- Email: [email protected]
On detailed thermal response modeling of vertical greenery systems as cooling measure for buildings and cities in summer conditions
Energy 115 (2016) 1055e1068
Contents lists available at ScienceDirect
Energy journal homepage: www.elsevier.com/locate/energy
On detailed thermal response modeling of vertical greenery systems as cooling measure for buildings and cities in summer conditions * Toma z Suklje , Saso Medved, Ciril Arkar Laboratory for Sustainable Technologies in Buildings, Faculty of Mechanical Engineering, University of Ljubljana, Slovenia
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 February 2016 Received in revised form 24 August 2016 Accepted 28 August 2016
Vertical greenery systems (VGSs) are becoming a common architectural element in urban environments. In addition to the aesthetics of VGSs, impacts on building's energy demand and heat island mitigation in cities has been identified. In the present study, experimental results of thermal response and properties of VGSs with vertical leaf area index (LAIV) equal to 6.1 and 7.2 are presented. Experimental results show that VGSs can impact up to 34 K lower surface temperatures of a façade, while maintaining air temperatures in the VGSs' canopies close to ambient temperatures. Properties of the VGSs were used as a basis as well as an input for a detailed mathematical model of the thermal response of a building envelope with a VGS. The validated mathematical model was used for parametrical analysis of the impact of thermal resistance of a building envelope on the cooling potential of the VGSs. The results show that the cooling effect is more significant for less insulated façades, and that a VGS can be modeled as an independent urban cooling element. Finally, a parametrical model of the latent heat flux of a VGS was developed and can be used as a boundary condition in urban heat island studies. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Green façade Evaporative cooling Urban heat island Volumetric cooling power Adaptive building envelope Adaptive façade
1. Introduction In the past, the primary role of greenery in the urban environments was its aesthetics [1]. Recently, the importance of greenery in urban environments has been increased due to the identified impact on the mitigation of the heat island effect [2,3] and buildings' energy demand, when applied on roofs [4,5] or façades [6]. The latter enable a greater potential for energy savings as façades represent the marginal share of a building envelope surface in urban environments. Due to the various architectural approaches, VGSs can be divided into direct and indirect VGSs, living wall, and double-skin façades with plants [7,8]. It has been proven that VGSs impact temperature and flow conditions at a building envelope boundary due to evaporative cooling, the shading of a façade, the selective absorption of solar radiation and structural properties of VGSs [9,10]. In addition, due to the adaptive properties of VGSs, greenery represents a convenient way to upgrade a static building envelope into an adaptive one [11], which is a rapidly developing research area [12,13]. The thermal response of VGSs is most frequently evaluated with
* Corresponding author. E-mail addresses: [email protected], [email protected] (T. Suklje). http://dx.doi.org/10.1016/j.energy.2016.08.095 0360-5442/© 2016 Elsevier Ltd. All rights reserved.
in situ measurements of temperatures in consecutive layers from a façade's surface towards the environment [14]. Results from the rez et al. [15] showed that VGSs whole-year survey carried out by Pe are most effective in summer conditions. Furthermore, due to the evaporative cooling and shading of the façade, surface temperatures of the façade with a VGS are up to 15 K lower in comparison to the reference façade [6,7,16,17]. Moreover, in the thermal response study of double-skin façade with plants, Stec et al. [6] reported that greenery is more efficient in the reduction of air temperature than conventional shades; however, according to findings from Fang et al. [17], it does not reduce the overheating of a double-skin façade. Conclusively, the lower surface temperature of a façade with a VGS can cause a 20% reduction of cooling demand [6]. In addition, lower surface temperatures prolong the life expectancy of a façade and reduce maintenance costs [16]. Wong et al. [14] showed that a VGS as an independent heat island mitigation measure does not reduce the urban heat island (UHI) effect significantly. A similar conclusion can be drawn from in situ experimental studies on air temperatures in the vicinity of a VGS [7,16]. In contrast, results from a comprehensive CFD study under the assumption of the application of the greenery on all urban surfaces from Vidrih [18] show significant reductions of air temperatures in the urban area. In order to model the thermal response of a building envelope
1056
T. Suklje et al. / Energy 115 (2016) 1055e1068
with a VGS, information on a VGS must be provided, such as density (characterized with the leaf area index (LAI) [19]), transmitivity [6], coverage factor or view factors and evapotranspiration (ET), defined as the amount of water evaporated from the surface of leaves [19]. Stec et al. [6] developed a lumped-capacitance model of the double skin façade with plants and validated it with laboratory experiments. For quantification of evapotranspiration, the Pennman-Monteith method [19] was used and compared with a proposed simplified method for ET. The transmitivity of the VGS was determined experimentally, though LAI was not evaluated. Most recently Flores Larsen et al. [20] adopted Stec's [6] model and upgraded the simplified model of the ET. Kontoleon and Eumorfopoulou [21] also developed a lumpedcapacitance model of VGS, validated it with results from previous research [22], and carried out a parametrical analysis of an impact of a façade orientation and proportion of a leaf coverage on the VGS's cooling performance. The evapotranspiration, i.e. latent heat flux, was not explained. Susorova et al. [23] developed a similar mathematical model, where the latent heat flux has been accounted for with an empirical relation for leaf temperature, based on the greenery properties available in the literature. VGSs in urban environments can be modeled in CFD tools either as a temperature boundary condition or a heat flux boundary condition based on the latent heat flux of a VGS [18]. The latter can be formed as the VGS's volumetric cooling power boundary condition, which is rarely mentioned in literature, according to Gromke et al. [24]. The majority of available studies focused on experiments, less so on numerical modeling, and few use experiments to validate mathematical models, thus providing general applicability. To the authors' knowledge, no research provides both a detailed validated mathematical model and experimentally quantified properties of VGS. The latter is essential for evaluation of radiation and latent heat transfer as two of the most influential physical processes in a VGS. In the present study, we have developed such a mathematical model and validated it with in situ experimental results of the thermal response of the VGS with two densities (LAIV equal to 6.1 and 7.2). Measured properties of VGSs, such as transmitivity, leaf area index and view factors, were used as model inputs. Furthermore, the model's applicability was demonstrated with a parametrical analysis of the VGSs' impact on a building envelope and vice-versa. In addition, the validated mathematical model was
used for the development of the parametrical model of the latent heat flux of the VGSs that can be used as a boundary condition in the UHI mitigation research. 2. Experiment The experiment, as depicted in Fig. 1a, has been designed based on the experience from the previous study [11] and upgraded with latest findings from the literature review. The experiment was located in the vicinity of the city center of Ljubljana, Slovenia (46.12059, 14.49543). 2.1. Plant selection and VGS design For the VGS, the plant Phaseolus Vulgaris L was chosen because it is adapted to the local climate, is fast growing, and has a thick canopy. The VGS was applied to the façade indirectly, forming an 8 cm air gap (microclimatic layer) with the façade (Fig. 1b). A supporting construction for the greenery was designed to allow an assembly of the one- and two-layered VGS (Fig. 1c and d), the canopy density thus varies, and enable a preparation of the back-up VGSs. Maximal availability of irrigation water was enabled with an automated irrigation system placed in the containers. 2.2. Experimental setup The central component of the experimental setup was a thermostated test cell. The interior air temperature of the test cell was maintained at 23.5 ± 0.5 C. A south-orientated façade of the test cell was divided by a non-transparent high-reflective wind barrier into two parts: the left side served as a reference façade and on the right side the VGS was installed (Fig. 2c). The wind barrier was also placed on the opposite side of the VGS, preventing wind-driven ambient air from entering the air gap between the VGS and the façade (Fig. 2c) [11]. The sensor distribution on the test facility is outlined in Fig. 2. Global solar radiation on the vertical surface was measured with the Kipp&Zonen CM11-P pyranometer (non-linearity ± 0.6%, temperature dependence ± 1%, tilt error ± 0.25%). The Kipp&Zonen CG1 pyrgeometer (non-linearity ±1%, temperature dependence ±2%) was used for measurements of downward long-wave radiation. Ambient air temperature (RTD ±0.5%) and relative humidity (AHLBORN FHA 646 E1 ±2% RH), and wind velocity (Fischer 451214 ± 0.3 m/s) were measured at a height 1.3 m in the front of
Fig. 1. A front view of the experiment with the thermo-stated test cell with and without the VGS (a), a side-view of the VGS positioned at the thermo-stated test cell (b), a schematic of the one-layered (c) and the two-layered VGS (d).
T. Suklje et al. / Energy 115 (2016) 1055e1068
1057
Fig. 2. A side view (a), a front view (b) and a top view (c) scheme of the experimental setup. Dimensions are presented in millimeters.
the façade. Air temperature in the VGS qVGS,j and in the air gap qag,j was measured with T-type thermocouples (±0.25 K), shielded from net radiation heat flux, while the surface temperature of the façade qse,j was measured with K-type thermocouples (±0.25 K). The surface temperature of the VGS was measured with the IR camera FLIR B335 (7.5e13 mm, ± 2%). Heat flux at the interior surface of the test cell was measured with the AHLBORN FQ90118 sensors (measurement accuracy ± 5%). Air temperature in the thermo-stated test cell was measured with T-type thermocouples (measurement accuracy ± 0.25 K). All sensors were properly calibrated before the installation in the experimental setup and connected to data acquisition units (AHLBORN Almemo 2290-8 and Agilent 34970A) capturing data every 30 s. Measurements were carried out in summer 2014.
2.3. Results Results from the experimental evaluation of the thermal response of the one- and the two-layered VGS are presented in terms of temperatures in three consecutive layers (Fig. 2a). In Fig. 3b, a comparison of surface temperatures of the façade with and without the one-layered VGS is shown for meteorological data in the corresponding test period (Fig. 3a). It can be ascertained that surface temperatures of the façade without the VGS are up to 29 K higher in the daytime and up to 2 K lower in the night-time. Furthermore, surface temperatures of the façade with the VGS are up to 9 K higher than temperatures in the air gap, as shown in
Fig. 3c. Temperatures in the air gap and in the VGS canopy are about the same and from 3 to 7 K higher than ambient air temperatures in the daytime. Conversely, temperatures in the air gap differ from air temperatures in the VGS canopy in the night-time. The latter are up to 1.5 K lower, whereas ambient air temperatures are lower by an additional 2 K. The difference in the surface temperatures of the façade with and without the two-layered VGS is even greater compared to the one-layered VGS (up to 34 K as depicted in Fig. 4b). Surface temperatures of the façade are up to 5.6 K higher in the daytime and up to 1 K lower in the night-time compared to temperatures in the air gap. Furthermore, temperatures in the VGS are up to 1.5 K higher than temperatures in the air gap. What is more, from Fig. 4c, it is evident that temperatures in the thicker VGS canopy are closer to ambient air temperatures in the daytime and up to 2.5 K higher in the nighttime. 2.4. Experimentally determined properties of the VGS At this point, the presented methods and results of evaluation of properties of the VGS are either used as a basis for a development of a mathematical model or used as inputs. 2.4.1. VGS-temperature based properties The measured surface temperatures of the VGS with the IR camera have been analyzed with the ThermaCAM Researcher Professional 2.9 software [25]. Average surface temperatures of the VGS have been calculated for a 0.5 m 0.5 m area (Fig. 5), 1 m above
1058
T. Suklje et al. / Energy 115 (2016) 1055e1068
Fig. 3. Meteorological data (a), surface temperatures of the façade with and without the one-layered VGS (a), temperatures in the VGS canopy, the air gap and surface temperatures of the façade behind the VGS (b).
