Professional Documents
Culture Documents
Analysis and Design of Steel Deck - Concrete Composite Slabs PDF
Analysis and Design of Steel Deck - Concrete Composite Slabs PDF
Analysis and Design of Steel Deck - Concrete Composite Slabs PDF
DOCTOR OF PHILOSOPHY
in
Civil Engineering
W. S. Easterling, Chairman
R. M. Barker
E. G. Henneke
S. M. Holzer
T. M. Murray
October, 1997
Blacksburg, Virginia
Keywords: composite slabs, direct method, iterative method, finite element model, long span, resistance factor
ANALYSIS AND DESIGN OF
STEEL DECK – CONCRETE COMPOSITE SLABS
by
Budi R. Widjaja
Dr. W. S. Easterling, Chairman
Department of Civil Engineering
(ABSTRACT)
As cold-formed steel decks are used in virtually every steel-framed structure for
composite slab systems, efforts to develop more efficient composite floor systems continues.
Efficient composite floor systems can be obtained by optimally utilizing the materials, which
includes the possibility of developing long span composite slab systems. For this purpose, new
deck profiles that can have a longer span and better interaction with the concrete slab are
investigated.
Two new mechanical based methods for predicting composite slab strength and behavior
are introduced. They are referred to as the iterative and direct methods. These methods, which
accurately account for the contribution of parameters affecting the composite action, are used to
predict the strength and behavior of composite slabs. Application of the methods in the
analytical and experimental study of strength and behavior of composite slabs in general reveals
that more accurate predictions are obtained by these methods compared to those of a modified
version of the Steel Deck Institute method (SDI-M). A nonlinear finite element model is also
developed to provide additional reference. These methods, which are supported by elemental
tests of shear bond and end anchorages, offer an alternative solution to performing a large
number of full-scale tests as required for the traditional m-k method. Results from 27 composite
slab tests are compared with the analytical methods.
Four long span composite slab specimens of 20 ft span length, using two different types
of deck profiles, were built and tested experimentally. Without significantly increasing the slab
depth and weight compared to those of composite slabs with typical span, it was found that these
long span slabs showed good performance under the load tests. Some problems with the
vibration behavior were encountered, which are thought to be due to the relatively thin layer of
concrete cover above the deck rib. Further study on the use of deeper concrete cover to improve
the vibrational behavior is suggested.
Finally, resistance factors based on the AISI-LRFD approach were established. The
resistance factors for flexural design of composite slab systems were found to be φ=0.90 for the
SDI-M method and φ=0.85 for the direct method.
In Memory of my Father
and
In Love of my Mother
ACKNOWLEDGMENTS
I am most grateful to Dr. W. Samuel Easterling for his continuous support, guidance and
friendship throughout my graduate study at Virginia Polytechnic Institute and State University
(Virginia Tech). I would also like to express my sincere appreciation to the members of the
research committee, Dr. R. M. Barker, Dr. E. G. Henneke, Dr. S. M. Holzer and Dr. T. M.
Murray. Special thanks goes to Dr. R. M. Barker for his valuable discussion on the resistance
factors and to Dr. T. M. Murray for his valuable discussion on floor vibrations.
I gratefully acknowledge financial support from the National Science Foundation, under
research grant no. MSS-9222064, the American Institute of Steel Construction, the American
Iron and Steel Institute, Vulcraft and Consolidated System Incorporated. Further, material for
test specimens was supplied by BHP of America, TRW Nelson Stud Welding Division and
United Steel Deck. My sincere thanks is also for the Steel Deck Institute for the Scholarship
Award that I received for my research and very special thanks to Mr. and Mrs. R. B. Heagler for
their very warm hospitality during my visit at the SDI annual meeting in Florida. Mr. Heagler
also keeps me updated with new technical issues and developments in the SDI.
I would also like to thank to Dr. M. Crisinel and Dr. B. J. Daniels for the access to use
the COMPCAL program at the Ecole Polytechnique Federale de Lausanne, Switzerland. They
also allowed me to use the drawings for the elemental tests.
To all my friends in the Civil Engineering Department and especially those at the
Structures and Materials Laboratory of Virginia Tech, I extend my appreciation for their support,
discussion and friendship. I am particularly indebted to Joseph N. Howard for his immeasurable
help in performing the vibration tests on the long span slabs. Special thanks goes to Dennis W.
