Sie sind auf Seite 1von 22

Accepted Manuscript

On failure modes and design of multi-bolted frp plate in structural joints

S. Russo

PII: S0263-8223(18)34730-5
DOI: https://doi.org/10.1016/j.compstruct.2019.03.048
Reference: COST 10776

To appear in: Composite Structures

Received Date: 31 December 2018


Revised Date: 20 February 2019
Accepted Date: 12 March 2019

Please cite this article as: Russo, S., On failure modes and design of multi-bolted frp plate in structural joints,
Composite Structures (2019), doi: https://doi.org/10.1016/j.compstruct.2019.03.048

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ON FAILURE MODES AND DESIGN OF MULTI-BOLTED FRP PLATE IN
STRUCTURAL JOINTS

S. Russo*
(*) Iuav University of Venice, Tolentini n.191, 30123, Venice, Italy; russo@iuav.it

Abstract
Technical recommendations for design of all FRP structures are by now available and recognized.
Inside this frame, the design of all FRP bolted connections with steel bolts has a not negligible
position due to its importance in application, especially when bolted connection have, or could
have, a significant role in term of bearing capacity and hierarchy between structural parts. The
design strength formula for bolted FRP connections - proposed in European and national normative
and guidelines - define different approaches in function of the numbering of bolts, the geometrical
characteristics and typology of each joint and action, that could be out or in plane. Nevertheless, the
specific case of multi-bolted (MB) or high-multi-bolted (HMB) FRP plate used in structural joints
has been not yet clearly faced, nor from experimental and numerical approach, nor for failure mode.
Besides also the strength design approach becomes complex in presence of MB connections. By the
way, this research try to investigate specifically this point starting from recent experimental based
results dedicated to the structural performance of MB and HMB FRP plates used in connections,
and in comparing them with expected ultimate strengths and available design formula. In comparing
available formula with experimental outputs - and in verifying their reliability/compatibility - the
proposed investigation shows also some thought onto the potential oversize/undersize of this kind
of structural joints. The research is exclusively dedicated to the MB FRP plates produced by
pultrusion process and used in joints subjected to only in plane action.

Keywords: multi-bolted plates; FRP pultruded material; design criteria; failure modes; in plane action;
complex FRP joints

1. Introduction
The bolted connections in all FRP structures are for many reasons preferable than the glued solution
that needs a very attentive action - especially in the construction phases - that could not be always
guaranteed. Currently, the structural design approach for all FRP bolted connections is faced by
many technical recommendations produced by specific countries, as well British recommendations,
[1], Italian ones, [2] , from American, [3-4], from Dutch [5] and Germany [6]. Besides specific
European normative for the experimental test [7], and a technical report dedicated to the structural

1
design guidelines, [8] - intended also as a preliminary document toward the specific Eurocode for
FRP constructions - are available.
Obviously, also the private professional producer market is characterized by many technical
suggestions, whose some of the more dominant are the following cited from [9] to [12]. As well as
the literature proposes by now a very high quantity of specific papers and books, among which are
mentioned the following from [13] to [17].
Nevertheless, the specific case of the multi-bolted (MB) or high-multi-bolted (HMB) FRP plates
used in all FRP structural connections subjected to in plane action, is not treated in the above
mentioned documents; this is the reason why that aspect is here focused. The proposed
investigation’s topic is indeed around the influence of an high number of bolts, and their
configuration and distribution, in failure mode and structural strength design approach, both still
unknown.
About MB FRP connections, some results already published are dedicated to the stress distribution
in presence of traction applied to the centroid of the bolt configuration [18], as well as some other
researchers have only proposed potential approach [19], or are supported only through and
experimental overview, [20], or dedicated to specific item, [21]. In fact, they cannot help concretely
the specific investigation here proposed.
By the way, the concept of MB and HMB joint is not yet encoded in term of quantity and
configuration, [22]. Moreover, the available failure criteria for all FRP bolted connections simplify
the approach too much – as proposed in the two most important recommendations showed in [2]
and in [8]. The consequence is the absence of any positive support when connections are complex –
characterized by high number of bolt and not symmetric configuration - as they are often in the
effective applications.
More in detail, those technical recommendations propose design formula for simplified connections
with a limited number of bolts and a very simple action, as a traction’s load, perfectly applied to the
center of the bolt’s configuration; but these conditions are almost never realistic.
Viceversa, in most of the real applications made by FRP, the bolted connections in beam-column
frames or bridges or strut and tie spatial constructions, are subjected to an in plane moment that is
not easy applicable.
That is with the consequence that when joints are MB or HMB type, the effective stresses will be
among all the bolts - and then to each bolt - in a way not yet studied or never deeply understood.
Moreover the effective stress distribution will influence also the related failure modes.
More specifically, the proposed investigation - mainly managed through a comparison between
experimental results and different available closed formula - it is exclusively in the frame of joints
with steel bolts, FRP plate used as connection between structural elements, and in plane action; both
plates and structural element are FRP material made and produced by means pultrusion process. As
a directly consequence, the failure and strength design approach will exclude steel bolts and will be
FRP side.

