Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

Applied Energy 266 (2020) 114776

Contents lists available at ScienceDirect

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

Optimizing the thermal performance of window frames through aerogel- T


enhancements
Jason Paulos1, Umberto Berardi1
Ryerson University, 350 Victoria street, Toronto, M5B 2K3, ON, Canada

HIGHLIGHTS

• The thermal transmittance of 48 commercially-available high-performance frames is investigated.


• Aluminium, fibreglass, PVC, and wood-composite window frames are compared and contrasted.
• Improvements of the frames by inserting aerogel granules in their cavities are investigated numerically.
• Filling the cavities with the aerogel would reduce the thermal transmittance by up to 29%.
• Window frames with a thermal transmittance lower than 0.5 W/m K are proposed.
2

ARTICLE INFO ABSTRACT

Keywords: Windows are often considered the weakest point in building envelopes, especially in buildings with high
Aerogel window-to-wall ratios, where the windows exhibit much higher thermal transmittance than the opaque portion
Aerogel-enhanced material of the walls. The poor thermal performance of windows can largely be attributed to their frames which have
Window frames higher thermal transmittance than the glazing portions. This study investigates the thermal transmittance of 48
Finite element method
commercially-available high-performance aluminium, fibreglass, polyvinylchloride (PVC) and wood-composite
Thermal transmittance
window frames. Then, it focuses on the possible improvements of each frame by inserting aerogel in the frame
cavities. Several modifications of the frame cavities are assessed through two-dimensional numerical modelling
done according to the ISO 10077-2 standard. This research concludes that filling existing empty cavities of
window frames with aerogel granules could reduce the frame thermal transmittance by 4–29% depending on the
frame type. Moreover, the complete filling of the cavities with aerogel can further reduce the thermal trans-
mittance by 35%. Finally, for each investigated material, window frames with a thermal transmittance as low as
0.5 W/m2 K are proposed.

1. Introduction represents the weak element of the building envelope, as typical win-
dows have a thermal transmittance (U-value) ranging from 2 W/m2 K to
Buildings globally account for 32% of total global final energy use, 4.5 W/m2 K; even in high-performance windows, the U-value typically
51% of global electricity consumption, and 33% of black carbon ranges from 0.8 W/m2 K to 1.5 W/m2 K [3], a value much higher than
emissions [1]. Moreover, in many developed countries, buildings are the U-value of the opaque portion of the wall.
responsible for consuming more than 40% of the total energy [1]. It is estimated that windows alone contribute from 25% to 40% of
Consequently, in the context of energy saving, the building sector has the heat loss through a typical building assembly [4,5]. With modern
attracted increasing attention globally. buildings having higher window-to-wall ratios, the high U-values of
The increasing demand for energy-saving within the building sector windows exacerbate heat losses and poor building energy perfor-
requires to find new solutions for improving the thermal properties of mances. Accordingly implementing highly insulated windows is critical
the building envelope. Not surprisingly, recent policies and codes often for reducing building energy consumption [6], while guaranteeing
prescribe highly insulated envelope, which depending on the climate, lighting and visual comfort [7], balancing the several requirements of
should have a thermal transmittance of walls and roofs ranging from window systems [8].
0.15 W/m2 K to 0.30 W/m2 K [2]. In this scenario, fenestration often One of the most critical elements in windows is its frame. Over the

E-mail address: uberardi@ryerson.ca (U. Berardi).


1
Address: Ryerson University, 350 Victoria street, Toronto, ON, Canada.

https://doi.org/10.1016/j.apenergy.2020.114776
Received 18 November 2019; Received in revised form 3 February 2020; Accepted 27 February 2020
Available online 19 March 2020
0306-2619/ © 2020 Elsevier Ltd. All rights reserved.
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

Nomenclature IGU Insulated Glazing Unit


NFRC National Fenestration Rating Council
λ thermal conductivity PHI Passive House Institute
CEN European Committee for Standardization PUR Rigid Polyurethane Foam
CFD Computational Fluid Dynamics PVC Polyvinyl chloride
EPDM Ethylene Propylene Diene Terpolymer Uf-value Frame U-value
EPS Expanded Polystyrene Insulation XPS Extruded Polystyrene Insulation
FEM Finite Element Method

last few decades, the design of window frames has been subject to be significantly mitigated by implementing some features within the
significant innovations, ranging from new materials to new designs. cavity itself, such as inserting either thin vinyl webbings perpendicular
Nowadays, wood frames are typically used in residential constructions to the direction of the heat flow or polyurethane (PU) spray foam in-
and are known to have high strength and reasonably good thermal sulation [13]. The authors noted that although adding PU foam in-
performance. However, wood frames are vulnerable to cracking, sulation within the cavity was an effective measure to improve the
warping and moisture ingress, so modern frames have often looked at frame performance, it prevented frames from being recyclable and af-
other materials. Aluminium frames have become common thanks to the fected the weep hole punching operation during the frame fabrication
low cost of the row material, its lightness and strength. Aluminium [13].
window frames dominate the window market in high-rise buildings Fang et al. also examined several thermal performance measures on
where stronger frames are required. However, the aluminium is highly a window frame through Finite Element Modelling (FEM) simulations.
conductive, and it has the poorest thermal performance among com- The authors investigated six frames with different variables: rebate
monly used framing materials. height, frame width, thickness and material, insulation in the frame
Another alternative is represented by polymer frames, generally cavities, and increasing frame design complexity, as shown in Fig. 1.
made from extruded vinyl, polyvinyl chloride (PVC), un-plasticized The study found that increasing the rebate height, from 10 mm to
polyvinyl chloride (uPVC) or glass fibre reinforced polyester (GFRP). 24 mm of a wood frame, which measured 50 × 40 mm with poly-
PVC frames are hollow and contain several subdivided air-filled cav- urethane foam insulation within the cavities, decreased the Uf-value
ities. However, they have limitations when it comes to size and weights from 1.50 W/m2 K to 1.14 W/m2 K. With a similar wood frame, in-
as they are not inherently strong and require steel reinforcement within creasing the width from 40 mm to 60 mm, with a rebate height of
the frame cavity. GFRP has been shown to be eight times stronger than 16 mm, the Uf-value decreased from 1.80 W/m2 K to 1.15 W/m2 K.
PVC [5], to be less prominent to absorb moisture, to corrode or to suffer Decreasing the thickness of an aluminium frame without a thermal
UV degradation. However, the thermal conductivity of GFRP is almost break from 3 mm to 0.5 mm reduced the Uf-value from 2.70 W/m2 K to
double that of wood and PVC. Consequently, a perfect material for 0.49 W/m2 K. The study also explored the addition of polyurethane,
frames does not exist and often the most high-performance frames have urethane, glass wool, and vermiculite flake insulation within the six
explored the option of combined materials, such as vinyl-clad alumi- cavities of a polypropylene frame. Polyurethane foam (λ = 0.024 W/
nium or aluminium-clad wood. m K) filled cavities had the lowest Uf-value of 1.45 W/m2 K, while
The thermal transmittance of window frames depends on many cavity with vermiculite flakes (λ = 0.068 W/m K) had the highest Uf-
factors that influence their performance, including the geometries, value of 2.0 W/m2 K [14].
cavity spacing, and surface emissivity [9,10]. Numerous studies have A study led by Asdrubali et al. focused on the effect of the geometric
focused on determining the impact of individual modifications on and surface emissivity of frame cavities on the overall frame thermal
window frames, but limited studies have combined several modifica- performance [15]. CFD simulations using the ISO 10077-2 standard
tions together [11,12]. In this study, the thermal transmittance of 48 procedures were conducted on a thermally broken aluminium frame.
commercially-available high-performance aluminium, fibreglass, PVC The frame was also simulated with five additional EPDM gaskets, four
and wood-composite window frames are investigated, together with the placed between the glazing and the frame and one placed within the
possibilities for their improvement by inserting aerogel-enhanced pro- central cavity. The study found that the additional gaskets reduced the
ducts. thermal conductivities of the air gaps which in turn decreased the Uf-
The next section presents a literature review of the topic of this value from 2.54 W/m2 K to 2.26 W/m2 K [15]. This model was then
paper. Then, section 3 presents the selected frames together with the validated with a guarded hot box experiment, finding that the gaskets
methodology adopted for the two-dimensional simulations according to reduced the Uf-value from 2.36 W/m2 K to 2.195 W/m2 K. Additional
the ISO 10077-2 standard. Section 4 reports the results, which are then simulations were also conducted modelling the frame cavities with a
discussed in Section 5. Final remarks are reported in Section 6. 0.06 emissivity coating: in this last case, the Uf-value to decrease from
2.54 W/m2 K to 2.27 W/m2 K. Combining the five additional gaskets