the ground, considering the VGS's emissivity of 0.983 according to reference [26]. Average surface temperatures were then compared with air temperatures in the VGS canopy qVGS over four days (Fig. 6). Based on the comparison, it has been ascertained that temperatures are approximately the same (qs,VGS ¼ qVGS). 2.4.2. Canopy density of the VGS The density of the VGS canopy is most commonly evaluated with a leaf area index (LAI), which is defined as a one-sided surface area of leaves projected on a horizontal plane [19]. Based on results from previous VGS thermal response studies [9,11], it has been ascertained that the density of a VGS could be evaluated with a projection of leaves on a vertical plane and called a vertical leaf area index (LAIV). This was measured with a Li-Cor LAI-2200 plant canopy analyzer in 10 points at the same heights as temperatures in the air gap (Fig. 2a). Measurements were carried out before sunrise at the beginning, in the middle and at the end of the test period. The average LAIV of the one- and the two-layer VGS was 6.1 ± 0.5 and 7.2 ± 0.6, respectively. 2.4.3. Transmitivity Transmitivity of the VGSs tVGS has been determined with simultaneous measurements of solar irradiation in front of the
VGSs (reference) and behind the VGSs (Fig. 2). For the latter, the Kipp&Zonen CM11-P pyranometer was placed on the vertical guide. The guide was placed in three different positions alongside the façade. Measurements took place every hour during the daytime. With the use of statistical methods, it has been ascertained that the transmitivity of analyzed VGS is independent of meteorological parameters. Based on that finding, transmitivity has been determined as the mean of all measured points. The transmitivity of the one- and two-layered VGS tVGS was 0.17 ± 0.02 and 0.09 ± 0.02, respectively. 2.4.4. View factor A view factor of the façade through a VGS F has been determined experimentally, using a Nikon DS3200 digital single-lens reflex (DSLR) camera. Photographs were taken at the height of 1.3 m in front of the VGS. At the backside of the VGS, a white screen was placed in order to distinguish between the VGS and background. In the post-process, the background was cropped (Fig. 7a and c) using Corel PHOTO software [27]. Later, pictures were converted to binary (Fig. 7b and d) using Matlab [28]. Similarly to the transmitivity, the view factor was found to be independent of meteorological parameters, thus calculated as the ratio of black pixels to the sum of all pixels. The view factor of the one- and the two-layered VGS F was
T. Suklje et al. / Energy 115 (2016) 1055e1068
1059
Fig. 4. Meteorological data (a), surface temperatures of the façade with and without the two-layered VGS (b), temperatures in the VGS canopy, the air gap and surface temperatures of the façade behind the VGS (c).
0.042 and 0.005, respectively. 3. Mathematical model
Fig. 5. Average surface temperature of the VGS qs,VGS has been calculated for a 0.5 m 0.5 m area, 1 m above the ground, considering the VGS's emissivity of 0.983, measured on 11th August 2014.
For a mathematical model, one-dimensional heat transfer was assumed, considering experimentally determined properties of the VGSs. Based on the comparison of air temperatures in the VGS canopy and surface temperatures of the VGS (Section 2.4.1), it can be assumed that the VGS could be modeled as a homogeneous layer with no thermal resistance. The thermal capacity of the one- and the two-layered VGS is assumed to be equal to a 2 mm and a 4 mm thick water layer [29], respectively. Furthermore, in the previous studies, it was ascertained that the air temperature in the VGS canopy is not correlated with wind velocity [9,11]; thus, the VGSs can be considered to be an impermeable homogenous layer. The installation of a VGS impacts the boundary conditions at the façade. Boundary conditions are therefore considered in four planes: (i) exterior and (ii) interior surface of VGS; (iii) interior and (iv) exterior surface of the façade (Fig. 8). The façade of the thermostated test cell (Figs. 1a and 2a) consists of four layers with thermophysical properties presented in Table 1. Energy balance equations for n nodes of the considered VGS and façade surfaces are presented in Eq. (1). Within façade layers (Table 1) heat conduction is considered with the spatial
T. Suklje et al. / Energy 115 (2016) 1055e1068
35
35
30
30 θ [°C]
θ [°C]
1060
25 20
20 15
15
4/8/14 8:00
4/8/14 14:00
8/8/14 8:00
4/8/14 20:00
40
35
35
30 θ [°C]
θ [°C]
25
30 25
8/8/14 14:00
8/8/14 20:00
12/8/14 14:00
12/8/14 20:00
25 20
20
11/8/14 8:00
11/8/14 14:00
15
11/8/14 20:00 12/8/14 8:00 θVGS
θs,VGS
Fig. 6. Comparison of surface temperatures of the VGS and air temperature in the VGS canopy during four days.
asol,VGS an absorptivity of the VGS. The latter was determined according to the measured spectral reflectivity of the plant Phaseolus Vulgaris L [31]:
discretization according to reference [30].
ð1Þ
rVGS $dVGS $cp;VGS $
dqVGS ¼ q_ a;sol;VGS q_ IR;net;VGSamb dt q_ IR;net;VGSse q_ conv;VGSamb
Z
rsol;VGS ¼
q_ conv;VGSag q_ lat
rair $dair $cp;air $
ð3Þ
rref ;1 $dref ;1 $cp;ref ;1 $
ðnÞ
Z
Eðl; TÞrVGS ðlÞdl 1:1
¼ 0:232;
(3)
Eðl; TÞdl
assuming the emissivity of a black body at the temperature of 5800 K [32]. Considering the transmitivity of leaves to be negligible, the absorptivity of the VGS is, therefore, equal to 0.768. Through the VGS transmitted solar radiation is then absorbed by the façade surface with the absorptivity asol,se equal to 0.75.
dqse ¼ q_ a;sol;se q_ IR;net;seamb dt þq_ IR;net;VGSse q_ conv;se lref ;1 $ qse qref ;1
q_ a;sol;se ¼ tVGS $asol;se $Gglob;90
dqsi ¼ lref ;4 $ qref ;4 qsi q_ conv;si dt
rref ;4 $dref ;4 $cref ;4 $
(1)
3.1. Short-wave radiation heat transfer The absorbed solar radiation on the exterior surface of the VGS has been accounted for with the following expression:
q_ a;sol;VGS ¼ ð1 tVGS Þ$asol;VGS $Gglob;90
0:4
0:4
dqag ¼ q_ conv;VGSag þ q_ conv;se dt
ð2Þ
1:1
(2)
where Gglob,90 is global solar radiation on a vertical surface, tVGS is a measured transmitivity of VGS as presented in Section 2.4.3 and
(4)
3.2. Long-wave radiation heat transfer For net long-wave radiation heat transfer at the exterior surface of the VGS view, factors F from Section 2.4.4 and the emissivity of the VGS εIR,VGS equal to 0.983 [26] are considered together with the temperature of the VGS and downward long-wave radiation heat flux q_ IR;down .
4 q_ IR;down q_ IR;net;VGSamb ¼ ð1 FÞ$εIR;VGS s$TVGS
(5)
In a case in which experimental data of a downward long-wave · radiation heat flux qIR;down on a vertical plane (b ¼ 90 ) was not available, it was calculated as proposed in Eq. (6):
T. Suklje et al. / Energy 115 (2016) 1055e1068
1061
Fig. 7. Post-processed and binary pictures of one-layered ((a) and (b)) and two-layered ((c) and (d)) VGS.
Fig. 8. Boundary planes (i to iv) and heat fluxes considered in the mathematical model.
taken from Meteonorm database [34].
Table 1 Thermo-physical properties of the façade of the test cell. Layer
dref [m]
lref [W/(mK)]
rref [kg/m3]
cp,ref [J/(kgK)]
1 2 3 4
0.008 0.150 0.075 0.004
1.1 0.034 0.13 1.1
1800 30 450 1800
880 1200 1000 880
1 cos b 4 1 þ cos b 4 $Tground þ Tsky ; q_ IR;down ¼ s$ 2 2
1 Tsky ¼ Tamb $ εsky þ 0:8 1 εsky cc 4
εsky ¼ 0:711 þ 0:005$Tdew þ 0:73$
Tdew 2 100
(7)
(8)
Net long-wave heat transfer between the VGS and the façade is modeled as the two-surface enclosure, hence:
(6)
where Tground is ground temperature, which is assumed to be equal to the ambient air temperature Tamb, and Tsky is the sky temperature. The latter is calculated in respect to the ambient air temperature Tamb, the cloud coverage cc and the emissivity of the sky εsky (Eq. (7)). The emissivity of the sky εsky is determined with Eq. (8) [33]. The cloud coverage cc and dew temperatures Tdew data is
q_ IR;net;VGSse ¼ 1ε
4 T4 s TVGS se IR;VGS
εIR;VGS
1 þ þ ð1FÞ
1εIR;se εIR;se
;
(9)
where in addition to the presented variables in this section, εIR,se is the emissivity of the façade's surface, which has been determined experimentally and equals 0.98. Finally, net long-wave radiation heat transfer between the
T. Suklje et al. / Energy 115 (2016) 1055e1068
1062
façade and the ambient is given by:
4 q_ IR;down q_ IR;net;seamb ¼ F$εIR;se s$Tse
g ¼ 0:665$103 $patm (10)
Latent heat flux of the VGS is determined with the water evaporation mass-flow-rate m_ w and the latent heat rlat equals to 2490 kJ/kg [6].
q_ lat ¼ m_ w $rlat
D q_ net q_ ground þ rair cp;air ð
rlat D þ g 1 þ rras
pv;sat pv Þ ra
(12)
As the VGS is a vertical façade element, net radiation heat exchange has been evaluated in the vertical plane, hence neglecting radiation heat transfer with grounds q_ ground . The variables in the right-hand side of Eq. (13) are calculated using Eq. (2), Eq. (5), and Eq. (9).
q_ net ¼ q_ a;sol;VGS q_ IR;net;VGSamb q_ IR;VGSse
(13)
A surface resistance of leaves rs, which describes the resistance of water vapor transfer through the transpiring canopy, is evaluated with respect to the stomatal resistance of a well-illuminated leaf rl and density of the VGS. The latter is characterized by LAIV as presented in Section 2.4.2.
rS ¼
rl 0:5$LAIV
(14)
An aerodynamic resistance of leaves ra, which governs the heat and water vapor transfer from leaves into the ambient air, changes with the wind velocity v and its measuring position relative to the VGS zm, as well as the measuring position of ambient air relative humidity relative to the VGS zh:
! zm 2 3 d zom =
ra ¼
$ln
! zh 2 3 d zoh =
ln
qamb 4098 0:6108$exp q17:27$ amb þ237:3 ðqamb þ 237:3Þ
2
(17)
Variables pv and pv,sat are partial water vapor pressure and saturation water vapor pressure, respectively. They are determined with respect to ambient air temperature qamb and relative humidity RH in accordance with reference [19].
(11)
A process of water evaporation in plants is named evapotranspiration ET. ET in the VGSs was modeled with the Penman-Monteith equation [19]:
ET ¼
A slope of the saturation vapor pressure curve D is evaluated in respect to ambient air temperature qamb [19]:
D¼
3.3. Latent heat transfer
(16)
;
k2 v
(15)
while the remaining variables are calculated in accordance with reference [19], with respect to the thickness of the VGS canopy d. The values are presented in Table 2. Psychrometric constant g is calculated in respect to the atmospheric pressure patm, given in Table 2: Table 2 Constants for evaluation of ET.