2
The FRP plates investigated where produced by pultrusion process, as already mentioned, with the
external load orthogonal respect the fiber’s orientation; each final FRP plate, because of its high
thickness, was then made as a sum of two glued thinner FRP pultruded plates.
The FRP connections analyzed are of two types, with different dimensions, bearing capacity and
number of bolts; the first one with twelve bolts, the second one with eighteen, in both cases without
symmetric configuration in the plate, [23], [24]. For both type of plates the bolts are located in two
different zones on the top and on the bottom, so with an intermediate part - between them - without
holes. This scenario about the bolted FRP connection confirms for more reasons the complexity of
the proposed research and the need to face such kind of joints with new specific design formula and
failure criteria. Besides, others significant invetsigations have been dedicated to specific cross
section, [25], special technical solutions, [26] and on the specific effect in failure of beam-column
joints, [27].
Nevertheless, for the same reason, the right need to try to simplify this field is confirmed because of
the too high level of geometrical, static and mechanical parameters. Contemporary, there is also the
aim to address in the more proper way the designer and the professional structural engineers who
want to use complex bolted FRP connection.
So, this study wants to extract from previous researches on similar item, [23-24], [28], the first
tentative of understanding the limitation of the current formula and failure criteria respect the cases
of MB and HMB plate inside all FRP joints.
From the point of view of the reason way and origin of the two type of investigated connections,
they have been directly assumed from an all FRP strut and tie spatial structure built as a temporary
shelter and calculated in conservative way, [28].
This emergency structure had built in 2009 in a seismic zone – and is still there - in absence of
specific recommendation for seismic zone and considering the local and global second order effect
in pultruded FRP elements as dominant principles for structural design at ultimate limit state.
Finally, the comparison proposed mainly between experimental results and available formula
evidenced many information about critic points and low reliability of the available structural design
approach with consequent overestimation and/or underestimation of the expected strengths.

2. State of the art


Following, a very short but inclusive overview of the general design approach about the all FRP
connections with steel bolts and exclusively in presence of in plane action, is proposed. The
synthesis is specifically in function of the type of joint and action, the material’s characteristic, the
production technology and available design formula.
As much known, the type of connection influences the structural design approach in presence of in
plane action. By the way, a column-beam joint is basically strongly different than a beam-beam one,
in the case of in line position or the orthogonal one. By the way, the role of the intermediate
structural element, that will be a plate or angle is not negligible respect the structural behavior of

3
the joint in its overall. Indeed, unless the adhesive solution that remain the more strength, the angles
are less performant , see [4] and [21] - respect the use of the plate, due to also the orthotropic
characteristics of the FRP pultruded material.
As well as we can have a joint only subjected to traction load or a column-column joint or a
connection between a column and its foundation on the ground. In all the cited case we have
substantially joints for which the in plane moment position is sufficiently easy to deduce.
Nevertheless, at the same time, no one of the cited joints is complex as the similar one indicated in
Figure 1. In that case the plate, that is FRP laminated, then is a convergence of five structural all
FRP pultruded elements that generate a complex in plane stress distribution through 28 steel bolts,
[28].

Figure 1. Example of MB FRP plate in complex all FRP joint with steel bolts, [23]

Besides, the geometrical suggestions provided in function of the dimension of the diameter bold and
plate’s thickness with consequent limitations from the edges of the plates are strategic. This point is
particularly complicated due to the too much high quantity of available recommendations not
always between them harmonized. An only indicative and limited overview is presented in Table 1,
having defined as reference the symbols reported in Figure 2, detail a), related to the failure mode in
presence of a very simplified connection. In detail – see Table 1 - the term p1 and p2 are equal
respectively to the bolt’s distance in direction of the load and the bolt’s distance in the opposite
orthogonal direction, while t is the thickness of the FRP, d and d0 are respectively the bolt and hole
diameter; the e1 and e2 terms reported in table 1 are again indicated in Figure 2. However, the same
Table 1 shows also similarity of interest respect bolted joints with other materials.

Limitations Geometrical Geometrical Geometrical


recommendation Compatibility Compatibility
FRP Steel Wood
with b = 2e2
a = e1
Bolt diameter d>tmin 1.5<a/d<3 t>2.4 cm
tmin<d<1.5tmin t = d(3+8d)>2.4 cm

4
Hole diameter d<d0+1mm D<d0+2mm D = d0
(d0)
Distances p1>4d 25t<p1<3d >5d parallel toi fibres
between holes p2>4d
L>2.8d 5d orthogonal to fibres

distances from edges end e1>4d 2<b/d<3.5 end e1>10d (parallel)


side e2>2db end e1>5d (orthogonal)
e2> 5d (parallel)
e2>3d (orthogonal)
Table 1: Type of joints and material with suggested geometrical limitation, [7].