Huffman and Brett N. Farmer for their constant help and cheerful support during my research
work at the Structures Lab.
Last but certainly not the least, I am thankful to my wife, Surjani, for being a constant
source of inspiration and encouragement. She is a wonderful wife and friend.
TABLE OF CONTENTS
ABSTRACT.......................................................................................................................... ii
DEDICATIONS.................................................................................................................... iv
ACKNOWLEDGMENTS .................................................................................................... v
TABLE OF CONTENTS...................................................................................................... vi
LIST OF FIGURES............................................................................................................... ix
LIST OF TABLES ................................................................................................................ xi
LIST OF NOTATIONS ........................................................................................................ xii
Chapter 1. Introduction
1.1. Motivation and Scope of the Research ......................................................................... 1
1.2. Organization of this Report .......................................................................................... 3
vi
3.2. Review of Methods of Prediction of Composite Slab Strength by Means of
Semi-Empirical Formulations and Simple Mechanical Models ................................... 19
3.3. SDI-M Method.............................................................................................................. 27
3.4. Iterative Method............................................................................................................ 27
3.5. Direct Method............................................................................................................... 34
3.6. Comparison of Calculated and Test Results................................................................. 35
3.7 Concluding Remarks..................................................................................................... 40
vii
Chapter 6. Reduction Factor, ϕ
6.1. General.......................................................................................................................... 71
6.2. Review of Probabilistic Concepts of Load and Resistance Factor Design .................. 71
6.2.1. Reliability Index ................................................................................................ 72
6.2.2. AISC LRFD Approach for the Resistance Factor ............................................. 74
6.2.3. AISI LRFD Approach for the Resistance Factor .............................................. 75
6.3. Statistical Data.............................................................................................................. 76
6.3.1. Material Factor, M ............................................................................................ 77
6.3.2. Fabrication Factor F .......................................................................................... 78
6.3.3. Professional Factor, P........................................................................................ 79
6.3.4. Load Statistic..................................................................................................... 79
6.4. The Resistance Factor................................................................................................... 80
6.5. Concluding Remarks..................................................................................................... 82
References............................................................................................................................. 85
VITA ..................................................................................................................................... 96
viii
LIST OF FIGURES
ix
4-10. Slab-15: (a) Load vs. mid-span deflection. (b) Load vs. end-slip............................... 50
4-11. Slab-21: (a) Load vs. mid-span deflection. (b) Load vs. end-slip............................... 51
4-12. Composite slab strength: FEM vs. experimental ........................................................ 51
x
LIST OF TABLES
xi
LIST OF NOTATIONS
Fs + Fst
= (Eqn.(3-24))
0.85 f c ' b
b = section width
C = resultant of concrete compressive force
c = depth of the neutral axis of composite section
Dn = nominal value of dead load
d = distance of the steel deck centroid to the top surface of the slab (effective depth)
= length of each segment
dL, dL i = elongation of the bottom fiber of concrete slab of segment i
A
F = minimum anchorage force (Chapter 3) = f y A s − webs − A bf , (Eqn.(3-8))
2
xii
Fs , Fst = tensile force in the steel deck resulted from the effect of shear bond and end
anchorages respectively
Fs,lim it = upper limit of Fs
f s,max , f s,min
f yc = corrected steel deck yield stress due to concrete casting and shoring
xiii
Ln = nominal value of live load
M m,SDI , M m,Direct
= means of material factor with regard to the SDI and Direct method, respectively
M nc , M nd = nominal moment capacity: phase-1 and phase-2, respectively
pf = probability of failure
xiv
T1 , T2 , T3 = forces acting in top flange, web and bottom flange of steel deck
t = steel deck thickness
tm = mean of steel deck thickness = µ t
u1d = nodal displacement of steel deck beam element in d.o.f.-1 direction (horizontal)
V, VR , VQ
= coefficients of variation of material factor with regard to the SDI and Direct
method, respectively
Vu = ultimate shear capacity
yc = horizontal projection of y d
β = reliability index
ε , ε cu = concrete strain, concrete strain at the peak compressive stress
xv
λ , λ R , λ Q = log-normal mean: general, resistance, load effect
D D
Ψ = γ D n + γ L / 1.05 n + 1 , (Eqn.(6-18))
Ln Ln
∫ M m ds
Ωi = i (Eqn.(3-22))
∫ M m ds
L
xvi
CHAPTER 1
INTRODUCTION
Chapter 1 . Introduction 1
3. To provide structural designers with analytical means by which they can verify design
calculations.