2. Literature review

2.1. Frame performance measures

One of the first comprehensive studies about window frames dates y


back to 1992 when Beck and Arasteh explored various commercially-
available vinyl frames. Using an infrared imaging radiometer and 2D H
FEM simulations, they found that frames with the lowest U-value (U
frame-value or Uf-value) increase the conductive heat transfer path
length by minimizing the cross-sectional area in the direction of heat x
transfer [13]. This result was achieved by increasing the height of the
frame profiles, although high frames usually result in large hollow W
cavities, and enhance more convective heat transfer within the frame
cavities. The study by Beck and Arasteh found that this problem could Fig. 1. Cross-section of a frame illustrating definitions.

2
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

with the low emissivity coating further decreased the Uf-value to between the NFRC and the ISO standards. Beyond different boundary
2.07 W/m2 K [15]. conditions (exterior temperature equals to −18 °C in the NFRC versus
Bossche et al. studied several strategies for improving window 0 °C in the ISO 10077), the standards also differ on how the insulated
frames [16]. Numerical analyses were carried out on Belgian double- glazing units (IGUs) are modelled: the NFRC 100 simulates the actual
glazed aluminium, vinyl, and wood frames using FEM simulations. The IGU and the real spacer within the frame, whereas the ISO 10077-2
study found that the most effective measures for the aluminium frame simulates a calibration panel with thermal conductivity of 0.035 W/m K
were the addition of insulation between the glazing and frame, the in lieu of the IGU and the spacer.
division of the thermal break forming cavities of 6 mm deep, and the There are several uncertainties surrounding the accuracy of dif-
addition of insulation (λ = 0.035 W/m K) as a the thermal break, ferent methodologies with respect to determining the thermal trans-
which reduced the Uf-value by 11%, 13% and 16% respectively [16]. mittance of window frames. A study conducted by Gustavsen et al. [24]
The least effective measures were dividing the central gasket into compared Uf-value results between FEM and CFD simulations on a
smaller cavities, extending the thermal break from 34 mm to 55 mm simple square-shaped PVC, and a thermally broken aluminium extru-
and changing the polyamide thermal break conductivity from 0.30 W/ sion with one air cavity. Simulations were modelled on the individual
m K to 0.17 W/m K, which reduced the Uf-value by less than 5%. The vertical and horizontal sections of the extrusion with a separated si-
combination of all previous measures decreased the Uf-value from mulation conducted on a four-sided extrusion. The study found that the
2.78 W/m2 K to 1.47 W/m2 K, a 56% reduction in the thermal trans- FEM results for the vertical sections of the extrusion compared well to
mittance of the frame [16]. The study also analysed a steel-reinforced the CFD results, with a difference ranging between 2% and 5% [24].
vinyl frame and found that deepening the frame from 90 mm to However, a larger discrepancy was found with horizontal sections, as
120 mm and adding insulation in the central cavities reduced the Uf- FEM models suggested 8% to 11% higher results than CFD models.
value from 1.503 W/m2 K to 1.067 W/m2 K. For wood frames, in- Finally, the authors concluded that FEM simulations can be used with
creasing the width of the frame from 68 mm to 108 mm and replacing good accuracy when calculating Uf-value [24].
the solid hardwood profile with a composite, layered profile with A subsequent study by Gustavsen et al. compared the simulation
polyurethane insulation reduced the Uf-value from 1.640 W/m2 K to outputs between FEM and CFD on two PVC and aluminium horizontal
0.584 W/m2 K [16]. frames [25]. The frames were modelled with six configurations each;
Lechowska et al. explored different approaches to improve the two configurations involving CFD and four configurations involving
thermal performance of a window frame without changing the frame FEM. The ISO 15099 procedures were used to model the boundary
material and shape [17]. Using a triple-glazed operable PVC frame, conditions, and for the radiation and natural convention in the air
their study focused on placing polyurethane foam insulation and low- cavities for the FEM simulations. The study found that the FEM results
emissivity coatings within the internal cavities. Five numerical models compared reasonably well with the CFD simulations with a maximum
were simulated according to the ISO 10077-2 [18], and three models difference of 13% between the four frames and six configurations [25].
were tested in a guarded hot box according to the ISO 8990 standard Using a standard modelling configuration of CFD with fluid flow and a
[19]. Initial simulations concluded that filling the frame with two and FEM with a cavity separation yielded a maximum difference of 8.5%.
four insulation insertions reduced the Uf-value of the baseline model The accuracy was found to decrease with high-performing frames with
without insulation (1.098 W/m2 K) up to 1.034 W/m2 K (5% reduction) complex geometries [25].
and 0.998 W/m2 K (9% reduction) respectively [17]. These simulations Gustavsen et al. also compared FEM and CFD thermal transmittance
were then validated through a guarded hot box experiment with a results on five high-performing and complex frames: three wood
maximum difference in test results no greater than 3.3%. A fourth frames, one aluminium frame, and one PVC frame. Numerical simula-
model was simulated with polyurethane foam insulation placed in tions were conducted on the head, sill, and jamb of each frame in ac-
every cavity with the result of lowering the Uf-value to 0.804 W/m2 K cordance with the ISO 10077-2. The study found that all head and sill
(i.e. 27% reduction). A final model simulated the PVC frame with a low FEM results had slightly higher Uf-value compared to CFD, having a
emissivity coating of 0.06 applied to the surfaces within the cavities; difference ranging from 0.1% to 10.4% [26]. However, FEM results for
this improvement reduced the Uf-value to 0.792 W/m2 K [17]. jambs were significantly higher than that of CFD, with a variance be-
Gustavsen et al. conducted a market review of the best-performing tween 7.9% and 23.6%. This discrepancy between the FEM and CFD
window frames commonly found in Europe [20]. Overall, 56 frames results was lower for the frames with higher Uf value values. The au-
with different geometries, materials, and styles were presented. The Uf- thors noted that the large discrepancy for jambs was due to the sim-
value ranged from 0.63 W/m2 K to 1.14 to W/m2 K. Materials used for plified 2D equations for radiation and convection heat transfer in the
the high-performing window frames included wood, wood composites FEM modelling [26]. A guarded hot box apparatus was then used to
with aluminium, PVC and aluminium thermally broken frames. The determine the total window thermal transmittance for all the frames.
most common features inserted within the frame cavity were highly The study found that CFD simulations best aligned with hot box mea-
thermal insulating materials, including rigid polyurethane (PUR), surements, but FEM had conservative values, differing from 5% to 11%
polyethylene (PE) and extruded polystyrene (XPS). to hot-box measurements [26].
Gustaven et al. also reviewed the calculation procedures between
2.2. Heat transfer modelling the ISO 15099 and 10077-2 for determining the frame thermal trans-
mittance: this review found that these standards are really similar and
Several standards define the calculation, experimental and numer- they differ only in the heat transfer calculations through air cavities
ical simulation procedures to determine the U-value of windows and [27]. The authors concluded that while the ISO 15099 is considered to
their components. These include the ISO 15099 Thermal performance of take into account be 3D conduction and radiation, the ISO 10077-2
windows, doors and shading devices – Detailed calculations [21] and the method underestimates radiation and convection and overestimates
ISO 10077‐2 Thermal performance of windows, doors and shutters – Cal- conduction.
culation of thermal transmittance – Part 2: Numerical method for frames Baldinelli and Bianchi analysed the total window thermal trans-
[18]. In North America, other standards include the NFRC 100 Proce- mittance measurement differences between FEMs, CFDs and guarded
dure for Determining Fenestration Product U‐factors [22] and the CSA- hot box experiments on a double-glazed solid wood and thermally-
A440 North American Fenestration Standard / Specification for windows, broken aluminium frame [10]. The study found that the simple solid
doors, and skylights [23]. All these standards are intended for being used wood frame had the lowest discrepancy between the FEMs, CFDs, and
within computer software but they use different assumption in the hot box measurements. Aluminium had the highest discrepancy be-
modelling. For example, the interior film coefficient greatly varies tween all three testing methods, with FEM and CFD calculations