Figure 2. Reference joint, (a), and failure modes, (b), (c) and (d); [7]

More generally, about the influence due to the material, the current recommendations distinguish
with evidence between the use of an all FRP joint – in which also the bolts are in FRP material –
than the case, more common, in which all the structural elements are in FRP materials while the
bolts are always made by steel, as in the cases here investigated. Therefore, independently from the
clearance – that also strongly affects the bearing performance of the joint, [22] – a bolted joint
entirely made by FRP is characterized by a hierarchy between strength very different then that
activated when steel bolts are considered. In this last case, all the failures will be FRP side, even
though the expected ultimate strength value in FRP will be not reached due to the limit of
deformability of the FRP which has a fragile behavior. Besides, the production technology could
strongly influence the way of failure of the FRP plates. Indeed, in presence of pultrusion process -
that generates a more performing axis from the mechanical point of view - the direction of the
applied load than the dominant fibers orientation strongly affects the structural behavior. The other
productions technologies as well the laminated and bag molding ones consequently have an
important role in the effective mechanical performance of joints. Some of the most significant
recommendations in function of the fiber orientations with suggested geometrical compatibility are
considered in [1], [2] and [8]. The type of connection, as already mentioned, even if in the
exclusively frame of the in plane action, can strongly influence the stress distribution in each joint.
In detail Figure 3 shows some of the most used joints both in scientific investigation and real
applications. The joint proposed in the first row are all characterized by the presence of angles -
which are always keaknesse respect bolted joint with plate or bonded ones - and with two boundary
conditions, as well as perfectly clamped and semi rigid.

5
Bank 4 Mosallam 1 Mosallam 3 Mottram 1 Mottram 7

Russo 1 Smith 4 Feroldi-Russo A Feroldi-Russo B Russo 2

Figure 3. Type of all FRP bolted joints

Also in the row at the bottom the joint by Smith, [30], is in fact composed by element, angle and
bolts, while the joint by Feroldi, [23-24], provided plate instead of angles but with the same
structural function intended as connection between different structural elements. Beside the joint
with angles will be characterized also by the local potential failure in the flange and semi flange
since that they will be always subjected to out of plain action. Besides that is a confirmation about
the limited strength and stiffness when the bolted connection si realized by means steel or FRP
angles.
That could generate a reduced stress in the bolted zone but an increasing of stress in the flange that
represents the more vulnerable and weakness part in open FRP shape elements. Vice versa, in the
joint in which the intermediate part is by plate the flange will be protected but with an increment of
stresses in bolted zone.
In detail in ‘Bank4’ and both ‘Mosallam’ , respectively presented in [13] and [18], the bolt with
open wide flange profiles are used and are both made by FRP; in both Mottram type the bolts are
made by steel and the profiles have wide flange, [14 and 22], but in one case also the bonded
technic is used; the Smith, [30] option is with steel bolt and open FRP cross section with narrow
flange.
Nevertheless, the typology of the complex all FRP bolted in joints by means MB plate could be
more wide and extensive than that shortly presented in Figure 3 (‘Russo2’), [25]. By the way the
joint indicated in Figure 4 is a connection between a compressed (on the top) and tensile element
(on the bottom) in an all FRP spatial strut and tie covering structure, [25]. In detail, that joint is
realized through sixteen bolt made by steel and a laminate FRP plate as intermediate element of
connection between the two elements that are their self a consequence of a built-up cross section
made by four “C” FRP pultruded shape as that indicated in Figure 3, (Russo1).

6
Figure 4. Example of MB FRP joint between two built-up FRP column and an FRP plate, [23]

Therefore in this case the FRP plate plays an important role since that it is not only an element of
the joint through bolts, but also an effective structural element whose area absorbs a lot of in plane
actions derived directly from the inclined compressed element and from the horizontal tensed one.

3. Failure Mode Criteria


The recognized failure modes in presence of the in plane actions are the following: a) Net-tension,
b) Pin-bearing , c) Shear-out, d) cleavage and e) Bolt-shear, whose the last one could be excluded
when, as in the case here investigated, the bolts will be always made by steel with consequent
collapse FRP side. These simplified failure modes with the related schemes are partially synthesized
in Figure 2. Nevertheless the potential failure modes are considered separately, as if they could
manifest without an effective overlapping phase or a close interaction between themselves. Indeed
in the more complex joints also the common sense makes real that the singular failure modes are
combined and mixed , [31], and cannot be traced back to a failure modes related to a single bolt as
indicated in Figure 1 or in presence of two bolts as presented in [8]. All that said, the Tsai and Hill
general criteria, [32] and the Hashin one, more detailed and dedicated to fiber and matrix failure
modes, [31], are those mainly known and used. So, having defined a generic orthotropic lamina in
presence of a stress x with an angle  respect the longitudinal direction, the stress in the principal
material direction will be the following:

 L   x cos 2  (1)

 T   x sin 2  (2)