3.4. Convective heat transfer In the study, convective heat transfer is evaluated with Newton's cooling law, considering specific airflow conditions at each analyzed surface with a specific convective heat transfer coefficient using Eq. (18)e(21). The VGS canopy density impacts convective heat transfer through the extension of the heat exchange surface. Therefore, a convective heat transfer coefficient is calculated based on LAIV. In addition, aerodynamic resistance ra (Eq. (15)) and thermo-physical properties of the air are included as proposed by Ayata et al. [36]:
cp;air $rair hconv;VGSamb ¼ LAIV$ ra
(18)
The convective heat transfer coefficient at the interior surface of the VGS is determined based on the Nusselt number Nu, characteristic length D, which is assumed to be equal to twice the width of the air gap between the VGS and the building envelope (Fig. 2a), and the thermal conductivity of air lair:
hconv;VGSag ¼ Nu$
lair D
(19)
The Nusselt number Nu for mixed convection at the interior surface of the VGS is determined with an expression proposed by Stanghellini [37] and used in the VGS thermal response study by Stec et al. [6]. The VGS acts as an impermeable layer [9]; consequently, no-wind conditions are considered in the air gap. For nonwind conditions in urban settlements, it is assumed that the wind velocity is lesser than 0.2 m/s [38]. Thus, the Reynolds number Re is calculated for the air velocity equal to 0.2 m/s. In addition to the forced convection, the buoyancy-driven convection is taken into account with the inclusion of the Grashof number Gr in the following equation:
0:25 Nu ¼ 0:37$ Gr þ 6:416$Re2
(20)
The convective heat transfer at the façade surface is assumed to be buoyancy-driven and is evaluated with expression [39]:
1=3 hconv;se ¼ 1:31$ qse qag
(21)
At the interior surface of the thermo-stated test cell a constant convective heat transfer coefficient hconv,si is assumed. According to reference [40], value 7.7 W/(m2K) accounts for convective and radiation heat transfer.
Parameter
Value
Reference/comment
εIR,VGS [] rl [s/m] k [] zm, zh [m] d [m] p [kPa]
0.983 100 0.41 2.0/1.7 0.3/0.6 97.7
[26] [19,35] [19] measured for the one-layered/the two-layered VGS
3.5. Experimental results and the model validation
[19]
The presented mathematical model was implemented in Simulink [41], where the measured meteorological parameters
T. Suklje et al. / Energy 115 (2016) 1055e1068
1063
Fig. 9. Calculated and measured temperatures of the one-layered VGS (a), the air gap (b) and the exterior surface of the façade (c); heat flux at the interior surface of the test cell (d).
1064
T. Suklje et al. / Energy 115 (2016) 1055e1068
Fig. 10. Calculated and measured air temperatures of two-layered VGS (a), the air gap (b) and the exterior surface of the façade (c); heat fluxes at the interior surface of the test cell (d).
T. Suklje et al. / Energy 115 (2016) 1055e1068
(Figs. 3a and 4a) together with the VGS properties were used as the inputs. Numerical results were then validated with experimental results of three consecutive temperature nodes (air temperature of the VGS qVGS, the temperature of the air gap qag and the surface temperature of the façade qse e Fig. 2a) and heat flux at the interior surface of the test cell q_ i . Calculated and measured air temperatures of the one-layered (Fig. 9a) and the two-layered (Fig. 10a) VGS are in good agreement; however, maximal calculated daily temperatures are up to 3.5 K lower than the measured ones. Better agreement of the calculated and measured results can be found for temperatures of the air gap, where temperature difference does not exceed 2 K (Figs. 9b and 10b) during the daytime and 2.5 K during the nighttime. Furthermore, the calculated surface temperatures of the façade with the VGS differ from measured values by no more than 3 K, for the one-layered (Fig. 9c) as well as for the two-layered VGS (Fig. 10c). Calculated temperatures in all considered temperature nodes of the two-layered VGS on 11th July 2014 are considerably lower than measured ones due to the stormy weather. Finally, calculated and measured heat fluxes on the interior surface of the test cell with the one-layered (Fig. 9d) and the twolayered (Fig. 10d) VGS were compared. It has been ascertained that the calculated results do not exceed measured results by more than the measurement uncertainty accounted for dynamic conditions (±11.7%). The measurement uncertainties associated with instrumentation and experimental procedures have an effect on numerical results. To investigate these effects, a measurement uncertainty analysis was carried out. Due to the complexity of the mathematical model, a Monte Carlo method was used for the propagation of distributions [42,43]. For the analysis, a uniform probability density function (PDF) was associated with the input sources, which covered the measured meteorological data as well as LAIV, tVGS, and interior air temperature qi. Results of the analysis are shown in Figs. 9 and 10 as the measurement uncertainty intervals.
4. Parametrical analysis 4.1. Impact of VGS on thermal response of a façade The validated thermal response model of the façade with the two-layered VGS has been used to quantify a cooling potential of the VGS with respect to the thermal resistance of a façade. The cooling potential was evaluated with a comparison of heat fluxes at
1065
an interior surface of a façade with and without the VGS for meteorological conditions presented in Fig. 3a. In the analysis, a façade with the thermal resistance equal to 5, 2.1 and 0.6 m2 K/W was considered. The thermal resistance of a façade is chosen based on the typical U-vales of a building envelope in Eastern and Southern Europe. That is for a well-insulated low-energy building (0.19 W/m2K), a moderately insulated building (0.45 W/m2K) and a non-insulated building (1.3 W/m2K). The interior temperature was set at a constant value of 25 C. Fig. 11 shows the heat fluxes at the interior surface of a façade without and with the VGS for the selected period. It can be concluded that peak heat flux at the interior surface of a façade with a thermal resistance of 0.6 m2 K/W can be reduced by a maximum of 41.2 W/m2. Considerably lower reductions of peak heat flux can be identified (up to 12.2 W/m2) in the case of a façade with the thermal resistance of 2.1 m2 K/W. A heat flux is further lessened (as much as 5 W/m2) in the case of a façade with the highest thermal resistance among compared façades. Relatively to a reference façade, peak heat fluxes of façades with the VGS have all been decreased by 77%. The analysis of the average heat flux at the interior surface of a façade with the VGS shows a reduction of 11.2, 3.9 and 1.5 W/m2, respectively. It has been ascertained that the VGS, in addition to the neutralization of heat gains from the exterior environment, acts as an effective cooling measure regardless of the thermal resistance of a façade. The effect is more pronounced for less thermally insulated façades. From Fig. 11, it is also evident that a negative heat flux at the interior surface of a façade with the VGS is greater than one at a reference façade regardless of the thermal resistance of a façade. Based on the analysis of temperatures in the supporting layer of a façade, it has been concluded that the observed phenomena is a consequence of the heat accumulation.
4.2. Cooling effect of the VGS on the urban environment In order to provide a general model of the latent heat flux of the VGS, an impact of thermo-physical and optical properties of a façade on the latent heat flux was parametrically investigated. For this purpose, the validated thermal response model of the façade with the VGS was used with measured meteorological data (Figs. 3a and 4a) and supplemented with the test reference year (TRY) for Ljubljana, Slovenia from July to August [34]. Due to a difference in a position of the wind velocity measurements in TRY and the experiment, a correction of wind velocity is taken into account with
60 50
q̇ si [W/m2]
40 30 20 10 0 -10 -20 R=5_ref
R=2.1_ref
R=0.6_ref
R=5_VGS
R=2.1_VGS
R=0.6_VGS
Fig. 11. Heat fluxes at the interior surface of a façade with and without the two-layered VGS, altering the thermal resistance between 5, 2.1 and 0.6 m2 K/W.
T. Suklje et al. / Energy 115 (2016) 1055e1068
1066
Fig. 12. Latent heat flux of the two-layered VGS installed at a façade with the thermal resistance of 5, 2.1 and 0.6 m2 K/W (a) and the solar absorptivity 0.3, 0.75 and 0.9 (b).
Eq. (22), as proposed by Nottrott et al. [44].
vcorr ¼ 0:17$v þ 0:099
(22) C.
The interior temperature was considered constant at 25 The thermal resistance of a façade was considered to be equal as in the previous section, whereas values of the façade's solar absorptivity equal to 0.3, 0.75 and 0.9 were chosen. Results from the parametrical investigation are presented in Fig. 12. As can be seen from Fig. 12a, the thermal resistance of a façade impacts the latent heat flux of the VGS insignificantly. The same can be concluded for the impact of the façade's solar absorptivity (Fig. 12b). Conclusively, the VGS can be modeled as an independent urban cooling element in the UHI mitigation research. 4.2.1. Parametrical model of a latent heat flux of the VGS A parametrical model was developed in relation to the significant variables. A range of the validity of the model is presented in Eq. (23). In addition to the ambient air temperature and relative humidity and wind velocity, net radiation heat flux has been taken into account. Based on the findings from the previous section, net Table 3 Coefficients of parametrical model of latent heat flux. LAIV
b1
b2
b3
b4
b5
b6
b7
b8
6.1 7.2
15610.9 16349.7
106.49 111.18
0.18 0.19
5.03 5.71
0.03 0.03
2.79 4.66
0.648 0.581
0.0000 0.0002
radiation heat flux is given by Eq. (24), considering only absorbed solar radiation and net long-wave radiation heat flux with the ambient surroundings. 2 þ b4 $RH þ b5 $RH 2 þ b6 $v q_ lat ¼ b1 þ b2 $Tamb þ b3 $Tamb 2 þb7 $q_ net þ b8 $q_ net
8 282K T 9 amb 305K > > > > < 36% RH 86% = 0:2m=s v 3:5m=s . > >
> > : ; 30W m2 q_ net 625W m2 4 q_ IR;down q_ net ¼ asol;VGS $G90;glob εIR;VGS s$Ts;VGS
(23)
(24)
Coefficients in Eq. (23) were determined separately for the onelayered and the two-layered VGS and are presented in Table 3. Based on the comparison of the latent heat flux of the one- and the two-layered VGS calculated with the numerical and the parametrical model (presented in Fig. 13), it has been ascertained that proposed parametrical models are suitable. As expected, results show that the latent heat flux of the two-layered VGS is more intensive compared to the one-layered VGS. However, the difference does not exceed 10% at the highest of the latent heat flux. Consequently, the one-layered VGS is the more favorable choice to two-layered VGS as it takes less space with the comparable cooling effect. Models in such a form can be used by the urban physics researchers for the evaluation of the cooling potential of an
Fig. 13. Comparison of the latent heat flux of the one-layered (a) and the two-layered (b) VGS calculated with the numerical and the parametrical model.
T. Suklje et al. / Energy 115 (2016) 1055e1068
1067
Fig. 14. Plot of a latent heat flux of the two-layered VGS in relation to the ambient air temperature, relative humidity and net radiation heat flux for wind velocity 0.5 m/s (a) and 2 m/s (b).
arbitrary VGS in arbitrary urban areas. The parametrical model of the two-layered VGS was also used to demonstrate the relation between latent heat flux and the independent variables. In Fig. 14, a consecutive surface presents the relation of the latent heat flux to ambient air temperature and relative humidity at discrete values of the net radiation heat flux and wind velocity. At the constant value of the other independent variables, the latent heat flux varies less than 10 W/m2 due to the change in wind velocity. Furthermore, a 100 W/m2 change in net radiation heat flux increases latent heat flux by 58 W/m2. Moreover, a change within the given range (from 284 K to 304 K with 5 K step) of ambient air temperature reflects an increase of latent heat flux from 6 W/m2 to 34 W/m2. Finally, latent heat flux reduces from 92 W/m2 to 6 W/m2 increasing ambient air relative humidity from 35% to 85% with a 10% step.
