Applied Energy 199 (2017) 281–293

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

A comparative study on thermal performance evaluation of a new

double skin façade system integrated with photovoltaic blinds
Yongqiang Luo, Ling Zhang ⇑, Xiliang Wang, Lei Xie, Zhongbing Liu, Jing Wu, Yelin Zhang, Xihua He
College of Civil Engineering, Hunan University, Changsha 410082, PR China

h i g h l i g h t s

A new double skin façade using photovoltaic blinds was proposed and studied.

Experiment and simulation method were used for comparative study.

Influence of different system ventilation modes and blind parameters were analyzed.

Thermal performance of proposed façade and standard double skin façade was compared.

a r t i c l e i n f o a b s t r a c t

Article history: The glazing façade is embraced by architects, but this configuration may result in huge energy consump-
Received 23 February 2017 tion. This research proposed a new double skin façade using photovoltaic (PV) blinds as a shading device
Received in revised form 6 April 2017 (named PVB-DSF), which could realize multi-function of power generation, solar penetration reduction
Accepted 3 May 2017
and flexible daylighting control. The purpose of this comparative study is to demonstrate the superb ther-
Available online 10 May 2017
mal performance of PVB-DSF. Experimental rig was built at hot-summer and cold-winter zone of China.
The first stage comparative study was conducted to evaluate system thermal performance under the
Keywords:
effects of ventilation modes, PV-blind angle and PV-blind spacing. The second stage study was conducted
Double skin façade
Photovoltaic blinds
to compare thermal performance between PVB-DSF and standard DSF. A validated numerical model was
Thermal performance used to describe standard DSF. The results suggested the operation of natural ventilation mode and indi-
Solar heat gain cated the evident influence of PV-blind spacing on system performance. The comparison study further
Comparative study demonstrated that PVB-DSF can save about 12.16% and 25.57% of energy in summer compared with con-
ventional DSF with and without shading blinds. The insulation performance of PVB-DSF is shown by its
daily average heat transfer coefficient which was as low as 2.247.
Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction Currently, double glass window or double skin façade (DSF) is

widely applied to commercial and public buildings around world
Building envelope plays a vital role in protecting the indoor due to its better thermal performance. DSF refers to a building
environment and controlling indoor and outdoor space interac- façade covering one or more levels with multiple glazed skins, sep-
tions [1], but also is a major reason for huge building energy con- arated by an air gap, with the common attribute of controllable
sumption in terms of space cooling and heating [2]. External wall shading system and airflow within the cavity between the skins
and glass façades (or windows) are two kinds of building envelope. of the façade [8,9]. Some DSF structure has combined with over-
Extensive researches indicated that conventional window or glaz- hangs which proved to be energy efficient [10].
ing façade is responsible for about 40% of the total building energy A number of experiments, numerical simulations or case stud-
consumption [3–7]. It is the modern architectural aesthetics rather ies were implemented to understand the energy saving potential
than the superior system performance that facilitates the popular- of standard DSF. Different research conclusions about energy sav-
ity of conventional glazing facade. Actually, the ever increasing ing potential of DSF may be different in various climate zones
building energy consumption is partially the results of massive applications, and those reported results of energy saving potential
installation of glazing facade without energy saving considerations. are ranging from negative to 50%. A comparative study by Chan
et al. [11] showed that DSF using reflective glazing as the outer
⇑ Corresponding author. pane can cut up to 26% of building cooling energy annually. How-
E-mail addresses: [email protected] (Y. Luo), [email protected] (L. Zhang). ever, the payback time is as long as 81 years. In another study on

http://dx.doi.org/10.1016/j.apenergy.2017.05.026
0306-2619/Ó 2017 Elsevier Ltd. All rights reserved.
282 Y. Luo et al. / Applied Energy 199 (2017) 281–293

the thermal performance of DSF based on the ESP-r simulation, the conducted using TRNSYS software to analyze and optimize a-Si
results indicated that DSF can save 20% of heating energy demand and c-Si PV window. Results showed that maximum saving for
compared with single-skin façade [12]. The field experimental cooling energy occurs during September, while maximum incre-
study by Xu and Ojima showed that 10–15% and 20–30% of energy ment in heating energy was found in December. Two configura-
could be saved by DSF in summer and winter respectively [13]. tions of semi-transparent PVT system were presented by Shyam
From extensive literatures reading and learning, it is implied that and Tiwari [29], analysis indicated that for 30 years life time and
DSF may not be the best option for building energy conservation 4% interest rate, the unit cost of electricity was $0.016 for overall
in every region [14] considering climatic features and other factors. thermal saving and $0.109 for exergy saving respectively.
Apart from the advantages of DSF, two big challenges con- Although PV glazing façade has exhibited its energy saving
fronted by conventional DSF were summarized by Ghaffarianho- potential in cutting cooling and heating load of buildings [30], day-
seini in a review study [1]. One of them is the initial cost lighting control in daily operation seems unattainable. On one
concerning system design, construction as well as high cleaning, hand, whether it is a-Si or c-Si PV-DSF, daylight transmission
operating, inspection and maintenance costs in comparison to through PV cells is unchangeable. This is the reason that some
the conventional façades [1]. Another challenge is the high risk of studies are trying to improve indoor visual comfort level by opti-
unacceptable performance including the overheating problem in mizing the daylight and solar transmission through PV module
summer [15,16] and other economic factors [12]. For the past dec- [30,31]. On the other hand, we noticed that conventional DSF usu-
ade, advanced materials and new structures were applied to fix the ally can shield part of direct solar radiation by placing louvers in
mentioned problems of glazing façade and most of the structures the air cavity. However, the conventional shading device can be
are almost evolved and inherited from the prototype of conven- easily heated up to 60 °C in hot summer [32] and then become a
tional double glass façade. There are about 8 types of materials heat radiation source. Phase change material (PCM) were adopted
used in the newly developed glazing façade, which are insulating to buffer the heat transfer process [33]. Shen and Li [34,35]
and phase change material (PCM) material, reflecting material, designed a pipe-embedded DSF system using natural cooling
electrochromic material, thermochromic material, photovoltaic sources like a cooling tower or underground soil. The pipe-
material, water film based material and other materials [17]. When embedded DSF is proven to be at least 20% more energy efficient
those materials are used in glazing façade, they can largely reduce than conventional DSF according to their CFD simulations, but this
solar heat gain and heat loss in summer and winter respectively. system performance is obtained based on the supply of cool water.
Among those materials, c can not only improve system thermal This study presents a new DSF structure with built-in PV binds
performance, but also transform parts of solar energy into direct working as a shading device, power generator as well as the ther-
electricity. Different PV cells may have their specific property mal performance enhancer. The purpose of this study is going to
and some limitations currently, but this technology is deemed as demonstrate superb thermal performance of PVB-DSF by checking
one promising solution for the energy problem, especially in the the influence of operation modes, PV-blinds parameters and com-
building sector. paring with standard DSF system under various energy indexes.
In order to enhance glazing system performance with little Because glazing façade suffered a more serious problem of heat
damage to the transparency of PV glazing facade, the PV materials gain in summer conditions rather than heat loss in winter condi-
and glazing structure should be fairly designed and optimized. tions, this research is focused on comparative study of PVB-DSF
There are 4 kinds of PV materials used to model semi- system in cooling season. Some researchers previously have real-
transparent PV glazing system. They are Crystallic silicon (c-Si), ized the possibility of combing PV cells and venetian blinds to
Amorphous silicon (a-Si), Dye sensitized solar cells (DSSC), and shade solar irradiance while converting parts of radiation into elec-
Organic photovoltaics (O-PVs). The later three kinds of materials tricity [36,37]. But there are two major limitations within their
are semi-transparent in nature. Therefore, a-Si, DSSC and O-PVs works: (1) the PV cells are directly attached onto the surface of
can be directly used as the external skin of PV-DSF structure to blinds [38] which is not good for heat dissipation of PV modules;
realize multi-function of daylighting, heat gain reduction and (2) although system optical [38] and electricity generation
power generation. However, c-Si as an opaque PV cell, if it is to [38,39] analysis were provided, system thermal performance and
be used to manufacture semi-transparent PV glazing, has to be uni- full sized experimental investigations about DSF integrated PV-
formly installed on the surface of external glass while leaving some blinds are still not reported. Based on those two considerations,
room for sun light penetration through the uncovered regions this research proposed a new DSF integrated with PV-blinds
[18,19]. (named as PVB-DSF) and explored the system thermal perfor-
Semi-transparent PV-DSF structures by c-Si and a-Si cells mance by both experiment and simulation. The basic contribution
received more attentions from researchers gradually [20–22] and of present study is to provide better solution for lowering heat gain
they even have been applied in some demonstration projects. of glazing façade while using PV cells for power generation. The
Researchers from Hong Kong conducted experimental test and system thermal performance is going to be demonstrated and
energy simulation on a new type of see-through (a-Si) semi- discussed in detail.
transparent PV module as double glazed window [23] which is
similar to the system studied by Peng et al. [24,25]. It was con-
cluded from the study that this new PV window can curtail heat 2. Description of PVB-DSF system
gain by 47.8% and 38.9% compared with single clear glazing win-
dows and double-pane windows under Hong Kong climate condi- In essence, the proposed new glazing façade adopted venetian
tion. Moreover, this structure can achieve equivalent thermal blinds made from photovoltaic as shading device sandwiched by
performance as low-E window. Besides, Myong and Jeon [26] external and internal glass pane. Because this structure integrates
demonstrated the another effective and efficient bifacial TBC a- PV-blinds with DSF, it is called as PVB-DSF in short within this
Si:H PV cells. As for the application of semi-transparent PV in cold research. Fig. 1 depicts a sketch of PVB-DSF installed on a south-
season, Taffesse et al. [27] derived a periodic modeling method for facing wall. In order to deliver a clear presentation of this system,
energy simulation of semitransparent photovoltaic thermal both lateral section view and top view from section A-A are pro-
Trombe wall system. The proposed computation method was used vided respectively by Fig. 1(a) and (b). Dimension information
to optimize the parameters of this PV envelope. In Skandalos’s about the PVB-DSF is listed in Table 1. The so called PV-blinds
work [28], the optical, thermal and electrical simulations were are made of a-Si PV cell in the shape of the narrow and long slats,
 Y. Luo et al. / Applied Energy 199 (2017) 281–293 283