7
 LT   x sin  cos (3)

with L, T and LT the corresponding available stresses. Considering now that the maximum stress
and strain theory is related to the maximum work one - that is itself related to the yield criterion
proposed by Tsai and then specialized for an anisotropic material as specified by Hill - the stress in
the principal material direction could be determined through the following Tsai-Hill Theory, [30]:

cos 2  cos 2   sin 2  sin 4  sin 2   cos 2  1


    (4)
 2
LU  2
LU  2
TU  2
LTU  x2

With LU longitudinal tensile strength, TU transverse tensile strength and LTU shear strength
respectively. While the plane stress failure criteria for unidirectional fiber composites have been
proposed by Hashin [31].Besides, from a general point of view we have to consider also that in the
bolted joints the holes create also a weakness specific zone that could be considered as an induced
damage and could be also assumed as an inclusion, [31], with consequent hole-size effect. The
related failure criteria, divided in the point stress and average stress approaches are both based in
the following relation to calculate the normal stress y along the x axis in front of a single hole with
radius R (see again figure 1 detail a) as reference):

   R    R 6  R 8  
2 4
R
 y x,0  2     3   kT  35   7    (5)
2   x  x   x   x   

4. Strength Design Formula


The strength design formula available for the in plane action in bolted all FRP connections are
strictly related to the failure modes previously indicated in Figure 2 for one bolt and in fact – except
in [8] – they does not take into account of the bolt’s number. In reference with [8], [3] and Figure 2
for the adopted symbols, they are the following:

Vsd  Cr
Vsb.d  (6-10)
n
where Vsb,d is the force transmitted by each bolt, Vsd is the design force, and Cr and n represent
respectively the coefficient of load redistribution and the number of bolts. About the net tension
failure mode, following is reported the related formula (for 5 to 90 degree):

 fTt ,Rd  w  nd   t
1
Vsd  (7-11)
Ktc
where the Ktc is a stress concentration factor equal to 3.75, fTt,Rd is the tensile design strength of the
FRP material, while w is indicated in Figure 2, t is the thickness of the plate and d is the bolt
diameter. About the pin-bearing failure we have the following:

1
Vsb, d   fT ,br, Rd  db  t (8-12)
Kcc

8
Where Vsb,d is the beraing force transmitted by each bolt, fT,br,Rd is the transvers design strength for
pin-bearing failure with to90° , Kcc is a stress concentration factor equal to (d0/d)2. Besides
for 0°<<5° the relation of the pin-bearing failure will be the following:

1
Vsb.d   fT ,br,Rd  db  t (9-13)
K cc
While for the shear out failure we have the following formula:

Vsb.d  fV , Rd  2e  d   t (10-14)

where fV,Rd is the design shear strength.

5. Available results on MB or HMB FRP joints


About column-beam-plate connection type in presence of only in plane actions, some results are
available, [23-24] while about specific investigation on MB connections, the researches are very
few, [21], and mainly dedicated only to simplified cases as well as a traction stress applied at the
center of the bolts configuration, [8]. By the way, Figure 5 shows the general setup of the two
complex joints with MB FRP plate subject to in plane action object of this investigation. The
mechanical characteristics of the FRP pultruded and glued plate are indicated in Table 2; the
mechanical and strengths values have been determined following the test procedures indicated in
[7]. Each steel bolt is characterized by two resistance cross sections (see again Figure 5) since that
each plate is used as an intermediate element inside two open “C” pultruded shape that generate
then each built-up structural element.
More in detail, the experimental program provided test on beam-column joint subjected to in plane
moment applid through the load directly applied to the column, as reported in Figre 5 for joint A
and B. Besides, each structural joint as been instrumented with strain gauges for local measurement
of deformation to better understand the rel strain distribution in the MB FRP plates. Then each
beam-column joint has been clamped to the ground favoring the moment-rotation as indicated in
Figure 6.

Joint 1 – Detail A

9
Joint 2 – Detail B
Figure 5. Test setup and detail

Mechanical characteristics/Strengths Symbols Values


I Longitudinal Tensile Modulus of Elasticity Ey 27 GPa
II Transverse Tensile Modulus of Elasticity Ez=Ex 10.5 GPa
III Transverse Shear Modulus of Elasticity Gxz 3.5 GPa
IV In plane Shear Modulus of Elasticity Gxy=Gyz 2.5 GPa
V Major Poisson’s ratio νxy = νyz 0.25
VI Minor Poisson’s ratio νxz 0.12
VII Bulk weight density of FRP γ 1850 (kg/m3)
VIII Longitudinal Tensile strength σyt 430 (MPa)
IX Transverse Tensile Strength σxt = σzt 45 (MPa)
X Longitudinal Compressive Strength σyc 330 (MPa)
XI Transverse Compressive Strength σxc = σzc 80 (MPa)
XII Shear Strength τxy = τyz ;τxz 45 (MPa)60 (MPa)

