x, y, z) θ e θ i θ e (1) Energies 2018, 11, x FOR PEER REVIEW 4 of 13 where f Rsi (x, y, z) is factor at internal surface at point (x, y, z), θ si (x, y, z) is where frsi(x, at y, z) internal is surface at factor point at (x, y, internal z), θ i is surface internal at point (x,, y, z), θsi(x, y, z) is θ e is external. internal surface at point (x, y, z), θi is internal, θe is external To. compare experimental simulation results, five assessment points were selected from fixed To window compare because experimental it is less affected simulation by variables, results, such five assessment air leakage points or were operating selected hardware, from that were fixed omitted window frombecause simulation. it is less affected The assessment by variables, points such were as air selected leakage or from operating center-of-glazing, hardware, frame, that were omitted bottomfrom edge of simulation. glazing, The which assessment tends topoints be more were vulnerable selected from to condensation center-ofglazing, edge due frame, to bottom stratification. edge of Theglazing, edge-of-glazing which tends wasto defined be more as vulnerable area of glazing to than top condensation than top edge due to stratification. The edge-of-glazing was defined where multi-dimensional heat transfer effects dominate [9]. Generally, it is defined as region of as area of glazing where multi-dimensional heat transfer effects dominate [9]. Generally, it is view area less than 63.5 mm from sight line [27]. In this study, assessment points of edge defined as region of view area less than 63.5 mm from sight line [27]. In this study, portion assessment were determined points of to be edge 2 cm portion from were sight determined line according to be 2 cm to from Korean sight national line according stard to [28], as described Korean in Figure national 2. stard [28], as described in Figure 2. Figure Figure The The tested tested window scheme assessment assessment points. points Experimental Investigations 2.2. Experimental Investigations The test method was based on methodology of calibrated hot box in accordance with KS The F 2295 test [29]. method KS F 2295 wasis based Korean on national methodology stard of for calibrated test method hot of dew boxcondensation in accordance for with KS F windows, 2295 [29]. which KS F 2295 specifies is Korean test window national size, stard details for of a surround test method panel, of dew conditions condensation of for windows, chambers, which among specifies or parameters. test window This stard size, follows detailsks of F a 2278 surround for panel, airflow conditions conditions of of chambers, chambers among or surface parameters. rmal resistance This stard of a test follows window KS[30]. F 2278 KS for F 2278 is airflow Korean conditions national of chambers stard describing surface rmal method resistance used to test of window a test window rmal resistance [30]. KS F(U-value) is Both Korean stards national use same hot box method while using different environmental conditions. stard describing method used to test window rmal resistance (U-value). Both stards use During hot-box experiment, test window was enclosed between warm cold climate same hot box method while using different environmental conditions. chambers. The window was mounted in a 350 mm thick surrounding panel with a rmal resistance During of 4.3 m 2 K/W. hot-box Figures experiment, 3 4 display test schematic window of was experimental enclosed between configuration warm cold view climate chambers. of The window window was hot box. mounted The hot inbox a 350 featured mm thick natural surrounding convection, panel while with cold a rmal box featured resistance of 4.3forced m 2 K/W. convection. Figures The 3 airflow 4 display was adjusted schematic in accordance of experimental with KS F 2278 configuration to produce average view of window surface rmal in hot resistance box. The of hot 0.11 box ± 0.02 featured m 2 K/W natural 0.05 convection, ± 0.02 m 2 K/W while on surface cold box of featured hot box forced convection. cold The box, airflow respectively. was adjusted The indoor accordance air with was KS Fset 2278 to 25 to produce C, an average outdoor air surface rmal resistance to of C, ± following 0.02 m 2 K/W Korean 0.05 design ± 0.02stard m 2 K/Wfor onpreventing surfacecondensation of hot box in cold box, apartment respectively. buildings The [28]. indoor In this airstard, three was climate set regions to 25 C, are specified outdoor for air condensation to risk assessment based on average daily minimum outdoor air of coldest month: 15 C, following Korean design stard for preventing condensation in apartment buildings [28]. 20 C for region I, 15 C for region II, 10 C for region III. Table 1 presents conditions of In this stard, three climate regions are specified for condensation risk assessment based on hot cold boxes. Type T rmocouples were used for measuring internal surface average s daily minimum of window. outdoor The air accuracy of rmocouple of coldest was month: ±0.5 C. 20 C for region I, 15 C for region II, 10 C for region III. Table 1 presents conditions of hot cold boxes.