Fig. 1. Schematic diagram of PVB-DSF system.

Table 1 can be slightly different for varying market price and Exchange
The system structural parameters of PVB-DSF. Rate of Dollar. The material of PV-blind is a-Si (amorphous Silicon)
Parameters Values which is the same as the materials used in Shyam’s research on
Width of external and internal glass 1.1 m PVT system [29]. Therefore, the energy payback time on exergy
Height of external and internal glass 0.94 m basis for PV-blind system is about 7.7 years [29]. Considering the
Thickness of external and internal glass 0.006 m system difference, the conclusion about energy payback from
Depth of air flow duct 0.25 m Shyam’s research cannot directly used for analysis of PVB-DSF
Width of air inlet and outlet 1.1 m
but it provided a useful guidance to investigation of energy pay-
Height of air inlet and outlet 0.15 m
Number of PV blinds 18 back time of PVB-DSF, which is going to be conducted for further
Width of PV module 0.025 m study on PVB-DSF.
Thickness of PV module 0.003 m Horizontal comparison can help understanding PVB-DSF sys-
Spacing between adjacent blinds 0.045 m
tem. Therefore, Table 2 compares PVB-DSF with conventional
Slat angle 45°
DSF, semi-transparent a-Si PV-DSF and c-Si glazing façade from
aspects of external glass, cavity and internal glass. Those 4 struc-
tures can shield glazing from direct solar irradiance at different
which is shown by Fig. 2. Those blinds are connected in parallel by
approaches to reduce indoor heat gain in hot summer. The conven-
the conductor through the positive and negative electrode of cells.
tional DSF utilize aluminum venetian blinds, while semi-
Besides, those PV cells have to be linked to an electrical load to
transparent a-Si and c-Si PV-DSF and PVB-DSF use PV cells to block
complete a circuit. The air inlet and outlet in Fig. 1(a) are designed
most of solar radiation. Considering multi-function of PV blinds,
to realize ventilation in the cavity to remove excessive heat from
this request that blinds spacing should be slightly larger than the
PV blinds. The cost for PV-blind is about $5.49 per module (for year
width of slats to avoid partial shading among PV-blinds. The prob-
of 2017) and $98.89 per manufactured PVB-DSF (model specifica-
lem of partial shading among PV cells should be prevented to lower
tion of Fig. 2). The cost for installation is not included and this cost
the risk of overheating, power mismatching and loss [40,41].
Therefore, the blinds spaced 4.5 cm from each other in system
design.
The philosophy of semi-transparent PV-DSF is to convert a por-
tion of possible harmful solar radiation in summer into electric
power, leaving rest of thermal energy either dissipated to cavity

Table 2
Structure comparison among different glazing systems.

External pane Cavity Internal pane

Conventional Clear glass or Aluminum made Clear glass
DSF vacuum glass venetian blinds
Semi-transparent Amorphous None Clear glass
a-Si PV-DSF silicon cell or inward
window
Semi-transparent Partially covered None Clear glass
c-Si PV-DSF crystalline silicon
cell
PV blinds-DSF Clear glass or Photovoltaic Clear glass
vacuum glass blinds
Fig. 2. Photograph of photovoltaic blinds.
284 Y. Luo et al. / Applied Energy 199 (2017) 281–293

by ventilation or penetration into the indoor environment. It is a or sealed to realize different ventilation modes. Fig. 3 showcases
big step evolved from conventional DSF and this idea is co- two ventilation modes used in the presented study, in which 4
shared by a-Si and c-Si PV-DSF and PVB-DSF. But PVB-DSF can DC fans were installed for forced ventilation. The measured datum
stand out by providing flexible daylighting control through PV were collected and recorded by a Data Logger 34972A. The time
blinds angle. And this study is going to verify its system thermal step for data screening and recording is 5 min. Detailed informa-
performance in cooling season as well. tion about experimental instruments are provided in Table 3.