Table 2. Mechanical characteristics and declared strength for FRP plates

More specifically, the first joint, Detail A of Figure 5, has the bolt diameter equal to 12 mm with
hole diameter of 13 mm, while the second joint, Detail B, has bolt diameter equal to 14 mm with
hole diameter of 15 mm. In the two joints steel washer for each bolt connection have been
used.With specific reference to the position of the plate inside the joint at ultimate limit state, the
detail on the right of Figure 6 shows the evident rotated part of the plate of the top then presumably
the bottom part which partially remain straight due to the type of connection. Globally the shear
deformation and the rotation of the plate as a consequence of local failure mode between bolt and
FRP plate in each hole appear evident. A comparison about the moment-rotations curves carried
through numerical simulation and experimental results already discussed, [23-24], is again proposed
in Figure 7.

10
Figure 6. Performance of the complex joint at serviceability and ultimate state, [21], [22]

The curves (Fig.7) are related also to the joint indicated in Figure 5 and their tendency confirms the
influence of the type of joint, the geometry and the presence of the MB plate in simply and complex
joint. In detail, the comparison shows first of all as the presence of the plate (joint 2), as
intermediate part between structural element, seems to guarantee more performance in rotation
value in comparison than that obtained with angles solution ( see also references [4] and [21]).
Besides, above all, the moment values result sensibly higher than those with angles, both bolted and
bonded.

22
Bank 4
20
Mosallam 1
18

Mosallam 3
16

14 Mottram 1
Moment [kNm]

12 Mottram 7

10
Smith 1

8
Smith 4
6
Nodo
Experimental
4 Joint A
A_sperimentale

Nodo
Experimental
2 Joint B
B_sperimentale
Joint B
0
0 0.025 0.05 0.075 0.1 0.125 0.15 0.175 0.2
Rotation [rad]

11
Figure 7. Moment-Rotation curves for different joints (indicated in Figure 3)

The local failure modes of the two structural joint at the maximum load, coincide in fact, as
expected, with the failure modes of the two MB FRP plates relieved after collapse, as showed in
Figure 8 and in Figure 9 respectively for Joint 1 and 2.

LOADING DIRECTION
LOADING DIRECTION

JOINT 1

Figure 8. Complex failure modes of the MB FRP plates at the collapse of the Joint 1, [21]

1 2 3

4 5 6
LOADING DIRECTION LOADING DIRECTION

12
JOINT 2

Figure 9. Complex failure modes of the MB FRP plates at the collapse of the Joint 2, [21]

Besides, the investigation on the mechanical performance of MB or HMB complex joints implies
also the understanding of the centroid position of each plate in function of the bolt configuration
and the effective moment in plane distribution. This difference is made even more complicated by
the fact that we have two different - and not symmetrical - bolted zones separated by a zone without
holes. Indeed, as easily recognizable through the set up in its overall (see again Figure 9) the two
plates are in both cases subjected to in plane action with a subdivision of the consequent in plane
moment between the two bolted zone. The effective strain distribution in plain during the test has
been studied through strain gauges applied in the MB FRP plates as indicated in Figure 10.

Figure 10. Strain’s position in each MB FRP plate

6. Discussion
The analysis of the available models and failure criteria respect the effective stress distribution in
MB joints here examined, allows a comparison in the frame of the expected strength than the
13
strengths effectively measured. Also about the real failure mode respect the conventional ones, and
then on the potential presence of new failure criteria, some first evaluation is possible. Following
the first related discussions and analysis are presented.
6.1 Failure mode’s recognition
The proposed complex failure mode are different for the Joint Type 1 and 2 even though a kind of
similarity in failure modes is recognizable. Respect the simplified schemes indicated in Figure 2
and current recommendations [3], [8] and the real failure indicated in Figure 9, the survey and the
comparison show new overlapped and mixed failure modes, [31]. It is sufficiently evident that no
one of the failure modes evidenced in Figure 2 is recognizable now through the real crisis processes
of the MB FRP plates. A first interpretation of the failure modes relieved indicates that not one of
known simple failure mode indicated in Figure 2 appeared, but if anything is evident that failure
modes are new and intended as a consequence of the overlapping and combination of the singular
simple known failure mode. Indeed the detail a) of the Joint 1 on the top of Figure 8 shows that all
the bottom bolted zone is subjected – in its overall - to a kind of shear deformation with interaction
between multiple shear- out failure combined with the bearing and , partially, cleavage ones. This
first interpretation is represented by means schemes in Figure 11.

