Mechanisms of Block Shear Failure of Bolted Connections
Mechanisms of Block Shear Failure of Bolted Connections
Mechanisms of Block Shear Failure of Bolted Connections
Scholars' Mine
Drew D. A. Clements
Recommended Citation
Teh, Lip H. and Clements, Drew D. A., "Mechanisms of Block Shear Failure of Bolted Connections" (2012).
International Specialty Conference on Cold-Formed Steel Structures. 3.
https://scholarsmine.mst.edu/isccss/21iccfss/21iccfss-session9/3
This Article - Conference proceedings is brought to you for free and open access by Scholars' Mine. It has been
accepted for inclusion in International Specialty Conference on Cold-Formed Steel Structures by an authorized
administrator of Scholars' Mine. This work is protected by U. S. Copyright Law. Unauthorized use including
reproduction for redistribution requires the permission of the copyright holder. For more information, please
contact [email protected].
Twenty-First International Specialty Conference on Cold-Formed Steel Structures
St. Louis, Missouri, USA, October 24 & 25, 2012
Abstract
This paper examines the mechanisms for block shear failure postulated in the
design provisions specified in the North American Specification for the Design
of Cold-formed Steel Structural Members 2007 and AS/NZS 4600:2005 Cold-
formed Steel Structures. It explains that there is only one feasible mechanism for
the limit state of conventional block shear failure, that which involves shear
yielding and tensile rupture. It proposes an equation that provides more accurate
results compared to the code equations in predicting the block shear capacities
of bolted connections in steels having minimal strain hardening. A resistance
factor of 0.8 for the proposed equation is computed with respect to the LRFD
approach given in the North American cold-formed steel specification.
Introduction
1
Senior Lecturer, School of Civil, Mining & Environmental Engineering,
University Of Wollongong, Wollongong, NSW 2500, AUSTRALIA.
2
Graduate Structural Engineer, Hatch, 25 Atchinson Street, Wollongong, NSW
2500, AUSTRALIA. Formerly Honours Student, School of Civil, Mining &
Environmental Engineering, University Of Wollongong, Wollongong, NSW
2500, AUSTRALIA.
585
586
Based on the mechanism identified as the only feasible one and the active shear
resistance planes discussed in the companion paper (Clements & Teh 2012), this
paper presents an equation for determining the block shear capacities of bolted
connections in cold-formed steel sheets. The steel materials used in the
experimental tests had low strain hardening, minimising the “noise” caused by
shear strain hardening on the evaluation of alternative equations.
in which Fy is the yield stress, Agv is the gross shear area, Fu is the tensile
strength, Ant is the net tensile area, and Anv is the net shear area. The regions
corresponding to these areas as defined by the code are shown in Figure 1.
Equation (1) represents the block shear failure by shear yielding and tensile
rupture, while Equation (2) postulates the simultaneous shear and tensile
ruptures mechanism.
Consider the connected end of a flat member shown in Figure 2 that is subjected
to a concentric load and is restrained from out-of-plane failure modes. Leaving
out the pure net section tension failure mode and the bearing failure mode from
the present discussion, there are essentially only two possible failure modes for
the connected end. If the connection shear length (denoted by en in Figure 2) is
relatively short, it will fail by “shear out” of each bolt, a distinct failure mode
illustrated in Figure 3(a).
Section E5.1 of Supplement No. 2 to the North American Specification for the
Design of Cold-Formed Steel Structural Members (AISI 2010) specifies the
shear out capacity Psop of the two-bolt connection in Figure 2 to be
Even at an aspect ratio that is slightly lower than the threshold ratio, a block
shear failure by shear yielding and tensile rupture is still possible as shown in
Figure 4, where the shear-out deformations were over-run by the shear yielding
and tensile rupture mechanism. The change-over in the failure mode took place
when yielding started in the tensile net section between the two bolt holes,
589
where tensile rupture eventually took place. Yielding of the tensile net section
took place following shear strain hardening along the shear-out paths.