Efficient composite floor systems can be obtained by optimally utilizing materials, which
includes the possibility of developing long span composite slab systems. These long span
systems require investigation of new deck profiles that can be used to provide an adequate
interaction with the concrete slab. However, with the dependency of steel manufacturers on full-
scale slab tests, a substantial number of tests have to be performed to develop a new deck profile.
Therefore, from the manufacturer point of view, an alternative that can reduce the required
number of full-scale tests is desirable. This can be achieved by using analytical means supported
by elemental tests that are less expensive than the full-scale tests. Many kinds of analytical
means are now being made available due to development in the past decade, particularly in the
area of nonlinear analysis. By the same means, structural designers will have analytical tools to
cross-examine the design calculations. Current design formulations, such as the m and k method
(Schuster 1970, Porter et al 1976), do not sufficiently describe the physical behavior of
composite slabs. The only way structural designers can verify the design calculation based on
load tables generated by the m and k method is to look back into the experimental test results.
Depending upon the application, these analytical tools may range from a simple hand calculation
to a special purpose nonlinear finite element code.
As a continuation of on-going research in the area of composite slabs, with the same
motivations as mentioned above, this study has been conducted. New deck profiles, which
enable the deck to span longer than the typical spans currently used, are investigated. By
introducing a longer span floor system some filler beams can be eliminated along with their
connections to the girders. This results in more economical floor systems.
To establish a profile suitable for long spans, analytical models are developed to predict
the behavior of the new slab prior to any experimental tests. Two mechanical based models and
a finite element model are introduced. These models require knowledge of interaction properties
of some components of composite slabs. Hence, elemental tests for the shear bond and end
anchorages are performed. These analytical models, along with the elemental tests, offer an
alternate solution to the full scale tests that are required for the current design procedures.
Additionally, resistant factors, φ, for flexure design of composite slabs are also sought. The
current resistant factors, φ, for composite slab design (Standard for 1992) were taken from the
Chapter 1 . Introduction 2
steel or concrete design specifications. Therefore, it is desired to obtain these factors based on
test results and refined analytical studies of composite slabs.
Chapter 1 . Introduction 3
CHAPTER 2
ELEMENTAL TESTS
2.1. General
Composite slab behavior is a function of interactions among the components of the slab.
Two of the most important interactions that significantly affect the slab behavior are: (1) the
shear bond interaction at the interface of steel deck and concrete and (2) the interaction among
the concrete, steel deck and end anchorages at the supports. Therefore, two types of elemental
tests were conducted in this study: shear bond and end anchorage. The purpose of these tests is
to study more closely the strength and behavior of shear bond interaction and end anchorages.
These tests will also provide interaction data required for the numerical analysis that will be
described in detail in Chapter 3 and Chapter 4. Elemental tests used in this study are similar to
the push-out and pull-out tests by Daniels (1988).
2.2. Review of Research on Elemental Tests for Shear Bond and End Anchorages
The shear bond, or m-k, method requires a substantial number of performance tests for
the shear bond regression line, plus additional flexure tests if flexural failure occurs within the
range of parameters tested. The problem becomes more pronounced with the recent findings of
other parameters that have significant impact on the strength of composite slabs, such as load
pattern, end anchorages and additional reinforcing bars. This finding drastically increases the
number of performance tests the manufacturers have to perform (Daniels and Crisinel 1987,
1993; Patrick 1990; Patrick and Bridge 1990; Patrick and Poh 1990; Bode and Sauerborn 1992).
For shear bond tests, a pair of additional frames is added to induced lateral force (Fig. 2-
4). The lateral force is applied by tightening the nuts in the rods. This lateral force is to simulate
internal pressure that is developed on the interface between the deck and the concrete. Load
cells were installed in the lateral frames, as indicated in Fig. 2-4, to measure the magnitude of the
lateral load applied.
