You are on page 1of 13

CIB-W18/46-15-4

INTERNATIONAL COUNCIL FOR RESEARCH AND INNOVATION


IN BUILDING AND CONSTRUCTION

WORKING COMMISSION W18 - TIMBER STRUCTURES

CONNECTIONS AND ANCHORING FOR WALL AND SLAB ELEMENTS


IN SEISMIC DESIGN

M Schick
T Vogt
W Seim

University of Kassel, Department of Structural Engineering

GERMANY

MEETING FORTY SIX


VANCOUVER
CANADA
AUGUST 2013
Connections and anchoring for wall and slab
elements in seismic design

Michael Schick
Tobias Vogt
Werner Seim
University of Kassel, Department of Structural Engineering, Chair for Building
Rehabilitation and Timber Engineering, Kurt-Wolters-Straße 3, 34125 Kassel, Germany.

1 Introduction
This paper focus on the consideration of over-strength for connections and
interconnections of light frame elements used for wall and slab structures.
If ductility and energy dissipation under cyclic loading are utilised in the case of
earthquake design, it is indispensable to ensure that all the elements outside the ductile
zone remain elastic. Figure 1.1a shows a typical situation of a timber-framed wall element.
A square element with no vertical loading was chosen for simplification. While the wall
element must reach its load-bearing capacity, the anchoring of the wall remains more or
less elastic. Failure modes of the anchoring, such as stud failure (see Fig. 1.2a) or tie down
rupture (see Fig. 1.2b), are typically brittle.
The engineer follows the load flow from the top to the foundation of the building by using
force-based design methods. Consequently, the anchoring and the wall element itself are
both designed for the same action Ed if there are no further reflections. The characteristic
value for resistance Rk and – assuming an identical partial safety factor for resistance – the
design values Rd are then both on the same level. The hidden reserves of the ductile
element may now cause preliminary failure of the anchoring before any energy dissipation
can be activated.

(a) (b) (c)


Figure 1.1: Test set-up of wall elements (a), hysteretic curves of anchoring (b) and wall
element (c) from testing

1
(a) stud failure (b) tie down rupture
Figure 1.2: Brittle failure modes

Figures 1.1b and 1.1c illustrate that the anchoring should always be designed for a
sufficient over-strength to ensure that there is no brittle failure before the capacity of the
wall element is reached. The specification of EC 8 [1] is “sufficient over-strength” for this
case.

2 State-of-the-art
2.1 Design of light frame elements
The dimensioning of light framed timber walls and connections is based on the lower
bound theorem of the theory of plasticity for the wall and on the upper bound theorem for
the connections.
An application of the lower bound theorem within the theory of plasticity is the design
method for light frame walls according to EC5 [2]. The deformation capacity is provided
here by many single fasteners placed to connect the sheathing materials to the studs. The
maximum load capacity is reached when the fasteners achieve their maximum load
capacity. It is inherent to the application of the lower bound theorem that the maximum
load lies above the load bearing resistance calculated.
The characteristic resistance of dowelled type connections could be described by a function
in relation to the yielding moment My of the fastener and the embedding strength fh of the
connected parts.
R k = R k (M y,k , f h,k ) (1)
Yielding moment and embedding strength which are given as characteristic values in EC 5
are based on experimental data. The appraisal of the characteristic values for the yielding
moment for circular sections needs the diameter d of the fastener and the tensile strength
fu,k of the steel. The pure mechanical approach was modified to consider the bending angle
at a joint slip of 15 mm (see Blass et al. [3]).
M y,k = 0.3 ⋅ d 2.6 ⋅ f u,k (2)

2
Embedding strength values fh for solid timber in nailed connections are defined as
=
f h,k 0.082 ⋅ d ⋅ρ2k (3)
where ρk is the density of the timber. The characteristic values of embedding strength for
oriented strand board (OSB) can be calculated by
f h,k = 0.65 ⋅ d −0.7 ⋅ t 0.1 (4)
where t is the thickness of the sheathing material. Values for gypsum fibre board (GFB)
can be taken from product information or general approval as
7 ⋅ d −0.7 ⋅ t 0.9 .
f h,k = (5)
The characteristic value for the load-bearing resistance of timber-framed wall elements is
mainly defined by the connection between the sheathing board and the timber framing,
R c,k ⋅ bi
Rk = (6)
s
where Rc,k is the characteristic resistance of one single fastener, bi is the length of the wall
element and s is the spacing of the fasteners. Other failure modes, such as shear failure of
the board, tensile and compression failure of the stud, failure in anchoring, or the transfer
of compression forces, can be excluded by a balanced design and detailing.