From the parametrical analysis of the impact of the thermal resistance and the solar absorptivity of a façade on the latent heat flux of the VGS, it can be concluded that neither is influential. Based on these findings the parametrical model of the latent heat flux of the VGS was developed and can be used either as a heat flux or a volumetric cooling power boundary condition in CFD modeling. Furthermore, it has been ascertained that the one-layered VGS is the more favorable choice in comparison to the two-layered VGS as it takes less space with the comparable cooling effect. Conclusively, the validated thermal response model of the VGS can be used to evaluate the cooling potential of an arbitrary VGS in a diverse urban areas. In the future, an expanded study of the overall performance of a building with the VGS should be performed, and applicability of the parametrical models demonstrated in CFD tools.
5. Conclusions
Nomenclature
This paper presents an experimental and numerical investigation of the thermal response of the VGS. Considerable attention has been paid to the experimental quantification of the properties of the VGS including transmitivity, view factors, and foliage density, which is rarely mentioned in the literature, to the authors' knowledge. Determined properties of the VGSs were used for the design of the mathematical model of the thermal response of the VGSs as well as inputs for the model. In contrast to the existing thermal response models of VGSs available in the literature, the current one provides repeatability and traceability of the results, which is due to the detailed description of the mathematical model and inputs. In the experimental survey, the evaluation of temperature conditions and properties of the VGS with LAIV densities of 6.1 and 7.2 are studied. Experimental results show that the VGS can have an impact up to 34 K lower surface temperatures of the façade, while maintaining air temperatures in the VGS canopy close to ambient air temperatures. The validated mathematical model has been used for the parametrical study of the cooling potential of the VGS. In the analysis that included a façade with different thermal resistances, it has been concluded that peak heat fluxes at the interior surface of compared façades with the VGS can be reduced by 77%. Furthermore, it has been ascertained that the VGS acts as an effective cooling measure for a building. In addition, the cooling effect appears to be inversely proportionate with the thermal resistance of a façade.
cc d D ET F Gglob,90 h k p q_ R RH ra rl rlat rs T
cloud coverage [] thickness [m] characteristic length [m] evapotranspiration [kg/s] view factor [] global solar radiation on vertical [W/m2] convective heat transfer coefficient [W/(m2K)] von Karman constant [] pressure [kPa] heat flux [W/m2] thermal resistance [m2K/W] ambient air relative humidity [%] aerodynamic resistance [s/m] stomatal resistance [s/m] latent heat [kJ/kg] surface resistance of leaves [s/m] absolute temperature [K]
Greek symbols a absorptivity [] ε emissivity [] q temperature [ C] l thermal conductivity [W/(mK)] r density [kg/m3] r reflectivity []
1068
T. Suklje et al. / Energy 115 (2016) 1055e1068
Subscripts a absorbed ag air gap amb ambient air atm atmospheric conv convective corr corrected dew dew point temperature e exterior exp experimental i interior IR long-wave radiation lat latent pm parametrical model ref reference façade net net num numerical s surface sat saturated se exterior surface of the test cell si interior surface of the test cell sol solar radiation sky sky v water vapor VGS vertical greenery system w water References €hler M. Green facadesda view back and some visions. Urban Ecosyst [1] Ko 2008;11(4):423e36. http://dx.doi.org/10.1007/s11252-008-0063-x. [2] Vidrih B, Medved S. Multiparametric model of urban park cooling island. Urban For Urban Green 2013;12(2):220e9. http://dx.doi.org/10.1016/ j.ufug.2013.01.002. [3] Price A, Jones E, Jefferson F. Vertical greenery systems as a strategy in urban heat island mitigation. Water Air Soil Pollut 2015;226(8):1e11. http:// dx.doi.org/10.1007/s11270-015-2464-9. [4] Bevilacqua P, et al. Plant cover and floristic composition effect on thermal behaviour of extensive green roofs. Build Environ 2015;92:305e16. http:// dx.doi.org/10.1016/j.buildenv.2015.04.026. [5] Tsang SW, Jim CY. Theoretical evaluation of thermal and energy performance of tropical green roofs. Energy 2011;36(5):3590e8. http://dx.doi.org/10.1016/ j.energy.2011.03.072. [6] Stec WJ, Paassen AHC, Maziarz A. Modelling the double skin façade with plants. Energy Build 2005;37(5):419e27. http://dx.doi.org/10.1016/ j.enbuild.2004.08.008. [7] Perini K, Ottele M, Fraaij AL, Haas EM, Raiteri R. Vertical greening systems and the effect on air flow and temperature on the building envelope. Build Environ 2011;46:2287e94. M, Perini K, Fraaij ALA, Haas EM, Raiteri R. Comparative life cycle [8] Ottele analysis for green façades and living wall systems. Energy Build 2011;43(12): 3419e29. http://dx.doi.org/10.1016/j.enbuild.2011.09.010. [9] Suklje T, Arkar C, Medved S. The local ventilation system coupled with the indirect green façade: a preliminary study. Int J Des Nat Ecodyn 2014;9(4): 314e20. http://dx.doi.org/10.2495/DNE-V9-N4-314-320. [10] Jim CY. Cold-season solar input and ambivalent thermal behavior brought by climber greenwalls. Energy 2015;90(1):926e38. http://dx.doi.org/10.1016/ j.energy.2015.07.127. [11] Suklje T, Medved S, Arkar C. An experimental study on a microclimatic layer of a bionic façade inspired by vertical greenery. J Bionic Eng 2013;10(2):177e85. http://dx.doi.org/10.1016/S1672-6529(13)60213-9. stola D, Hensen JLM. Framework for assessing [12] Kasinalis C, Loonen RCGM, Co the performance potential of seasonally adaptable facades using multiobjective optimization. Energy Build 2014;79:106e13. http://dx.doi.org/ 10.1016/j.enbuild.2014.04.045. stola D, Hensen JLM. Climate adaptive building [13] Loonen RCGM, Tr cka M, Co shells: state-of-the-art and future challenges. Renew Sustain Energy Rev 2013;25:483e93. http://dx.doi.org/10.1016/j.rser.2013.04.016. [14] Wong NH, Tan AYK, Tan YP, Wong NC. Energy simulation of vertical greenery systems. Energy Build 2009;41:1401e8. rez G, Rinco n L, Vila A, Gonza lez JM, Cabeza LF. Green vertical systems for [15] Pe
[16] [17]
[18]
[19] [20]
[21]
[22]
[23]
[24]
[25] [26]
[27] [28] [29] [30] [31]
[32] [33] [34] [35]
[36]
[37]
[38]
[39]
[40]
[41] [42]
[43]
[44]
buildings as passive systems for energy savings. Appl Energy 2011;88: 4854e9. Wong NH, et al. Thermal evaluation of vertical greenery systems for building walls. Build Environ 2010;45:663e72. Fang W, Xiaosong Z, Junjie T, Xiuwei L. The thermal performance of double skin façade with Tillandsia usneoides plant curtain. Energy Build 2011;43(9): 2127e33. http://dx.doi.org/10.1016/j.enbuild.2011.04.021. Vidrih B. Impact of global climate changes and urbanization on local climatic condition in cities. PhD thesis. Ljubljana: University of Ljubljana, Faculty of Mechanical Engineering; 2011. Allen RG, Pereira LS, Raes D, Smith M. FAO irrigation and drainage Paper No. 56: crop evapotranspiration. 2001. Flores Larsen S, Filippín C, Lesino G. Modeling double skin green façades with traditional thermal simulation software. Sol Energy 2015;121:56e67. http:// dx.doi.org/10.1016/j.solener.2015.08.033. Kontoleon KJ, Eumorfopoulou EA. The effect of the orientation and proportion of a plant-covered wall layer on the thermal performance of a building zone. Build Environ 2010;45(5):1287e303. http://dx.doi.org/10.1016/ j.buildenv.2009.11.013. Eumorfopoulou EA, Kontoleon KJ. Experimental approach to the contribution of plant-covered walls to the thermal behaviour of building envelopes. Build Environ 2009;44(5):1024e38. http://dx.doi.org/10.1016/ j.buildenv.2008.07.004. Susorova I, Angulo M, Bahrami P, Brent S. A model of vegetated exterior facades for evaluation of wall thermal performance. Build Environ 2013;67: 1e13. http://dx.doi.org/10.1016/j.buildenv.2013.04.027. Gromke C, Blocken B, Janssen W, Merema B, van Hooff T, Timmermans H. CFD analysis of transpirational cooling by vegetation: case study for specific meteorological conditions during a heat wave in Arnhem, Netherlands. Build Environ 2015;83(0):11e26. http://dx.doi.org/10.1016/j.buildenv.2014.04.022. FLIR. ThermaCAM researcher professional 2.9. 2005. Lopez A, Valera DL, Pena A. Determining the emissivity of the leaves of nine horticultural crops by means of infrared thermography. Sci Hortic Amst 2012;137. http://dx.doi.org/10.1016/j.scienta.2012.01.022. 10e10. CorelCorporation. CorelPHOTO-PAINT. 2014. MathWorks. MATLAB R2014a. 2013. Holm D. Thermal improvement by means of leaf cover on external walls d a simulation model. Energy Build 1989;14:19e30. Davies MG. Building heat transfer. Chichester, England: John Wiley & Son; 2004. Gutierrez MR, Escalante JAE, Rodriguez MTG. Canopy reflectance, stomatal conductance, and yield of Phaseolus vulgaris L. and Phaseolus coccinues L. under saline field conditions. Int J Agric Biol 2005;7(3). Incropera FP, DeWitt DP, Bergman, Lavine. Fundamentals of heat and mass transfer. sixth ed. , New York: Wiley; 2006. TRNSYS16. Mathematical reference, vol. 5; 2006. Meteonorm. Global meteorological database for applied climatology. 2004. Malys L, Musy M, Inard C. A hydrothermal model to assess the impact of green walls on urban microclimate and building energy consumption. Build Environ 2014;73(0):187e97. http://dx.doi.org/10.1016/j.buildenv.2013.12.012. Ayata T, Tabares-Velasco PC, Srebric J. An investigation of sensible heat fluxes at a green roof in a laboratory setup. Build Environ 2011;46(9):1851e61. http://dx.doi.org/10.1016/j.buildenv.2011.03.006. Stanghellini C. Mixed convection above greenhouse crop canopies. Agric For Meteorol 1993;66(1e2):111e7. http://dx.doi.org/10.1016/0168-1923(93) 90085-V. Niachou K, Hassid S, Santamouris M, Livada I. Experimental performance investigation of natural, mechanical and hybrid ventilation in urban environment. Build Environ 2008;43(8):1373e82. http://dx.doi.org/10.1016/ j.buildenv.2007.01.046. Blanco JM, Arriaga P, Rojí E, Cuadrado J. Investigating the thermal behavior of double-skin perforated sheet façades: part A: model characterization and validation procedure. Build Environ 2014;82(0):50e62. http://dx.doi.org/ 10.1016/j.buildenv.2014.08.007. EN ISO 13786. In: Thermal peroformance of building components - dynamic thermal characteristics - calculation methods. Brussels: European commitee for standardization; 1999. Matlab&Simulink. Mathwroks. 2014. Asuero AG, Gonza lez AG. Estimation of the uncertainty of inHerrador MA, direct measurements from the propagation of distributions by using the Monte-Carlo method: an overview. Chemom Intell Lab Syst 2005;79(1e2): 115e22. http://dx.doi.org/10.1016/j.chemolab.2005.04.010. Couto PRG, Damasceno JC, Oliveira SP. Monte carlo simulations applied to uncertainty in measurement. In: Chan VWK, editor. Theory and applications of monte carlo simulations. Intech; 2013. Nottrott A, Onomura S, Inagaki A, Kanda M, Kleissl J. Convective heat transfer on leeward building walls in an urban environment: measurements in an outdoor scale model. Int J Heat Mass Transf 2011;54(15e16):3128e38. http:// dx.doi.org/10.1016/j.ijheatmasstransfer.2011.04.020.