2.5
(2) 3
4.75
7.25
12
1.5 2.5
(3) 2
0.5
6
12
9.25 3
(4)
6
1
3.75 7.125 1.5 0.5
12.875
9 2.5
(5) 4.5
1
tension rod
instrumented with
strain gages
clevis
specimen
Load cells
nuts to adjust
internal pressure
specimen
load cell
Figure 2-4. Shear bond specimen with frames for lateral force
7.0
5.0
4.0
3.0
2.0
1.0
0.0
0.00 0.20 0.40 0.60 0.80 1.00
SLIP (in)
14.0
12.0
SHEAR STRESS (psi)
10.0
8.0
6.0
4.0
2.0
0.0
0.00 0.20 0.40 0.60 0.80 1.00
SLIP (in)
puddle
welds
hydraulic ram
load cell
tension rod
specimen
In EA1 group of specimens, in which the studs were welded through the deck, the typical
response of load vs. slip shows relatively ductile plateau. The failure was due to steel deck
tearing and pilling in front and behind the studs, respectively. In EA2 group of specimens, the
fact that strength of the specimens was considerably lower than in the EA1 was because the studs
were not welded through the deck. Another cause was the relatively short distance of the steel
deck puddle weld to the end of the deck (1.5 in). Therefore, the behavior of EA2 specimens are
similar to those of EA3, where ductile plateau can not be maintained as soon as the deck tearing
12.0
10.0
8.0
LOAD (kips)
6.0
4.0
2.0
0.0
0.00 0.05 0.10 0.15 0.20 0.25 0.30
SLIP (in)
7.0
6.0
5.0
LOAD (kips)
4.0
3.0
2.0
1.0
0.0
0.00 0.05 0.10 0.15 0.20 0.25 0.30
SLIP (in)
3.1. General
One of the purposes of developing simple mechanical based methods for composite slab
strength is to provide tools suitable for design purposes. Methods based on this model have been
developed worldwide in the past two decades (Stark 1978, Patrick 1990, Stark and Brekelmans
1990, Heagler et al. 1991, Bode & Sauerborn 1992, Easterling and Young 1992, Patrick and
Bridge 1994). Despite the complex nature of interactions inside composite slab systems, the
methods have demonstrated good performance in predicting the slab strength. In contrast to the
so-called m-k method, these methods do not rely heavily on full-scale test results, which becomes
the main advantage of the methods.
In this study, two new methods based on simple mechanical model are developed. The
methods are based on partial connection theory. Unified formulation for the studded and non-
studded slabs and inclusion of shear bond strength at the steel deck-concrete interface offer
advancements to the SDI method (Heagler et al. 1991). In comparison to the method developed
by Patrick (1990), the remaining strength of the steel deck beyond the shear bond transfer
strength is considered. On the other hand, clamping forces at the supports are neglected due to
the fact that at the supports, the slab rests on the tip of the supporting beams.
The first of the two new methods is an iterative procedure, in which the slab strength is
calculated based on the location of the critical cross section, i.e., the location of the concrete
crack that initiates shear bond failure. With this method, the ultimate strength and response
mρd
Vu = bd + k fc ' (3-1)
L'
where Vu = ultimate shear capacity obtained from experimental test, b = unit width of the slab, d
= slab effective depth, measured from the compression fiber to the centroid of the steel deck,
ρ = A s bd , L' = shear span length, f c ' = concrete compressive strength, A s = steel deck
cross sectional area per unit width, m and k are parameters shown in Fig. 3-1, obtained by
regression on the values obtained from full scale tests.
REGRESSION
LINE
m
REDUCED
REGRESSION
LINE
m’
Vu
bd f c '
k’
ρd
L fc '
given by:
N b = k. f c '. h b . b (3-3)
N b
M p ' = 1.25M p 1 - ≤ Mp (3-4)
A s . f y
where M u = ultimate bending moment capacity, M p = steel deck plastic moment capacity, b =
slab unit width, f y = steel deck yield stress, d = repeat definition, k, and h b are explained in
Fig. 3-2. Equation (3-4) is a bi-linear simplification of a nonlinear relation between M p ' and
k. f c '
hb Nb
z
fy
Na Mp '
fy
Mp '
Mp
1.0
Nb
M p ' = 1.25 M p 1 - ≤ Mp
0.8 A sfy
0.6
0.4
0.2
In 1991, the Steel Deck Institute (SDI) launched an alternative formulation to predict the
strength of composite slabs for design purposes (Heagler et al. 1991, 1992, 1997, Easterling and
Young 1992). These design procedures were based on research conducted at Virginia
Polytechnic Institute and State University and West Virginia University sponsored by the SDI.