As the aspect ratio increases beyond the threshold ratio, block shear failure can
only be due to shear yielding and tensile rupture since the tensile strains are
always more critical than the shear strains. An example of such a failure mode is
shown in Figure 3(b), where tensile rupture took place in the net section between
the two bolt holes. This theoretical exposition is borne out by extensive
experimental tests (Hardash & Bjorhovde 1985, Seleim & LaBoube 1996, Huns
et al. 2006).
Obviously, at an aspect ratio that is sufficiently lower than the threshold ratio,
the shear-out failure mode governs. There is therefore no aspect ratio at which a
block shear failure occurs by the shear rupture and tensile yielding mechanism
postulated in Equation (3).
The present exposition does not account for the situation in which bolt hole
deformations are such that shear rupture could precede tensile rupture. However,
for the specimens tested by Seleim & LaBoube (1996) in which the bearing
failure took place before the block shear failure, the mechanism was still shear
yielding and tensile rupture. In these cases, the strength limit state was actually
bearing failure rather than block shear failure. (It was not possible for the
bearing failures to have followed the block shear failures, but the opposite must
have ensued when the tests were continued well past the ultimate bearing
capacities, resulting in the reduction of the shear resistance area of each block.)
590
It is possible for a connection with a very high aspect ratio in which the shear
resistance dominates to undergo a second limit load following the tensile rupture
that is higher than the first limit load. However, for the purpose of the present
work, the block shear capacity of a bolted connection is defined as the maximum
load preceding the tensile rupture, as represented by the limit load in Figure 5.
limit load
140
120
Applied Load (kN)
100
80
60
40
20
0
0 2 4 6 8 10
Displacement (mm)
Figure 5 Definition of block shear capacity
591
Material and geometric properties for determining the block shear capacity
2. Most evaluated equations use the gross area for computing the yield
resistance component of a block shear capacity. Since in reality the
gross area is not wholly available, such an approach tends to
overestimate the yielding resistance.
As the two factors may offset each other, Equation (1) was often found to
provide the most reasonable (albeit significantly varied) results compared to the
other evaluated equations.
The authors note the experimental evidence of Franchuk et al. (2003), which
suggests that the actual shear failure planes lie midway between the gross and
the net shear planes indicated in Figure 1. The location of these so-called active
shear planes, defined in Figure 6, has been confirmed through geometric and
material nonlinear contact finite element analysis (Clements & Teh 2012).
For the sake of conservatism and simplicity, the present work uses the yield
stress in determining the shear resistance component of the block shear capacity.
For cold-reduced high-strength sheet steels including the G450 steel materials
used in the present work, the ratio of ultimate tensile strength to yield stress can
be significantly below 1.10.
592
n 1 1
Lav Lgv r d h
2 4
n d
Lgv r h 2
Aav 2 Lav t
Figure 6 Tensile and shear resistance planes defined in the present work
Based on the preceding expositions and the conclusion of Teh & Gilbert (2012)
regarding the effect of in-plane shear lag on the tension capacity of a net section,
the block shear failure strength of a bolted connection should be computed from
0 .9 0 . 1 d
R n 0.6 F y Aav Fu A nt
p 2
(5)
in which the active shear area Aav is determined from the length of the active
shear planes shown in Figure 6. The variable d denotes the bolt diameter and p2
the bolt spacing in the tensile resistance plane, defined in Figure 7.
593
Equation (5) incorporates an in-plane “shear lag factor” proposed by Teh &
Gilbert (2012) in determining the net section tension capacity. The shear lag
factor accounts for the fact that the tensile stresses are not uniformly distributed
across the net section, which has a significant effect on the tension capacity of
bolted connections in cold-reduced sheet steel.
Test materials
Two nominal thicknesses were used in the present work, being 1.5 mm and 3.0
mm. The average base metal thicknesses tbase, yield stresses Fy, tensile strengths
Fu and elongations at fracture over 15 mm, 25 mm and 50 mm gauge lengths 15,
25 and 50, and uniform elongation outside fracture uo of the steel materials as
obtained from six 12.5 mm wide tension coupons are shown in Table 1. Tensile
loadings of all coupons and bolted connection specimens are in the direction
transverse to the rolling direction of the G450 sheet steel. The tension coupon
tests were conducted at a constant stroke rate of 1 mm/minute resulting in a
strain rate of about 2 10 4 per second prior to necking.