Figure 11. Schemes of the potential new and overlapped local, global and interacted failure modes
Besides that appears all in coherence with the ‘loading direction’ of the external load (Figs 8 and 9).
For the Joint 2, due to the different of bolt’s number between the two bolted zone, the crisis
phenomenon is concentrated on the top – when the bolt’s number is lower - with multiple shear-out
failure mode and a kind of new only vertical cleavage phenomenon that seems to be similar to the
timber MB connections. More generally, the Joint 1, due to the symmetry between bolt’s number
even if with different configuration, appears more useful for a first tentative of understanding the
new failure criteria in MB plates. By the way, the common approach, intended to a single bolt, has
to be abandoned respect a more complex failure criteria in which all the singular bolted zone are
involved. Indeed the maximum moment in plane at the bottom of the Joint 1 generated an expulsion
of the bolted zone - in opposition to the in plane rotation induced by the external load - compatible
also with the fiber direction and the vulnerability of the matrix in compression. The different
hypothesis about the potential centroid position and the related transferred in plane action derived
from the external action, are respectively indicated in Figure 12 and 13.

14
Figure 12. Hypothesis of centroid positions in MB FRP plates

Figure 13. Action’s distribution derived from the experimental in plane moment for two centroid , [21]

The potential adoption of the hypothesis indicated in this figure with different centroid strongly
influence the consequent structural assumption and above all, influence also the entity of the load
respect the two bolted zone. In this frame, should be also reasonable to adopt only the hypothesis of
one geometrical center (see Figure 5, detail A and B).
Nevertheless - if we can accept that we could consider only one center of each plate - is also true
that the moment transferred through the external load (see figures 8 and 9) in this hypothesis will
be in any case supported first by the bolted area at the top – even with lower value of the in plane
moment - and only then also by the bolted area at the bottom, with the higher value of the in plane
moment induced by the external action.

6.2 Strength’s Comparison


The effective experimental results obtained in [24], in term of local strain between bolts, could be
used to better understand which one hypothesis between them showed in Figures 12 and 13 is more
15
reliable and realistic. The experimental stress values in the FRP plates derived directly from the
experimental strains measured through strain gauges (see previous Figure 10) - using the E values
of Table 2 - are now reported in Tables 3 and 4 respectively for the first stiffness’s loss (SLS) and in
correspondence of the joint’s collapse (ULS).

Strain position exp x10-6 (np) Ey (MPa) exp(y)(MPa) Ez/x (MPa) exp(z/x) (MPa)
Joint Type 1
CH1 244.49 27 6.60 10.5 2.57
CH2 -239.83 27 -6.47 10.5 -2.52
CH3 -529.71 27 -14.3 10.5 -5.56
CH5 -60.13 27 -1.62 10.5 -0.63
CH6 -16.21 27 -4.48 10.5 -1.74
Mean value horizontal
Mean value vertical
Joint Type 2
CH1 -519.43 27 -14.02 10.5 -5.45
CH2 448.98 27 12.12 10.5 4.71
CH3 -123.10 27 -3.32 10.5 -1.29
CH4 361.74 27 9.77 10.5 3.79
CH5 758.51 27 20.47 10.5 7.96
CH6 -1516.33 27 -40.94 10.5 -15.92
Mean value horizontal
Mean value vertical
Table 3. Experimental in plane stress in the FRP plates at the first loss of stiffness (SLS) of the joints

Strain position expx10- (np) Ey (MPa) exp(y) MPa) Ez/x (MPa) exp(z/x) (MPa)
Joint Type 1
CH1 232.19 27 6.27 10.5 2.44
CH2 -187.75 27 -5.07 10.5 -1.97
CH3 -1272.86 27 -34.37 10.5 -13.36
CH6 -1241.96 27 -33.53 10.5 -13.04
Mean value horizontal
Mean value vertical
Joint Type 2
CH1 -4786.38 27 - 129.23 10.5 -50.26
CH2 -1778.09 27 - 48.01 10.5 -18.67
CH3 -2000.49 27 - 54.01 10.5 -21.00
CH4 - 416.82 27 - 11.25 10.5 -4.37
CH5 72.13 27 1.95 10.5 0.76
CH6 -2483.46 27 - 67.05 10.5 -27.07
Mean value horizontal
Mean value vertical
Table 4. Experimental in plane stress in the FRP plates at the collapse (ULS) of the joints

16
In the frame of the ultimate limit state of the investigated Joints (Table 4), it is evident that the
stress level is higher in Joint 2 than Joint 1, as expected. Besides, is also clear the prominence of the
compression values of stress (identified in the tables through the ‘-‘ sign).That appears coherent
with the type of joint and the applied external action. Besides, for both joints the level of stress is
sensitively lower than the declared values of stress showed in Table 2. After the stress calculation
through the experimental strain values, could be also of interest to compare those values with the
ones intended as a consequence of the expected action distribution in each bolt due to the in plane
moment in presence of high bolts number. That expected distribution, as a sum of horizontal and
inclined components induced by the external action indicated in Figure 5, is shown in Figure 13 for
both type of plates, whose values for the more stressed bolts (related to the numbering reported in
Figure 11) are indicated in Table 5.
Joint Type/ Px (N) Py (N) PR (N)
single bolt
Joint 1
7 -1.82E+04 -4.13E+04 45131.73
8 -1.82E+04 0.00E+00 18186.06
9 -1.82E+04 4.13E+04 45131.73
10 2.31E+04 -4.13E+04 47335.47
11 2.31E+04 0.00E+00 23119.39
12 2.31E+04 4.13E+04 47335.47
Joint 2
1 -2.73E+04 -3.54E+04 44.683.66
2 -2.73E+04 0.00E+00 27300.5
3 -2.73E+04 3.54+04 44683.66
4 3.59E+04 -3.54E+04 50736.24
5 3.59E+04 0.00E+00 35867.16
6 3.59E+04 3.54E+04 50736.24
Table 5 Values of loads (see Fig. 12 for bolt position)