Contents lists available at ScienceDirect
Energy journal homepage: www.elsevier.com/locate/energy
On detailed thermal response modeling of vertical greenery systems as cooling measure for buildings and cities in summer conditions * Toma z Suklje , Saso Medved, Ciril Arkar Laboratory for Sustainable Technologies in Buildings, Faculty of Mechanical Engineering, University of Ljubljana, Slovenia
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 February 2016 Received in revised form 24 August 2016 Accepted 28 August 2016
Vertical greenery systems (VGSs) are becoming a common architectural element in urban environments. In addition to the aesthetics of VGSs, impacts on building's energy demand and heat island mitigation in cities has been identified. In the present study, experimental results of thermal response and properties of VGSs with vertical leaf area index (LAIV) equal to 6.1 and 7.2 are presented. Experimental results show that VGSs can impact up to 34 K lower surface temperatures of a façade, while maintaining air temperatures in the VGSs' canopies close to ambient temperatures. Properties of the VGSs were used as a basis as well as an input for a detailed mathematical model of the thermal response of a building envelope with a VGS. The validated mathematical model was used for parametrical analysis of the impact of thermal resistance of a building envelope on the cooling potential of the VGSs. The results show that the cooling effect is more significant for less insulated façades, and that a VGS can be modeled as an independent urban cooling element. Finally, a parametrical model of the latent heat flux of a VGS was developed and can be used as a boundary condition in urban heat island studies. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Green façade Evaporative cooling Urban heat island Volumetric cooling power Adaptive building envelope Adaptive façade
1. Introduction In the past, the primary role of greenery in the urban environments was its aesthetics [1]. Recently, the importance of greenery in urban environments has been increased due to the identified impact on the mitigation of the heat island effect [2,3] and buildings' energy demand, when applied on roofs [4,5] or façades [6]. The latter enable a greater potential for energy savings as façades represent the marginal share of a building envelope surface in urban environments. Due to the various architectural approaches, VGSs can be divided into direct and indirect VGSs, living wall, and double-skin façades with plants [7,8]. It has been proven that VGSs impact temperature and flow conditions at a building envelope boundary due to evaporative cooling, the shading of a façade, the selective absorption of solar radiation and structural properties of VGSs [9,10]. In addition, due to the adaptive properties of VGSs, greenery represents a convenient way to upgrade a static building envelope into an adaptive one [11], which is a rapidly developing research area [12,13]. The thermal response of VGSs is most frequently evaluated with
* Corresponding author. E-mail addresses: [email protected], [email protected] (T. Suklje). http://dx.doi.org/10.1016/j.energy.2016.08.095 0360-5442/© 2016 Elsevier Ltd. All rights reserved.
in situ measurements of temperatures in consecutive layers from a façade's surface towards the environment [14]. Results from the rez et al. [15] showed that VGSs whole-year survey carried out by Pe are most effective in summer conditions. Furthermore, due to the evaporative cooling and shading of the façade, surface temperatures of the façade with a VGS are up to 15 K lower in comparison to the reference façade [6,7,16,17]. Moreover, in the thermal response study of double-skin façade with plants, Stec et al. [6] reported that greenery is more efficient in the reduction of air temperature than conventional shades; however, according to findings from Fang et al. [17], it does not reduce the overheating of a double-skin façade. Conclusively, the lower surface temperature of a façade with a VGS can cause a 20% reduction of cooling demand [6]. In addition, lower surface temperatures prolong the life expectancy of a façade and reduce maintenance costs [16]. Wong et al. [14] showed that a VGS as an independent heat island mitigation measure does not reduce the urban heat island (UHI) effect significantly. A similar conclusion can be drawn from in situ experimental studies on air temperatures in the vicinity of a VGS [7,16]. In contrast, results from a comprehensive CFD study under the assumption of the application of the greenery on all urban surfaces from Vidrih [18] show significant reductions of air temperatures in the urban area. In order to model the thermal response of a building envelope
1056
T. Suklje et al. / Energy 115 (2016) 1055e1068
with a VGS, information on a VGS must be provided, such as density (characterized with the leaf area index (LAI) [19]), transmitivity [6], coverage factor or view factors and evapotranspiration (ET), defined as the amount of water evaporated from the surface of leaves [19]. Stec et al. [6] developed a lumped-capacitance model of the double skin façade with plants and validated it with laboratory experiments. For quantification of evapotranspiration, the Pennman-Monteith method [19] was used and compared with a proposed simplified method for ET. The transmitivity of the VGS was determined experimentally, though LAI was not evaluated. Most recently Flores Larsen et al. [20] adopted Stec's [6] model and upgraded the simplified model of the ET. Kontoleon and Eumorfopoulou [21] also developed a lumpedcapacitance model of VGS, validated it with results from previous research [22], and carried out a parametrical analysis of an impact of a façade orientation and proportion of a leaf coverage on the VGS's cooling performance. The evapotranspiration, i.e. latent heat flux, was not explained. Susorova et al. [23] developed a similar mathematical model, where the latent heat flux has been accounted for with an empirical relation for leaf temperature, based on the greenery properties available in the literature. VGSs in urban environments can be modeled in CFD tools either as a temperature boundary condition or a heat flux boundary condition based on the latent heat flux of a VGS [18]. The latter can be formed as the VGS's volumetric cooling power boundary condition, which is rarely mentioned in literature, according to Gromke et al. [24]. The majority of available studies focused on experiments, less so on numerical modeling, and few use experiments to validate mathematical models, thus providing general applicability. To the authors' knowledge, no research provides both a detailed validated mathematical model and experimentally quantified properties of VGS. The latter is essential for evaluation of radiation and latent heat transfer as two of the most influential physical processes in a VGS. In the present study, we have developed such a mathematical model and validated it with in situ experimental results of the thermal response of the VGS with two densities (LAIV equal to 6.1 and 7.2). Measured properties of VGSs, such as transmitivity, leaf area index and view factors, were used as model inputs. Furthermore, the model's applicability was demonstrated with a parametrical analysis of the VGSs' impact on a building envelope and vice-versa. In addition, the validated mathematical model was
used for the development of the parametrical model of the latent heat flux of the VGSs that can be used as a boundary condition in the UHI mitigation research. 2. Experiment The experiment, as depicted in Fig. 1a, has been designed based on the experience from the previous study [11] and upgraded with latest findings from the literature review. The experiment was located in the vicinity of the city center of Ljubljana, Slovenia (46.12059, 14.49543). 2.1. Plant selection and VGS design For the VGS, the plant Phaseolus Vulgaris L was chosen because it is adapted to the local climate, is fast growing, and has a thick canopy. The VGS was applied to the façade indirectly, forming an 8 cm air gap (microclimatic layer) with the façade (Fig. 1b). A supporting construction for the greenery was designed to allow an assembly of the one- and two-layered VGS (Fig. 1c and d), the canopy density thus varies, and enable a preparation of the back-up VGSs. Maximal availability of irrigation water was enabled with an automated irrigation system placed in the containers. 2.2. Experimental setup The central component of the experimental setup was a thermostated test cell. The interior air temperature of the test cell was maintained at 23.5 ± 0.5 C. A south-orientated façade of the test cell was divided by a non-transparent high-reflective wind barrier into two parts: the left side served as a reference façade and on the right side the VGS was installed (Fig. 2c). The wind barrier was also placed on the opposite side of the VGS, preventing wind-driven ambient air from entering the air gap between the VGS and the façade (Fig. 2c) [11]. The sensor distribution on the test facility is outlined in Fig. 2. Global solar radiation on the vertical surface was measured with the Kipp&Zonen CM11-P pyranometer (non-linearity ± 0.6%, temperature dependence ± 1%, tilt error ± 0.25%). The Kipp&Zonen CG1 pyrgeometer (non-linearity ±1%, temperature dependence ±2%) was used for measurements of downward long-wave radiation. Ambient air temperature (RTD ±0.5%) and relative humidity (AHLBORN FHA 646 E1 ±2% RH), and wind velocity (Fischer 451214 ± 0.3 m/s) were measured at a height 1.3 m in the front of
Fig. 1. A front view of the experiment with the thermo-stated test cell with and without the VGS (a), a side-view of the VGS positioned at the thermo-stated test cell (b), a schematic of the one-layered (c) and the two-layered VGS (d).
T. Suklje et al. / Energy 115 (2016) 1055e1068
1057
Fig. 2. A side view (a), a front view (b) and a top view (c) scheme of the experimental setup. Dimensions are presented in millimeters.
the façade. Air temperature in the VGS qVGS,j and in the air gap qag,j was measured with T-type thermocouples (±0.25 K), shielded from net radiation heat flux, while the surface temperature of the façade qse,j was measured with K-type thermocouples (±0.25 K). The surface temperature of the VGS was measured with the IR camera FLIR B335 (7.5e13 mm, ± 2%). Heat flux at the interior surface of the test cell was measured with the AHLBORN FQ90118 sensors (measurement accuracy ± 5%). Air temperature in the thermo-stated test cell was measured with T-type thermocouples (measurement accuracy ± 0.25 K). All sensors were properly calibrated before the installation in the experimental setup and connected to data acquisition units (AHLBORN Almemo 2290-8 and Agilent 34970A) capturing data every 30 s. Measurements were carried out in summer 2014.
2.3. Results Results from the experimental evaluation of the thermal response of the one- and the two-layered VGS are presented in terms of temperatures in three consecutive layers (Fig. 2a). In Fig. 3b, a comparison of surface temperatures of the façade with and without the one-layered VGS is shown for meteorological data in the corresponding test period (Fig. 3a). It can be ascertained that surface temperatures of the façade without the VGS are up to 29 K higher in the daytime and up to 2 K lower in the night-time. Furthermore, surface temperatures of the façade with the VGS are up to 9 K higher than temperatures in the air gap, as shown in
Fig. 3c. Temperatures in the air gap and in the VGS canopy are about the same and from 3 to 7 K higher than ambient air temperatures in the daytime. Conversely, temperatures in the air gap differ from air temperatures in the VGS canopy in the night-time. The latter are up to 1.5 K lower, whereas ambient air temperatures are lower by an additional 2 K. The difference in the surface temperatures of the façade with and without the two-layered VGS is even greater compared to the one-layered VGS (up to 34 K as depicted in Fig. 4b). Surface temperatures of the façade are up to 5.6 K higher in the daytime and up to 1 K lower in the night-time compared to temperatures in the air gap. Furthermore, temperatures in the VGS are up to 1.5 K higher than temperatures in the air gap. What is more, from Fig. 4c, it is evident that temperatures in the thicker VGS canopy are closer to ambient air temperatures in the daytime and up to 2.5 K higher in the nighttime. 2.4. Experimentally determined properties of the VGS At this point, the presented methods and results of evaluation of properties of the VGS are either used as a basis for a development of a mathematical model or used as inputs. 2.4.1. VGS-temperature based properties The measured surface temperatures of the VGS with the IR camera have been analyzed with the ThermaCAM Researcher Professional 2.9 software [25]. Average surface temperatures of the VGS have been calculated for a 0.5 m 0.5 m area (Fig. 5), 1 m above
1058
T. Suklje et al. / Energy 115 (2016) 1055e1068
Fig. 3. Meteorological data (a), surface temperatures of the façade with and without the one-layered VGS (a), temperatures in the VGS canopy, the air gap and surface temperatures of the façade behind the VGS (b).