3. Research methods 3.2. Experiment plan

Both experiments and simulation were used for the compara- A series of comparative experiments were conducted to reveal
tive study on thermal performance evaluation of PVB-DSF and system thermal behavior under different ventilation modes and
standard DSF. There are two stages of comparative study. The first system parameter settings. According to basic judgment and con-
stage is to compare thermal performance of PVB-DSF under differ- clusions from relevant studies about semi-transparent PV-DSF
ent operation modes and blinds parameters. The second stage is to [24], non-ventilation mode in summer condition performances
compare PVB-DSF and standard DSF structure. behave poorer than ventilation mode. Therefore, we excluded the
In order to demonstrate superb thermal performance of PVB- experiment of non-ventilation mode of PVB-DSF, while concen-
DSF system, the thermal behavior of PVB-DSF and conventional trated on natural and forced ventilation mode. A series of experi-
DSF should be investigated under the identical conditions, includ- ments were implemented from July to September. Typical days
ing the same environment and indoor air temperature and solar were selected for comparative study to ensure the results and con-
radiation. The thermal response of PVB-DSF system was tested in clusions are safe and valid for PVB-DSF system. There are two
an experiment in Changsha, China and the simulation results of major aspects in the experimental plan:
standard DSF under the same inputting parameters were used for
comparison. The adopted simulation model of DSF was previously (a) Comparison between natural and forced ventilation mode.
validated through field experiment test right in Changsha, China, As showed in Fig. 3, both inlet and outlet louver are used
which can largely ensure the validity of this comparison study. in natural ventilation mode, while 4 DC fans were placed
at the outlet to provide forced ventilation for cavity. Three
3.1. Experiment setup consecutive-day measurements of temperatures and solar
radiation were recorded from July 12th to 14th for natural
The experimental study was launched to evaluate thermal per- ventilation mode test, and July 24th to 26th for forced ven-
formance of PVB-DSF system in summer conditions. A series of tilation mode test.
tests and experiments were carried out from Jun. 2016 to Sep. (b) The core structure of PVB-DSF is PV blinds which need fur-
2016 at Changsha, Hunan province, China. The rig is located in a ther investigations. As showed by Fig. 4, PV-blinds angle is
typical region of Hot-Summer and Cold-Winter zone of China, changed from 30° to 45° and 60° with slats spacing
which is suitable for testing the capability of PVB-DSF for overheat- unchanged (4.5 cm). Then system performance was tested
ing prevention compared to conventional DSF. The PVB-DSF was for the effects of PV-blinds angle from Sep 7th and 8th. In
installed onto the south-facing wall. Thermocouples PT100 was addition, a single day system test was also conducted in
used to measure the temperatures of the glazing system, tradi- which blinds spacing is changed from 2.5 cm to 4.5 cm and
tional wall surfaces as well as ambient and indoor air tempera- 3.5 cm with slats angle of 45°.
tures. Fig. 1 exhibits the placement of temperature sensors and
solar pyranometers in the experiment. Two sets of solar pyranome- 3.3. Parallel comparative study with conventional DSF
ters were used to record outdoor global and diffuse solar irradiance
on a vertical plane as well as the indoor side south-facing vertical Sections 3.1 and 3.2 are designed to investigate how system
plane (near to internal glass of PVB-DSF). There is air inlet and out- performance is affected by operational modes and system param-
let set at the bottom and top position of external glass. The air inlet eters through field experimental test. In order to fairly evaluate
and out can be naturally ventilated, mechanical (forced) ventilated the thermal performance of PVB-DSF, a conventional DSF system

Fig. 3. Experiment picture of different ventilation modes of PVB-DSF system.

 Y. Luo et al. / Applied Energy 199 (2017) 281–293 285

Table 3
The key information of experimental instruments and their specifications.

Experimental equipment Manufacturer and model Sensitivity and/or technical data Measurement error
Solar pyranometers Jinzhou Sunshine Meteorological TBD-1 for diffuse radiation: 9.804 lV/(W/m );
2
Non-linearity  ±2%
Science and TBQ-2 for total radiation: 9.341 lV/(W/m2); (at 1000 W/m2)
Technology Co., Ltd. (TBD-1 and Output signal: 0–20 mA
TBQ-2)
Data logger Agilent KEYSIGHT TECHNOLOGY Data Logger: 34972A LXI Data Acquisition/Switch The minimum resolutions
(34972A LXI Data Acquisition/ unit: The Plug-In Modules: (1) scanning speed: are 1 lV and 0.1 °C
Switch unit and 34908A 40- 60 ch/s, (2) Maximum Input: 300 V; 1 A; 50 W;
Channel
Single-Ended Multiplexer)
Thermocouples PT 100 RTD thermocouples Temperature range: 50 to 200 °C ±0.2 °C
DC fans Taiwan SUNON Co., Ltd. (SUNON/ Power: 19 W, Voltage: AC 220 V, Current: 0.1 A,
SF11025AT) Volume: 112 m3/h, Operation environment: 10 to 60 °C

with and without venetian blinds was chosen as a reference fore, this simulation model can be safely adopted and used for
system. A recently published and validated numerical model of the purpose of comparison between PVB-DSF and standard DSF
conventional DSF system was adopted to deliver simulation under the same conditions.
results, which could be used to be compared with the experiment The basic strategy of the zonal model is to split DSF system into
results of PVB-DSF. The compared DSF and PVB-DSF are shared several zones and building governing differential equations for
with identical outside and inside air temperature as well as solar each zones. Five important governing equations describing two
radiation. Both systems are in forced ventilation mode. Three con- glass panes, one shading blind, and two air cavities are expressed
secutive days of measurements from Jul 8th to Jul 10th were used by Eqs. (1)–(5), where Teg,i, Tca1,i, Tbl,i, Tca2,i, Tig,i, are respectively
for this comparison study. the node i temperature of external glass, external cavity, blinds,
internal cavity, and internal glass pane (°C); Qeg,sol,i, Qbl,sol,i,
3.4. Simulation models of standard DSF used for comparative study Qig,sol,i, are the solar energy absorbed by external glass, blinds
and internal glass of DSF (W/m2); m is the air mass flow rate (kg/
For comparative study, the dynamical simulation model devel- s); C is specific capacity of air (J/(kg K)); h means the convective
oped and experimentally validated by recent research on standard or radiative heat transfer coefficient (W/m2 K); e and r are respec-
DSF was adopted [42]. This is an improved zonal model integrated tively the emissivity of glazing surface and Stefan-Boltzmann con-
with optical model [43] and airflow model was established. The stant; the subscript eg, ca1, bl, ca2, ig, in, out, sol are respectively
simulation accuracy is high according to the comparison results. the external glass pane, external air cavity, blinds layer,
Most importantly, the experiment location conducted by this study internal air cavity, internal glass, indoor space, ambient, and solar
is right in Changsha, China, the same as the present study. There- radiation.

Fig. 4. Experiment picture of different parameter settings of PV blinds in experiment.

286 Y. Luo et al. / Applied Energy 199 (2017) 281–293

dT eg;i the heat gain or heat loss of the façade, which can be calculated
ðmCÞeg;i ¼ Aeg;i hegout ðT out  T eg;i Þ
dt in this study by Eq. (6).
þ Aeg;i hegca1 ðT ca1;i  T eg;i Þ þ Aeg;i eegout rðT 4out  T 4eg;i Þ Ethg ¼ Gsg þ Erg þ Ecg ð6Þ
þ Aeg;i eegbl r ðT 4bl;i  T 4eg;i Þ
where Ggs is solar radiation penetrating the façade and received by
þ Aeg;i eegig rðT 4ig;i  T 4eg;i Þ þ Q eg;sol;i ð1Þ indoor space which can be harvested by the solar pyranometer
installed vertically near the internal glass (Fig. 1); Egr is the long
dT ca1;i wave radiant heat exchange between internal glass surface and
ðmCÞca1;i ¼ Aca1;i hegca1 ðT eg;i  T ca1;i Þ indoor space surfaces; Egc is the convective heat flux. The item Egr
dt
þ Aca1;i hca1bl ðT bl;i  T eg;i Þ and Egc can be calculated by Eqs. (7) and (8) [46]. The radiation heat
transfer coefficient hr and the convective heat transfer coefficient
þ ðmca1;i þ mðca1;ca2Þ;i ÞðC ca1;i1 T ca1;i1  C ca1;i T ca1;i Þ
for internal glass hc can be calculated by Eqs. (9) and (10) [47].
ð2Þ The parameter e is the surface emissivity and r is Boltzmann con-
stant. The convective heat transfer coefficient used in this study is
dT bl;i cited from other researches of glazing facades [47]. It is valid for
ðmCÞbl;i ¼ Abl;i hblca1 ðT ca1;i  T bl;i Þ þ Abl;i hblca2 ðT ca2;i  T bl;i Þ
dt the present study because the convective heat transfer coefficient
þ Abl;i eegbl rðT 4eg;i  T 4bl;i Þ is heavily linked to the surface material.
þ Abl;i eigbl rðT 4ig;i  T 4bl;i Þ þ Q bl;sol;i ð3Þ Erg ¼ hr ðT ig  AUSTÞ ð7Þ

