Professional Documents
Culture Documents
Test For Seismic Qualification
Test For Seismic Qualification
Test For Seismic Qualification
Abstract
The qualification of post-installed anchors for use in seismic environments in the U.S.
has been addressed independently by a number of different groups, including ICBO
Evaluation Service, Inc., the Structural Engineers Association of Southern California
(SEAOSC), ACI, and the telecommunications industry. Some of these methods have
been derived from approaches developed in the nuclear industry to account for a variety
of possible events (earthquake, explosion, impact).
The primary focus of most of the test methods current and proposed is the response of
the installed anchor to external cyclic loading, tension and shear. With the exception of
the NEBS criteria, which consists of shake-table testing, strain rate effects are not taken
into account. Two methods, the Provisional Test Method developed by ACI Committee
355 and the German nuclear standard developed by the Deutsches Institut für
Bautechnologie (DIBt), explicitly consider damage to the concrete in the form of a static
crack passing through the anchor location. The SEAOSC criteria provides a comparison
of the post-installed anchor with an “equivalent” cast-in-place headed anchor, and results
in load-displacement information (stiffness degradation, total slip) for cyclic loading
throughout the entire load range (to failure). Results are presented for one anchor tested
to three of these criteria: ICBO ES AC01 Method 2, SEAOSC, and the German nuclear
standard. Conclusions are drawn regarding the effectiveness of the respective test
methods.
Strong ground motion associated with earthquakes can be defined in terms of strain rate
− 5 − 2
( 10 < ε < 10 ) , number of cycles (typically < 30) and displacement (from several
centimeters to a meter or more). Previous studies indicate that the strain rates associated
Inertial forces generated in structures by strong ground motion are more difficult to
characterize. Historically, the design of structures for earthquake resistance has focused
on collapse prevention. For this purpose, design forces were defined as a percentage of
the building mass, typically at levels that are far below expected inertial forces. Implicit
in this approach is an expectation of structure overload with attendant member yielding
and stiffness degradation.
More recently, design methods are derived from a multi-level performance concept (life-
safety, damage limitation, continued operability, etc.) that places greater emphasis on
design for ‘real’ force levels. Material resistances are simultaneously derived to
represent ‘real’ ultimate strengths.1 Clearly, expectations of force and displacement
demands beyond the elastic range continue to form the basis for fixed base (non-
isolated) seismic design of structures.
2. History
Since the 1970s, attention had been focused in within the U.S. nuclear industry on
encouraging ductile failure of anchorages through the design process.2 While generally
practical for the design of cast-in-place anchorages, this approach is often difficult to
implement in the case of post-installed anchors, most of which were/are not designed to
fail in a ductile manner. A nation-wide review of as-built conditions at U.S. nuclear
facilities in the early 1990s focused on best-guess estimations of anchor static capacity
as a means of retroactively qualifying anchorages for seismic loads. Combined with the
severe loading criteria that had been established for nuclear construction, it was believed
that this approach contained sufficient conservatism to avoid further seismic
qualification testing of anchors.
tension
30%Fy Ni = 1/2(Ns-Nm)+Nm
15%Fy Nm = 25%Nu
16%Fy
Vs= 50%Vu
12%Fy
8%Fy
Vi = 1/2(Vs-Vm)+Vm
4%Fy
Vm= 25%Vu
shear
shear
cycles
10 cycles
200 100
80 30
30
30
The telecommunications industry likewise began to review the seismic requirements for
post-installed anchors in connection with the installation of then state-of-the-art digital
switching installations in the early 1980s. Both static and shake table testing of specific
Prior to 1997, testing of post-installed anchors for seismic performance was not common
practice outside of the nuclear and telecom industries. Based on long-standing tradition,
post-installed anchors were routinely listed by the International Conference of Building
Officials Evaluation Service (ICBO ES) as suitable for wind and seismic loading based
on static load tests in uncracked, unreinforced concrete specimens. Additionally, as late
as 1987, increases in allowable loads of 33% for seismic loads were granted in
conformance with applicable sections of the Uniform Building Code pertaining to short-
duration loading.
1. The number of cycles is significantly reduced. This was done to reduce the
probability of fatigue failure in the test.