Again, as partially expected, the maximum values of load intended as resultant PR of the horizontal
(x) and orthogonal component (y), (fig.13) have been reported for Joint 2 (see table 5). Besides the
higher value of component is always in compression, in y direction than x one. The maximum value
relieved is equal to 50736 N. The comparison between the declared values of strength (Table 2) and
the experimental ones indicated in Table 4 is now reported in Table 6. The declared values resulted
always higher than the experimental one except in the case of the transverse tensile strength
calculated in using the Ey value for which we obtained the stress value equal to 129.23 MPa respect
the declared one equal to 45 MPa. Besides the shear experimental strengths resulted enormously
lower than the declared one. The comparison in using design formulas expressed in the previous
chapter 4 is indicated in Table 7.

Type of Strength Joint 1 (MPa) Joint 2 (MPa)


Transverse tensile strength declared 45 45
Transverse tensile strength 6.27 129.23
Experimental (with Ey)
Transverse Tensile strength 2.44 50.26
Experimental (with Ex/Ez)
Transverse compressive strength declared 80 80
17
Transverse compressive strength experimental 34.37 54.01
(with Ey)
Transverse compressive strength experimental 13.36 50.26
(with Ex/Ez)
Shear strength declared 45 45
Shear strength experimental 0.88 0.65
Table 6. Strength comparison; numerical and experimental results

The load comparison with the same approach indicated for the previous comparison in the frame of
strengths showed in Table 7 is proposed for both joints using in detail formulas 12, 13 and 14.
Type of force Joint 1 (kN) Joint 2 (kN)
Px (exp) 35.4 35.9
Py (exp) 16.5 35.4
PR (resultant, exp) 39.07 50.37
Formula (8) with declared value 6.91 8.23
45 MPa
Formula (8) with declared value 12.27 14.63
80 MPa
Formula (8) with experimental 0.96 8.78
stress transverse tensile value
Formula (9) with declared value 50.61 64.40
Formula (9) with experimental 5.25 23.64
stress transverse compression
value
Formula (10) with declared value 31.72 37.12
45 MPa
Formula (10) with declared value 42.30 49.50
60 MPa
Formula (10) with experimental 0.620 0.54
shear strength
Table 7. Force comparison; numerical and experimental results

A first analysis of the outcomes reported in Table 7 indicates that the different between declared
and experimental values remains not negligible and that for Joint 2 the distance between the two
type of value is reduced. Moreover, the effective real measured stresses remain very low and that
point is compatible with the failure mode never activated strength side.

7. Conclusions
On the base of the obtained experimental and numerical results, the following final considerations
can be proposed:
- The current failure modes available for all FRP bolted connections are not applicable/reliable in
presence of MB and HMB FRP plates used in structural joints in presence of in plane action.

- The current structural design formula are not applicable and reliable in presence of MB FRP plate
used in joint and in any case nor in MB all FRP connections.

18
- In complex and MB connections, it appears that the mode of failure related to net-tension and
cleavage are excluded in favor of a dominant interaction between bearing and shear-out, that
acquire a feature of chain/combination mode of failure developed by vertical and horizontal bolt’s
lines in function of the load orientation.

- The experimental stress distribution derived directly from the experimental strain values indicated
that the stress level at the collapse is always considerably lower than the declared strengths.
- The crisis of the two MB FRP plates investigated occurred for very low values of strength. From
one side that point could be considered as expected, but from another side it means that the two
joints are cover dimensioned. In other words, it means also that currently the structural design
approach for these type of joint is uselessly too conservative.
- The comparison between experimental and calculated strength than that the declared ones shows
that the experimental values appeared always lower. In detail that difference is more evident on
shear strength.
- The comparison between calculated loads and loads derived from experimental strains, showed
with evidence the very low reliability of the available design formula and the need to propose new
strength design approach for MB plate in complex joint in presence of in plane action.
-The design criteria specifically dedicated to MB or HMB FRP plates are still in fact uncovered,
both in available researches and technical recommendation. The obtained results could help in
cover also this aspect that is strategic in approaching all FRP bolted structural joint.
- Failure criteria appear still reliable and solid only in presence of very simple all FRP bolted
connections.

