IOP Conference Series: Materials Science and Engineering

PAPER • OPEN ACCESS

Flexural strength improvement for structural glass: a numerical study

To cite this article: R Hin et al 2020 IOP Conf. Ser.: Mater. Sci. Eng. 849 012083

View the article online for updates and enhancements.

This content was downloaded from IP address 193.187.112.116 on 30/05/2020 at 18:23

ICONBUILD & RCCE 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 849 (2020) 012083 doi:10.1088/1757-899X/849/1/012083

Flexural strength improvement for structural glass: a numerical

study

R Hin1,4 , K Cheng1, V Han1, F Bernard2, C Seang1, V Keryvin3, J-C Sanglebœuf4

1
Institute of Technology of Cambodia, Russian Federation Blvd., P.O. Box 86, Phnom Penh, Cambodia.
2
Institut National des Sciences Appliquées de Rennes, 20 avenue des Buttes de Coësmes 35708 Rennes France
3 Institut de Recherche Dupuy de Lôme, University of South Brittany, UMR CNRS 6027, F-56321 Lorient,

France
4 Institute of Physics of Rennes, University of Rennes 1, UMR CNRS 6251, F-35042 Rennes, France

Abstract. Glass is generally known as a fragile material. It is sensible to the cracks created from
manufacturing or contact damage. The strength of a perfect glass without crack could reach 10 GPa.
By mean of strengthening such as thermal tempering, glass can be safely use for building as
architectural elements and very limited to the structural elements. The authors have been developing
glass strengthening methods and structural design for large scale glass beam. Some influencing
factors are considered: material, premature crack effect, geometry of sample and bolt. The
mechanical behaviour of glass is modelled as elastic-plastic material, which show significant results
in glass-bolt contact problem. The crack length, size and position provide information of a critical
angle that allow to govern the crack effect in the beam connection.

1. Introduction
In construction building, the usage of glass material increases remarkably for architectural elements and
also structural elements. We can see some innovative glass structures such as the Apple Glass Cube canopy
in New York [1], the renovated pavilion of the Eiffel Tower’s first floor [2], etc. To be a primary structural
element, glass need to sustain safely with high load and the glass connections must be able to transfer the
load to other elements of building. In the flexural strength of a glass beam, the connection strength is
predominant over the bearing capacity of the beam itself. According to the previous studies [3], the load-
carrying of glass connection is still limited because of the local failure at the contact areas. Fractography
analyses by using SEM observation shown that the interaction between the glass and metal assembly excites
a flaw on the lateral surface of the glass’s hole. Furthermore, the distribution of principle stresses is also an
important factor. A study on the geometrical optimization of the bolt and hole shape shows that, the conical
bolt with chamfered glass hole presents the maximum stresses 1.5 times higher than the stiff cylindrical bolt
with cylindrical glass hole [4]. Some experiments on a monolithic annealed glass shown in figure 1 and
figure 2 confirmed that the origin of failure is initiated from where the principle stresses reach the maximum
value, that is at the angle of 90° from the contact loading direction. Underneath the glass-bolt contact, a high
compressive stress field is generated where the favourable densification phenomena could occur [5]. On the
other hand, an observation on the hole’s surface shows that the drilling process induced some crack and
important roughness on the lateral face of the hole. The machining with cemented carbide or diamond bits
causes chatter cracks similar to a scratch event on a brittle material [6]. The crack diameter is in a range of

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
ICONBUILD & RCCE 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 849 (2020) 012083 doi:10.1088/1757-899X/849/1/012083

300-400µm [7]. This machining default is the main factor controlling the strength of glass-bolt connection.
The polishing [8] and ion exchange [9] processes could decrease the crack size and thus improve the
connection strength. The authors aim to study on the effects of connection conditions on the bearing
capacity of the glass beam.

Figure 1. Glass drilling process.

Figure 2. Glass-Bolt connection test: (a) testing Figure 3. Glass hole strengthening
assembly, (b) tested sample, (c) failure observation. by chemical tempering.

2. Methodology
The current studies are conducted to understand the behaviour of the glass-bolt connection and a way to
improve its capacity. Some influencing factors are considered: material, premature crack effect, geometry
of sample and bolt. Base on the 2-dimensional finite element analysis, the results of simulations such as the
intermediate (middle) principal stress, are compared for a given loading.