dT ca2;i Ecg ¼ hc ðT ig  T in Þ ð8Þ

ðmCÞca2;i ¼ Aca2;i hblca2 ðT bl;i  T ca2;i Þ
dt
þ Aca2;i hca2ig ðT ig;i  T ca2;i Þ hr ¼ r  e  ðAUST 2 þ T 2ig ÞðAUST þ T ig Þ ð9Þ
þ mca2;i ðC ca2;i1 T ca2;i1  C ca2;i T ca2;i Þ
þ mðca1;ca2Þ;i ðC ca1;i T ca1;i  C ca2;i T ca2;i Þ ð4Þ hc ¼ 2:92  jT ig  T in j1=4 ð10Þ
where Tig (K) is the internal surface average temperature of building
dT ig;i envelope, AUST (K) is the area-weighted average temperature of
ðmCÞig;i ¼ Aig;i higca2 ðT ca2  T ig;i Þ þ Aig;i higin ðT in  T ig;i Þ
dt indoor surfaces.
þ Aig;i eegig rðT 4eg;i  T 4ig;i Þ The dynamic overall heat transfer coefficient k (W/m2 K) is used
þ Aig;i eigbl rðT 4bl;i  T 4ig;i Þ þ Aig;i einig rðT 4in  T 4ig;i Þ to evaluate the thermal insulating performance. The lower value of
k means better insulation performance. Eq. (11) gives the calcula-
þ Q ig;sol;i ð5Þ tion formula, which is cited from another research on PV glazing
Inputting data for simulation include indoor and outdoor air façade by Peng et al. [48]. It should be noted that the defined
temperature and solar radiation. In order to ensure a fair compar- parameter k is different from common k-value. Conventionally,
ative study between PVB-DSF and standard DSF, the inputting the k-value is obtained from the steady-state condition, which is
parameters used for DSF are from measurements in experiment not suitable for current comparative study. Considering the low
of PVB-DSF. The model is solved by finite difference method. thermal inertia of glazing façade, dynamic overall heat transfer
This is a widely used and validated numerical method which coefficient k used in this comparative study is feasible. By using
could reach better simulation accuracy than lumped model and this parameter, the insulating performance of two glazing systems
less CPU time than CFD method [44]. For this study, the improved can be compared dynamically or by an average value.
zonal method presented by Wang et al. [42] which combined the Erg þ Ecg
almost modeling equations of classical zonal method and air flow k ð11Þ
T in  T out
modeling pattern from work by Tanimoto and Kimura [45]. This
numerical model was validated by experimental test conducted Dynamic solar heat gain coefficient is defined and used in this
in hot summer and cold winter zone of China, which can largely comparative study, as a dimensionless parameter to show the heat
ensure the validity of using this model for the comparison study gain by a transparent envelope under unit intensity of solar radia-
in this study. tion. Eq. (12) gives the calculation formula where Gv (W/m2) is the
From the model validation part delivered by Wang et al. [42], instantaneous solar radiation incident on vertical façade, which is
the simulated dynamic temperatures at external glass, blinds and also cited from previous relevant study [48]. It also should be noted
internal glass are very close to the results from experiment, but that the difference between Eq. (12) and conventionally used def-
the calculated air temperature in the cavity cannot be accurately inition of SHGC. Conventionally, SHGC also is obtained from the
predicted at noon time. Because what this study concerned is the steady-state condition. But in order compare two glazing system
heat flux into indoor space through DSF, the simulation error at dynamically, Eq. (12) is used in this study and it also proved to
cavity air temperature does not influence the implementation of be feasible in Peng’s study of PV glazing system [48].
following work.
Ethg E g þ Erg þ Gsg
SHGC  ¼ c ð12Þ
Gv Gv
3.5. Thermal performance evaluation indexes
There is another important parameter named direct solar trans-
Because of the transparency features of glazing façade, specific mission se which can directly reflect the solar gain through glazing
indexes are needed for thermal performance evaluation. Usually, façade. This is also a dynamic parameter which totally obtained
instantaneous heat flux through glazing façade, dynamic overall from the measurement of indoor side and outdoor side solar pyra-
heat transfer coefficient k, and dynamic solar heat gain coefficient nometer in experiment.
(SHGC) are important parameters used as performance evaluators.
Gs
The instantaneous heat fluxes Egth penetrating through glazing se ¼ ð13Þ
Gv
façade (PVB-DSF or conventional DSF) system can directly reflect
 Y. Luo et al. / Applied Energy 199 (2017) 281–293 287

Table 4 PV-blinds parameters were experimentally investigated and com-

Average relative errors for calculated parameters in different modes. pared. This part is going to investigate how different operation
Egth SHGC k RSHGC modes and blinds parameters can influence system thermal perfor-
Mode 1 0.04279 0.04724 0.04281 0.04984 mance. Secondly, the conventional DSF with and without shading
Mode 2 0.04618 0.05033 0.04928 0.05303 blinds was chosen to be compared with PVB-DSF. This part of com-
parison is going to demonstrate the thermal performance of PVB-
DSF.
In order to deliver a comprehensive evaluation of PVB-DSF, one
more parameter for glazing performance evaluation was intro- 4.1. Different ventilation modes of PVB-DSF
duced, named relative solar heat gain coefficient (RSHGC). Param-
eter RSHGC means the surplus heat gain of researched glazing The glazing part as well as the PV modules in PVB-DSF absorb
façade to a reference façade system under the same condition. This and store a portion of solar energy at daytime. This will lead to sys-
reference system can be another glazing system or opaque wall tem temperature rise, which could be beneficial for winter condi-
system. Because there are two glazing system compared in this tions but fatal for both the systematical electric and thermal
study and considering the frequently used concept of window wall performance in summer conditions. Therefore, different ventilation
ratio (WWR), an opaque brick wall with the same thickness of glaz- modes should be adopted for winter and summer conditions. It
ing façade was selected as reference system for the calculation of was shown by the experimental studies of semi-transparent PV-
RSHGC. This parameter therefore can reflect thermal performance DSF that PV glazing façade should operate under ventilation mode
difference between glazing and brick wall. RSHGC will have a con- in summer and non-ventilation mode in winter [48,50]. We can
nection to the design of WWR. For example, a window with high safely extend this conclusion to PVB-DSF system using energy flow
RSHGC has to be designed with lower WWR to prevent heat loss analysis. When the cavity of PVB-DSF is closed and isolated from
or heat gain. The current research objective is PVB-DSF system ambient air flow, it can better accumulate thermal energy in win-
and the discussion about WWR is not included. ter. However, we can only ensure that, for summer conditions, ven-
Ethg  Eref tilation mode is better than non-ventilation mode. We need further
RSHGC ¼ th
ð14Þ exploration to understand which kinds of ventilation modes are
Gv
suitable for PVB-DSF under hot summer and cold winter zone of
China.
3.6. Uncertainty analysis Mechanical ventilation mode (mode 1) and natural ventilation
mode (mode 2) was chosen for comparative study. Three consecu-
An uncertainty analysis was implemented based on the mea- tive days test data was used and processed for the comparison of
surement errors. The measuring errors for each instrument are mode 1 and mode 2. Evaluation indexes of SHGC, k and RSHGC
provided in Table 3 but the error like SHGC or k cannot be directly were adopted in this sub-section. Fig. 5 shows daily averaged SHGC
obtained. The uncertainty analysis is to give the credibility of those of PVB-DSF under mode 1 and mode 2. Both SHGC of mode 1 and 2
calculated parameters derived from other measured data. can show very similar system thermal performance of natural and
If parameter y is the function of x1, x2, . . . xn by the form of Eq. mechanical ventilation mode of PVB-DSF. From comparison,
(15), then the relative error of y is linked to the errors from xi which mechanical ventilation (mode 1) is slightly better than natural ven-
can be calculated by Eq. (16) [49]. If parameter y is the function of tilation mode (mode 2). It is observed that their difference have a
x1, x2, . . . xn as the form of Eq. (17), the relative error of y can be cal- connection with the weather condition. For example, day 2 is a
culated by Eq. (18). The errors of Egth, SHGC, k, and RSHGC are calcu- sunny day without a breeze and the difference caused by used
lated for three different modes. Instantaneous values were used to DC fans in mode 1 is obvious.
calculate the relative errors and then these error values were aver- Fig. 6 shows daily averaged k of PVB-DSF under mode 1 and
aged. Those errors are closely related to the measured data and their mode 2 which is an important parameter for insulation perfor-
average values are listed in Table 4. Those errors are controlled mance evaluation. According to Fig. 6, even the highest averaged
within 6% which can be an indication of the reliability of the mea- k of PVB-DSF in summer is 2.976 which is much lower than the
sured data. average k-value of semi-transparent PV-DSF [48]. The k-value of
mode 1 could be slightly lower than the mode 2 which can indicate
x1 x2    xm
y¼ ð15Þ a better insulation performance.
xmþ1 xmþ2    xn
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u
dy u Xm dxi 2 X n 
dxi
2
¼ t i¼1 þ  ð16Þ
y xi i¼mþ1
xi