3
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

differing by 10% and CFD and hot box measurements differing by 14% studies regarding glazing systems with silica aerogel, both involving
[10]. A study conducted by RDH Building Engineering compared the laboratory development [39] and in-field experimental analysis
thermal transmittance of several double- and triple-glazed vinyl and [40,41], analyzing the thermal, energy, lighting, and acoustic perfor-
fibreglass frames using both the NFRC 100 and ISO 10077-2 calculation mance of these systems [42,43].
methods [28]. This study found some differences between the two Unfortunately, no study has considered the thermal enhancement of
standards, with NFRC Uf-values being from 5% lower to 24% higher window frames with aerogel granules. This research aims to study the
compared to ISO values [28]. This study concluded that each standard impact of placing granular aerogel within the internal cavities of high-
has strengths and weaknesses, however, the FEM numerical approach is performing window frames.
generally reliable.

2.3. Aerogel 3. Methodology

Aerogel is known to be one of the lightest materials available in the 3.1. Investigated window frames
market as a result of its open-pore microstructure. Pores can occupy
between 85% and 99.9% of total aerogel volume with a resulting bulk In this research, simulations were conducted on a variety of con-
density between 70 kg/m3 and 150 kg/m3. This microstructure gives figurations and framing materials. All the selected frames are manu-
aerogel a high compression strength of 300 kPa and a tensile strength as factured within North America and Europe and are designed for cold
low as 16 kPa, making it a very brittle material. This highly porous and temperate climates. Frames with a thermal transmittance lower
microstructure has a low solid skeleton conductivity and small pore than 1.60 W/m2 K were selected to ensure that the investigated frames
sizes, less than the mean free path of air that reduces heat transfer were already high performing, while the impact of their enhancement
through gaseous convection and conductivity. Aerogels achieve a low with aerogel could be fully observed. All frames were tripled glazed and
thermal conductivity as a result of its microstructure, often described as contained either EPS, XPS or polyurethane foam insulation within the
“beads on a string” or “pearls on a necklace” [29]. Aerogels are con- central air cavities. Figs. 2 and 3 illustrate some of the selected frame
sidered to be non-flammable and non-reactive [30]. The thermal con- profiles.
ductivity around 0.01–0.02 W/m K has suggested the use of aerogels Ten thermally broken aluminium frames were analysed. All frames
within glazing cavities, with a reduction in the heat loss as high as 60% had an insulating rebate bar placed underneath the IGU and used
compared to conventional IGUs [31,32]. polyamide as a thermal break. The thermal transmittance ranges of the
Currently, commercially-available aerogel products have a mean aluminium frames varied from 0.64 W/m2 K to 1.15 W/m2 K.
thermal conductivity around 0.015 W/m K [33], although several stu- All 18 fibreglass frames were tripled glazed and contained either
dies have tried to enhance further the performance of aerogel products. EPS or polyisocyanurate rigid foam insulation within one or multiple
Neugebauer et al. lowered the thermal conductivity of granular aerogel cavities. Frame profiles had an average wall thickness of 2.3 mm with a
through compaction and reached a thermal conductivity of 0.013 W/ minimum glass content of 60%. The thermal transmittance of the fi-
m K when the granules were compressed by 59% [34]. Granule particle breglass frames varied from 0.82 W/m2 K to 1.60 W/m2 K.
sizes also influence the thermal conductivity. Gao et al. found that PVC frames have several chambers that contained different amounts
particles larger than 3 mm had a thermal conductivity of 0.021 W/m K, of EPS insulation within the cavities. A majority of these frames were
whereas particles smaller than 0.5 mm had a 20% lower thermal con- either PHI or PHIUS certified. The simulated thermal transmittance
ductivity [31]. range of the PVC frames varies from 0.71 W/m2 K to 0.88 W/m2 K.
Silica aerogel is naturally hydrophilic as a result of the silanol polar Ten wood composite frames made from spruce and having an ex-
groups that has a high affinity for water. However, aerogel is typically ternal aluminium cladding were selected. The frames were designed
treated to make the pore surface hydrophobic. This reduces the risk of either for triple or quadruple glazed and contained EPS, PUR or XPS
aerogel being damaged by the water surface tension [35,36]. insulation within the frame. All wood frames had a low thermal
In synthesis, aerogel is considered to be one of the highest per- transmittance ranging from 0.60 to 0.93 W/m2 K.
forming insulating materials, as shown by comparing its property with
that of other common materials (Table 1).
The thermal properties of aerogel are known to be stable over a long 3.2. Numerical methodology
period of time. Ihara et al. focused on a study that exposed aerogel to
different moisture fluctuation cycles and found that moisture has an Two-dimensional heat transfer analysis of the frames was performed
overall negligible impact on aerogel granules [35]. using the software THERM version 7.6. THERM is a common software
Several recent studies have investigated the possibility of introdu- used to solve two-dimensional steady-state conduction, convection, and
cing aerogel within the glass cavities and have monitored real perfor- radiation heat transfer [44]. The software THERM was developed by
mance over long periods of time with negligible reductions. For ex- the Lawrence Berkley National Laboratory and adopts a 2D heat
ample, Berardi presented the development of a monolithic aerogel transfer FEM approach [44]. Two-dimensional simulations are con-
glazed window [37] and investigated the benefits of these windows in a sidered an efficient and accurate method to obtain results compared to
building energy retrofit project [38]. Similarly, researchers from the CFD simulations and hot box testing [45]. In fact, a CFD model can take
University of Perugia (Italy) have presented multiple characterization approximately 500 times longer than performing two-dimensional
conduction run in THERM. This approach allows testing several design
Table 1 iterations that may not be noticeable in lengthy and expensive physical
Thermal conductivities of traditional and advanced insulation materials. laboratory experiments. The software generates automatically a
meshing algorithm on the modelled cross-section, using the Quadtree
Material Thermal Conductivity (W/m K)
method, to calculate the thermal transmittance and surface tempera-
Mineral wool 0.034–0.045 tures in each mesh quadrant. An error estimator within the software can
Glass wool 0.031–0.043 determine the percentage error energy norm, related to the gradient of
Expanded polystyrene (EPS) 0.029–0.055 heat flux, and refine the mesh based on the error level that can be
Extruded polystyrene (XPS) 0.029–0.055
controlled by the user. THERM has several advantages as a tool to solve
Polyurethane foam (PUR) 0.020–0.029
Silica aerogel 0.012–0.020 complex heat transfer problems, and in fact, it is the software adopted
in several of the studies cited in Section 2.1.

4
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

aluminium fibreglass PVC wood-composite


AL-01 FG-04 PVC-02 WD-04

AL-06 FG-06 PVC-06 WD-06

AL-10 FG-09 PVC-10 WD-10

Fig. 2. Pictures of some of the frames investigated in this paper. Codes refer to the design as reported in Tables 4–7.

3.3. Material properties and boundary conditions for the other cases. The second case simulated the existing insulation
with EPS (λ = 0.035 W/m K), PUR (λ = 0.024 W/m K), and aerogel
The material properties used in the numerical simulations, in- (λ = 0.012 W/m K). Results served to compare the impact of aerogel
cluding the thermal conductivity and the emissivity, are listed in with other common insulation. A third case modelled the insulation
Table 2. Values were obtained by manufacturer technical specifications with the cavities completely filled by the aerogel. Additional cases were
or according to the predetermined inputs indicated in the ISO 10077-2 modelled for the aluminium and the PVC frames as they contain several
standard [18]. cavities that were filled with the aerogel granules progressively.
Table 3 reports the boundary conditions and the associated surface
resistances used in the simulations as defined by the ISO 10077-2. Fa-
cing surfaces, edges and junctions have reduced radiation and con- 3.5. Model implementation
vection heat transfer which leads to increased surface resistance. These
surfaces have been defined as internal reduced radiation boundaries Frames were modelled following the procedures outlined in the
and are determined based on segment dimensions that are less than routine “Calculating fenestration product performance” available in
30 mm. The cutting plane of the bottom of the frame to the adjacent WINDOW 7 and THERM 6 [46]. According to this methodology, the
wall assembly and the top of the calibration panel were considered glazing portion was substituted with a calibration insulation panel
adiabatic boundaries. material that extends 190 mm above the frame sightline. The calibra-
tion panel eliminates any thermal bridging effects caused by the spacer
and glass. Frame accessories, such as locks, setting blocks, corner keys,
3.4. Frame modification with aerogel non-continuous hardware and weep holes were not modelled.
Some frames were also modelled with an IGU in order to evaluate
The frame specimens and their cross-sectional components were the impact on the glazing part over the total window thermal trans-
modified by placing aerogel within the cavities for each frame type, as mittance. The IGU was comprised of 4 mm thick glass with cavities
shown for example in Fig. 4. The first simulated case modelled the containing a gas fill of 90% argon and 10% air. The space between the
frames as per manufacturer drawings and was used to compare results glass panes varied by 11–16 mm and was determined based on the

5
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

aluminium frames
PVC frames
fibreglass frames
wood-composite frames

Fig. 3. Some frame profiles of aluminium, fibreglass, PVC, and wood-composite frames.

frame dimensions. The centre of glazing U-value ranged from 0.94 to foam insert and bf is the projected length of the frame section. The
1.09 W/m2 K and was calculated using WINDOW [44]. A high-per- thermal transmittance of the overall fenestration product was calcu-
forming warm edge Swissspacer was implemented. The spacer was lated as [47]:
10 mm high and contained a stainless steel backing foil, a primary
polyurethane seal and a silicone secondary seal. The thermal trans- Ugv A gv + Uf Af + l
UT =
mittance of the simulated frames was calculated as [18]: AT (2)