2.2 Statistics
It is necessary to transfer mean values into characteristic values to ensure that a predefined
safety level is achieved. This applies to material strength as well as to the load-bearing
resistance of structural elements. Figure 2.1a shows the standard deviation of a basic
population. The Gaussian distribution
− (x −µ x )2
1 2⋅σ x
=
f (x) ⋅e (7)
σx ⋅ 2 ⋅ π
can be used to describe a symmetrical distribution of samples.
The standard deviation σx of the basic population is crucial for the form of the distribution;
the smaller the standard deviation, the more slender the curves become. The mean value of
the basic population is μx.
Statistics in engineering practice are mostly applied on sample testing. On the basis of the
results of a number of n sample tests, the mean value x̄n of the random sample is
ascertained by
1 n
xn = ∑ xi .
n i =1
(8)

For simplification, statistical assessment in civil engineering assumes symmetrical


distribution. The standard deviation again describes the form of the frequency distribution.
The standard deviation Sx depending on sample tests is defined by
n
1
Sx = ⋅ ∑ (x i − x n ) . (9)
n − 1 i =1

3
Based on the definition of the confidence interval, EC 0 [4] defines characteristic values
for load-bearing resistance x0.05 as material properties by
Sx
x k = x ⋅ (1 − k n ⋅ ) (10)
x
where kn is a factor to determine the p-fractile of the deviation. The value of kn depends on
whether Sx is equal to σx or not and on the number n of the samples. kn is 1.64 for an
infinite number of test results. Figure 2.1b shows the approach to get characteristic out of
mean values for an infinite series assuming a normal distribution.

(a) (b)
Figure 2.1: Normal frequency distribution of a population (a) and a random sample (b)

2.3 Over-strength
The connections with nails or staples are the part of the light framed shear walls which is
able to dissipate energy in case of an earthquake impact due to the plastic deformation
behaviour of the steel. To reach the plastic behaviour of the shear wall, it is necessary to
ensure that all the other parts do not exhibit preliminary failure.
If the load-bearing capacity of a wall element is taken as the action imposed on the
anchoring and on the connection, then the definition is
R a,0.05 ≥ R w ,0.95 (11)
with
R w ,0.95 =γRd ⋅ R w ,k (12)

according to Figure 1.1 with the over-strength factor γRd. It is also obvious from Figure 1.1
that the calculation of γRd must account for material and mechanical over-strength (“hidden
reserves”) and statistics.
While the over-strength values for steel structures and structural concrete structures are
specified between 1.1 and 1.35, there is no specific information about the over-strength
factors for timber structures in EC 8.
The Canadian standard contains over-strength factors for structural timber construction as
those documented and explained by Mitchell et al. [5]. Over-strength is defined there as a
product of partial factors, considering statistics, hardening effects, and the level of
simplification for the modelling and the load-bearing resistance. The over-strength factor is
defined as 1.7 for nailed shear walls.

4
Jorissen and Fragiacomo [6] derive over-strength factors for dowelled type connections.
The over-strength factor is defined as a product of partial factors considering the
approximation of the analytical formulas used to predict the strength properties and the
material safety factor.
Over-strength factors for different types of specimens are derived between 1.18 and 2.08.
The average value of the over-strength factor is calculated as 1.6.
According to the approach suggested by Jorissen and Fragiacomo, Brühl and Kuhlman [7]
proposed an over-strength factor for beams under bending stress with yielding links of
1.28.
Sustersic et al. [8] determined the same over-strength factors for typical steel connectors
with nails and self-tapping screws as those used for the anchoring of cross laminated
timber (CLT) wall elements. Over-strength is again defined there as a product of partial
factors with values reaching from 1.182 to 2.119.

3 Experimental investigation
Comprehensive testing on connection units, wall elements and anchoring units was carried
out at the University of Kassel. These experimental investigations were part of a project on
the optimisation of wall and slab elements for multi-storey buildings (see
www.optimberquake.eu).

3.1 Tests on connection units


Table 3.1 shows the testing programme and the material parameters of the connection
units. Nails with diameters of 2.5 mm, 2.8 mm and 3.1 mm and staples with a diameter of
1.53 mm were used for the connections. OSB panels, respectively GFBs, with a thickness
of 10 mm and 18 mm were connected to solid construction timber (SCT) of 110 mm x
75 mm.

5
Table 3.1: Connection units – testing programme and material parameter

1)
m: monotonic, c: cyclic according to ISO 16670 [12]
2)
mean values – for OSB according to [9], for GFB according to [10]
3)
mean values from sample testing

The loading protocol of EN 26891 [11] with a rate of 0.2 mm/s was used for the monotonic
tests. The cyclic tests were carried out according to ISO 16670 [12]. The yielding moment,
pull-out resistance of the nails and staples and density of SCT was determined
experimentally. The mean values for embedding strength are taken from internal reports.
The material values are documented in Table 3.1. See Table 4.1 for the test results.
Detailed information about the test series is documented in Deliverable 2A of the research
project Optimberquake [13].

