Professional Documents
Culture Documents
Michael Engelmann, Dipl.-Ing.; Bernhard Weller, Prof. Dr.-Ing. Technische Universitt Dresden Institute of Building
Construction, Dresden, Germany. Contact: michael.engelmann@tu-dresden.de
DOI: 10.2749/101686616X14555428759000
y,v
destructive bending load of Fmax (see
Table 2). Finally, destructive bending 12 34
tests were executed at P0 = 50 kN. The Adhesive joint
tests were performed at room tempera- v y, bottom
tures of 2628C in the testing facility. Detail in figure 2 Spiral cable 24 mm
(mm) z,w in a groove
Materials and Specimens F 2,3 top
Active side Passive side
wz,A F/2 F/2 wz,B
Three specimens with a span of 9 m P; wP
Specimen 1; 3
P0 = 50 kN
P0 = 75 kN cross-section and run below the contin-
P0 = 100 kN
Equ. 2 uous lines. The load-bearing behaviour
was equal to the unbroken situation
60
P (kN)
P0 = 100 kN
Discussion
0.001
(rad)
Post-Tensioning
During the post-tensioning process,
the cable load introduces compressive
0.003
P0 = 50 kN
P0 = 75 kN
Stress at top edge P0 = 100 kN
0
Equ. 1
mean (N/mm2)
S
y,v
12 34
(1)
15
vy,bottom
Evaluation range e14,bottom where
20
100
Specimen 1
Specimen 2 preliminary design. Consequently, an
Specimen 3
P0 = 0 kN upscaling up to 9 m in span is feasible
80
P0 = 50 kN and confirms the results presented in
P0 = 75 kN
P0 = 100 kN Ref. [17].
In concrete structures, the cable loads
60
F (kN)
simultaneously. In a post-calculation
the relation of cable load and piston
Evaluation range
travel was also determined. Figure 3a
0
Specimen 1
Specimen 2 The unbonded tendon is fixed at each
Specimen 3 end and at the connectors. This means
0.02
P0 = 0 kN
P0 = 50 kN that the strain of the cable needs to
P0 = 75 kN comply with the deflection of the
0.01
Fig. 4: Results of all non-destructive bending tests: (a) increase in bending during displace-
ment controlled test; (b) twist in cross-section during bending
Compared with the results in Table 1 and beam theory in Eq. (1). The results for
Fig. 3b where values of 0.099 N/(mm2 kN) specimens 1 and 3 drift apart consider-
for specimen 2; 0.139 N/(mm2 kN) and ably at cable loads P > 80 kN. During
0.166 N/(mm2 kN) for specimens 1 and this stage (see Fig. 3b) the highest twist
3 are given, a deviation of 72%, 101% values were recorded, meaning that
and 120% is present. In terms of speci- additional bending about the minor axis
men 2, this is caused by the proximity of is present, which is not covered by the
(2)
the strain gauges to the connector that analytical model. However, during cable
causes a local stress field. This does not loads P < 60 kN the linear analytical
comply with the assumptions of linear result overlaps the experimental results. where
lowing aspects:
0
P0 = 75 kN
P0 = 100 kN in a decreased distance between the
Increasing twist
ends of the beam. It is suspected that
the edge tolerance in specimen 2 was
20
S
y,v minor axis. An increase in lateral
Stress at top edge
12 34 deflection because of this effect was
40
Resuld of all non-destructive tests vy,bottom reported in Refs. [7, 11] as well.
14,bottom
z,w The strain of the blocking material
0 5 10 15 20 was not taken into account. This
wz (mm) leads to an increased reduction in
the length of the glass structure
Fig. 5: Results of all non-destructive bending tests: (a) change in cable load during bending (LG). However, this deformation
at alternating initial cable loads; (b) mean stress at bottom and top edges of specimen 1 as is considerably smaller than the
an example at alternating initial cable loads strain of the cable and is considered
to have only a minor effect on the
system.
Specimen 1 Specimen 2 Specimen 3
As shown in Fig. 3b all beams
P0 (kN) 53.4 51.0 50.1
undergo lateral deflection, which is
Fmax.1 (kN) 78.8 71.6 80.8 not considered in the analysis. This
wz(Fmax.1) (mm) 14.84 26.13 15.07 will decrease the distance between
the anchors and the connectors,
Ffail.1 (kN) 74.4 71.2 75.6
which determines the length of the
wz.fail.1 (mm) 20.42 38.90 23.34 cable.
Failure pane no. 2 2 and 3 2 and 3 Equations (1) and (2) ignore the
Fmax.2 (kN) 66.00 43.60 52.40 dead load of the specimens.
wz(Fmax.2) (mm) 14.43 36.16 13.95
It is therefore concluded that the lat-
Ffail.2 (kN) 62.40 43.60 48.40
eral deflection influences the results,
wz.fail.2 (mm) 25.76 44.24 28.34 which needs to be studied in a more
Failure pane no. 3 and 4 4 1 precise numerical analysis by includ-
ing the local effects at the anchors and
Note. All results were recorded at 5 Hz. Slopes were calculated using a linear regression within the
evaluation range where R2 is the coefficient of determination. connectors. A linear elastic analytical
model is limited to the linear stage.