L p2D U 'p b p where UT and UGV are the thermal transmittance of the total window
Uf = and projected vision glass respectively, AT, Agv and Af represent the
bf (1)
total, projected vision glass and frame area respectively, lψ and ψ are
where UF denotes the thermal transmittance of the frame, Lp2D is the the thermal linear transmittance coefficients between the glazing and
thermal coupling coefficient, U′p is the thermal transmittance of the the frame.
foam insert that replaces the glazing component, bp is the length of the

6
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

Table 2 4. Results
Thermal conductivity of materials in simulated frame sections.
Group Material Thermal Conductivity Emissivity This section discusses the simulation results for aluminum, fi-
(W/m K) breglass, PVC and wood-composite frames. While it could not be “evi-
dent” in the original submission, the first step of this study was the
Frame Aluminium (Si alloys) 160 0.3
validation of the adopted numerical approach. Tables 4 (aluminum), 5
Fibreglass (UP-resin)a 0.300 0.9
PVC, rigid 0.170 0.9
(fibreglass), 6 (PVC) and 7 (wood) report in the fifth column (case 1:
Softwood 0.110–0.130 0.9 original frame), the values of the manufacture declared data (“Mfr.
Steel 50 0.8 declared”) and those of the simulated one, as obtained in the present
Thermal break, Elastomeric foam, flexiblea 0.030 0.9 study. These columns show the validation of the adopted numerical
sealants & EPDM 0.250 0.9 approach. In fact, these columns consistently show extremely small
weather Polyamide (nylon)a 0.170–0.250 0.9 differences of the numerical simulated results always below 0.05 W/
stripping PVC, flexible (PVC-P) 40% 0.140 0.9
m K (and often as small as 0.01 W/m K). As such, there was typically
softener
Silicone, pure 0.350 0.9 only a 1% to 3% maximum difference between the manufacture results
a
and the numerical ones. The small difference between the experimental
Insulation Aerogel 0.012 0.9
Expanded Polystyrene 0.035 0.9
results as declared by the manufacture and the numerical results of the
Extruded Polystyrenea 0.029 0.9 FEM modelling done according to the ISO10077, confirms the full va-
Polyisocyanurate 0.024 0.9 lidity of the adopted methodology and the validity of the simulations.
PU (Polyurethane), rigid 0.024 0.9

Glass & Spacer Argonb 0.029 0.9


Calibration Panel 0.035 0.9 4.1. Aluminum frames
Glass 1.000 0.9
Polyurethanea 0.025 0.9 Table 4 shows the changes in thermal transmittance for the alu-
Stainless Steel Barab 0.190 0.9
minum frames with different insulation types and the several iterations
a of aerogel enhancements. Results show that replacing PUR with aerogel
Obtained by manufacturer/passive house report.
b
Equivalent thermal conductivity. can decrease the Uf-value by 4% to 22%. Aerogel is advantageous for
the frames higher than 100 mm, such as AL-04, AL-05, AL-09, and AL-
10. Those frames that showed a slight improvement in their thermal
Table 3
Simulation boundary conditions and associated surface resistance [18]. performance when the aerogel replaced the PUR (AL-01, AL-02 and AL-
07), had the slimmest cavities and contained the least amount of in-
Boundary Temperature (°C) Surface resistance (m2 C/W)
sulation. The frames with cavities completely filled with aerogel re-
Adiabatic – infinity sulted in a decrease of the Uf-value from 0.1% to 13%. The frames
External 0 0.04 larger or higher than 115 mm exhibited the greatest improvement as
Internal 20 0.13 larger heights can accommodate more additional insulation. In the case
Internal reduced radiation 20 0.20 of AL-04, the thermal breaks with hollow cavities were filled with

Case One Case Two Case Three Case Four

Fig. 4. Example of an aluminium (above) and PVC (below) frame case study with the iterations considered to improve its performance by adding aerogel in several
cavities: areas highlighted in teal represent aerogel and grey areas represent unventilated air cavities.

7
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

Table 4
Aluminium frame simulation results.

aerogel, which resulted in an additional 2–11% reduction in Uf-value 4.2. Fibreglass frames
for both fixed and operable frame configurations. Overall, the alu-
minum frames decreased their original Uf-value between 5–35% for Table 5 compares the impact of different types of insulations and
fixed frames, and 18–34% for operable frames. The profiles with the aerogel iterations for the fiberglass frames. Replacing the PUR with
lowest Uf-value among all the implemented cases were AL-09 (0.46 W/ aerogel decreased the Uf-value from 2% to 9% for both fixed and op-
m2 K) and AL-06 (0.52 W/m2 K). erable frames. Similarly to aluminum frames, the profiles that improved

8
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

Table 5
Fibreglass frame simulation results.

(continued on next page)

9
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

Table 5 (continued)

the most were those with the largest frame and cavity heights, i.e. the by vinyl webbings. Replacing the PUR with aerogel reduced the U-value
frames FG-03, FG-08 and FG-15. Although adding insulation to profiles by 5–8% for operable frames. Those frames having a high of 72 mm had
shorter than 50 mm was advantageous, the improvement with aerogel a comparable U-value to the frames higher than 110 mm, despite these
compared to common types of insulation was negligible for these last frames having more insulation. When all the cavities containing
smaller frames. The FG-03 frame shows that completely filling the insulation were completely filled with aerogel, the Uf-value reduced by
cavity can reduce the Uf-value between 2% and 11%. The greatest 9–20% for fixed frames and by 12–18% for operable frames. The frames
improvements were obtained in the tallest (83 mm) and slimmest that benefitted the most from this change were those which had irre-
profiles (40 mm), e.g. FG-08 and FG-14 respectively. The profiles with gular insulation geometries, such as PVC-01 and PVC-06. Completely
the lowest Uf-value for fixed frames were FG-08 (0.66 W/m2 K) and FG- filling each cavity within the frame, excluding the rebate cavity, was
06 (0.80 W/m2 K). not as effective, as it only decreased the Uf-value by 2–6% from cases of
filling some cavities only. However, introducing the aerogel in all the
4.3. PVC frames cavities can decrease the frame Uf-value between 23% and 32%. The
profiles that demonstrated the lowest Uf-value were PVC-01 (0.51 W/
Table 6 shows the changes in thermal transmittance for PVC frames. m2 K) for fixed frames and PVC-04 (0.60 W/m2 K) for operable ones.
In these cases, the insulation does not decrease the Uf-value compared
to aluminum and fibreglass frames, due to the cavities being separated

10
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

Table 6
PVC frame simulation results.