2.1. Material
The common glass used in building is soda-lime-silica glass. Its macroscopic mechanical behaviour is
known as isotropic elastic linear with Young modulus of 72 GPa and Poisson ratio of 0.21 [10]. The glass’s
strength is not intrinsic property because of the surface’s flaws from which the glass failure is almost
unpredictable. Though, a microscopic observation shows some permanent deformation without failure on a
scratched glass’s surface. There are two mechanisms involving in this deformation: densification and shear
flow. According to Rouxel et al. [11], the initial densification threshold pressure of the soda-lime-silica
glass is 3 GPa and the saturation level of densification is 6.3%. The shear flow threshold was identified by
comparing the numerical simulations with a constitutive model of Keryvin et al. [12] and the experimental
load-displacement curves and imprints of nano-indentations and nano-scratches. The equivalent yield
strength is found to 3.81 GPa [13]. Due to the difficulty of numerical solution for a fully modelled behaviour
of glass and small amount of the densification effect comparing to the global mechanical responses, the
authors simplify the mechanical behaviour of the glass to be elastic linear and perfectly plastic with the yield
strength of 3.81 GPa. The contact behaviour between glass and metal bolt is a hard contact with friction
coefficient of 0.2.

2
ICONBUILD & RCCE 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 849 (2020) 012083 doi:10.1088/1757-899X/849/1/012083

2.2. Glass strengthening effect

Thermal and chemical tempering strengthens the glass by introducing the residual stresses in the glass
volume. A chemical tempering by ion exchange modifies the composition of the glass from the surface to a
deep less than 100µm ([14], [15]). This gradient of properties is assumed to have less effect on the study.
The only one effect of the strengthening is the residual stresses. The surface’s compressive stresses induced
by thermal and chemical strengthening are respectively 100 MPa and 250 MPa with the respective depths
of one-fifth of the glass thickness and 90µm [16].

2.3. Geometry and numerical modelling

The glass specimen’s size is 150mm x 300mm x 12mm with two holes of diameter of 12mm drilled along
the centre axis and the diameter of bolt is 10 mm. The crack’s opening is fixed a = 0.050 mm and the its
depth, b = [0.000 : 0.750] mm, is a parameter for glass strength observations. Another parameter from the
premature crack is its position comparing to the loading direction, that is studied between 0° to 90°. Since
the geometry and loading are symmetric for the axes perpendicular to the loading direction, the sample and
bolt are modelled as a perforated square contacted to a circle. The meshes are gradually changed size from
0.170 mm to 6.00 mm with a 4-node bilinear plane stress quadrilateral, reduced integration, hourglass
control, CPS4R mesh type.

3. Results and discussion

3.1. Mechanical properties of glass
The figure 4 shows the relationship between loading value and the corresponding maximum Mises stress in
the sample. The properties of glass are discussed: purely elastic and elastic perfectly plastic. Prior to the
shear threshold (3.810 GPa), the Mises stress-force responses of both cases are identical. However, from
that point forward, it is obvious that the very small amount of plastic zone (figure 5.a.) need to be taken into
account in the glass material modelling. In addition, we observed that the middle principal stresses illustrate
the tensile stress field that would initiate the failure more clearer than the Mises stress (figure 5.b. and figure
6).

Figure 4. Comparison of Mises stress - Force responses of

different behaviours of the glass sample. Loading by the imposed
displacement on the perimeter of bolt.

3
ICONBUILD & RCCE 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 849 (2020) 012083 doi:10.1088/1757-899X/849/1/012083

(a) Mises stresses, the maximum (b) Intermediate principal stress, the
value is 3810 MPa. maximum value is 253 MPa.

Figure 5. Stresses distributions in the glass for elastic-plastic behaviour with yield
strength of 3810 MPa. The images in (a) and (b) are captured at the same time
increment. The positive value of Mid. Principal stresses could be illustrated more
clearer by graded more colours.

3.2. Effect of the premature crack

The parametric studies of the position and the depth of premature cracking along the perimeter of the glass
hole are presented in figure 6 and figure 7. At a given load, the maximum value of the middle stresses
presented in the glass increases importantly with the greater depth of the premature crack opening. On the
other hand, the position of the crack comparing to the loading direction show two groups of tendency. The
presence of a single crack at a vicinity of the contact area at an angle less than 30° has less effects on the
glass failure. The region between 15° to 30° is where there are transitions of tensile and compressive zone
due to the stress fields of the Hertzian contact and crack tip. The value of 30° is the critical angle which
depends on the bolt-hole diameter ratio and the crack size and depth.

Figure 6. Intermediate (Mid.) principal stress distribution at load = 6kN on samples with
different crack depth b. The crack openings are equal: a = 0.05 mm.

4
ICONBUILD & RCCE 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 849 (2020) 012083 doi:10.1088/1757-899X/849/1/012083

800 700
10kN
Maximum Mid. Principal Stress [N/mm2]

Maximum Mid. Principal Stress [N/mm2]

10kN
700 6kN 600
6kN
600 2kN 2kN
500
500
400
400
300
300
200
200
100 100

0 0
0 15 30 45 60 75 90 0.000 0.250 0.500 0.750
Crack position [°] Crack depth [mm]

Figure 7. Observation on the maximum value of intermediate (Mid.) principal stresses for
samples with different factors: (left) crack position of a 0.250 mm depth comparing to the
loading direction, (right) cracks depth at 90° crack position.