3.2 Full-scale tests


The wall elements for full-scale testing consist of studs with a cross-section of 140 mm x
60 mm and a distance of 625 mm between the studs and bottom and top rails with a cross-
section of 140 mm x 85 mm. OSB panels and GFBs with thicknesses of 10 mm and 18 mm
on one or both sides were used for the sheathing.
The OSB panels are connected with nails with a diameter of 2.8 mm and a distance
between the fasteners of 75 mm. The GFBs are connected with staples with a diameter of
1.53 mm and a spacing of 75 mm. A vertical load of 10 kN/m was applied. The loading
protocol from ISO 21581 [14] with a rate of 1.0 mm/s for the monotonic tests comes into
play. The cyclic tests were carried out according to ISO 21581 and CUREE (basic loading
history) [15]. The testing programme is shown in Table 3.2. The results of these tests are
given in Table 4.2. Detailed information about the test series is documented in Deliverable
2B of the research project Optimberquake [16].

6
Table 3.2: Full-scale wall elements – testing programme

4 Statistical evaluation of test results


4.1 Relation of characteristic values and test results
Characteristic values calculated according to EC 5 are based on the assumption that 95% of
the basic population will reach these characteristic values. This applies to structural
elements as well as connections. On the other hand, there will be a distribution of test
results with a mean value – in most cases – higher than expected and a form more or less
slender compared to the distribution of calculated values (see Figure 4.1). Three steps are
proposed to explain the difference between the characteristic values according to EC 5 and
the mean values from testing. Firstly, the expected mean values of the connections have to
be calculated according to EC 5 by using the material properties as mean values (see Table
3.1).
R ∗m = R ∗m (f h,m , M y,m ) (13)

The “hidden reserves” can then be specified as the difference between R ∗m from the
calculation and Rexp,m from the test results.
Finally, the quantile value can be calculated depending on the standard deviation and the
number of test results.

7
(a) (b)
Figure 4.1: Capacity as calculated and from testing for connections (a) and
wall elements (b)
Figure 4.1 shows the distribution of calculated values and the distribution of the
experimental data with
Rk design value according to code provisions
R*m mean value of resistance by using the mean values of material properties
Rexp,m mean value of capacity from testing
Rexp,0.95 95% quantile from testing
In order to get the Rexp,0.95 values from characteristic values, according to the design codes,
the over-strength factor γRd is introduced:
γ Rd = γ mat ⋅ γ mech ⋅ γ 0.95 (14)
where γmat considers the spread between the characteristic values according to design
provisions and the calculated values using the material properties as mean values:
R ∗m
γ mat = . (15)
Rk
The mechanical effects are considered by the factor γmech. Mechanical effects resulting in
an increase of the load-bearing capacity are mainly due to friction and pull-out resistance
of the fastener. Furthermore, the underestimation due to the application of the lower bound
theorem is included within the mechanical effects.
R exp,m
γ mech = . (16)
R ∗m
In order to consider 95% quantile values, γ0.95 is defined as
Sx
γ 0.95 =1 + k n ⋅ (17)
R exp,m
resulting from
Sx
= R exp,m ⋅ (1 + k n ⋅
R exp,0.95 ) (18)
R exp,m

8
4.2 Connection units
Table 4.1 depicts the test results, characteristic values according to EC 5 and the mean
values calculated based on the European yielding model by using the mean values of
material properties. The over-strength factors are calculated according to section 4.1.
The γ0.95 values are calculated based on the assumption that the standard deviation is
known for the basic population.
Table 4.1: Test results and over-strength factors for connection units

4.3 Wall elements


The results of the testing series on wall elements are documented in Table 4.2. The
capacity was calculated based on a plastic model (see section 2.1).
R ∗m ⋅ bi
R ∗m,w = (19)
s
where
R ∗m calculated capacity for the fasteners according to EC 5 by using the mean values of
the material properties
bi length of the wall element
s spacing of the fasteners
Wall elements of test series IV are tested with an eccentric anchorage, thus, the test results
are not completely comparable with the test results of other test series. Therefore, the
factor γmech was determined for every single test for this series. For the other test series,
γ0.95 was determined according to EC 0 based on the assumption of a normal distribution
and of a known standard deviation.