2. The peak load level was reduced to 150% of the maximum allowable design load,
which in turn is limited to 133% of the static design load. Taken together, these
limits typically result in a peak load on the anchor equal to twice the static allowable
value, or roughly one-half of ultimate. Given that ultimate strength in tension for
most post-installed anchors is limited by concrete cone breakout, this often
represents a significant reduction from the CSA Standard (60% Fy, bolt).
In 1991 ICBO ES ceased authorizing the use of post-installed anchors for use in tensile
zones. This issue was not re-visited in the context of seismic loading, and hence the
seismic qualification tests for both mechanical and bonded anchors6 are performed in
uncracked, unreinforced concrete specimens.
Concurrent with the development of seismic qualification testing at ICBO, the Structural
Engineers Association of Southern California (SEAOSC) proposed a test standard based
on the assumption that the historical provisions for cast-in-place anchors in the Uniform
1. monotonic tension loading to failure in 1.5 mm wide parallel crack (crack width
constant over depth of test member);
2. 15 tension load cycles ( f ≤ 1 Hz ) in 1.5 mm wide parallel crack; and
3. 10 crack opening and closing cycles (1.0 mm to 1.5 mm) with a constant tension
load applied to the anchor, followed by loading to failure with the crack width held
constant at 1.5 mm.
In addition, 15 shear load cycles in a 1.0-mm parallel crack are performed, followed by
shear loading to failure.
The extreme crack widths required (an earlier variant required a maximum crack width
of 2.0 mm) coupled with relatively severe pass/fail criteria make this the most rigorous
test method currently in existence.
The Hilti HDA belongs to the class of self-undercutting undercut anchors (see Fig. 3).
The anchor is offered in four sizes (M10, M12, M16, and M20) and two shear sleeve
variants (preset and through-set, see Fig. 4). Equipped with an ISO Grade 8.8 bolt, the
HDA is proportioned to exhibit
bolt failure at static tension
ultimate for concrete strengths
greater than 14 MPa and where
full development of the
concrete cone breakout strength
Figure 3 – HDA Self-Undercutting Anchor is afforded. The HDA was
The HDA M12 is equipped with a 12-mm ISO 8.8 bolt and has an effective anchoring
depth of 100 mm. The outside diameter of the HDA M12 is 21 mm.
Five tests each in cyclic tension and shear were conducted in 22 MPa (cylinder strength)
normal weight concrete. In addition, five static tension and shear tests to failure were
performed to establish reference values. All testing was conducted in accordance with
ASTM E488-90.
Table 1 - HDA Allowable Stress Design The average ultimate static tension capacity
Seismic - ICBO ES of the HDA M12 as determined in the
reference tests (all tests resulted in steel
f'c = 2,500 psi
failure) was 70.5 kN with a COV of 0.05%.
Anchor ASD Seismic ASD Seismic Accordingly, the maximum test load for the
Tension Shear seismic qualification test was set at 70.5/2 =
(k) (kN) (k) (kN) 35.3 kN. This translates to a steel stress of
HDA-P M10 3.5 15.6 2.2 9.6 approximately 418 MPa or 65% Fy (52% Fu).
HDA-T M10 3.5 15.6 6.2 27.7 [Note: The allowable earthquake load for
HDA-P M12 5.3 23.6 3.2 14.1 this anchor would be set at 70.5/3 = 23.5
kN.] The anchor survived the seismic test
HDA-T M12 5.3 23.6 6.8 30.3
with all residual strength tests resulting in
HDA-P M16 9.5 42.3 5.6 25.1
steel failure. The maximum peak
HDA-T M16 9.5 42.3 12.2 54.5 displacement was 1.0 mm.
The resulting allowable tension and shear values for seismic loading are shown in Table
1.
Testing of the HDA per SEAOSC Standard Method of Cyclic Load Test for Anchors in
Concrete or Grouted Masonry was conducted at Consolidated Engineering Laboratories
in Oakland, California. Loads were applied with a 98 kN capacity hydraulic actuator
equipped with an in-line load cell and servo-controlled hydraulics. LVDTs were used to
Table 2 - HDA SEAOSC Test Result Summary measure displacement.