X
n
y¼ xi ð17Þ
i¼1

dy X n
dxi
¼ ð18Þ
y i¼1
xi

4. Results and discussion

Both experiment and simulation approaches were used to con-

duct a comparative study on thermal performance evaluation of
PVB-DSF system. Firstly, different system ventilation modes and Fig. 5. SHGC comparison between mode 1 and mode 2.
288 Y. Luo et al. / Applied Energy 199 (2017) 281–293

summer conditions. Based on this result, PV-blinds parameters

such as blinds angle and spacing were changed in experiment to
investigate the effect of those parameters on system thermal per-
formance under natural ventilation mode.
Single day test was implemented for each situation. When PV
blinds angle is the research target, blinds spacing remains 4.5 cm.
When PV blinds spacing is changed in experiment, blinds angle
keeps as 45°. The experimental picture is shown by Fig. 4. A set
of solar heat gain analysis indexes (se, SHGC, and RSHGC) was
adopted to deliver comprehensive evaluation.

4.2.1. Influence of PV blinds angle

In order to investigate the effect of PV blinds angle on system
thermal performance, three separate tests were implemented with
blinds angle at 30°, 45° and 60° while keeping blinds spacing at
4.5 cm. Fig. 8 shows direct solar transmission of PVB-DSF under
different blinds angle settings. Experimental data indicated that
Fig. 6. Comparison between mode 1 and mode 2 in terms of parameter k. the blinds angle has little effect on se. The similar results also
can be viewed from indexes of SHGC and RSHGC by Figs. 9 and
10. Those curves are pretty close to each other and the effect of
blinds angle is implicit. This can be explained by the large blind
spacing. In order to avoid partial shading of PV blind, the blind
spacing has to be set at 4.5 cm to prevent power loss. But the

Fig. 7. RSHGC comparison of mode 1 and mode 2.

In order to deliver a more direct understanding of PVB-DSF, we

installed temperature sensors on a south-facing traditional brick
wall (240 mm thickness with 10 mm mortar layer on both external
and internal surface of brick wall) shown as Fig. 1 to realize the cal-
culation of relative SHGC (RSHGC). Fig. 7 compares the daily aver-
Fig. 8. Solar transmission of PVB-DSF under different PV-blinds angle with blinds
aged RSHGC of mode 1 and mode 2. It is shown that mode 1 and 2 spacing of 4.5 cm.
has similar RSHGC. The results indicated that the average heat gain
of PVB-DSF is about 45–55% larger than the traditional brick wall
under unit intensity of solar radiation in summer conditions in 1.0
China. It is clearly shown that mode 1 outperforms mode 2 in three Blinds angle 30 degree
0.9 Blinds angle 45 degree
consecutive days in RSHGC comparison.
Blinds angle 60 degree
In short summary, both natural and mechanical ventilation 0.8
mode of PVB-DSF can reach satisfying thermal performance in
0.7
terms of SHGC, k-value and RSHGC. And those two kinds of venti-
lation methods don’t show obvious differences on three important 0.6
SHGC

evaluation indexes. Mechanical ventilation otherwise consumes 0.5

extra fan power to accelerate the air movement in the cavity. In
practical operation, natural ventilation mode is suggested for 0.4
PVB-DSF working in hot summer and cold winter zone of China. 0.3
Using fans with larger ventilation volume certainly could improve
system thermal performance, but extra electric power consump- 0.2
tion makes it an uneconomical option. 0.1

0.0
00:00 06:00 12:00 18:00 00:00
4.2. Different parameters settings of PV blinds
Time
After the comparative study on natural and mechanical ventila- Fig. 9. SHGC of PVB-DSF under different PV-blinds angle with blinds spacing of
tion modes, natural ventilation mode is suggested to be used in 4.5 cm.
 Y. Luo et al. / Applied Energy 199 (2017) 281–293 289

1.4 0.30
Blinds angle 30 degree Blinds spacing 2.5 cm
1.3 Blinds angle 45 degree Blinds spacing 3.5 cm
1.2 Blinds angle 60 degree 0.25 Blinds spacing 4.5 cm
1.1
1.0
0.20
0.9

DSHGC
0.8
RSHGC

0.7 0.15
0.6
0.5 0.10
0.4
0.3
0.05
0.2
0.1
0.0 0.00
00:00 06:00 12:00 18:00 00:00 00:00 06:00 12:00 18:00 00:00

Time Time

Fig. 11. Solar transmission of PVB-DSF under different PV-blinds spacing.

Fig. 10. RSHGC of PVB-DSF under different PV-blinds angle with blinds spacing of
4.5 cm.

1.2
Blinds spacing 2.5 cm
Table 5 1.1 Blinds spacing 3.5 cm
Solar heat gain analysis of PVB-DSF under different PV-blind parameters.
1.0 Blinds spacing 4.5 cm
Adjusted se SHGC RSHGC
parameters
0.9
0.8
PV-blinds angles (blinds spacing 30° 0.1263 0.3284 0.3314
remains 4.5 cm) 45° 0.1112 0.3111 0.3237 0.7
SHGC

60° 0.1133 0.3151 0.3188

0.6
PV-blinds spacing (blinds angle 2.5 cm 0.0531 0.2015 0.2443
0.5
remains 45°) 3.5 cm 0.0908 0.2596 0.3061
4.5 cm 0.1112 0.3112 0.3112 0.4
0.3
0.2
influence of the blind angle under such condition has minor impact 0.1
on system thermal performance. 0.0
Because the cooling load of glazing façade at noon time is 00:00 06:00 12:00 18:00 00:00
always high, therefore we calculated the average solar transmis- Time
sion, SHGC and RSHGC respectively using the data from 11:00 to
14:00 to fairly compare and analyze the influence of PV blinds Fig. 12. SHGC of PVB-DSF under different PV-blinds spacing.
angle in numeric values. The calculated values are listed in Table 5.
It was shown that solar transmission can be 0.1263, 0.1112 and
0.1133 respectively for blinds angle of 30°, 45° and 60°. The PV 1.5
Blinds spacing 2.5 cm
blinds with larger angle can block more solar heat transmission Blinds spacing 3.5 cm
and the figures in Table 5 can show the right tendency but not that 1.0 Blinds spacing 4.5 cm
evident. The blinds spacing is still too large, and those test results
were obtained from three single-day measurements which may be
0.5
influenced by weather conditions.
RSHGC