the ground, considering the VGS's emissivity of 0.983 according to reference [26]. Average surface temperatures were then compared with air temperatures in the VGS canopy qVGS over four days (Fig. 6). Based on the comparison, it has been ascertained that temperatures are approximately the same (qs,VGS ¼ qVGS). 2.4.2. Canopy density of the VGS The density of the VGS canopy is most commonly evaluated with a leaf area index (LAI), which is defined as a one-sided surface area of leaves projected on a horizontal plane [19]. Based on results from previous VGS thermal response studies [9,11], it has been ascertained that the density of a VGS could be evaluated with a projection of leaves on a vertical plane and called a vertical leaf area index (LAIV). This was measured with a Li-Cor LAI-2200 plant canopy analyzer in 10 points at the same heights as temperatures in the air gap (Fig. 2a). Measurements were carried out before sunrise at the beginning, in the middle and at the end of the test period. The average LAIV of the one- and the two-layer VGS was 6.1 ± 0.5 and 7.2 ± 0.6, respectively. 2.4.3. Transmitivity Transmitivity of the VGSs tVGS has been determined with simultaneous measurements of solar irradiation in front of the
VGSs (reference) and behind the VGSs (Fig. 2). For the latter, the Kipp&Zonen CM11-P pyranometer was placed on the vertical guide. The guide was placed in three different positions alongside the façade. Measurements took place every hour during the daytime. With the use of statistical methods, it has been ascertained that the transmitivity of analyzed VGS is independent of meteorological parameters. Based on that finding, transmitivity has been determined as the mean of all measured points. The transmitivity of the one- and two-layered VGS tVGS was 0.17 ± 0.02 and 0.09 ± 0.02, respectively. 2.4.4. View factor A view factor of the façade through a VGS F has been determined experimentally, using a Nikon DS3200 digital single-lens reflex (DSLR) camera. Photographs were taken at the height of 1.3 m in front of the VGS. At the backside of the VGS, a white screen was placed in order to distinguish between the VGS and background. In the post-process, the background was cropped (Fig. 7a and c) using Corel PHOTO software [27]. Later, pictures were converted to binary (Fig. 7b and d) using Matlab [28]. Similarly to the transmitivity, the view factor was found to be independent of meteorological parameters, thus calculated as the ratio of black pixels to the sum of all pixels. The view factor of the one- and the two-layered VGS F was
T. Suklje et al. / Energy 115 (2016) 1055e1068
1059
Fig. 4. Meteorological data (a), surface temperatures of the façade with and without the two-layered VGS (b), temperatures in the VGS canopy, the air gap and surface temperatures of the façade behind the VGS (c).
0.042 and 0.005, respectively. 3. Mathematical model
Fig. 5. Average surface temperature of the VGS qs,VGS has been calculated for a 0.5 m 0.5 m area, 1 m above the ground, considering the VGS's emissivity of 0.983, measured on 11th August 2014.
For a mathematical model, one-dimensional heat transfer was assumed, considering experimentally determined properties of the VGSs. Based on the comparison of air temperatures in the VGS canopy and surface temperatures of the VGS (Section 2.4.1), it can be assumed that the VGS could be modeled as a homogeneous layer with no thermal resistance. The thermal capacity of the one- and the two-layered VGS is assumed to be equal to a 2 mm and a 4 mm thick water layer [29], respectively. Furthermore, in the previous studies, it was ascertained that the air temperature in the VGS canopy is not correlated with wind velocity [9,11]; thus, the VGSs can be considered to be an impermeable homogenous layer. The installation of a VGS impacts the boundary conditions at the façade. Boundary conditions are therefore considered in four planes: (i) exterior and (ii) interior surface of VGS; (iii) interior and (iv) exterior surface of the façade (Fig. 8). The façade of the thermostated test cell (Figs. 1a and 2a) consists of four layers with thermophysical properties presented in Table 1. Energy balance equations for n nodes of the considered VGS and façade surfaces are presented in Eq. (1). Within façade layers (Table 1) heat conduction is considered with the spatial
T. Suklje et al. / Energy 115 (2016) 1055e1068
35
35
30
30 θ [°C]
θ [°C]
1060
25 20
20 15
15
4/8/14 8:00
4/8/14 14:00
8/8/14 8:00
4/8/14 20:00
40
35
35
30 θ [°C]
θ [°C]
25
30 25
8/8/14 14:00
8/8/14 20:00
12/8/14 14:00
12/8/14 20:00
25 20
20
11/8/14 8:00
11/8/14 14:00
15
11/8/14 20:00 12/8/14 8:00 θVGS
θs,VGS
Fig. 6. Comparison of surface temperatures of the VGS and air temperature in the VGS canopy during four days.
asol,VGS an absorptivity of the VGS. The latter was determined according to the measured spectral reflectivity of the plant Phaseolus Vulgaris L [31]:
discretization according to reference [30].
ð1Þ
rVGS $dVGS $cp;VGS $
dqVGS ¼ q_ a;sol;VGS q_ IR;net;VGSamb dt q_ IR;net;VGSse q_ conv;VGSamb
Z
rsol;VGS ¼
q_ conv;VGSag q_ lat
rair $dair $cp;air $
ð3Þ
rref ;1 $dref ;1 $cp;ref ;1 $
ðnÞ
Z
Eðl; TÞrVGS ðlÞdl 1:1
¼ 0:232;
(3)
Eðl; TÞdl
assuming the emissivity of a black body at the temperature of 5800 K [32]. Considering the transmitivity of leaves to be negligible, the absorptivity of the VGS is, therefore, equal to 0.768. Through the VGS transmitted solar radiation is then absorbed by the façade surface with the absorptivity asol,se equal to 0.75.
dqse ¼ q_ a;sol;se q_ IR;net;seamb dt þq_ IR;net;VGSse q_ conv;se lref ;1 $ qse qref ;1
q_ a;sol;se ¼ tVGS $asol;se $Gglob;90
dqsi ¼ lref ;4 $ qref ;4 qsi q_ conv;si dt
rref ;4 $dref ;4 $cref ;4 $
(1)
3.1. Short-wave radiation heat transfer The absorbed solar radiation on the exterior surface of the VGS has been accounted for with the following expression:
q_ a;sol;VGS ¼ ð1 tVGS Þ$asol;VGS $Gglob;90
0:4
0:4
dqag ¼ q_ conv;VGSag þ q_ conv;se dt
ð2Þ
1:1
(2)
where Gglob,90 is global solar radiation on a vertical surface, tVGS is a measured transmitivity of VGS as presented in Section 2.4.3 and
(4)
3.2. Long-wave radiation heat transfer For net long-wave radiation heat transfer at the exterior surface of the VGS view, factors F from Section 2.4.4 and the emissivity of the VGS εIR,VGS equal to 0.983 [26] are considered together with the temperature of the VGS and downward long-wave radiation heat flux q_ IR;down .
4 q_ IR;down q_ IR;net;VGSamb ¼ ð1 FÞ$εIR;VGS s$TVGS
(5)
In a case in which experimental data of a downward long-wave · radiation heat flux qIR;down on a vertical plane (b ¼ 90 ) was not available, it was calculated as proposed in Eq. (6):
T. Suklje et al. / Energy 115 (2016) 1055e1068
1061
Fig. 7. Post-processed and binary pictures of one-layered ((a) and (b)) and two-layered ((c) and (d)) VGS.
Fig. 8. Boundary planes (i to iv) and heat fluxes considered in the mathematical model.
taken from Meteonorm database [34].
Table 1 Thermo-physical properties of the façade of the test cell. Layer
dref [m]
lref [W/(mK)]
rref [kg/m3]
cp,ref [J/(kgK)]
1 2 3 4
0.008 0.150 0.075 0.004
1.1 0.034 0.13 1.1
1800 30 450 1800
880 1200 1000 880
1 cos b 4 1 þ cos b 4 $Tground þ Tsky ; q_ IR;down ¼ s$ 2 2
1 Tsky ¼ Tamb $ εsky þ 0:8 1 εsky cc 4
εsky ¼ 0:711 þ 0:005$Tdew þ 0:73$
Tdew 2 100
(7)
(8)
Net long-wave heat transfer between the VGS and the façade is modeled as the two-surface enclosure, hence:
(6)
where Tground is ground temperature, which is assumed to be equal to the ambient air temperature Tamb, and Tsky is the sky temperature. The latter is calculated in respect to the ambient air temperature Tamb, the cloud coverage cc and the emissivity of the sky εsky (Eq. (7)). The emissivity of the sky εsky is determined with Eq. (8) [33]. The cloud coverage cc and dew temperatures Tdew data is
q_ IR;net;VGSse ¼ 1ε
4 T4 s TVGS se IR;VGS
εIR;VGS
1 þ þ ð1FÞ
1εIR;se εIR;se
;
(9)
where in addition to the presented variables in this section, εIR,se is the emissivity of the façade's surface, which has been determined experimentally and equals 0.98. Finally, net long-wave radiation heat transfer between the
T. Suklje et al. / Energy 115 (2016) 1055e1068
1062
façade and the ambient is given by:
4 q_ IR;down q_ IR;net;seamb ¼ F$εIR;se s$Tse
g ¼ 0:665$103 $patm (10)
Latent heat flux of the VGS is determined with the water evaporation mass-flow-rate m_ w and the latent heat rlat equals to 2490 kJ/kg [6].
q_ lat ¼ m_ w $rlat
D q_ net q_ ground þ rair cp;air ð
rlat D þ g 1 þ rras
pv;sat pv Þ ra
(12)
As the VGS is a vertical façade element, net radiation heat exchange has been evaluated in the vertical plane, hence neglecting radiation heat transfer with grounds q_ ground . The variables in the right-hand side of Eq. (13) are calculated using Eq. (2), Eq. (5), and Eq. (9).
q_ net ¼ q_ a;sol;VGS q_ IR;net;VGSamb q_ IR;VGSse
(13)
A surface resistance of leaves rs, which describes the resistance of water vapor transfer through the transpiring canopy, is evaluated with respect to the stomatal resistance of a well-illuminated leaf rl and density of the VGS. The latter is characterized by LAIV as presented in Section 2.4.2.
rS ¼
rl 0:5$LAIV
(14)
An aerodynamic resistance of leaves ra, which governs the heat and water vapor transfer from leaves into the ambient air, changes with the wind velocity v and its measuring position relative to the VGS zm, as well as the measuring position of ambient air relative humidity relative to the VGS zh:
! zm 2 3 d zom =
ra ¼
$ln
! zh 2 3 d zoh =
ln
qamb 4098 0:6108$exp q17:27$ amb þ237:3 ðqamb þ 237:3Þ
2
(17)
Variables pv and pv,sat are partial water vapor pressure and saturation water vapor pressure, respectively. They are determined with respect to ambient air temperature qamb and relative humidity RH in accordance with reference [19].
(11)
A process of water evaporation in plants is named evapotranspiration ET. ET in the VGSs was modeled with the Penman-Monteith equation [19]:
ET ¼
A slope of the saturation vapor pressure curve D is evaluated in respect to ambient air temperature qamb [19]:
D¼
3.3. Latent heat transfer
(16)
;
k2 v
(15)
while the remaining variables are calculated in accordance with reference [19], with respect to the thickness of the VGS canopy d. The values are presented in Table 2. Psychrometric constant g is calculated in respect to the atmospheric pressure patm, given in Table 2: Table 2 Constants for evaluation of ET.