Aknowledgement
The author is grateful to Francesca Feroldi for her initial investigation in the same research’s field
and specific first evaluation. Thanks also to the technicians of the Lab of Strength of Materials at
the Iuav University of Venice (Italy)

References
[1] Clarke JL, editor. Structural Design of Polymer Composites: EUROCOMP design Code and
Handbook, London: E &Fe N Spon; 1996
[2] National Research Council of Italy, Guide for the Design and Construction of Structures made
of FRP pultruded elements, CNR DT 205, 2007, Rome, October 2008
[3] ASCE Manuals for Composite Materials, Vol 1-3, 1984
[4] Design guide for FRP composite connections. Manual of practice (MOP) 102, ASCE 2011,
Reston, VA, 624.

19
[5] CUR 96, Fiber Reinforced Polymers in Civil Load Bearing Structures ( Dutch
Recommendation, 2003)
[6] DIN 13121 Structural Polymer Components for Building and Construction ( Germany, August
2010).
[7] UNI 2003. Reinforced plastics composites specifications for pultruded profiles. Part 2- Method
of test and general requirements”. EN13706-2-2003.
[8] Prospect for New Guidance in the Design of FRP. European Commission, JRC Science for
Policy Report, Report EUR 27666 EN, Support to the implementation, harmonization and further
development of the Eurocode, 2016.
[9] ACMA Pre-Standard for Load and Resistance Factor Design of Pultruded Fiber Polymer
Structures (American Composites Manufactures Association, November 2010)
[10] Strongwell, Structural Design Manual, 1999
[11] Design manual, Fiberline Composites, 2003
[12] Design of FRP Structures in seismic zone, Topglass, 2014
[13] Bank LC., Composites for construction-structural design with FRP materials. Whiley, 2006
[14] Mottram JT, Zheng Y., State of the art review on the design of beams-to-columns connections
for pultruded structures. Composite Structure Journal 1996; 35:387-401
[15] Introduction to Composite Material Design, Ever J. Barbero, CRC Press, 2010

[16] Zureick A., Shih, N., Local buckling of fiber-reinforced polymeric structural members under
linearly-varying edge loading — Part 1. Theoretical formulation, Composite Strcutures, vol.41,
1998
[17] Pecce M, Cosenza E., Local buckling curves for the design of FRP profiles, Thin Walled
Structure Journal, 2000, 37; 207-222.
[18] Ascione L., Feo L., Mosallam A.S., Stress analysis of multi bolted joints for FRP pultruded
composite structures, Composite Structure Journal 94, 2012
[19] Camanho P.P., Mattews F.L., Stress analysis and strength prediction of mechanically fastened
joints in FRP: A Review, Composite Part A, 28(6), 529-547.
[20] C.T. McCarthy, M.A. McCarthy, V.P. Lawhor, Progressive damage analysis of multi bolted
joints with variable bolt-hole clearance, Composite part B Engineering, vol.36, Issue 4,June 2005,
pages 290-305.
[21] A preliminary numerical and experimental investigation on the shear stress distribution on
multi-row bolted FRP joints. MECH RESH COM, 164-168, 37, 2010.

[22] Mottram JT, Girao Coelho AM., A Review of the behavior and analysis of bolted connections
and joints in pultruded fiber reinforced polymers. Material Design Journal, 2015;74:86-107.
[23] Feroldi F., Russo S., Structural behavior of All –FRP beam column plate bolted joints. Journal
of Composites for Construction (ASCE), 2016.

20
[24] Feroldi F., Russo S., Mechanical Performance of Pultruded FRP Plates in Beam to Beam
Connections. Journal of Composites for Construction (ASCE), 2017.
[25] Wu C., Bai Y., Connections of tubular GFRP wall studs to steel beams for building
construction. Compos B Eng 2016, 95, 64-75.

[26] Zhang ZJ, Bai Y., Xiao X., Bonded Sleeve Connections for Joining Tubular Glass Fiber–
Reinforced Polymer Beams and Columns: Experimental and Numerical Studies. Compos Constr
2018, 22(4), 04018019.

[27] Ascione F., Lamberti M., Razaqpur AG, Malagic M., Pseudo-ductile failure of adhesively
joined GFRP beam-column connections: An experimental and numerical investigation. Compos
Constr, 864-873, 200, 2018.

[28] Russo S., Experimental and finite element analysis of a very large pultruded FRP structure
subjected to free vibration.Com Struct 2012; 94: 1097-105.ISSN: 0263-8223.
[29] Russo S., Damage assessment of GFRP pultruded structural elements, Composite structures,
96; 661-669.
[30] Hart-Smith LJ, The key to designing efficient bolted composites Joints, Composites, 25(8),
835-837, 1994.
[31] Russo S., First investigation on mixed cracks and failure modes in multi-bolted FRP plates.
Composite Structures Journal 154 (2016) 17-30.
[32] Tsai, S. W. and Wu, E. M. (1971). A general theory of strength for anisotropic
materials. Journal of Composite Materials. vol. 5, pp. 58–80.

21

Das könnte Ihnen auch gefallen