Besides, the average SHGC are 0.3284, 0.3111 and 0.3151

respectively for blinds angle of 30°, 45° and 60°. The effect of blinds 0.0
angle also cannot be clearly shown by SHGC because those three
tests were obtained under different weather conditions. However, -0.5
by adopting RSHGC and using traditional brick wall as reference,
the value of RSHGC for blinds angle of 30°, 45° and 60° are
0.3314, 0.3237 and 0.3188 accordingly. This newly defined index -1.0
can not only present an easily understood evaluation, but also
can filter some disturbance factors to show a more objective com- -1.5
parison result. 00:00 06:00 12:00 18:00 00:00
Time
4.2.2. Influence of PV blinds spacing Fig. 13. RSHGC of PVB-DSF under different PV-blinds spacing.
Another three single-day tests were conducted for PVB-DSF
with different PV blinds spacing of 2.5 cm, 3.5 cm and 4.5 cm while
keeping blind angle at 45°. Figs. 11–13 respectively showcase the weather conditions were slightly different in three single-day tests.
effect of blinds spacing on solar transmission, SHGC and RSHGC. The thermal performance improvement by changing blinds spac-
It is evident that from those three indexes blinds spacing have ing from 4.5 cm to 2.5 cm is obvious. According to Table 5, the solar
the significant influence on system thermal performance although transmission of 4.5 cm case is nearly double of 2.5 cm case. In
290 Y. Luo et al. / Applied Energy 199 (2017) 281–293

addition, PVB-DSF with blind-spacing of 2.5 cm can reduce about solar radiation and diffuse radiation on a vertical plane of a south-
0.1 SHGC compared to 4.5 cm case. As for RSHGC, the heat gain facing wall were measured.
by PVB-DSF with blinds spacing of 2.5 cm, 3.5 cm and 4.5 cm is lar- Temperature sensors were attached at different parts of PVB-
ger than traditional brick wall by 24.43%, 30.61% and 31.12% DSF, and the internal surface temperature of the internal glass pane
respectively. is one important part, which can directly affect the instantaneous
Specifically, Fig. 13 shows the negative value of RSHGC in the heat gain, SHGC as well as human thermal comfort. Fig. 15 shows
early morning, which can be explained by the relatively low tem- temperature profile comparison of internal surface of PVB-DSF,
perature of glazing façade and the calculation formula of RSHGC. conventional DSF with and without shading blinds. The compar-
Because the specific heat capacity of glass façade is much smaller ison clearly indicates that conventional DSF can be 1–1.5 °C higher
than a brick wall, the internal surface temperature of glazing than PVB-DSF at noon time but the difference is not evident for
façade can be easily influenced by ambient temperature at night using shading blinds or nor.
and early morning and resulted as relatively low surface tempera- In summer conditions, the direct solar transmission through
ture. The internal surface temperature of traditional brick wall is glazing façade is the major cause of high cooling load. Fig. 16 pre-
higher than glazing system so as to the instantaneous heat gain sents the comparison of solar penetration through PVB-DSF and
on summer morning. According to Eq. (14), RSHGC is the calcula- standard DSF system. The experiment and simulation results show
tion of difference heat gain between PVB-DSF and traditional wall. that conventional DSF without shading blinds will be suffered from
Therefore the RSHGC in early morning could be negative. In order high direct solar transmission by 60–75 W/m2, while solar pene-
to take clear numerical comparison of the influence of PV-blind tration through PVB-DSF and conventional blinds embedded DSF
spacing, the averaged value of solar transmission, SHGC and RSHGC were only 33–35 W/m2 with maximum level of 45 W/m2 under
are listed in Table 5. the same weather conditions. This comparison refers that the
Based on above analysis, on one hand it was found that larger installation of shading blinds can reduce about 45–53% of direct
angle of PV-blind leads to larger solar heat gain but the compara- solar heat gain in summer.
tive analysis by different indexes showed that thermal perfor-
mance of PVB-DSF is slightly sensitive to PV-blind angle. On the
other hand, wider blind spacing of PV-blind also leads to higher
values of SHGC and RSHGC which is harmful for glazing façade in 36 PVB-DSF
summer conditions. And the change of PV-blind spacing has a more 35 Conventional DSF
obvious impact on thermal performance of PVB-DSF system. Conventional DSF without blinds
34
33

4.3. Comparison between PVB-DSF and standard DSF 32

Temperature ( C)
o

31
A number of researchers are proposing new and effective glaz- 30
ing façade to reduce energy consumption in building sector. The 29
conventional DSF is seen as the benchmark for many new proposed 28
glazing façades to demonstrate their merits and analyze system 27
energy saving potential quantitatively. In a newly proposed water 26
pipe-embedded DSF by Shen and Li [34,35], an effectiveness coef- 25
ficient was used for thermal performance evaluation by taking con- 24
ventional DSF as a reference system. Therefore, in this present
23
study of PVB-DSF, we chose conventional DSF system for the com- 2016/7/8 2016/7/8 2016/7/9 2016/7/9 2016/7/10 2016/7/10 2016/7/11
prehensive thermal performance comparison. Time (min)
The PVB-DSF system was experimentally tested in three consec-
utive testing days under mechanical ventilation mode. Experimen- Fig. 15. Comparison of internal surface temperature of PVB-DSF and conventional
DSF.
tal period is from Jul. 8th to Jul. 10th 2016. The indoor and outdoor
temperature and radiation were measured. Those measured tem-
peratures and solar radiation were used as input data for system
simulation program of DSF which is shown by Fig. 14. Both global
Solar radiation penetrating into indoor space (W/m )
2

80
46 350 PVB-DSF
Solar radiation on south-facing wall (W/m )
2

75
44 Tin Conventional DSF
70
Tout 300 Conventional DSF without blinds
42 65
Gt,v
40 60
Gdif,v 250
38 55
Temperature ( C)

50
36
o

200 45
34
40
32 35
150
30 30
28 25
100
26 20
15
24 50 10
22 5
20 0 0
2016/7/8 2016/7/8 2016/7/9 2016/7/9 2016/7/10 2016/7/10 2016/7/11 2016/7/8 2016/7/8 2016/7/9 2016/7/9 2016/7/10 2016/7/10 2016/7/11
Time Time (min)

Fig. 14. Indoor and outdoor thermal and radiation environment in summer. Fig. 16. Comparison of solar transmission through PVB-DSF and conventional DSF.
 Y. Luo et al. / Applied Energy 199 (2017) 281–293 291

165 PVB-DSF
Conventional DSF
150 Conventional DSf without blinds
135
120
Heat flux (W/m )
2

105
90
75
60
45
30
15
0
2016/7/8 2016/7/8 2016/7/9 2016/7/9 2016/7/10 2016/7/10 2016/7/11
Time (min)

Fig. 17. Comparison of instantaneous heat gain of PVB-DSF and conventional DSF.
Fig. 19. Comparison of k-value of PVB-DSF and conventional DSF.

1.2
PVB-DSF
Conventional DSF 1.2
PVB-DSF
1.0 Conventional DSF without blinds 1.1 Conventional DSF
Conventional DSF without blinds
1.0
0.8 0.9
0.8
SHGC

0.7
RSHGC

0.6
0.6
0.4 0.5
0.4
0.2 0.3
0.2
0.0 0.1
2016/7/8 2016/7/8 2016/7/9 2016/7/9 2016/7/10 2016/7/10 2016/7/11
0.0
Time (min) 2016/7/8 2016/7/8 2016/7/9 2016/7/9 2016/7/10 2016/7/10 2016/7/11
Time (min)
Fig. 18. SHGC comparison between PVB-DSF and conventional DSF.
Fig. 20. RSHGC comparison between PVB-DSF and conventional DSF.