3.4. Convective heat transfer In the study, convective heat transfer is evaluated with Newton's cooling law, considering specific airflow conditions at each analyzed surface with a specific convective heat transfer coefficient using Eq. (18)e(21). The VGS canopy density impacts convective heat transfer through the extension of the heat exchange surface. Therefore, a convective heat transfer coefficient is calculated based on LAIV. In addition, aerodynamic resistance ra (Eq. (15)) and thermo-physical properties of the air are included as proposed by Ayata et al. [36]:
cp;air $rair hconv;VGSamb ¼ LAIV$ ra
(18)
The convective heat transfer coefficient at the interior surface of the VGS is determined based on the Nusselt number Nu, characteristic length D, which is assumed to be equal to twice the width of the air gap between the VGS and the building envelope (Fig. 2a), and the thermal conductivity of air lair:
hconv;VGSag ¼ Nu$
lair D
(19)
The Nusselt number Nu for mixed convection at the interior surface of the VGS is determined with an expression proposed by Stanghellini [37] and used in the VGS thermal response study by Stec et al. [6]. The VGS acts as an impermeable layer [9]; consequently, no-wind conditions are considered in the air gap. For nonwind conditions in urban settlements, it is assumed that the wind velocity is lesser than 0.2 m/s [38]. Thus, the Reynolds number Re is calculated for the air velocity equal to 0.2 m/s. In addition to the forced convection, the buoyancy-driven convection is taken into account with the inclusion of the Grashof number Gr in the following equation:
0:25 Nu ¼ 0:37$ Gr þ 6:416$Re2
(20)
The convective heat transfer at the façade surface is assumed to be buoyancy-driven and is evaluated with expression [39]:
1=3 hconv;se ¼ 1:31$ qse qag
(21)
At the interior surface of the thermo-stated test cell a constant convective heat transfer coefficient hconv,si is assumed. According to reference [40], value 7.7 W/(m2K) accounts for convective and radiation heat transfer.
Parameter
Value
Reference/comment
εIR,VGS [] rl [s/m] k [] zm, zh [m] d [m] p [kPa]
0.983 100 0.41 2.0/1.7 0.3/0.6 97.7
[26] [19,35] [19] measured for the one-layered/the two-layered VGS
3.5. Experimental results and the model validation
[19]
The presented mathematical model was implemented in Simulink [41], where the measured meteorological parameters
T. Suklje et al. / Energy 115 (2016) 1055e1068
1063
Fig. 9. Calculated and measured temperatures of the one-layered VGS (a), the air gap (b) and the exterior surface of the façade (c); heat flux at the interior surface of the test cell (d).
1064
T. Suklje et al. / Energy 115 (2016) 1055e1068
Fig. 10. Calculated and measured air temperatures of two-layered VGS (a), the air gap (b) and the exterior surface of the façade (c); heat fluxes at the interior surface of the test cell (d).
T. Suklje et al. / Energy 115 (2016) 1055e1068
(Figs. 3a and 4a) together with the VGS properties were used as the inputs. Numerical results were then validated with experimental results of three consecutive temperature nodes (air temperature of the VGS qVGS, the temperature of the air gap qag and the surface temperature of the façade qse e Fig. 2a) and heat flux at the interior surface of the test cell q_ i . Calculated and measured air temperatures of the one-layered (Fig. 9a) and the two-layered (Fig. 10a) VGS are in good agreement; however, maximal calculated daily temperatures are up to 3.5 K lower than the measured ones. Better agreement of the calculated and measured results can be found for temperatures of the air gap, where temperature difference does not exceed 2 K (Figs. 9b and 10b) during the daytime and 2.5 K during the nighttime. Furthermore, the calculated surface temperatures of the façade with the VGS differ from measured values by no more than 3 K, for the one-layered (Fig. 9c) as well as for the two-layered VGS (Fig. 10c). Calculated temperatures in all considered temperature nodes of the two-layered VGS on 11th July 2014 are considerably lower than measured ones due to the stormy weather. Finally, calculated and measured heat fluxes on the interior surface of the test cell with the one-layered (Fig. 9d) and the twolayered (Fig. 10d) VGS were compared. It has been ascertained that the calculated results do not exceed measured results by more than the measurement uncertainty accounted for dynamic conditions (±11.7%). The measurement uncertainties associated with instrumentation and experimental procedures have an effect on numerical results. To investigate these effects, a measurement uncertainty analysis was carried out. Due to the complexity of the mathematical model, a Monte Carlo method was used for the propagation of distributions [42,43]. For the analysis, a uniform probability density function (PDF) was associated with the input sources, which covered the measured meteorological data as well as LAIV, tVGS, and interior air temperature qi. Results of the analysis are shown in Figs. 9 and 10 as the measurement uncertainty intervals.
4. Parametrical analysis 4.1. Impact of VGS on thermal response of a façade The validated thermal response model of the façade with the two-layered VGS has been used to quantify a cooling potential of the VGS with respect to the thermal resistance of a façade. The cooling potential was evaluated with a comparison of heat fluxes at
1065
an interior surface of a façade with and without the VGS for meteorological conditions presented in Fig. 3a. In the analysis, a façade with the thermal resistance equal to 5, 2.1 and 0.6 m2 K/W was considered. The thermal resistance of a façade is chosen based on the typical U-vales of a building envelope in Eastern and Southern Europe. That is for a well-insulated low-energy building (0.19 W/m2K), a moderately insulated building (0.45 W/m2K) and a non-insulated building (1.3 W/m2K). The interior temperature was set at a constant value of 25 C. Fig. 11 shows the heat fluxes at the interior surface of a façade without and with the VGS for the selected period. It can be concluded that peak heat flux at the interior surface of a façade with a thermal resistance of 0.6 m2 K/W can be reduced by a maximum of 41.2 W/m2. Considerably lower reductions of peak heat flux can be identified (up to 12.2 W/m2) in the case of a façade with the thermal resistance of 2.1 m2 K/W. A heat flux is further lessened (as much as 5 W/m2) in the case of a façade with the highest thermal resistance among compared façades. Relatively to a reference façade, peak heat fluxes of façades with the VGS have all been decreased by 77%. The analysis of the average heat flux at the interior surface of a façade with the VGS shows a reduction of 11.2, 3.9 and 1.5 W/m2, respectively. It has been ascertained that the VGS, in addition to the neutralization of heat gains from the exterior environment, acts as an effective cooling measure regardless of the thermal resistance of a façade. The effect is more pronounced for less thermally insulated façades. From Fig. 11, it is also evident that a negative heat flux at the interior surface of a façade with the VGS is greater than one at a reference façade regardless of the thermal resistance of a façade. Based on the analysis of temperatures in the supporting layer of a façade, it has been concluded that the observed phenomena is a consequence of the heat accumulation.
4.2. Cooling effect of the VGS on the urban environment In order to provide a general model of the latent heat flux of the VGS, an impact of thermo-physical and optical properties of a façade on the latent heat flux was parametrically investigated. For this purpose, the validated thermal response model of the façade with the VGS was used with measured meteorological data (Figs. 3a and 4a) and supplemented with the test reference year (TRY) for Ljubljana, Slovenia from July to August [34]. Due to a difference in a position of the wind velocity measurements in TRY and the experiment, a correction of wind velocity is taken into account with
60 50
q̇ si [W/m2]
40 30 20 10 0 -10 -20 R=5_ref
R=2.1_ref
R=0.6_ref
R=5_VGS
R=2.1_VGS
R=0.6_VGS
Fig. 11. Heat fluxes at the interior surface of a façade with and without the two-layered VGS, altering the thermal resistance between 5, 2.1 and 0.6 m2 K/W.
T. Suklje et al. / Energy 115 (2016) 1055e1068
1066
Fig. 12. Latent heat flux of the two-layered VGS installed at a façade with the thermal resistance of 5, 2.1 and 0.6 m2 K/W (a) and the solar absorptivity 0.3, 0.75 and 0.9 (b).
Eq. (22), as proposed by Nottrott et al. [44].
vcorr ¼ 0:17$v þ 0:099
(22) C.
The interior temperature was considered constant at 25 The thermal resistance of a façade was considered to be equal as in the previous section, whereas values of the façade's solar absorptivity equal to 0.3, 0.75 and 0.9 were chosen. Results from the parametrical investigation are presented in Fig. 12. As can be seen from Fig. 12a, the thermal resistance of a façade impacts the latent heat flux of the VGS insignificantly. The same can be concluded for the impact of the façade's solar absorptivity (Fig. 12b). Conclusively, the VGS can be modeled as an independent urban cooling element in the UHI mitigation research. 4.2.1. Parametrical model of a latent heat flux of the VGS A parametrical model was developed in relation to the significant variables. A range of the validity of the model is presented in Eq. (23). In addition to the ambient air temperature and relative humidity and wind velocity, net radiation heat flux has been taken into account. Based on the findings from the previous section, net Table 3 Coefficients of parametrical model of latent heat flux. LAIV
b1
b2
b3
b4
b5
b6
b7
b8
6.1 7.2
15610.9 16349.7
106.49 111.18
0.18 0.19
5.03 5.71
0.03 0.03
2.79 4.66
0.648 0.581
0.0000 0.0002
radiation heat flux is given by Eq. (24), considering only absorbed solar radiation and net long-wave radiation heat flux with the ambient surroundings. 2 þ b4 $RH þ b5 $RH 2 þ b6 $v q_ lat ¼ b1 þ b2 $Tamb þ b3 $Tamb 2 þb7 $q_ net þ b8 $q_ net
8 282K T 9 amb 305K > > > > < 36% RH 86% = 0:2m=s v 3:5m=s . > >
> > : ; 30W m2 q_ net 625W m2 4 q_ IR;down q_ net ¼ asol;VGS $G90;glob εIR;VGS s$Ts;VGS
(23)
(24)
Coefficients in Eq. (23) were determined separately for the onelayered and the two-layered VGS and are presented in Table 3. Based on the comparison of the latent heat flux of the one- and the two-layered VGS calculated with the numerical and the parametrical model (presented in Fig. 13), it has been ascertained that proposed parametrical models are suitable. As expected, results show that the latent heat flux of the two-layered VGS is more intensive compared to the one-layered VGS. However, the difference does not exceed 10% at the highest of the latent heat flux. Consequently, the one-layered VGS is the more favorable choice to two-layered VGS as it takes less space with the comparable cooling effect. Models in such a form can be used by the urban physics researchers for the evaluation of the cooling potential of an
Fig. 13. Comparison of the latent heat flux of the one-layered (a) and the two-layered (b) VGS calculated with the numerical and the parametrical model.
T. Suklje et al. / Energy 115 (2016) 1055e1068
1067
Fig. 14. Plot of a latent heat flux of the two-layered VGS in relation to the ambient air temperature, relative humidity and net radiation heat flux for wind velocity 0.5 m/s (a) and 2 m/s (b).
arbitrary VGS in arbitrary urban areas. The parametrical model of the two-layered VGS was also used to demonstrate the relation between latent heat flux and the independent variables. In Fig. 14, a consecutive surface presents the relation of the latent heat flux to ambient air temperature and relative humidity at discrete values of the net radiation heat flux and wind velocity. At the constant value of the other independent variables, the latent heat flux varies less than 10 W/m2 due to the change in wind velocity. Furthermore, a 100 W/m2 change in net radiation heat flux increases latent heat flux by 58 W/m2. Moreover, a change within the given range (from 284 K to 304 K with 5 K step) of ambient air temperature reflects an increase of latent heat flux from 6 W/m2 to 34 W/m2. Finally, latent heat flux reduces from 92 W/m2 to 6 W/m2 increasing ambient air relative humidity from 35% to 85% with a 10% step.
From the parametrical analysis of the impact of the thermal resistance and the solar absorptivity of a façade on the latent heat flux of the VGS, it can be concluded that neither is influential. Based on these findings the parametrical model of the latent heat flux of the VGS was developed and can be used either as a heat flux or a volumetric cooling power boundary condition in CFD modeling. Furthermore, it has been ascertained that the one-layered VGS is the more favorable choice in comparison to the two-layered VGS as it takes less space with the comparable cooling effect. Conclusively, the validated thermal response model of the VGS can be used to evaluate the cooling potential of an arbitrary VGS in a diverse urban areas. In the future, an expanded study of the overall performance of a building with the VGS should be performed, and applicability of the parametrical models demonstrated in CFD tools.