5 Energies 2018, 11, of 13 Type T rmocouples were used for measuring internal surface s of window. The accuracy of rmocouple was ±0.5 C. Energies 2018, 11, x FOR PEER REVIEW 5 of 13 Energies 2018, 11, x FOR PEER REVIEW 5 of 13 Figure 3. Schematic of experimental configuration. Figure 3. Schematic of experimental configuration. Figure 3. Schematic of experimental configuration. Figure 4. View of a test window in hot box. Figure 4. View of a test window in hot box. Figure 4. View of a test window in hot box. Table 1. Boundary conditions for test. Table 1. Boundary conditions for test. Chamber Temperature Table ( C) 1. Boundary Relative Humidity conditions (%) for Surface test. Resistance (m 2 K/W) Chamber Hot box Temperature 25.0 ± 1.0 ( C) Relative 50 Humidity ± 1.0 (%) Surface Resistance 0.11 ± 0.02 (m 2 K/W) Cold Hot box ± ± ± 0.02 Chamber Temperature ( C) Relative Humidity (%) Surface Resistance (m 2 K/W) Cold box 15.0 ± ± Numerical Hot box Procedure 25.0 ± ± ± Cold Numerical box Procedure 15.0 ± ± 0.02 The numerical simulations were performed using computer program Physibel Trisco 12.0w (2010) The for numerical 3D heat simulations transfer calculation were performed [31]. This using program computer is a rmal program analysis Physibel program Trisco for steadystate heat for transfer Procedure 3D heat calculations; transfer calculation s [31]. This are program determined is a rmal using analysis finite program difference for method. steady- 12.0w 2.3. Numerical (2010) state 3D heat program transfer operates calculations; in accordance s with EN are 673 determined (2011) for using glazing cavities finite difference ISO method The numerical simulations were performed using computer program Physibel Trisco 12.0w The (2012) 3D for program cavities operates in frames in [19,32]. accordance The with concept EN of 673 equivalent (2011) for rmal glazing conductivity cavities was ISO used to (2010) for 3D heat transfer calculation [31]. This program is a rmal analysis program for (2012) calculate for cavities heat transfer in frames in [19,32]. frame The cavities; concept heat of equivalent transfer analysis rmal was conductivity simplified was by replacing used to steady-state calculate heat cavity with heat transfer an transfer calculations; imaginary in solid frame material. cavities; s heat transfer are determined analysis was using simplified by finite replacing difference method. cavity The Thefour with 3Dsimulation an program imaginary cases operates solid with material. different in accordance boundary with conditions EN 673 used (2011) in this forstudy glazing are shown cavities in ISO Table The 2. An (2012) four indoor simulation forair cavities cases inwith frames of different 25 C [19,32]. boundary an The outdoor concept conditions air ofused equivalent in this of 15 study rmal C were are shown conductivity applied in was used Table in all to simulation 2. calculate An indoor cases air to heat match transfer experimental of in25 C frame conditions. an cavities; outdoor In air reality, heat transfer surface of 15 analysis rmal C were was resistance applied simplified by replacing of all simulation window cavity is cases dynamic with to match an imaginary depends experimental solid on material. conditions. indoor In reality, outdoor environmental surface rmal conditions. resistance of window is dynamic depends on indoor outdoor environmental conditions.