The internal surface temperature of glazing façade and direct

solar penetration shown by Figs. 15 and 16 are two basic compo- The comprehensive heat transfer coefficient k-value is a crucial
nents of daily heat gain. Fig. 17 shows the comparison of instanta- parameter to evaluate insulation performance of building enve-
neous heat flux among three different systems. If shading device is lope. Fig. 19 presents the dynamic k-value of PVB-DSF and conven-
not installed for conventional DSF, the instantaneous heat gain at tional DSF system. Those two system have similar k-value at night
noon time can reach as high as 125–135 W/m2, while heat flux but quite different at daytime. The k-value of conventional DSF
could be reduced to about 105 W/m2 if shading blinds are installed with and without venetian blinds reaches about 4.99 and 5.31
for glazing facade. And this index can be further curtailed to 75– respectively. However, k-value of PVB-DSF is around 3.5 at noon
85 W/m2 by PVB-DSF system. If daily totally cooling load of three time which exhibits a quite satisfying insulation performance. By
configurations were considered and compared, it turns out that the way, the daily average k-value is 2.247 which is much lower
conventional DSF with and without shading device consumes than the value of semi-transparent PV-DSF in previous study in
848.03 Wh/(m2day) and 1000.82 Wh/(m2day) respectively, which the k-value under ventilation mode is 4.6 [24]. This low k-
while PVB-DSF consumes only 744.94 Wh/(m2day). The thermal value of PVB-DSF is achieved by lower instantaneous heat flux in
performance enhancement brought by PVB-DSF is significant, fact.
because PVB-DSF can save about 12.16% and 25.57% of cooling In addition, the RSHGC comparison shown by Fig. 20 also indi-
power consumption in summer compared with conventional DSF cates PVB-DSF has an obviously better thermal performance than
with and without blinds. conventional DSF. The RSHGC of PVB-DSF, conventional DSF with
Fig. 18 shows instantaneous SHGC of PVB-DSF and conventional and without venetian blinds are 0.341, 0.441 and 0.549 respec-
DSF system. The SHGC of conventional DSF reaches 0.439 (using tively at noon time, which means the heat flux of PVB-DSF is
shading blinds) and 0.547 (without shading blinds) at noon time. 34.1% larger than traditional brick wall with identical thickness
And SHGC of PVB-DSF is only 0.339 which is 22.78% and 38.03% but this value reaches 44.1% and 54.9% for conventional DSF.
lower than conventional DSF with and without venetian blinds. In short summary of the comparison between PVB-DSF and con-
Three curves in Fig. 18 have a distinct descending order from con- ventional DSF, the internal surface temperature of glazing façade
ventional DSF without venetian blinds to PVB-DSF throughout Tig, solar penetration Gs, heat gain Eth, k-value, SHGC and RSHGC
entire three testing days. of three systems in summer conditions were listed in Table 6.
292 Y. Luo et al. / Applied Energy 199 (2017) 281–293

Table 6
Summary of thermal performance of PVB-DSF compared with conventional DSF (noon time).

System Tig (°C) Gs (W/m2) Eth (W/m2) k-value (W/m2 K) SHGC RSHGC
PVB-DSF 31.18 32.12 79.05 3.56 0.339 0.341
Conventional DSF using venetian blinds 32.83 36.29 102.22 4.99 0.439 0.441
Conventional DSF without venetian blinds 33.19 57.37 127.43 5.31 0.547 0.549

The measured and simulated data at noon time from 12:00 to 33–35 W/m2. The direct shading effect of PV blinds is not
13:00 were used and averaged, which can clearly showcase the evident compared with conventional venetian blinds, but
improved thermal performance of PVB-DSF from different evalua- the comparison about the instantaneous heat gain clearly
tion indexes. shows that conventional DSF with and without shading
device consumes 848.03 Wh/(m2day) and 1000.82 Wh/
(m2day) respectively, while PVB-DSF consumes only
5. Conclusions 744.94 Wh/(m2day). PVB-DSF can save about 12.16% and
25.57% of indoor cooling power consumption in summer
This study presented a new double skin façade using PV blinds compared with conventional DSF with and without shading
as a shading device in the cavity, which is called as PVB-DSF. This blinds.
glazing system was designed to provide a multi-function of power (4) The average k-value of conventional DSF with and without
generation, solar penetration reduction and flexible daylighting blinds reaches about 4.99 and 5.31 respectively. However,
control, which is unattainable for traditional PV façade. A compar- k-value of PVB-DSF is around 3.5 at noon time which exhi-
ative research approach was implemented to evaluate system ther- bits a quite satisfying insulation performance. The daily
mal performance. Evaluation indexes such as heat flux, k-value, average k-value of PVB-DSF is 2.247. This result is much
and dynamic solar heat gain coefficient (SHGC), direct solar trans- lower than the k-value of semi-transparent PV-DSF in which
mission and relative solar heat coefficient (RSHGC) were used to the value is 4.6.
deliver more comprehensive solar heat gain analysis. There are (5) SHGC of PVB-DSF at noon time is only 0.339 which is 22.78%
basically two major parts of comparative researches. Firstly, the and 38.03% lower than conventional DSF with and without
influences of different ventilation modes and blinds parameter set- venetian blinds. The RSHGC of PVB-DSF, conventional DSF
tings were studied by field test. Secondly, improved thermal per- with and without venetian blinds are 0.341, 0.441 and
formance of PVB-DSF was demonstrated by comparing with 0.549 respectively at noon time, which means the heat flux
standard DSF under a series of evaluation indexes. Both experi- of PVB-DSF is 34.1% larger than traditional brick wall with
mental measurement and simulation program were adopted in identical thickness but this value reaches 44.1% and 54.9%
the research content and some important results are listed here: for conventional DSF.

(1) k-value, SHGC and RSHGC were used to evaluate the system The newly proposed PV-blind double skin façade has fully
thermal performance under natural and mechanical ventila- demonstrated its thermal performance through comparative stud-
tion mode in summer condition for three consecutive days of ies. This PV façade system can generate electricity, block extra solar
testing. The value of k-value, SHGC and RSHGC under two radiation with satisfying thermal insulation. This glazing system
ventilation modes doesn’t show distinctive differences. The can contribute more to energy efficiency in building sector and
data analysis indicated that RSHGC can deliver a fair evalu- future research plans could be focused on system numerical simu-
ation despite of slight influence of different testing weather lation, economic analysis, environmental impact as well as system
conditions. But the forced ventilation mode slightly outper- optimizations.
forms natural ventilation mode. The k-value in mechanical
ventilation mode is slightly lower than the natural ventila-
tion mode. But considering the electric power consumption Acknowledgement
by DC fans, it is suggested that natural ventilation mode
should be adopted for the application of PVB-DSF in summer The work described in this paper is sponsored by the National
conditions. Natural Science Foundation of China (No. 51578221); International
(2) For one hand, the effect of PV blinds angle on system ther- Comparison Research on the Construction Mode of Green Town
mal performance is not evident by using indexes of solar (2015ZK2005). Prof. Ling Zhang would also like to acknowledge
transmission, SHGC and RSHGC, because of large blind spac- the financial support from the Collaborative Innovation Center
ing. On the other hand, the smaller PV blinds spacing with for Building Energy Conservation and Environment Control, Zhuz-
same blinds angle can reduce solar heat gain by different hou, Hunan Province, China. Dr. Yongqiang Luo and Prof. Ling
scale in experiment. The averaged SHGC at noon time for Zhang would like to thank the instructions and help provided by
PVB-DSF with different blinds spacing of 2.5 cm, 3.5 cm Prof. Youming Chen. And we also want extend our great gratitude
and 4.5 cm are 0.2015, 0.2596 and 0.3112 respectively. to a professional interpreter Lanlan Hu, for her particular language
Because the performance of PVB-DSF is less sensitive to help during the entire process of paper composing and revising.
PV-blind angle than PV-blind spacing and blind spacing is
fixed after system installation, the PV-blind spacing should References
be properly designed for balancing daylighting demand
and solar heat gain. [1] Ghaffarianhoseini A, Ghaffarianhoseini A, Berardi U, Tookey J, Li DHW,
Kariminia S. Exploring the advantages and challenges of double-skin façades
(3) In comparison between DSF and PVB-DSF, conventional DSF (DSFs). Renew Sustain Energy Rev 2016;60:1052–65.
is 1–1.5 °C higher than PVB-DSF at noon time. Conventional [2] Luo Y, Zhang L, Liu Z, Wang Y, Meng F, Wu J. Thermal performance evaluation
DSF without shading blinds suffered from high direct solar of an active building integrated photovoltaic thermoelectric wall system. Appl
Energy 2016;177:25–39.
transmission by 60–75 W/m2, while solar penetration
[3] Yin R, Xu P, Shen P. Case study: energy savings from solar window film in two
through PVB-DSF and standard DSF with blinds were only commercial buildings in Shanghai. Energy Build 2012;45:132–40.
 Y. Luo et al. / Applied Energy 199 (2017) 281–293 293