Table 3: Results of destructive bending tests
Bending
LP,0 length of the unloaded cable; LG change in cable length due Each beam was tensioned to an ini-
LP,w length of the loaded cable; deflection of the beam and tial cable load P0 for non-destructive
Le change in cable length due to NG(P) axial force in glass due to tests in order to observe the effect of
cable load; cable load < 0. cable load on the global load-bearing
100
Specimen 1 Failed panes:
Specimen 2 Additionally, the axial compressive
Specimen 3
Destructive test 1 s s
load from post-tensioning increases
80 Destructive test 2 this problem. Thus, an interaction of
Fmax 12 34 12 34
s s
behaviour was observed.
40
12 34 12 34
s s
system is statically indeterminate. To
12 34 12 34
handle this analytically, one has to split
the system into a determined part and
0
(3)
Specimen 1
where
40
Specimen 2
Specimen 3
Destructive test 1
Destructive test 2
Mi,j is the bending moment and Ni,j is
Fmax the axial force with i = 0 determined
30
Fig. 11: Bonding of the steel connectors. Cleaning and preparation of the surfaces (left); application of the adhesive (centre) and
positioning on the glass edge where spacers were used as a bonding device (right)
Acknowledgements
Fig. 12: Spannglass bridge for pedestrian bridge design competition at the IABSE The scientific project was sponsored by the
Conference 2015 in Nara, Japan.20 The international jury awarded the bridge design with German Federal Ministry of Economics and
the Outstanding Design Award (First prize) Technology and was executed with the help
of THIELE Glas Werk GmbH (Wermsdorf,
Germany) and KL-MEGLA GmbH (Eitorf,
to be installed. Additionally, a test Both packages are interconnected by Germany). Furthermore, PFEIFER Seil- und
installation of the whole bridge was stainless steel connectors, which are Hebetechnik GmbH (Memmingen, Germany)
performed before shipping the parts adhesively bonded to the glass edge. gave valuable support. Additionally, we
to Dsseldorf. They redirect the cable, take on the acknowledge the valuable technical support
steel frame holding the segmented of Friedrich-Siemens-Laboratorium and
Outlook: Future Bridge Design glass balustrades as well as the walk- Otto Mohr Labor of Technische Universitt
on glass deck and prevent a contact Dresden.
Options
between the two glass packages. The
The 2014 IABSE conference in Nara, deck forms an arch of walk-on glaz- References
Japan, hosted a bridge design compe- ing covered with anti-skid surface. This
tition. One of the entries was a deck [1] Weller B, Meier A, Weimar T, Menkenhagen
ensured that pedestrians as well as J., Koschecknick K. Hybride Glastrger als
bridge with a smooth curved flat rolling users may use the connection.20 Fassadenelemente. In Stahlbau Spezial kon-
arch-like structure for a bridge across struktiver Glasbau. Wiley-VCH Verlag, 2010;
the 26 m Saho River in Nara, Japan Conclusion and Summary pp. 4145. doi:10.1002/stab.201001306
(Fig. 12). A highly slender structure, 0.3 [2] Louter C. Fragile yet Ductile. Structural
m high at centre span and 1.1 m high During this study, the complete design Aspects of Reinforced Glass Beams. Dissertation,
at the ends was achieved using two process of an innovative glass beam Delft University of Technology, TU Delft, The
Spannglass beams made from three concept with post-tensioned reinforce- Netherlands, 2011.
sections, each forming the main load- ment leading to a pedestrian bridge [3] Bos FP, Veer FA, Hobbelmann GJ, Louter
bearing structure. Each beam is made was described. The bridge consisted PC. Stainless steel reinforced and post-tensioned
of two packages of heat-strengthened of a glass balustrade, glass bridge deck glass beams. In Proceedings of ICEM12 12th
laminated glass at a distance of 80 mm and two 9 m Spannglass beams. After International Conference on Experimental
to each other. The gap is used to guide several design studies, the final option Mechanics, Politecnico di Bari, Italy, 29 August
a 20 mm steel cable post-tensioned to emerged and was prepared for struc- 2 September, 2004.
150 kN to accommodate the tensile tural analysis. Balustrades and walk- [4] Louter PC, van Heusden JF, Veer FA,
stress and deflection from dead load. on glazing were detailed analytically Vambersky JNJA, de Boer HR, Versteegen J.
IABSE Publications
Digitised Archive
from 1929-1999
Did you know?
IABSE publications between 1929 - 1999 can be downloaded for FREE
from the IABSE Archive on the internet - by everyone.
Do full text searches in any part of the papers, or browse
by title, author or year.
Explore the Archive with over 80'000 pages and find a treasure
of information on research, projects and IABSE.
www.iabse.org/IABSE/Publications/Archive
PS: Conference Papers published after 1999 (over 3'500) are available via Ingenta.
Full Papers may be downloaded with a subscription or on a pay per view basis.
www.ingentaconnect.com/content/iabse