6 Energies 2018, 11, of 13 The four simulation cases with different boundary conditions used in this study are shown in Table 2. An indoor air of 25 C an outdoor air of 15 C were applied in all simulation cases to match experimental conditions. In reality, surface rmal resistance of window is dynamic depends on indoor outdoor environmental conditions. However, both experiment calculation of condensation risk are to be performed under static conditions. In or words, surface rmal resistance should be assumed constant. This study applied four different surface rmal resistances, each constituting a fixed surface resistance using simplified correlations for convection radiation heat transfer. The surface rmal resistances that were set as a condition of simulation were within margin of error of surface rmal resistances set by experiment except for edge of Case 4 (refer to Figure 5). Table 2. Simulation cases boundary conditions. Temperature Energies 2018, ( 11, C) x FOR PEER REVIEW Surface Thermal Resistance (m 2 K/W) 6 of 13 Case Reference External Internal External Internal However, both experiment calculation of condensation risk are to be performed under static conditions. In or 1 words, surface rmal resistance 0.11 should be assumed constant. This [8] [29] study applied four different surface rmal resistances, each constituting a fixed surface resistance [33,34] using simplified correlations for convection radiation Normal heat transfer. The surface 0.13 rmal resistances that were set 4as a condition 0.04 of [19] Reduced simulation radiation were within convection margin at edges of error 0.20 of surface rmal resistances set by experiment except for edge of Case 4 (refer to Figure 5). (a) (b) Figure 5. The treatment of boundaries according to ISO : (a) schematic of boundaries for Figure 5. The Case treatment 4; (b) shown of in red are boundaries local regions according of increased to ISO surface : resistance (a) due schematic to reduced of radiation boundaries for Case 4; (b) shown convection. in red are local regions of increased surface resistance due to reduced radiation convection. Table 2. Simulation cases boundary conditions. Temperature ( C) Surface Thermal Resistance (m Case 2 K/W) The surface Reference External rmal Internal resistances External of Cases 1 2 were Internal referenced from Korean national stards, Korean design stards for preventing condensation 0.11 in apartment buildings [8] [8] [29] KS F 2295 [29] test method of dew condensation for windows doors. There is a slight difference [33,34] between two stards in terms of external surface Normal rmal resistance The surface rmal [19] resistance of Cases 3 4 were referenced Reduced from radiation international convection stards: at edges 0.20 ISO 13788, ISO 6946, ISO ISO includes calculation methods for assessment of risk of internal The surface rmal resistances of Cases 1 2 were referenced from Korean national surface condensation, stards, Korean providing design stards surface for preventing resistances condensation for in assessment apartment buildings of surface [8] condensation KS on windows. F 2295 ISO[29] 6946 test method provides of dew condensation conventional for surface windows rmal doors. resistances There a slight for surfaces difference in contact with air. ISO between two specifies stards in calculation terms of method external surface for rmal resistance. transmittance The surface of rmal windows, doors, resistance of Cases 3 4 were referenced from international stards: ISO 13788, ISO 6946, shutters. The surface rmal resistances for surface condensation assessment of windows are ISO ISO includes calculation methods for assessment of risk of internal suggestedsurface in Annex condensation, B of ISOproviding In surface case resistances of internal for assessment surface rmal of surface resistance condensation of a window, heat transfer on windows. coefficients ISO 6946 provides are locally conventional lower, particularly surface rmal at resistances glazingfor near surfaces in frame, contact due to lower airflow with speed air. ISO different specifies radiation calculation view method factors for [16]. The rmal viewtransmittance factor is a of dimensionless windows, factor doors, shutters. The rmal resistances for condensation assessment of that determines how much of a surface is visible to anor surface. The edge-of-glazing closer to windows are suggested in Annex B of ISO In case of internal surface rmal resistance of a window, heat transfer coefficients are locally lower, particularly at glazing near frame, due to lower airflow speed different radiation view factors [16]. The view factor is a dimensionless factor that determines how much of a surface is visible to anor surface. The edgeof-glazing closer to frame surface has a higher view factor between m compared to centerof-glazing. This implies that projecting surfaces such as sill, jamb, head surfaces can affect radiation heat transfer in edge-of-glazing [22].