Cyclic Tension Results
Mean The test members consisted
Mean Ultimate Displacement @ of concrete blocks with a
Anchor Tension Load Failure
cylinder compressive
n (k) (kN) COV (in.) (mm) strength of between 20 MPa
1/2" X 4" A307 3 11.0 49.1 1.0% 0.2 5.0 and 24 MPa at the time of
HDA-P M10 x 100 5 11.4 50.8 0.4% 0.2 4.5 testing. The blocks were cast
5/8" X 4-1/2" A307 3 17.6 78.2 4.7% 0.2 4.3 with 1/2- (12 mm) and 5/8-
HDA-P M12 x 125 3 16.2 72.1 2.4% 0.3 7.7 inch (16 mm) A307 standard
5/8" X 4-1/2" A307 3 17.0 75.8 4.2% 0.2 5.9
hex head bolts placed at
1 minimum embedment per
HDA-P M16 x 190 4 20.3 90.4 0.2% 0.2 4.5
1997 UBC Table 19D, i.e., 4
inches (102 mm) and 4-1/2
Cyclic Shear Results inches (114 mm),
Mean
Mean Ultimate Displacement @ respectively. Block
Anchor Shear Load Failure dimensions were 152 cm x
n (k) (kN) COV (in.) (mm) 91 cm x 61 cm. They were
1/2" X 4" A307 3 5.2 23.1 7.9% 0.2 4.5
cast vertically, with the bolts
2 placed in the side forms to
HDA-P M10 x 100 3 6.0 26.7 0.3% 0.2 5.1
provide equivalent casting
5/8" X 4-1/2" A307 3 9.0 40.1 7.3% 0.3 6.6 conditions for all anchors.
3
HDA-T M10 x 100 3 15.1 67.0 0.2% 0.4 9.6 The HDA anchors were
1
value limited by actuator capacity subsequently installed in the
2
bolt only cured concrete blocks
3
shear sleeve engaged adjacent to the cast-in-place
bolts, with sufficient spacing
All tests resulted in steel failure at ultimate, with the exception of the HDA M16
anchors, which were not tested, to failure.
Test results are provided in Table 2. Sample load-displacement curves are shown in Fig.
5, and the allowable loads implied from the test data are given in Table 3.
Testing of the HDA for structural applications in German nuclear facilities was
conducted at the University of Stuttgart according to the DIBt Guideline For Evaluating
Anchor Fastenings For Granting Permission In Individual Cases According To The State
Structure Regulations Of The Federal States. Tests were conducted with three HDA
sizes, M10, M12 and M16. Three types of tension tests were conducted:
12
11
10
HDA M10 kN
9
8 1/2” A307 std
hex bolt HDA M10
7
6
5
4
3
2
1
0
0 0.05 0.10 0.15 0.20 0.25 in mm
Figure 5 – Sample SEAOSC Seismic Test Load- Figure 6 – DIBt Tension Cycle in 1.5
Displacement Curves mm Crack
result. The greatest differences in the allowable loads occur in shear, whereby the
increased number of cycles associated with the ICBO ES Method 2 test restricts the
shear value, essentially as a function of low-cycle fatigue. As discussed above, this is
true to the original concept of the Canadian standard on which the ICBO test is based.
Of greater interest, however, is the information to be derived from the test aside from
simple pass/fail results. Figure 5 shows typical load-slip curves derived from the
SEAOSC seismic cycle. Note that it is possible to derive stiffnesses from this data
throughout the entire load range that reflects a realistic numbers of cycles. While the
German standard provides similar information (see Fig. 6), the information provided
regarding the response at near-ultimate load levels is limited.
10. Conclusions
Documented anchor failures in past earthquakes are likewise in many cases attributable
to shear overload, as opposed to pullout.
Critical for an anchorage design that considers stiffness and material interaction in the
detail design is understanding the likely response history of the components being
considered. In the case of the anchor, such information can only be provided from
testing that mimics the essential components of strong motion and measures the required
response parameters in a way that is useful for design.10 While the recently published
ACI 355.2-00 test method11 includes cracking and is therefore a dramatic improvement
over previous criteria, it continues to be based on a low cycle fatigue loading regimen.
12. Summary
Three methods for qualification of anchors for seismic loading are currently in use in the
U.S. and Europe. Taken together, these methods encompass the effects of cyclic loading,
base material damage, and anchor overload. An undercut anchor has been tested using
the three standards, and the results provide a limited basis for evaluation of the test
methods. A test method that combines the best elements of the three current methods is
required to meet the requirements of future design codes.
References
*
It should be noted that a cycling crack test would more closely approximate the
conditions associated with a structure subjected to strong ground motion, although the
practical implications of such a test have not been adequately explored to date.