a
M n = R.A s . f y d - (3-5)
2
where
A sf y
a = (3-6)
0.85f c' b
N rQn
R = (3-7)
F
A
F = f y A s − webs − A bf (3-8)
2
studs per unit width of the slab, Q n = nominal shear stud strength, A webs , A bf = area of the
webs and bottom flange of the steel deck, respectively, per unit width of the slab. In the non-
studded slabs, the bending capacity of the slabs is predicted by using the moment at first yield,
which is given by:
M et = (T1e1 + T2 e 2 + T3 e 3 ) (3-9)
where T1 , T2 , T3 are the total forces of the top flange, web and bottom flange of the deck,
Nb
Ls = (3-10)
b. τ shear bond
where L s = shear bond length, N b = normal force developed in the concrete slab (see Fig. 3-4),
b = slab unit width, τ shear bond = shear bond strength at the interface between the steel deck and
concrete. In this case, the shear bond strength is determined from full-scale composite slab tests.
M Nb N b max
Nb = 0 LA
LB
LA LB Ls
Figure 3-4. Boundary curve based on the partial interaction theory
(Bode & Sauerborn 1992)
where f s = shear bond force per unit length, x = distance from the support to the section being
force varies along the slab, then a plot of M n vs. T (reinforcing force provided by the shear
bond, end anchorages, etc.) needs to be generated, as shown in Fig. 3-6, in order to form the
boundary curve for the slab load carrying capacity (Fig. 3-7). This concept is very similar to the
one introduced by Stark and Brekelmans (1990) (compare Fig. 3-6 to Fig 3-3) and Bode and
Sauerborn (1992) (compare Fig. 3-7 to Fig. 3-4). The critical section is then found by matching
up the boundary curve to the bending moment diagram due to the applied load, and the first point
to intersect with the bending moment capacity diagram is the critical location.
fs C
M M
T T
µκ R
x
Figure 3-5. Free body diagram of the forces acting in the composite slab section
(Patrick 1990, Patrick and Bridge 1994)
Mn
δ = 10
.
Figure 3-6. Plot of M n vs. T (Patrick 1990, Patrick and Bridge 1994)
B
A
M
A
Figure 3-7. Boundary curve for the ultimate bending moment capacity
(Patrick 1990, Patrick and Bridge 1994)
The procedure offers a good means that can take into account the shear bond and end
anchorage effect in the determination of the bending moment capacity based on the critical cross
(original steel deck yield stress) in Eqns. (3-5) to (3-9) with f yc (corrected steel deck yield stress
due to concrete casting), and (2) omission of the construction shoring effect in the f yc , thus in
this case, the slab is treated as if it were unshored. Tests on shored composite slabs revealed that
unconservative predictions using the SDI method could be resulted when the shoring effect was
included in this simple model.
qc
fs Fst
Two phases are considered in the analysis: phase-1, analysis of a composite cross section
in which the steel deck acts as a tensile member reinforcing the slab, and phase-2, analysis of the
steel deck as a flexural member. Phase-1 can be regarded as the composite action while phase-2
f1
fc '
c
h1 C(c, f1)
M
T(c, f1)
Fs (c, f1)
Fst (c, f1)
ft
f2 (c, f1)
Two independent variables have to be solved to determine the stress distribution on the
cross section. In Fig. 3-9, c and f1 are chosen as the independent variables. They can be solved
from the two equilibrium equations on the cross section: equilibrium of forces and equilibrium of
moments. The magnitude of Fs and Fst , however, depends upon the value of slip between the
concrete and steel deck which in turn depends on concrete strain at locations where these two
forces are acting. The result is a nonlinear relation between Fs or Fst and the concrete strain,
such that c and f1 are coupled together in a nonlinear system of equations. Therefore, an iterative
procedure is needed to solve for c and f1 . The iterations are performed for each cross section for
a given load level. The greater the number of cross sections considered the more accurate the
prediction of the location of the critical section.
different cross sections. This is illustrated by different values of f s,A and f s,B in Fig. 3-10b.