7 Energies 2018, 11, of 13 frame surface has a higher view factor between m compared to center-of-glazing. This implies that projecting surfaces such as sill, jamb, head surfaces can affect radiation heat transfer in edge-of-glazing [22]. Local variation of surface rmal resistance is reflected in ISO In Case 4, internal surface rmal resistance of 0.20 m 2 K/W was used for edges between two surfaces considering reduced radiation Energies 2018, 11, x FOR convection, PEER REVIEW as described in Figure 5. 7 of 13 This study investigated wher a small change of surface rmal resistance in numerical Energies 2018, 11, x FOR PEER REVIEW 7 of 13 Local variation of surface rmal resistance is reflected in ISO In Case 4, internal simulations can influence calculated s. In consideration of computational efficiency, surface rmal resistance of 0.20 m 2 K/W was used for edges between two surfaces considering lower arealocal of variation window, of surface whichrmal included resistance all is reflected measurement in ISO points In Case of 4, surface internal reduced radiation convection, as described in Figure 5. surface rmal resistance of 0.20 m in experiment This study (referinvestigated to Figurewher 2), was 2 K/W was used for edges between two surfaces considering a small adopted change asof surface simulation rmal resistance model. in Figures numerical 6 7 shows reduced radiation convection, as described in Figure 5. simulation simulations model can influence calculated s. assessment In consideration points. The of computational upper boundary efficiency, of This study investigated wher a small change of surface rmal resistance in numerical model lower area of window, which included all measurement points of surface in was 100 mm simulations from Point can influence C-1. No significant calculated s. In consideration gradientof occurred computational beyond efficiency, Point C-1 in experiment (refer to Figure 2), was adopted as simulation model. Figures 6 7 shows simulation y-axis direction. lower area model Table of 2window, presents which assessment included simulation all points. The upper cases measurement boundary with points of boundary of surface model was 100 conditions, in mm from Point Table 3 experiment (refer to Figure 2), was adopted simulation model. Figures 6 7 shows simulation presents C-1. material No significant properties. Thegradient rmoplastic occurred beyond rmal Point property C-1 in isy-axis fromdirection. manufacturer Table 2 [35]. model assessment points. The upper boundary of model was 100 mm from Point All or presents material rmal simulation cases properties with boundary are fromconditions, THERM Table 6.3/WINDOW 3 presents material 6.3properties. C-1. No significant gradient occurred beyond Point C-1 in y-axis direction. software Table 2 library The rmoplastic rmal property is from manufacturer [35]. All or material rmal properties ISO presents [19,25,26]. simulation cases with boundary conditions, Table 3 presents material properties. are from THERM 6.3/WINDOW 6.3 software library ISO [19,25,26]. The rmoplastic rmal property is from manufacturer [35]. All or material rmal properties are from THERM 6.3/WINDOW 6.3 software library ISO [19,25,26]. (a) (a) (b) (b) (c) Figure 6. Cross-sections of a simulation model: (c) (a) sill of a tilt turn window; (b) sill of a fix Figure 6. Cross-sections window; (c) meeting of a simulation rail. model: (a) sill of a tilt turn window; (b) sill of a fix window; (c) meeting Figure 6. rail. Cross-sections of a simulation model: (a) sill of a tilt turn window; (b) sill of a fix window; (c) meeting rail. Figure 7. 3D Simulation model assessment points. Figure 7. 3D Simulation model assessment points. Figure 7. 3D Simulation model assessment points.

8 Energies 2018, 11, of 13 Table 3. Material properties for numerical simulation. Group Material Thermal Conductivity (W/mK) Emissivity Reference Frame Glazing Spacer Steel, oxidized [25] EPDM (gaskets) [25] PVC [19] Glass [26] Low-e coating [26] Argon gas [26] Silicone sealant [19] Thermoplastic [35] Energies 2018, 11, x FOR PEER REVIEW 8 of Results Discussion Table 3. Material properties for numerical simulation Comparison of Temperatures Group Material Thermal Conductivity (W/mK) Emissivity Reference Steel, oxidized [25] Frame EPDM (gaskets) [25] PVC [19] Table 4 presents results of experimental tests numerical simulations. The absolute Glass [26] value of percent error Glazing waslow-e calculated coating for each1.00 comparison between 0.03 [26] tests simulations. Argon gas [26] Figure 8 compares internal surface between experimental measurements Silicone sealant [19] Spacer numerical simulation at each point. Thermoplastic For center-of-glazing frame, [35] calculated s were generally lower than measured ones. The edge-of-glazing had a higher calculated 3. Results Discussion than a measured. The measured at center-of-glazing was 22.8 C, calculated results 3.1. Comparison of simulation of Temperatures were lower than measured ones by approximately 1 C. In case of frame, Table 4 presents results wasof measured experimental to be tests approximately numerical simulations. 