9
Table 4.2: Test results and over-strength factors for wall elements

5 Proposal and outlook


Based on the experimental data, over-strength factors for connections and timber-framed
wall elements can be defined to ensure ductile behaviour. The γ values are comparatively
well distributed for all connection units and wall elements tested, hence, it seems
acceptable to use the mean values as determined for the partial factors to define over-
strength for connections and timber-framed wall elements.
In order to consider the spread between the capacity of the connection as calculated by
using the characteristic or mean values of material properties, the partial over-strength
factor is set as
γ mat =
1.30 . (20)
The “hidden reserves” based on mechanical effects which are not considered in design
codes but increase the capacity of the structural elements are rated by γmech. Based on the
results from wall testing (see Table 4.2), this factor can be defined as
γ mech =
1.33 . (21)
The factor considering the statistic effect to create 95% percentiles for the capacity of wall
elements can be set as
γ 0,95 =
1.28 . (22)

Multiplication of partial over-strength factors leads to


γ Rd = 1.3 ⋅1.33 ⋅1.28 ≅ 2.20 . (23)

10
The mechanical over-strength factor γmech can be decreased to 1.0 if the same mechanical
over-strength is expected for the wall element and connection. The over-strength then
decreases to
γ Rd= 1.30 ⋅1.0 ⋅1.28 ≅ 1.65 . (24)
This proposal is restricted to the sheathing materials and connections used for testing,
documented in Chapter 3. Additional testing and statistical assessment of test results from
literature will help to enlarge the application area.
It is possible that different over-strength factors could be set for different sheathing
materials and connections by increasing the number of test results.

Acknowledgements
The testing programme was part of the European research project Optimberquake, which is
currently being carried out with participants from Belgium, Italy and Germany. The project
is funded under the Cornet Programme of the European Union. The German part is
supported by the “aiF – Arbeitsgemeinschaft industrieller Gemeinschaftsförderung”. The
“ivth – Internationaler Verein für technische Holzfragen” is in charge of the financial
management. Co-funding comes from “Holzbau Deutschland”, “Qualitätsgemeinschaft
Holzbau und Ausbau”, “Verband der deutschen Holzwerkstoffindustrie”, Bundesverband
der Gipsindustrie”, “Deutscher Holzfertigbau-Verband” and the “Studiengemeinschaft
Holzleimbau”.

11
6 References
[1] DIN EN 1998-1: Eurocode 8: Design of structures for earthquake resistance – Part
1: General rules, seismic actions and rules for buildings; German version EN
1998-1:2004. April 2006.
[2] DIN EN 1995-1-1: Eurocode 5: Design of timber structures – Part 1-1: General –
Common rules and rules for buildings; German version EN 1995-1-1:2010-12.
December 2012.
[3] Blass H.J., Bienhaus A., Krämer V.; Effective bending capacity of dowel-type
fasteners. Proceedings PRO 22, International RILEM Symposium on Joints in
Timber Structures, pp. 71-80, 2001.
[4] DIN EN 1990: Eurocode: Basis of structural design; German version EN
1990:2012. December 2012.
[5] Mitchell D., Tremblay R., Karacabeyli R., Pauetre P., Saatciouglu M., Anderson
D.L.; Seismic force modification factors for the proposed 2005 edition of National
Building Code of Canada, Canadian Journal of Civil Engineering, pp. 308-327,
2003.
[6] Jorissen A., Fragiacomo M.; General notes on ductility in timber structures,
Engineering Structures, pp. 2987-2997, 2001.
[7] Brühl F., Kuhlmann U., Requirements on ductility in timber structures, CIB W18,
Alghero, 2011.
[8] Sustersic I., Fragiacomo, M., Dujic B.; Influence of connection properties on the
ductility and seismic resistance of multi-storey cross-lam buildings, CIB W18,
Alghero, 2011.
[9] Kessel M.H., Huse; Vergleichende Betrachtungen europäischer Bauprodukten-
normen mit nationalen Bestimmungen. Teilprojekt: Lochleibungsfestigkeit von
OSB-Platten. Abschlussbericht 3/1998, 1998.
[10] Polatschek T.M., Kessel M.H.; Zur Erweiterung der allgemeinen bauaufsichtlichen
Zulassung Z-9.1-434 von FERMACELL-Gipsfaserplatten – Expertise. Not
published, 2003.
[11] DIN EN 26891: Timber structures – Joints made with mechanical fasteners –
General principles for the deformation of strength and deformation characteristics
(ISO 6891:1983) – German version EN 26891:1991. July 1991.
[12] ISO 16670: Timber structures – Joints made with mechanical fasteners – Quasi-
static reversed-cyclic test method, 2003.
[13] Seim W., Vogt T.; Optimberquake - Deliverable 2A: Connection units – monotonic
and cyclic testing. Technical Report, University of Kassel, 2012.
[14] ISO 21581: Timber structures – Static and cyclic lateral load test method for shear
walls, 2010.
[15] Krawinkler H., Parisi F., Ibarra L., Ayoub A., Medina R.; Development of a Testing
Protocol for Woodframe Structures. CUREE Publication No. W-02, 2001.
[16] Seim W., Vogt T.; Optimberquake – Deliverable 2B: Timber framed wall elements
– monotonic and cyclic testing. Technical Report, University of Kassel, 2013.

12

You might also like