The shear bond force, Fs , acting on a cross section is the sum of f s from the end of the slab to
the particular cross section (represented by the shaded area in Fig. 3-10b). Figure 3-10c shows
the distribution of Fs along the slab. In the case of high strength shear bond, Fs can not be
fs f s,A
(a)
fs,B
slip
fs,A fs,B
(b) Fs
fs diagram
(c) f yc . A s
Fs limit
Partial interaction between the deck and the concrete is accounted for by limiting the
deck contribution to the capacity of the shear bond, such that after a certain phase, the steel deck
and concrete no longer have the same amount of strain at the interface. Hence, at any loading
point, strength contribution of the deck can not be greater than Fs as shown in Fig. 3-10c, so
that, as reinforcement for the concrete, the steel deck strength can be expressed as:
where Fs = shear bond force, ε s , E s and A s are, respectively, the strain, elastic modulus and
cross sectional area of the steel deck, Fs,lim it = limitation on the shear bond force based on the
shear bond force per unit length vs. slip data obtained from the elemental tests. Note that Fs,lim it
for a cross section does not have a constant value through the loading history, rather, forms a
function of slip at that location. Once the maximum normal stress in the steel deck reaches a
value of Fs,lim it / A s , slip starts to occur. Again, Fs,lim it can not exceed the strength of the steel
Fs,lim it ≤ f yc .A s (3-13)
The effect of the end anchorages, Fst , can be obtained upon the determination of the slip
of the slab relative to the beam at the location of the anchorages, i.e., at the support. Slip values
can be obtained by summing the elongation of the bottom fiber of the concrete for each element
or segment from the mid-span to the support, neglecting axial deformation of the steel deck.
To this end, both shear bond and end anchorage forces require determination of slip
along the slab. This creates a problem because the slip is not known in advance. Two
approaches can be pursued to overcome the problem. One is to apply a forward iteration
scheme, in which, the analysis proceeds by utilizing the values obtained from the last convergent
state. These values might not be correct for the current state, however, the forward iteration
scheme does not require additional iteration. The second approach is to use a backward iteration
scheme. In this scheme an additional iteration loop is introduced inside the current iteration loop
for c and f1 . Computationally, the approach is expensive.
In this study, a forward iteration scheme is applied with an assumed distribution of
bottom fiber elongation of the concrete slab along the length to reduce error introduced by this
integration scheme. The actual distribution of this elongation will have a parabolic shape as
shown in Fig. 3-11b. A simplified distribution by using a linear distribution as shown in Fig. 3-
xi
dL i = dL c (3-14)
L/2
in which, L = the span of the slab, dL i = elongation of bottom fiber of segment-i and dL c =
elongation of bottom fiber at the mid-span. Using Eqn. (3-14), the total slip at location x i can be
expressed as:
(a) L
L+dL
(b) dL diagram
(d) simplified
xi dL diagram
L/2 dLi dLc
Figure 3-11. Concrete bottom fiber elongation, dL, and slip diagrams
n dL c dL c
s i = ∑ dL i = (x i + x i +1 +...+ x n ) = (i + (i + 1)+...+n) d (3-15)
i=1 L/2 L/2
where s i = slip at location x i , n = total number of segments from the support to the mid-span, i
= sequence number of segment counted from the support, and d = length of each segment.
Substituting Eqn. (3-14) into Eqn. (3-15) for dL c , and replacing (i + (i + 1) +...+ n) in Eqn. (3-
n(n + 1) i(i − 1) 1
si = − dL i (3-16)
2 2 i
In phase-2 of the analysis, the remaining strength of the deck beyond its strength that has
been used for shear bond transfer is considered. This strength of the deck contributes additional
load carrying capacity and it is assumed that this action occurs through a non-composite type of
action. For this purpose, a deflection compatibility condition is assumed between the deck and
the concrete as illustrated in Fig. 3-12:
qc
dc
qd
ds
(b) non-composite action
ds = dc (3-17)
in which, d s = steel deck deflection, and d c = composite slab deflection. Additional strength
stemming from phase-2 of the analysis is contributed from the flexural strength of the deck and it
can be significant. The stress developed in the steel deck in conjunction with this additional
strength, however, can not be greater than the remaining strength available in the steel deck given
by:
shore removal, Fs (shear bond force), Fst (end anchorage force), and weld force, respectively. If
q d denotes the additional load carrying capacity, then the total load carrying capacity is simply:
q = qc + qd (3-19)
in which q c = load carrying capacity from phase-1 of the analysis (partially composite action).
Beyond this value, the deck is yielded and it deforms plastically without adding any contribution
on the load capacity.
Deflection of the slab can be computed simultaneously with the strength calculation. In
this part of analysis, however, there are additional assumptions required. The modulus of
elasticity of concrete is assumed unchanged and equal to its initial value, even though the
concrete is in an inelastic state in certain cross sections. Similar to the strength procedure, the
portion of the concrete stressed beyond the tensile stress limit is considered to be ineffective.