20 C atthe F-1absolute 19 C at F-2. value of percent error was calculated for each comparison between tests The simulation calculated of F-2 to be approximately simulations. C. For edge-of-glazing, Figure 8 compares internal surface between measured of E-1 was approximately 15 experimental measurements numerical simulation at each point. For center-of-glazing C frame, that ofcalculated E-2 wass approximately 16 C, indicating a were generally difference lower than of approximately measured ones. The 1 edge-of-glazing C. In Case 4, had a calculation higher calculated of E-1 E-2 was 15.6 C 16.0 than a measured. The measured at center-of-glazing was C, respectively; in this case, difference between measured calculated 22.8 C, calculated results of simulation were lower than measured ones by was approximately 5.4% 1 C. 1.9%, In respectively, case of frame, which was was smallest measured among to be approximately all cases. 20 C at F-1 19 C at F-2. The simulation calculated of F-2 to be approximately C. For Table edge-of-glazing, 4. Surface measured f Rsi from of E-1 tests was approximately simulations. 15 C that of E-2 was approximately 16 C, indicating a difference of approximately 1 C. In Case 4, calculation of E-1 E-2 was 15.6 C 16.0 C, respectively; in this case, difference between C-1 measured calculated E-1 was E-2 5.4% 1.9%, respectively, F-1 which was F-2 Case smallest T si ( among C) all f Rsi cases. T si ( C) f Rsi T si ( C) f Rsi T si ( C) f Rsi T si ( C) f Rsi Test Table Surface frsi from tests simulations C E E F F Case 3.9% * 2.4% 18.9% 9.3% 13.4% 6.8% 5.6% 3.2% 5.3% 2.9% Tsi ( C) frsi Tsi ( C) frsi Tsi ( C) frsi Tsi ( C) frsi Tsi ( C) frsi Test % 1 3.9% * 2.4% 2.4% 19.6% 18.9% 9.7% 9.3% 13.4% 14.0% 6.8% 7.1% 5.6% 5.6% 3.2% 3.2% 5.3% 4.7% 2.9% 2.6% Sim % 2.4% 19.6% 9.7% 14.0% 7.1% 5.6% 3.2% 4.7% 2.6% Sim. 6.1% 3.7% 11.5% 5.7% 7.0% 3.6% 10.2% 5.7% 10.5% 5.9% % % % % 7.0% % % % % 5.9% % 4 3.7% 5.4% 2.7% 1.9% 1.0% 11.7% 6.6% 11.1% 6.2% 6.1% 3.7% 5.4% 2.7% 1.9% 1.0% 11.7% 6.6% 11.1% 6.2% * Percent * Percent error error of of difference between measured calculated data. data. Figure 8. Measured calculated s at each assessment point. Figure 8. Measured calculated s at each assessment point.

9 Energies 2018, 11, of 13 Energies 2018, 11, x FOR PEER REVIEW 9 of Comparison of Temperature Factors Energies 2018, 11, x FOR PEER REVIEW 9 of 13 Figure 9 shows comparison between experimental measurements numerical simulations 3.2. Comparison for ffrsi of derived Temperature using Factors surface of each assessment point. The experimental results Figure show 9 that shows center-of-glazing comparison between had best experimental performance measurements in reducing condensation numerical risk, followed simulations by for frame frsi derived using edge-of-glazing. surface The difference of each in assessment performance point. of The those experimental three window regions results was show significant, that center-of-glazing approximately 0.1 had frsi between best performance each region. in The reducing simulation condensation results showed risk, a arelatively followed small by difference. frame The The edge-of-glazing. simulation agreed The difference with in experimental performance of results those three in window sense of providing regions was significant, same ranking approximately of performance 0.1 frsi at at between each assessment each region. point. The The simulation The maximum results difference showed in a frsi in fbetween Rsi relatively between small edge-of-glazing difference. The simulation frame, frame, agreed for for example example with between between experimental E-1 E-1 results F-1, F-1, was in was calculated sense of to to be be 0.04 providing 0.04 by by simulation. same This ranking This value value of performance corresponds at to each a to assessment a point. difference difference The maximum of approximately of difference 2 in C. frsi 2The C. The closest between closest jamb jamb to edge-of-glazing Point to Point E-1, E-1, as represented as represented frame, in Figure in for Figure example 10c, affects 10c, between affects E-1 F-1, of was E-1. of calculated The E-1. yellow Theto yellow be color