5. Conclusions
Nomenclature
This paper presents an experimental and numerical investigation of the thermal response of the VGS. Considerable attention has been paid to the experimental quantification of the properties of the VGS including transmitivity, view factors, and foliage density, which is rarely mentioned in the literature, to the authors' knowledge. Determined properties of the VGSs were used for the design of the mathematical model of the thermal response of the VGSs as well as inputs for the model. In contrast to the existing thermal response models of VGSs available in the literature, the current one provides repeatability and traceability of the results, which is due to the detailed description of the mathematical model and inputs. In the experimental survey, the evaluation of temperature conditions and properties of the VGS with LAIV densities of 6.1 and 7.2 are studied. Experimental results show that the VGS can have an impact up to 34 K lower surface temperatures of the façade, while maintaining air temperatures in the VGS canopy close to ambient air temperatures. The validated mathematical model has been used for the parametrical study of the cooling potential of the VGS. In the analysis that included a façade with different thermal resistances, it has been concluded that peak heat fluxes at the interior surface of compared façades with the VGS can be reduced by 77%. Furthermore, it has been ascertained that the VGS acts as an effective cooling measure for a building. In addition, the cooling effect appears to be inversely proportionate with the thermal resistance of a façade.
cc d D ET F Gglob,90 h k p q_ R RH ra rl rlat rs T
cloud coverage [] thickness [m] characteristic length [m] evapotranspiration [kg/s] view factor [] global solar radiation on vertical [W/m2] convective heat transfer coefficient [W/(m2K)] von Karman constant [] pressure [kPa] heat flux [W/m2] thermal resistance [m2K/W] ambient air relative humidity [%] aerodynamic resistance [s/m] stomatal resistance [s/m] latent heat [kJ/kg] surface resistance of leaves [s/m] absolute temperature [K]
Greek symbols a absorptivity [] ε emissivity [] q temperature [ C] l thermal conductivity [W/(mK)] r density [kg/m3] r reflectivity []
1068
T. Suklje et al. / Energy 115 (2016) 1055e1068
Subscripts a absorbed ag air gap amb ambient air atm atmospheric conv convective corr corrected dew dew point temperature e exterior exp experimental i interior IR long-wave radiation lat latent pm parametrical model ref reference façade net net num numerical s surface sat saturated se exterior surface of the test cell si interior surface of the test cell sol solar radiation sky sky v water vapor VGS vertical greenery system w water References €hler M. Green facadesda view back and some visions. Urban Ecosyst [1] Ko 2008;11(4):423e36. http://dx.doi.org/10.1007/s11252-008-0063-x. [2] Vidrih B, Medved S. Multiparametric model of urban park cooling island. Urban For Urban Green 2013;12(2):220e9. http://dx.doi.org/10.1016/ j.ufug.2013.01.002. [3] Price A, Jones E, Jefferson F. Vertical greenery systems as a strategy in urban heat island mitigation. Water Air Soil Pollut 2015;226(8):1e11. http:// dx.doi.org/10.1007/s11270-015-2464-9. [4] Bevilacqua P, et al. Plant cover and floristic composition effect on thermal behaviour of extensive green roofs. Build Environ 2015;92:305e16. http:// dx.doi.org/10.1016/j.buildenv.2015.04.026. [5] Tsang SW, Jim CY. Theoretical evaluation of thermal and energy performance of tropical green roofs. Energy 2011;36(5):3590e8. http://dx.doi.org/10.1016/ j.energy.2011.03.072. [6] Stec WJ, Paassen AHC, Maziarz A. Modelling the double skin façade with plants. Energy Build 2005;37(5):419e27. http://dx.doi.org/10.1016/ j.enbuild.2004.08.008. [7] Perini K, Ottele M, Fraaij AL, Haas EM, Raiteri R. Vertical greening systems and the effect on air flow and temperature on the building envelope. Build Environ 2011;46:2287e94. M, Perini K, Fraaij ALA, Haas EM, Raiteri R. Comparative life cycle [8] Ottele analysis for green façades and living wall systems. Energy Build 2011;43(12): 3419e29. http://dx.doi.org/10.1016/j.enbuild.2011.09.010. [9] Suklje T, Arkar C, Medved S. The local ventilation system coupled with the indirect green façade: a preliminary study. Int J Des Nat Ecodyn 2014;9(4): 314e20. http://dx.doi.org/10.2495/DNE-V9-N4-314-320. [10] Jim CY. Cold-season solar input and ambivalent thermal behavior brought by climber greenwalls. Energy 2015;90(1):926e38. http://dx.doi.org/10.1016/ j.energy.2015.07.127. [11] Suklje T, Medved S, Arkar C. An experimental study on a microclimatic layer of a bionic façade inspired by vertical greenery. J Bionic Eng 2013;10(2):177e85. http://dx.doi.org/10.1016/S1672-6529(13)60213-9. stola D, Hensen JLM. Framework for assessing [12] Kasinalis C, Loonen RCGM, Co the performance potential of seasonally adaptable facades using multiobjective optimization. Energy Build 2014;79:106e13. http://dx.doi.org/ 10.1016/j.enbuild.2014.04.045. stola D, Hensen JLM. Climate adaptive building [13] Loonen RCGM, Tr cka M, Co shells: state-of-the-art and future challenges. Renew Sustain Energy Rev 2013;25:483e93. http://dx.doi.org/10.1016/j.rser.2013.04.016. [14] Wong NH, Tan AYK, Tan YP, Wong NC. Energy simulation of vertical greenery systems. Energy Build 2009;41:1401e8. rez G, Rinco n L, Vila A, Gonza lez JM, Cabeza LF. Green vertical systems for [15] Pe
[16] [17]
[18]
[19] [20]
[21]
[22]
[23]
[24]
[25] [26]
[27] [28] [29] [30] [31]
[32] [33] [34] [35]
[36]
[37]
[38]
[39]
[40]
[41] [42]
[43]
[44]
buildings as passive systems for energy savings. Appl Energy 2011;88: 4854e9. Wong NH, et al. Thermal evaluation of vertical greenery systems for building walls. Build Environ 2010;45:663e72. Fang W, Xiaosong Z, Junjie T, Xiuwei L. The thermal performance of double skin façade with Tillandsia usneoides plant curtain. Energy Build 2011;43(9): 2127e33. http://dx.doi.org/10.1016/j.enbuild.2011.04.021. Vidrih B. Impact of global climate changes and urbanization on local climatic condition in cities. PhD thesis. Ljubljana: University of Ljubljana, Faculty of Mechanical Engineering; 2011. Allen RG, Pereira LS, Raes D, Smith M. FAO irrigation and drainage Paper No. 56: crop evapotranspiration. 2001. Flores Larsen S, Filippín C, Lesino G. Modeling double skin green façades with traditional thermal simulation software. Sol Energy 2015;121:56e67. http:// dx.doi.org/10.1016/j.solener.2015.08.033. Kontoleon KJ, Eumorfopoulou EA. The effect of the orientation and proportion of a plant-covered wall layer on the thermal performance of a building zone. Build Environ 2010;45(5):1287e303. http://dx.doi.org/10.1016/ j.buildenv.2009.11.013. Eumorfopoulou EA, Kontoleon KJ. Experimental approach to the contribution of plant-covered walls to the thermal behaviour of building envelopes. Build Environ 2009;44(5):1024e38. http://dx.doi.org/10.1016/ j.buildenv.2008.07.004. Susorova I, Angulo M, Bahrami P, Brent S. A model of vegetated exterior facades for evaluation of wall thermal performance. Build Environ 2013;67: 1e13. http://dx.doi.org/10.1016/j.buildenv.2013.04.027. Gromke C, Blocken B, Janssen W, Merema B, van Hooff T, Timmermans H. CFD analysis of transpirational cooling by vegetation: case study for specific meteorological conditions during a heat wave in Arnhem, Netherlands. Build Environ 2015;83(0):11e26. http://dx.doi.org/10.1016/j.buildenv.2014.04.022. FLIR. ThermaCAM researcher professional 2.9. 2005. Lopez A, Valera DL, Pena A. Determining the emissivity of the leaves of nine horticultural crops by means of infrared thermography. Sci Hortic Amst 2012;137. http://dx.doi.org/10.1016/j.scienta.2012.01.022. 10e10. CorelCorporation. CorelPHOTO-PAINT. 2014. MathWorks. MATLAB R2014a. 2013. Holm D. Thermal improvement by means of leaf cover on external walls d a simulation model. Energy Build 1989;14:19e30. Davies MG. Building heat transfer. Chichester, England: John Wiley & Son; 2004. Gutierrez MR, Escalante JAE, Rodriguez MTG. Canopy reflectance, stomatal conductance, and yield of Phaseolus vulgaris L. and Phaseolus coccinues L. under saline field conditions. Int J Agric Biol 2005;7(3). Incropera FP, DeWitt DP, Bergman, Lavine. Fundamentals of heat and mass transfer. sixth ed. , New York: Wiley; 2006. TRNSYS16. Mathematical reference, vol. 5; 2006. Meteonorm. Global meteorological database for applied climatology. 2004. Malys L, Musy M, Inard C. A hydrothermal model to assess the impact of green walls on urban microclimate and building energy consumption. Build Environ 2014;73(0):187e97. http://dx.doi.org/10.1016/j.buildenv.2013.12.012. Ayata T, Tabares-Velasco PC, Srebric J. An investigation of sensible heat fluxes at a green roof in a laboratory setup. Build Environ 2011;46(9):1851e61. http://dx.doi.org/10.1016/j.buildenv.2011.03.006. Stanghellini C. Mixed convection above greenhouse crop canopies. Agric For Meteorol 1993;66(1e2):111e7. http://dx.doi.org/10.1016/0168-1923(93) 90085-V. Niachou K, Hassid S, Santamouris M, Livada I. Experimental performance investigation of natural, mechanical and hybrid ventilation in urban environment. Build Environ 2008;43(8):1373e82. http://dx.doi.org/10.1016/ j.buildenv.2007.01.046. Blanco JM, Arriaga P, Rojí E, Cuadrado J. Investigating the thermal behavior of double-skin perforated sheet façades: part A: model characterization and validation procedure. Build Environ 2014;82(0):50e62. http://dx.doi.org/ 10.1016/j.buildenv.2014.08.007. EN ISO 13786. In: Thermal peroformance of building components - dynamic thermal characteristics - calculation methods. Brussels: European commitee for standardization; 1999. Matlab&Simulink. Mathwroks. 2014. Asuero AG, Gonza lez AG. Estimation of the uncertainty of inHerrador MA, direct measurements from the propagation of distributions by using the Monte-Carlo method: an overview. Chemom Intell Lab Syst 2005;79(1e2): 115e22. http://dx.doi.org/10.1016/j.chemolab.2005.04.010. Couto PRG, Damasceno JC, Oliveira SP. Monte carlo simulations applied to uncertainty in measurement. In: Chan VWK, editor. Theory and applications of monte carlo simulations. Intech; 2013. Nottrott A, Onomura S, Inagaki A, Kanda M, Kleissl J. Convective heat transfer on leeward building walls in an urban environment: measurements in an outdoor scale model. Int J Heat Mass Transf 2011;54(15e16):3128e38. http:// dx.doi.org/10.1016/j.ijheatmasstransfer.2011.04.020.