4.4. Wood frames PU to aerogel were WD-09 and WD-10, which had continuous insula-
tion placed on the exterior side of the frame. Contrarily to aluminum
Several simulations were then conducted on distinctive wood- and fibreglass frames, in wood frames, the height did not seem to affect
composite frames that contained insulation layers as thermal breaks significantly the thermal transmittance. This can be observed with
between layers of solid wood. Table 7 illustrates that replacing the PUR frame WD-05, which has the slimmest profile out of all wood frames
with aerogel can decrease the Uf-value for both fixed and operable (53 mm) and it has a comparable Uf-value to those frames as tall as
frame by 4–29%. The frames that exhibited the greatest benefits from 137 mm. The profiles that showed the lowest Uf-value after their

11
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

Table 7
Wood frame simulation results.

12
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

Table 8
Total window U-value simulation results.

13
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

enhancement with aerogel were WD-09 (0.44 W/m2 K) for fixed frames than that of aerogel (0.012 W/m K). On average, the aerogel resulted in
and WD-10 (0.51 W/m2 K) for the operable frames. an 8% reduction in U-value, compared to the PUR. Similar observations
were reported by Fang et al. [14], where the study focused on the
4.5. Impact of the aerogel on total window thermal transmittance impact of different insulation materials on a polypropylene frame.
Figs. 5–8 show the improvement from the existing Uf-value with
Table 8 displays the results of some selected frames that were different aerogel-enhancements for aluminium, fibreglass, PVC and
modelled with a triple-glazed IGU in order to evaluate the effect of the wood-composite frames. For all the frames, replacing the existing in-
different aerogel iterations on the total window performance. All the sulation with aerogel (case 2) was particularly beneficial for those
assessments were done assuming a window size of 1.48 m by 1.23 m. profiles that contained the most insulation. Similarly, completely filling
Simulations for the aluminum frame AL-09 and AL-10 show that re- the primary cavities with aerogel (case 3) was advantageous for those
placing the existing insulation with aerogel had the greatest benefit in frames with large cavities. As noted previously, the frame height was
reducing the U-value by 7%. The proceeding cases marginally reduced highly correlated to the highest decreases of the U-value. This is ob-
the U-value by 3% despite the Uf-value of the frame decreasing by over served with the frames AL-07, AL-08, and AL-09. This demonstrates the
28% and the frame area comprising 48% of the total window area. The importance of minimizing the air space between the insulation and the
linear thermal transmittance between the glass and frame remained the frame as filling the small tolerance of 2 mm with aerogel can further
same for all the iterations. Fibreglass frame FG-05 exhibited marginal reduce their Uf-value by an average of 4%, as it was observed for fi-
improvements to the overall window performance among all the breglass and PVC frames.
iterations. Despite a significant decrease in the U-value by implementing all
Replacing the existing insulation with the aerogel decreased the U- cases, the thermal coupling coefficient between the calibration panel
value by 1%. Completely filling the primary cavities further reduced the and the frame slightly changes for the different frames. Further analysis
U-value by an additional 1%. These iterations were more impactful on of temperature contour diagrams illustrates a minor variance for all
FG-08 when replacing the existing insulation and completely filling the iterations, as shown in Fig. 9. This figure illustrates the contours of the
cavity reduced the U-value by 3% and 6% respectively. Despite FG-05 heat flow superimposed on the frame profile. The figure demonstrates
having a 16% higher Uf-value than FG-08, FG-05 had a 15% lower total that the only significant change was the location of the highest recorded
window thermal transmittance. Simulations with the frame PVC-03 surface temperature. This suggests that as the frame lowers its U-value
found that replacing the existing insulation with aerogel only reduced through enhancements with aerogel, the rebate cavity and the glazing
the U-value by 2%. Filling the primary cavities with aerogel and placing stop become the weak points of the frame. Similar correlations were
the aerogel at the perimeter can further reduce the U-value by 1% and found with the PVC frames simulated by Lechowska et al. [17] with
2% respectively. Overall, the total window thermal transmittance could insulation in the rebate cavity.
be reduced from 0.87 W/m2 K to 0.829 W/m2 K by implementing all This suggests that as the frame lowers its U-value through en-
iterations. Amongst all the samples, the frame WD-05 with aerogel had hancements with aerogel, the rebate cavity and the glazing stop become
the least impact on the total window performance. Simulations with the the weak points of the frame. Similar correlations were found with the
IGU found that replacing the existing insulation with aerogel only re- PVC frames simulated by Lechowska et al. [17] with insulation in the
duced the U-value by 1.1%. However, replacing traditional insulation rebate cavity.
with aerogel for WD-09 decreased the U-value by almost 6%. Together with the numerical simulated approach, the authors also
conducted some experimental studies to provide insight into the feasi-
bility of integrating aerogel granules within the selected frames. The
5. Discussion
“fluid behaviour” of the aerogel granules was beneficial for filling the
small cavities (as shown in Fig. 10), that otherwise could be difficult to
5.1. Aerogel modified frame performance
achieve with traditional insulation materials. The assembling of the
frame was done using two fabrication methods: making a hole in an
The first step for improving all frames consisted of replacing existing
already assembled frame, already made into rectangular window shape,
insulation with EPS, PUR, and aerogel. For all the framing materials, a
or individually filling frame-bars and then joining them together. The
linear relationship between the different types of insulation and their
fabrication showed that aerogel granules can be easily recovered and
impact on the U-value was found, despite the thermal conductivity of
reused once the frames given the loose nature of the granules. In fact,
EPS (0.035 W/m K) and PUR (0.024 W/m K) being significantly higher

Fixed Frames Operable Frames


Fig. 5. U-value difference (in percentage reduction) compared to case 1 (the existing frame) for the several aluminium frames.

14
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

Fixed Frames Operable Frames


Fig. 6. U-value difference (in percentage reduction) compared to case 1 (the existing frame) for the several fiberglass frames.

Fixed Frames Operable Frames


Fig. 7. U-value difference (in percentage reduction) compared to case 1 (the existing frame) for the several PVC frames.

Fixed Frames Operable Frames


Fig. 8. U-value difference (in percentage reduction) compared to case 1 (the existing frame) for the several wood-composite fiberglass frames.

15
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

Case One Case Two Case Three Case Four


Fig. 9. Temperature contours of aluminium frame FG-05 with all iterations of aerogel-enhancements.