color of 0.04 of by jamb simulation. jamb in Figure in Figure This 10b value 10b represents corresponds lower lower to a compared difference to toof approximately orange color 2 of C. The or surfaces. closest jamb Radiation to Point exchange E-1, as represented occurs between in Figure 10c, jamb affects Point E-1. Since of E-1. The yellow color of jamb s of surface jamb in perpendicular Figure 10b represents to glazing lower is relatively low, compared as shown to in orange Figure color 10b, of radiation or at surfaces. Point E-1 Radiation is exchange reduced. occurs Convection is between also reduced jamb because Point E-1. jamb Since acts as an obstacle to to of airflow. jamb s surface perpendicular to glazing is relatively low, as shown in Figure 10b, radiation at Point E-1 is reduced. Convection is also reduced because jamb acts as an obstacle to airflow. Figure Figure Measured Measured calculated calculated factors factors (f (frsi) at each assessment point. Rsi ) at each assessment point. Figure 9. Measured calculated factors (frsi) at each assessment point. (a) (b) (c) (c) Figure D isorms: (a) isorms for a whole window; (b) (b) part part of of a window; (c) (c) isormal Figure 10. 3D isorms: (a) isorms for a whole window; (b) part of a window; (c) isormal lines lines Point E-1. lines Point E-1. Figure Figure shows differences in factor factor (frsi) (frsi) according according to to surface surface rmal rmal resistance. resistance. As As shown shown from comparison of of Case Case in in Figure Figure 11a d, 11a d, a small a small change change in in external external surface surface resistance resistance has has little little effect effect on on frsi of frsi of center-of-glazing, center-of-glazing, frame, frame, edgeedge-

10 indicates that a local change in surface rmal resistance of edge region has no significant effect on frsi of frame. As shown in Figure 11c,d, in contrast to center-of-glazing frame, edge-of-glazing consistently had higher frsi from simulated results than experimental results. The differences Energies between 2018, 11, results 382 of measurement simulations were less than 10%. The comparison of 10Cases of in Figure 11c,d shows effect of surface rmal resistance considering reduced radiation convection in edges. These conditions generated a difference of 0.02 frsi in calculation Figure 11 for shows edge-of-glazing, differencescorresponding to a factor (f Rsi ) according difference to of approximately surface rmal 1 C. resistance. As in Case As 4, shown accuracy from of predicting comparison of condensation Case 1 2 risk in Figure at 11a d, edge-of-glazing a small change could in be improved external most surface effectively resistance by using has little increased effect on local f Rsi surface of rmal center-of-glazing, resistance at frame, edges. For example, edge-of-glazing. relative to A comparison measurement of of Cases 1 3 indicates that a small change in internal frsi at Point E-1, simulation of Case 2 produces an error surface of approximately resistance generates 10%, while a change simulation of approximately of Case 4 reduces 0.03 f Rsi in error frame to less than edge-of-glazing, 3%, as shown in which Table corresponds 4. to a difference of approximately 1 C. (a) (b) (c) (d) Figure Figure Measured Measured calculated calculated factors factors (f (frsi) according to surface rmal Rsi ) according to surface rmal resistance: (a) resistance: center-of-glazing (a) center-of-glazing (C-1); (b) frame (C-1); (F-2); (b) (c) frame edge-of-glazing (F-2); (c) (E-1); edge-of-glazing (d) edge-of-glazing (E-1); (d) (E-2). edge-ofglazing (E-2). As shown in Figure 11a, for center-of-glazing, experimental measurements produced f Rsi values that were always higher than those in simulations. The agreement between measurements calculations was better for Cases 1 2 than for Cases 3 4; however, difference between two groups of cases was small (approximately 1%). There was no difference in Case 4, which considered reduced radiation convection in edges or junctions between two surfaces, Case 3, which did not consider se factors. This finding indicates that a small difference in surface rmal resistance has little effect on simulated f Rsi for center-of-glazing. As shown in Figure 11b, for frame, experimental results were always higher than those in simulations. The minimum deviation between measured calculated results for Point F-2 was 2.6% in Case 2. The frame also showed little difference between Cases 3 4. This finding indicates that a local change in surface rmal resistance of edge region has no significant effect on f Rsi of frame. As shown in Figure 11c,d, in contrast to center-of-glazing frame, edge-of-glazing consistently had higher f Rsi from simulated results than experimental results. The differences between results of measurement simulations were less than 10%. The comparison of Cases 3 4 in Figure 11c,d shows effect of surface rmal resistance considering reduced radiation convection in edges. These conditions generated a difference of 0.02 f Rsi in calculation