The tensile strengths in the direction transverse to the rolling direction of 1.5
mm and 3.0 mm G450 sheet steels obtained in the present work, rounded to the
nearest 5 MPa, are 6% and 10% higher than those obtained by Teh & Hancock
594
(2005) in the rolling direction. While Teh & Hancock (2005) did not provide the
ratios of ultimate tensile strength to yield stress, it is believed that the transverse
direction is associated with lower ratios. Any errors or offsetting effects arising
from the neglect of strain hardening in Equation (5) are thus minimised.
Two connection series were tested to investigate the accuracy of the code and
proposed equations in predicting the block shear capacities of simple bolted
connections in 1.5 mm and 3.0 mm G450 sheet steels. Series A comprised
connections having a single row of two bolts, and Series B connections having
two such rows, as shown in Figure 8. For each series of a given sheet thickness,
12 mm and 16 mm bolts were used. The bolt holes were nominally 1 mm larger
than the corresponding nominal bolt diameters. The bolts were only installed by
hand with minimal tightening, and no washers were used in all the tests.
Table 2 lists the relevant geometric dimensions and the test results of Series A
specimens (see Figure 8a for an example) which underwent the block shear
failure mode. All of them duly failed by the shear yielding and tensile rupture
mechanism. The variable dh denotes the nominal bolt hole diameter. Other
variables are defined in Figure 7.
Table 2 shows the ratios of the ultimate test load Pt to the block shear failure
strength Rn predicted by Equations (1), (2), (3) and (5). The first three are code
equations, while the last is proposed in the present work. As explained in the
preceding sections, Equations (2) and (3) do not represent the true mechanism of
block shear failures.
It can be seen from Table 2 that Equation (1), which is used in both cold-formed
steel design codes (AISI 2010, SA/SNZ 2005) for determining the block shear
failure load due to the shear yielding and tensile rupture mechanism,
consistently and significantly overestimates the block shear capacities of Series
A specimens.
Equation (3), which postulates the mechanism of shear rupture and tensile
yielding, overestimates the capacity of some specimens by almost 40%.
Equation (2), on the other hand, overestimates the block shear capacities by up
to 10% “only”. However, it should be noted that this equation postulates the
incorrect mechanism of simultaneous shear and tensile ruptures, while all the
specimens failed by the shear yielding and tensile rupture mechanism.
596
W p2 t e1 dh Pt/Rn
Spec
(mm) (mm) (mm) (mm) (mm) (1) (2) (3) (5)
CPD14 100 33 1.5 50 17 0.80 0.90 0.72 0.95
CPD15 100 33 3.0 50 13 0.90 0.93 0.82 1.01
CPD16 100 33 3.0 50 17 0.89 0.96 0.79 1.04
CPD18 120 40 1.5 50 17 0.86 0.96 0.79 1.00
CPD19 120 40 3.0 50 13 0.90 0.93 0.83 1.01
CPD20a 120 40 3.0 50 17 0.93 0.99 0.84 1.07
CPD20b 120 40 3.0 50 17 0.93 0.98 0.84 1.07
CPD22a 100 26 1.5 50 17 0.81 0.93 0.72 0.96
CPD22b 100 26 1.5 50 17 0.83 0.95 0.74 0.99
CPD23a 100 26 3.0 50 13 0.90 0.93 0.80 1.01
CPD23b 100 26 3.0 50 13 0.89 0.93 0.80 1.01
CPD24a 100 26 3.0 50 17 0.87 0.94 0.76 1.02
CPD24b 100 26 3.0 50 17 0.87 0.94 0.76 1.02
CPD26a 120 26 1.5 50 17 0.85 0.97 0.76 1.01
CPD26b 120 26 1.5 50 17 0.84 0.96 0.75 1.00
CPD27 120 26 3.0 50 13 0.91 0.94 0.81 1.02
CPD28a 120 26 3.0 50 17 0.91 0.98 0.79 1.06
CPD28b 120 26 3.0 50 17 0.89 0.96 0.77 1.04
CPD36 130 45 3.0 30 17 0.94 1.05 0.86 1.13
(Pt/Rn)av 0.88 0.95 0.79 1.02
Table 3 shows the outcomes for Series B specimens (see Figure 8b for an
example) which underwent the block shear failure mode. All of them duly failed
by the shear yielding and tensile rupture mechanism.