Fig. 10. Preparation of aerogel-modified frames: aerogel inserted into frames already made into rectangular window shape (left), or filled individually in frame-bars
that are then joined together (right).

upon opening the frames, the aerogel was easily removed from the 5.3. Frame optimization
frame. Very little to no aerogel was left stick on the surfaces and it was
easily removed by washing the frame, making this completely recycl- The frames included in this study were considered to be thermally
able. optimized, in terms of cavity depths, low emissivity coatings, thermal
breaks, and insulated cavities, while providing sufficient structural
strength, rigidity, and serviceability. However, applying aerogel within
5.2. Frame effect on overall window performance the frame cavities can further optimize the thermal performance of the
frame. Previous analysis showed that further optimization is required
In most cases, the aerogel-enhancements provided a reduction in the near the edge of the glass to decrease the thermal coupling coefficient.
overall U-value window ranging from 1% to 7%. For aluminium, fi- Gustavsen et al. recommend that aluminium frames should have a
breglass and PVC frame specimens, the overall improvements of the thermal break with a conductivity of 0.02 W/m K and an emissivity for
aerogel enhancement and its impact on the total window plateau after hollow frame surfaces of 0.05 [5]. Implementing both measures with
replacing traditional insulation with aerogel. This could be explained as aerogel insulated cavities could further reduce the Uf-value of AL-09 by
the linear thermal transmittance between the frame and IGU slightly 15% and achieve U-values as low as 0.41 W/m2 K. Other strategies
changes between all iterations. could include developing rebate insulation materials with thermal
Frames without insulation near the rebate cavity had positive linear conductivities lower than 0.038 W/m K. However, materials with such
thermal transmittance, indicating that a thermal bridge was occurring low thermal conductivities with sufficient strength are hard to find.
at the edge of the glass. This value is expected to increase for the IGUs Bossche et al. found that increasing the glazing thickness could also
with lower U-values. Accordingly, the full benefits of high-performing reduce the thermal transmittance of an aluminium frame by increasing
frames with aerogel can only be realized once the spacer has a U-value the heat transfer path [16]. This means that frames with quadruple
equal to or less than that of the centre of glass. Completely filling the glazing are effective as the aluminium has the strength to support the
frame with aerogel may not be optimal as this insulation should be additional weight.
placed at strategic points such as glazing stops and thermal breaks. For fibreglass frames, strategies for reducing the heat flux through
the glazing stop and rebate cavity were examined. As such, the best