11 Energies 2018, 11, of 13 for edge-of-glazing, corresponding to a difference of approximately 1 C. As in Case 4, accuracy of predicting condensation risk at edge-of-glazing could be improved most effectively by using increased local surface rmal resistance at edges. For example, relative to measurement of f Rsi at Point E-1, simulation of Case 2 produces an error of approximately 10%, while simulation of Case 4 reduces error to less than 3%, as shown in Table Conclusions Future Work This study analyzed effect of surface rmal resistance on simulation accuracy for condensation risk assessment of a high-performance window. The 3D simulations were performed under several surface rmal resistance conditions, simulated results were compared with experimental results. For center-of-glazing frame, 3D numerical simulations featured relatively good agreement with experimental results. A small change in surface rmal resistance has no significant effect on accuracy of condensation risk assessment for center-of-glazing or frame. However, relatively poor agreement was found for edge-of-glazing, which is more vulnerable to condensation. The accuracy of predicting condensation risk at edge-of-glazing was significantly improved using increased local surface rmal resistance with 3D simulation, reflecting reduced radiation convection at edges or junctions between two surfaces; error between measured calculated factors (f Rsi ) could be reduced to less than 3%. However, simulated prediction of condensation risk for edge-of-glazing was still not conservative. The simulation tended to underestimate condensation risk, which might be due to rmal effects of operating hardware in high-performance windows [36], which were omitted from heat transfer calculation in this study. As this study addressed one case of a PVC-framed triple-glazed window, furr investigation should be performed for or window types. Acknowledgments: This research was supported by a grant (18RERP-B ) from Residential Environment Research Program funded by Ministry of L, Infrastructure, Transport of Korean government. This research was also supported by a KETEP grant (No ) funded by Korean Government. Author Contributions: Seung-Yeong Song So Young Koo conceived presented idea; So Young Koo, Sihyun Park Jin-Hee Song performed simulations experiment; So Young Koo wrote manuscript; Seung-Yeong Song supervised project. Conflicts of Interest: The authors declare no conflict of interest. Nomenclature f Rsi U w W H θ i θ e θ si Temperature factor, dimensionless Window rmal transmittance, W/m 2 K Width of a window, m Height of a window, m Internal, C External, C Internal surface, C References 1. Gustavsen, A.; Grynning, S.; Arasteh, D.K.; Jelle, B.P.; Goudey, H. Key elements of materials performance targets for highly insulating window frames. Energy Build. 2011, 43, [CrossRef] 2. ISO Thermal Bridges in Building Construction Heat Flows Surface Temperatures Detailed Calculations; International Stard Organization: Geneva, Switzerl, Berthier, J. Diffusion de Vapeur Au-Travers des Parois: Condensation; Sciences du Batiment; CSTB REEF: Nantes, France, 1980; Volume DIN : Thermal Protection Energy Economy in Buildings Part 2: Minimum Requirements to Thermal Insulation; Deutsches Institut für Normung: Berlin, Germany, 2013.

12 Energies 2018, 11, of Dutch Building Code Online. Available online: (accessed on 6 November 2017). 6. ÖNORM B : Wärmeschutz im Hochbau Wasserdampfdiffusion und Kondensationsschutz (Water Vapor Diffusion Protection against Condensation); Austrian Stards: Vienna, Austria, Building Regulations 2000: Approved Document L1A, L1B, L2A, L2B; Ministry of Housing, Communities Local Government: London, UK, Ministry of L, Infrastructure Transportation; Korean Design Stard for Preventing Condensation in Apartment Buildings No ; Ministry of L, Infrastructure Transportation: Seoul, Korea, (In Korean) 9. National Fenestration Rating Council (NFRC). Determining Fenestration Product Condensation Resistance Values; NFRC ; National Fenestration Rating Council (NFRC): Greenbelt, MD, USA, Passive House Institute. Criteria Algorithms for Certified Passive House Components: Transparent Building Components Opening Elements in Building Envelope, verion 5.1; Passive House Institute: Darmstadt, Germany, Bergh, V.D.; Hart, R.; Jelle, B.P.; Gustavsen, A. Window spacers edge seals in insulating glass units: A state of art review future perspectives. Energy Build. 