597
W p2 t e1 dh Pt/Rn
Spec
(mm) (mm) (mm) (mm) (mm) (1) (2) (3) (5)
CQ2a 120 26 1.5 50 17 0.73 1.01 0.82 0.92
CQ2b 120 26 1.5 50 17 0.74 1.02 0.84 0.93
CQ3 120 26 3.0 50 13 0.85 1.00 0.89 1.00
CQ4 120 26 3.0 50 17 0.80 1.02 0.86 0.99
CQ5a 130 40 1.5 30 13 0.82 1.04 0.91 0.99
CQ5b 130 40 1.5 30 13 0.81 1.02 0.89 0.98
CQ6a 130 40 1.5 30 17 0.77 1.08 0.88 0.98
CQ6b 130 40 1.5 30 17 0.77 1.09 0.88 0.99
CQ7 130 40 3.0 30 13 0.89 1.07 0.96 1.07
CQ8 130 40 3.0 30 17 0.83 1.13 0.94 1.06
CQ9b 130 55 1.5 30 13 0.81 1.00 0.89 0.97
CQ10a 130 55 1.5 30 17 0.78 1.04 0.89 0.98
CQ10b 130 55 1.5 30 17 0.80 1.06 0.90 1.00
CQ11 130 55 3.0 30 13 0.87 1.02 0.94 1.03
CQ12 130 55 3.0 30 17 0.85 1.10 0.96 1.06
(Pt/Rn)av 0.81 1.05 0.90 1.00
As is the case with Series A specimens, Equations (1) and (3) consistently and
significantly overestimates the block shear capacities of Series B specimens.
The major reason is the use of the gross area in computing the tensile or shear
yielding resistance component of the block shear capacity. This effect is likely
to have been hidden to various extent in the experimental tests of bolted
connections in hot-rolled steel plates by considerable strain hardening due to the
very high ratios of ultimate tensile strength to yield stress (Fu/Fy). In certain
cases, it might have also been hidden by the much higher strain rates incurred
during the bolted connection tests compared to the tension coupon tests, or by
the bolt friction resistance.
598
Equation (2), which postulates the simultaneous shear and tensile rupture
mechanism, predicts lower capacities for Series B specimens compared to the
proposed Equation (5) despite its use of the tensile strength Fu rather than the
yield stress Fy in computing the shear resistance. The conservatism is due to the
over-reduced shear area Anv, the effect of which increases with increasing
number of bolt rows as the difference between the net and the active shear areas
widens while the shear resistance becomes more important relative to the tensile
resistance.
Equation (5), in conjunction with the active resistance planes defined in Figure
6, predicts the block shear capacities of Series A and B specimens with the
greatest accuracy. It was found that in order to achieve or exceed the target
reliability index 0 of 3.5 in the LRFD, a resistance factor of 0.83 is required.
Conclusions
The shear yielding and tensile rupture mechanism is represented by one equation
common to the North American and the Australasian cold-formed steel design
codes. This equation uses the gross shear area in determining the shear
resistance to block shear failure, and therefore overestimates the block shear
capacities of all specimens tested in the present work.
The other equation in the North American specification, which anticipates the
simultaneous shear and tensile rupture mechanism, overestimates the block
shear capacities of the single-row bolted connections, but underestimates those
of the double-row bolted ones.
The other equation in the Australasian standard, which anticipates the shear
rupture and tensile yielding mechanism, overestimates the block shear capacities
of all specimens tested in the present work.
The equation proposed in this paper, which is based on the shear yielding and
tensile rupture mechanism, and which uses the active shear resistance planes that
lie midway between the gross and the net shear planes, and incorporates an in-
plane shear lag factor, has been demonstrated to provide the most consistent and
accurate results in predicting the block shear capacities of the tested specimens.
599
References