16
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

performing modified aerogel frame (i.e. FG-08) was redesigned to [3] ASHRAE. Handbook – fundamentals. Atlanta: American Society of Heating,
emulate the glazing stop features of FG-05 while maintaining the Refrigerating and Air- Conditioning Engineers Inc.; 2017.
[4] Apte J, Arasteh D. Window-related energy consumption in the US residential and
original geometrical configuration. The design included glazing stops commercial building stock. Report LBNL-60146; 2006. p. 1–38.
that could accommodate aerogel insulation and different variations of [5] Gustavsen A, Grynninga S, Arasteh D, Jelle BP, Goudey H. Key elements of and
rebate depth and widths. Temperature contours of each variation material performance targets for highly insulating window frames. Energy Build
2011;43:2583–94.
suggested that increasing the rebate depth was the most effective [6] Acosta I, Campano MÁ, Molina JF. Window design in architecture: Analysis of
strategy, achieving U-values as low as 0.55 W/m2 K. A similar ap- energy savings for lighting and visual comfort in residential spaces. Appl Energy
proach was taken with frame PVC-03 which proved to be less effective 2016;168:493–506.
[7] Mangkuto RA, Rohmah M, Asri AD. Design optimisation for window size, orienta-
due to the vinyl webbings transferring heat throughout the frame. tion, and wall reflectance with regard to various daylight metrics and lighting en-
Similar to aluminium frames, Gustavsen et al. recommend a low ergy demand: A case study of buildings in the tropics. Appl Energy 2016;164:211–9.
emissivity for hollow frame surfaces of 0.05 [5]. By placing this [8] Sun Y, Wu Y, Wilson R. A review of thermal and optical characterisation of complex
window systems and their building performance prediction. Appl Energy
coating in the PVC-03 frame rebate cavity, the Uf-value could further
2018;222:729–47.
decrease by 11%, proving to be more beneficial than increasing the [9] Michaux G, Greffet R, Salagnac P, Ridoret JB. Modelling of an airflow window and
frame width and height. From studying the various wood composite numerical investigation of its thermal performances by comparison to conventional
frames, placing insulation on the exterior side of the frame proves to double and triple-glazed windows. Appl Energy 2019;242:27–45.
[10] Baldinelli G, Bianchi F. Windows thermal resistance: Infrared thermography aided
be the most effective configuration in retarding heat flow. Frames can comparative analysis among finite volumes simulations and experimental methods.
be further optimized by having a hybrid system of exterior insulation Appl Energy 2014;136:250–8.
and replacing a rectangular part of the central solid wood with [11] Aburas M, Soebarto V, Williamson T, Liang R, Wu Y. Thermochromic smart window
technologies for building application: A review. Appl Energy 2019;255:113522.
aerogel, similarly to WD-07. [12] Ibrahim M, Biwole PH, Wurtz E, Achard P. Limiting windows offset thermal bridge
losses using a new insulating coating. Appl Energy 2014;123:220–31.
6. Conclusions [13] Beck FA, Arasteh D. Improving The thermal performance of vinyl-framed windows.
Therm perform exter envel build V conf. 1992.
[14] Fang Y, Eames PC, Hyde TJ, Norton B. Complex multimaterial insulating frames for
A comparative evaluation of different insulation materials and windows with evacuated glazing. Sol Energy 2005;79:245–61.
several iterations of aerogel placement was conducted for 48 high- [15] Asdrubali F, Baldinelli G, Bianchi F. Influence of cavities geometric and emissivity
properties on the overall thermal performance of aluminum frames for windows.
performance commercially-available North American and European Energy Build 2013;60:298–309.
aluminium, fibreglass, PVC and wood-composite window frames. The [16] Van Den Bossche N, Buffel L, Janssens A. Thermal optimization of window frames.
study found that replacing the existing insulation with aerogel can re- Energy Procedia 2015;78:2500–5.
[17] Lechowska AA, Schnotale JA, Baldinelli G. Window frame thermal transmittance
duce the Uf-value for aluminium, fibreglass, PVC and wood-composite
improvements without frame geometry variations: An experimentally validated
frames by a maximum of 22%, 17%, 12% and 29% respectively. On CFD analysis. Energy Build 2017;145:188–99.
average, this represented an 8% thermal transmittance improvement [18] ISO 10077-2. Thermal performance of windows, doors and shutters—Calculation of
compared to traditional insulation materials such as rigid polyurethane thermal transmittance—Part 2: Numerical method for frames. Geneva:
International Organization for Standardization; 2017.
foam. Additional iterations of aerogel were performed involving com- [19] ISO 8990. Thermal insulation—Determination of steady-state thermal transmission
pletely filling each cavity and being inserted in small, irregular voids properties—Calibrated and guarded hot box. Geneva: International Organization
such as thermal breaks which may not be feasible with rigid insulation for Standardization; 1994.
[20] Gustavsen A, Jelle BP, Arasteh D, Kohler C. State-of-the-art highly insulating
due to tolerances required. This resulted in significant enhancements of window frames – research and market review. Oslo: State-of-the-Art; 2007.
frames which could further reduce the Uf-value by 13%. Amongst all [21] ISO 15099. Thermal performance of windows, doors and shading devices—Detailed