2013, 59, Shin, M.S.; Rhee, K.N.; Yu, J.Y.; Jung, G.J. Determination of equivalent rmal conductivity of window spacers in consideration of condensation prevention energy saving performance. Energies 2017, 10, 717. [CrossRef] 13. Hong, G.; Kim, D.D.; Kim, B.S. Experimental investigation of rmal behaviors in window systems by monitoring of surface condensation using full-scale measurements simulation tools. Energies 2016, 9, 979. [CrossRef] 14. Lechowska, A.; Schonotale, J. The window edge-of-glass region profile improvement by inserting a small additional glass pane A CFD study measurements. J. Build. Eng. 2015, 4, [CrossRef] 15. Arasteh, D.; Dariush, K.; Beck, F.A.; Stone, N.; Dupont, W.; Mathis, R.C.; Koenig, M. PHASE I Results of NFRC U-value procedure validation project. ASHRAE Trans. 1994, 100, Gustavsen, A.; Uvslokk, S.; Jelle, B.P. Numerical experimental studies of effect of various glazing spacers on window U-value glazing. In Proceedings of 7th Symposium on Building Physics in Nordic Countries, Reykjavik, Icel, June 2005; pp Gustavsen, A.; Arasteh, D.; Jelle, B.P.; Curcija, D.C.; Kohler, C. Developing low-conductance window frames: Capabilities limitations of current window heat transfer design tools. J. Build. Phys. 2008, 32, [CrossRef] 18. ISO Thermal Performance of Windows, Doors, Shading Devices Detailed Calculations; International Stard Organization: Geneva, Switzerl, ISO Thermal Performance of Windows, Doors Shutters Calculation of Thermal Transmittance Part 2: Numerical Method for Frames; International Stard Organization: Geneva, Switzerl, Gustavsen, A.; Goudey, H.; Arasteh, D.; Uvslokk, S. Experimental numerical examination of rmal transmittance of high performance window frames. In Proceedings of Thermal Performance of Exterior Envelopes of Whole Buildings XI International Conference, Clearwater Beach, FL, USA, 5 9 December Kohler, C.; Arasteh, D.; Mitchell, R. THERM simulations of window indoor surface s for predicting condensation. ASHRAE Trans. 2002, 109, Curcija, D.; Arasteh, D.; Huizenga, C.; Kohler, C.; Mitchell, R.; Bhari, M. Analyzing rmal performance of building envelope components using 2-D heat transfer tool with detailed radiation modeling. In Proceedings of 7th International IBPSA Conference, Rio de Janeiro, Brazil, August 2001; pp Bystronic Glass Homepage. TPS The Original from Inventor. Available online: (accessed on 6 November 2017). 24. Lawrence Berkeley National Laboratory. WINDOW/THERM 6.3 Manual Available online: (accessed on 6 November 2017). 25. Lawrence Berkeley National Laboratory. THERM Material Library, THERM Available online: (accessed on 6 November 2017).

13 Energies 2018, 11, of Lawrence Berkeley National Laboratory. WINDOW Glass Gas Library, WINDOW Available online: (accessed on 6 November 2017). 27. ASHRAE. ASHRAE Hbook of Fundamentals; ASHRAE: Atlanta, GA, USA, Ministry of L, Infrastructure Transportation; Korean Design Stard for Energy-Efficient Building Design, Notification No ; Korean Ministry of L, Infrastructure Transportation: Seoul, Korea, (In Korean) 29. Korean Stard Association. KS F 2295 Test Method of Dew Condensation for Windows Doors; Korean Stard Association: Seoul, Korea, (In Korean) 30. Korean Stard Association. KS F 2278 Stard Test Method for Thermal Resistance for Windows Doors; Korean Stard Association: Seoul, Korea, (In Korean) 31. Physibel. TRISCO Manual of Version 12.0w, Maldegem, Belgium, Available online: physibel.be (accessed on 6 November 2017). 32. DIN EN 673. Glass in Building Determination of Thermal Transmittance (U Value) Calculation Method; German Institute for Stardization: Berlin, Germany, ISO Hygrormal Performance of Building Components Building Elements Internal Surface Temperature to Avoid Critical Surface Humidity Interstitial Condensation Calculation Methods; International Stard Organization: Geneva, Switzerl, ISO Building Components Building Elements Thermal Resistance Thermal Transmittance Calculation Method; International Stard Organization: Geneva, Switzerl, LG Hausys. Accredited Test Report; LG Hausys: Seoul, Korea, (In Korean) 36. Hart, R.; Misiopecki, C.; Gustavsen, A.; Jelle, B.P.; Arasteh, D. Impacts of operating hardware on window rmal performance. In Proceedings of Building Enclosure Science & Technology (BEST ), Atlanta, GA, USA, 2 4 April by authors. Licensee MDPI, Basel, Switzerl. This article is an open access article distributed under terms conditions of Creative Commons Attribution (CC BY) license (