[4] Gao Y, Luo H, Zhang Z, Kang L, Chen Z, Du J, et al. Nanoceramic VO2 [27] Taffesse F, Verma A, Singh S, Tiwari GN. Periodic modeling of semi-transparent
thermochromic smart glass: a review on progress in solution processing. Nano photovoltaic thermal-trombe wall (SPVT-TW). Sol Energy 2016;135:265–73.
Energy 2012;1:221–46. [28] Skandalos N, Karamanis D. Investigation of thermal performance of semi-
[5] Richter B, Goldston D, Crabtree G, Glicksman L, Goldstein D, Greene D. How transparent PV technologies. Energy Build 2016;124:19–34.
America can look within to achieve energy security and reduce global [29] Shyam, Tiwari GN. Analysis of series connected photovoltaic thermal air
warming. Rev Mod Phys 2008;80:S1–S109. collectors partially covered by semitransparent photovoltaic module. Sol
[6] Parkin IP, Manning TD. Intelligent thermochromic windows. J Chem Educ Energy 2016;137:452–62.
2006;83:393–400. [30] Kapsis K, Athienitis AK. A study of the potential benefits of semi-transparent
[7] Vakiloroaya V, Samali B, Fakhar A, Pishghadam K. A review of different photovoltaics in commercial buildings. Sol Energy 2015;115:120–32.
strategies for HVAC energy saving. Energy Convers Manage 2014;77:738–54. [31] Miyazaki T, Akisawa A, Kashiwagi T. Energy savings of office buildings by the
[8] Shameri MA, Alghoul MA, Sopian K, Zain MFM, Elayeb O. Perspectives of use of semi-transparent solar cells for windows. Renew Energy
double skin facade systems in buildings and energy saving. Renew Sust Energy 2005;30:281–304.
Rev 2011;15:1468–75. [32] Manz H. Total solar energy transmittance of glass double facades with free
[9] Zhang T, Tan Y, Yang H, Zhang X. The application of air layers in building convection. Energy Build 2004;36:127–36.
envelopes: a review. Appl Energy 2016;165:707–34. [33] Silva T, Vicente R, Amaral C, Figueiredo A. Thermal performance of a window
[10] Ebrahimpour A, Maerefat M. Application of advanced glazing and overhangs in shutter containing PCM: numerical validation and experimental analysis. Appl
residential buildings. Energy Convers Manage 2011;52:212–9. Energy 2016;179:64–84.
[11] Chan ALS, Chow TT, Fong KF, Lin Z. Investigation on energy performance of [34] Shen C, Li XT. Solar heat gain reduction of double glazing window with cooling
double skin façade in Hong Kong. Energy Build 2009;41:1135–42. pipes embedded in venetian blinds by utilizing natural cooling. Energy Build
[12] Høseggen R, Wachenfeldt BJ, Hanssen SO. Building simulation as an assisting 2016;112:173–83.
tool in decision making. Energy Build 2008;40:821–7. [35] Shen C, Li X. Thermal performance of double skin façade with built-in pipes
[13] Xu L, Ojima T. Field experiments on natural energy utilization in a residential utilizing evaporative cooling water in cooling season. Sol Energy
house with a double skin façade system. Build Environ 2007;42:2014–23. 2016;137:55–65.
[14] Pasquay T. Natural ventilation in high-rise buildings with double facades, [36] Bahr W. A comprehensive assessment methodology of the building integrated
saving or waste of energy. Energy Build 2004;36:381–9. photovoltaic blind system. Energy Build 2014;82:703–8.
[15] Xue F, Li X. A fast assessment method for thermal performance of naturally [37] Charron R, Athienitis AK. Optimization of the performance of double-facades
ventilated double-skin façades during cooling season. Sol Energy with integrated photovoltaic panels and motorized blinds. Sol Energy
2015;114:303–13. 2006;80:482–91.
[16] da Silva Marques F, Gomes MG, Rodrigues AM. Measuring and estimating [38] Kim S-H, Kim I-T, Choi A-S, Sung M. Evaluation of optimized PV power
airflow in naturally ventilated double skin facades. Build Environ generation and electrical lighting energy savings from the PV blind-integrated
2015;87:292–301. daylight responsive dimming system using LED lighting. Sol Energy
[17] Gorgolis G, Karamanis D. Solar energy materials for glazing technologies. Sol 2014;107:746–57.
Energy Mater Sol C 2016;144:559–78. [39] Kang S, Hwang T, Kim JT. Theoretical analysis of the blinds integrated
[18] Park KE, Kang GH, Kim HI, Yu GJ, Kim JT. Analysis of thermal and electrical photovoltaic modules. Energy Build 2012;46:86–91.
performance of semi-transparent photovoltaic (PV) module. Energy [40] Bai J, Cao Y, Hao Y, Zhang Z, Liu S, Cao F. Characteristic output of PV systems
2010;35:2681–7. under partial shading or mismatch conditions. Sol Energy 2015;112:41–54.
[19] Xu S, Liao W, Huang J, Kang J. Optimal PV cell coverage ratio for semi- [41] Jung TH, Ko JW, Kang GH, Ahn HK. Output characteristics of PV module
transparent photovoltaics on office building façades in central China. Energy considering partially reverse biased conditions. Sol Energy 2013;92:214–20.
Build 2014;77:130–8. [42] Wang Y, Chen Y, Zhou J. Dynamic modeling of the ventilated double skin
[20] Wang M, Peng J, Li N, Lu L, Ma T, Yang H. Assessment of energy performance of façade in hot summer and cold winter zone in China. Build Environ
semi-transparent PV insulating glass units using a validated simulation model. 2016;106:365–77.
Energy 2016;112:538–48. [43] Wang Y, Chen Y. Modeling and calculation of solar gains through multi-glazing
[21] Wei L, Shen X. Energy performance comparison among see-through facades with specular reflection of venetian blind. Sol Energy 2016;130:33–45.
amorphous-silicon PV (photovoltaic) glazings and traditional glazings under [44] Jiru TE, Haghighat F. Modeling ventilated double skin façade—a zonal
different architectural conditions in China. Energy 2015;83:267–75. approach. Energy Build 2008;40:1567–76.
[22] Wang M, Peng J, Li N, Yang H, Wang C, Li X, et al. Comparison of energy [45] Tanimoto J, Kimura K-i. Simulation study on an airflow window system with
performance between PV double skin facades and PV insulating glass units. an integrated roll screen. Energy Build 1997;26:317–25.
Appl Energy 2017;194:148–60. [46] Jeong JW, Mumma SA. Simplified cooling capacity estimation model for top
[23] Zhang W, Lu L, Peng J, Song A. Comparison of the overall energy performance insulated metal ceiling radiant cooling panels. Appl Therm Eng
of semi-transparent photovoltaic windows and common energy-efficient 2004;24:2055–72.
windows in Hong Kong. Energy Build 2016;128:511–8. [47] Zanghirella F, Perino M, Serra V. A numerical model to evaluate the thermal
[24] Peng J, Lu L, Yang H. An experimental study of the thermal performance of a behaviour of active transparent façades. Energy Build 2011;43:1123–38.
novel photovoltaic double-skin facade in Hong Kong. Sol Energy [48] Peng J, Lu L, Yang H, Ma T. Comparative study of the thermal and power
2013;97:293–304. performances of a semi-transparent photovoltaic façade under different
[25] Peng J, Curcija DC, Lu L, Selkowitz SE, Yang H, Zhang W. Numerical ventilation modes. Appl Energy 2015;138:572–83.
investigation of the energy saving potential of a semi-transparent [49] Sardarabadi M, Passandideh-Fard M, Heris SZ. Experimental investigation of
photovoltaic double-skin facade in a cool-summer Mediterranean climate. the effects of silica/water nanofluid on PV/T (photovoltaic thermal units).
Appl Energy 2016;165:345–56. Energy 2014;66:264–72.
[26] Myong SY, Jeon SW. Efficient outdoor performance of esthetic bifacial a-Si: H [50] Peng JQ, Lu L, Yang HX, Han J. Investigation on the annual thermal
semi-transparent PV modules. Appl Energy 2016;164:312–20. performance of a photovoltaic wall mounted on a multi-layer facade. Appl
Energy 2013;112:646–56.