cases studied, aluminium frames resulted in the greatest improvement, calculations. Geneva: International Organization for Standardization; 2003.
[22] National Fenetration Rating Ccouncil (NFRC) 100. Procedure for determining fe-
with frames achieving Uf-values as low as 0.46 W/m2 K. Such measures nestration product U‐factors; 2014.
are typically difficult to achieve without increasing the frame width and [23] CSA-A440. North American Fenestration Standard/Specification for windows,
heights, adding more cavity divisions or changing the frame emissivity. doors, and skylights; 2019.
[24] Gustavsen A, Griffith BT, Arasteh D. Natural convection effects in three-dimensional
window. ASHRAE Trans 2001:1–11.
Declaration of Competing Interest [25] Gustavsen A, Arasteh D, Kohler C, Dalehaug A. Two-dimensional CFD and con-
duction simulations of heat transfer. ASHRAE Trans 2007;113:165–75.
[26] Gustavsen A, Talev G, Arasteh D, Goudey H, Kohler C, Uvslokk S, Jelle BP.
The authors declare that they have no known competing financial Experimental and numerical examination of the thermal transmittance of high
interests or personal relationships that could have appeared to influ- performance window frames; 2010.
ence the work reported in this paper. [27] Gustavsen A, Arasteh D, Curcija C, Kohler C. Developing low-conductance window
frames: capabilities and limitations of current window heat transfer design tools.
Build Phys 2008;32(2):131–53.
Acknowledgments [28] RDH Building Engineering Ltd. International Window Standards; 2014; A13. http://
www.hpo.bc.ca/files/download/Report/International-Window-Standards.pdf.
[29] Baetens R, Petter B, Gustavsen A. Aerogel insulation for building applications: A
The authors are thankful for the financial support of the Natural state-of-the-art review. Energy Build 2011;43:761–9.
Sciences and Engineering Research Council of Canada (NSERC) for the [30] Cuce E, Cuce PM, Wood CJ, Riffat SB. Toward aerogel based thermal super-
financial support through the Discovery Grant) and the Ontario insulation in buildings: A comprehensive review. Renew Sustain Energy Rev
2014;34:273–99.
Ministry of Research Innovation and Science (MRIS) for the ERA pro- [31] Gao T, Petter B, Ihara T, Gustavsen A. Insulating glazing units with silica aerogel
gram. granules : The impact of particle size. Appl Energy 2014;128:27–34.
[32] Berardi U. Aerogel-enhanced insulation for building applications. Nanotechnol eco-
efficient constr. Woodhead Publishing; 2019. p. 396–416.
Appendix A. Supplementary material [33] Berardi U, Nosrati R. Long-term behaviour of aerogel-enhanced insulating materials
under different aging laboratory conditions. Energy 2018;147:1188–202.
Supplementary data to this article can be found online at https:// [34] Neugebauer A, Chen K, Tang A, Allgeier A, Glicksman LR, Gibson LJ. Thermal
conductivity and characterization of compacted, granular silica aerogel. Energy
doi.org/10.1016/j.apenergy.2020.114776.
Build 2014;79:47–57.
[35] Ihara T, Jelle BP, Gao T, Gustavsen A. Aerogel granule aging driven by moisture and
References solar radiation. Energy Build 2015;103:238–48.
[36] Dorcheh SA, Abbasi MH. Silica aerogel; synthesis, properties and characterization. J
Mater Process Technol 2008;199:10–26.
[1] Berardi U. A cross country comparison of building energy consumption and their [37] Berardi U. The development of a monolithic aerogel glazed window for an energy
trends. Resour Conserv Recycl 2017;123:230–41. retrofitting project. Appl Energy 2015;154:603–15.
[2] Berardi U. Building energy consumption in US, EU, and BRIC countriesuilding en- [38] Berardi U. Aerogel-enhanced solutions for building energy retrofits: insights from a
ergy consumption in US, EU, and BRIC countries. Procedia Eng 2015;118:128–36. case study. Energy Build 2018;159:370–81.

17
J. Paulos and U. Berardi Applied Energy 266 (2020) 114776

[39] Zinzi M, Rossi G, Anderson AM, Carroll MK, Moretti E, Buratti C. Optical and visual Energy Build 2019;198:542–52.
experimental characterization of a glazing system with monolithic silica aerogel. [44] Mitchell R, Kohler C, Zhu L, Arasteh D, Simon V, John C, Charlie H. THERM 6.3/
Sol Energy 2019;183:30–9. WINDOW 6.3 NFRC simulation manual; 2016.
[40] Buratti C, Moretti E. Experimental performance evaluation of aerogel glazing sys- [45] Arasteh D, Berkeley L, Finlayson E. State-of-the-art software for window energy-
tems. Appl Energy 2012;97:430–7. efficiency rating and labeling properties for evaluating window performance.
[41] Buratti C, Moretti E. Glazing systems with silica aerogel for energy savings in ACEEE 1998 summer study energy effic build, Pacific Grove. 1998. p. 1–5.
buildings. Appl Energy 2012;98:396–403. [46] LBNL. Calculating fenestration product performance in Window 6 And Therm 6
[42] Cotana F, Pisello AL, Moretti E, Buratti C. Multipurpose characterization of glazing According to EN 673 and EN 10077; 2009. p. 1–14.
systems with silica aerogel: In-field experimental analysis of thermal-energy, [47] ISO 10077-1. Thermal performance of windows, doors and shutters — Calculation
lighting and acoustic performance. Build Environ 2014;81:92–102. of thermal transmittance – Part 1: General. Geneva: International Organization for
[43] Berardi U, Zaidi M. Characterization of commercial aerogel-enhanced blankets Standardization; 2017.
obtained with supercritical drying and of a new ambient pressure drying blanket.

18

You might also like