4. Summary and conclusion

The studies have shown some important results for understanding and improving the flexural behaviour of
structural glass. It is very important to model the plastic behaviour for glass material in the glass-bolt contact
problem. In order to observe or compare the stress field in glass material that present very limited plastic
zone, the intermediate principle stress is a better criterion comparing to the Mises stress. On the other hand,
the crack length is predominant in failure mechanisms, which could be minimize thanks to the compressive
residual stresses, e.g. in a tempered glass. The surface compressive stresses prevent crack’s growth unless
the external tensile stresses overpass their values. However, it is a challenge to model the effect of tempering
in glass. The profiles of the residual stresses at the edge of a tempered glass, especially around a hole, are
more appropriate with the 3-D modelling than the 2-D one. Another method to improve the glass connection
strength is to govern the contact surface between the bolt and the glass’s hole. The critical angle of 30° could
be modified since the Hertzian contact provides stresses distribution depending on the radii of the both
surfaces. It could be also supressed by introducing multiple contact surfaces along the hole’s perimeter.

Acknowledgments
The authors gratefully acknowledge financial support from the Royal Government of Cambodia.

References
[1] O’Callaghan J, Bostick C, 2012. The Apple Glass Cube: Version 2.0. Challenging Glass 3 –
Conference on Architectural and Structural Application of Glass.
[2] Leduc N, Raynaud J, Baldassini N, 2012. Project for the Eiffel Tower: Constructive Geometry.
Challenging Glass 3 – Conference on Architectural and Structural Application of Glass.
[3] Bernard F, Daudeville L, 2009. Point fixings in annealed and tempered glass structures: Modeling
and optimization of bolted connections. Eng. Struct. 31 946–55.
[4] Cheng K, Hin R, Han V, Bernard F, 2019. Development of Connection of Glass Beam: A Numerical
Study. Tech. Sci. Res. J. 6 (preprint-0609).
[5] Keryvin V, Charleux L, Hin R, Guin JP, Sanglebœuf JC 2017. Mechanical behaviour of fully
densified silica glass under Vickers indentation. Acta Mat. 129 492–9.

5
ICONBUILD & RCCE 2019 IOP Publishing
IOP Conf. Series: Materials Science and Engineering 849 (2020) 012083 doi:10.1088/1757-899X/849/1/012083

[6] Graham J, 1972. Damage induced by a sliding diamond – an approach to hard rock drilling. Rock
mechanics 4(4) 192-202.
[7] Kim HH, Chung S, Kim SC, Jee WH, Chung SC, 2006. Condition monitoring of micro-drilling
processes on glass by using machine vision. Proceedings of the ASPE Challenges at the Intersection
of Precision Engineering and Vacuum Technology, Pittsburgh, PA, USA, 1-2.
[8] Datsiou KC, Hall D, Overend M, 2018. Repair of soda-lime-silica glass. Glass Struct. Eng. 3 125-37.
[9] Nordberg ME, Mochel EL, Garfinkel HM, Olcott JS, 1964. Strengthening by ion exchange. J. Am.
Ceram. Soc. 47(5) 215-9.
[10] Yoshida S, Sanglebœuf JC, Rouxel T, 2005. Quantitative evaluation of indentation-induced
densification in glass. J. Mat. Res., 20(12), 3404-12.
[11] Rouxel T, Ji H, Keryvin V, Hammouda T, Yoshida S, 2008. Poisson’s Ratio and the Glass Network
Topology-Relevance to High Pressure Densification and Indentation Behavior. Adv. Mat. Res., 39
137-46.
[12] Keryvin V, Meng JX, Gicquel S, Guin JP, Charleux L, Sanglebœuf JC, Pilvin P, Rouxel P, Le Quilliec
G, 2014. Constitutive modeling of the densification process in silica glass under hydrostatic
compression. Acta Mat., 62(1) 250-7.
[13] Hin R, 2017. Comportement à l’indentation et à la rayure de verres métalliques et silicatés.
Matériaux. Université Rennes 1, NNT : 2017REN1S141ff. tel-01818097
[14] kistler SS, 1962, Stresses in glass produced by nonuniform exchange of monovalent ions. J. Am.
Ceram. Soc., 45(2) 59–68.
[15] Abrisa Technologies. 2011. Glass strengthening. Technical Report 877, Abrisa Technologies, Santa
Paula, CA 93060.
[16] Gy R, 2008, Ion exchange for glass strengthening. Mat. Sci. Eng. B: Solid-State Mat. for Adv. Tech.,
149(2) 159–165.
[17] Nielson JH, 2009. Tempered glass: bolted connections and related
Problems. PhD Thesis. Danmarks Tekniske Universitet.