Therefore, the cross sectional inertia of the concrete varies along the slab. The contribution of
steel deck stiffness to the slab stiffness is proportional to the degree of interaction between the
deck and the concrete. This degree of interaction is represented by the ratio of steel deck stress
to the corrected steel deck yield stress at the beginning of the analysis (after concrete casting and
shore removal). With this, the slab will have a non-prismatic effective cross section. The
deflection can then be computed by utilizing the unit load method for which the integration can
be performed numerically. The effective cross sectional inertia can be computed from:
Mm M m M m M m
δ = ∫L ds = ∫1 1 1 ds + ∫2 2 2 ds + ... + ∫n n n ds (3-20)
EI eff EI 1 EI 2 EI n
where δ is the mid-span deflection of the slab, M and M i ’s are moment functions along the slab
and at segment-i, respectively, due to the applied load, m and m i are moment functions along the
slab and at segment-i, respectively, due to a unit load at the mid-span, I i is the effective inertia
of segment-i and I eff is the average of the effective inertia of the slab. By assuming that the
cross sectional inertia does not vary within each segment, then Eqn. (3-20) can be reduced to:
where
∫ M m ds
Ωi = i (3-22)
∫ M m ds
L
with ∫ = integration over the segment, ∫ = integration over the entire length of the slab, M =
i L
bending moment function along the slab, and m = weighting function (bending moment caused
by the unit load).
0.85 fc '
C
y1 y2
M
Fst
Fs
The main advantage of the direct method is that the procedure of computation is non-iterative,
thus it is convenient for hand computation. The effects of shear bond and end anchorages can
M nc = Fs y1 + Fst y 2 (3-23)
where y1 , y 2 = the moment arm length of Fs and Fst , respectively, to the center of the
compressive stress block. The depth of the stress block is obtained from:
Fs + Fst
a = (3-24)
0.85f c' b
Equation (3-23) constitutes phase-1 of the analysis. Phase-2 of the analysis, the effect of the
flexural deck strength, is given by:
M nd = f y*S (3-25)
where fy* = the remaining deck strength, defined by Eqn. (3-18), and S = section modulus of the
steel deck. In contrast to the iterative method, the response history of the system can not be
obtained. The result only gives the nominal moment capacity. From Eqns. (3-23) to (3-25), it
can be noted that there is no distinction in the formulations whether the slab is studded or not.
The fact that the steel deck strength is limited to the shear bond action in the composite action
(phase-1) and the inclusion of the remaining strength of the deck represent a more realistic
physical interaction in composite slab. This gives a more accurate account for the changes in
steel deck strength such as shoring effect during the construction, etc.
SLAB DECK RIB STEEL EMBSM. OVER- SPAN END TOTAL DECK SHORING CONCR
# PROF. HT. THCK. TYPE HANG LENGTH ANCHR. DEPTH CONT. fc'
(in) (in) (ft) (ft) TYPE (in)
1 1 2 0.0345 1 1 9 S-5 4.5 C N 3180
2 1 2 0.0345 1 1 9 S-4 4.5 C N 3180
3 1 2 0.0345 1 1 9 S-3 4.5 C N 5170
4 1 2 0.0345 1 1 9 S-2 4.5 C N 5170
5 1 2 0.0345 1 1 9 W-7 4.5 C N 3340
6 1 2 0.0345 1 _ 9 W-7,P 4.5 C N 3340
7 1 2 0.0345 1 1 9 W-7 4.5 D N 3770
8 1 2 0.0345 1 _ 9 W-7,P 4.5 D N 3770
9 1 2 0.0470 2 1 9 S-3 4.5 C N 5300
10 1 2 0.0470 2 1 9 S-5 4.5 C N 5300
11 2 3 0.0355 3 1 10 S-3 5.5 C N 3750
12 2 3 0.0355 3 1 10 S-5 5.5 C N 3750
13 2 3 0.0355 3 1 10 W-7 5.5 D N 3370
14 1 2 0.0470 2 1 9 W-7 4.5 D N 3370
15 3 2 0.0335 _ 1 9 S-3 5.0 C Y 3800
16 3 2 0.0335 _ 1 9 S-6 5.0 C Y 3800
17 3 2 0.0335 _ 1 13 S-4 6.0 C Y 2780
18 3 2 0.0335 _ 1 13 W-6 6.0 D Y 2780
19 2 3 0.0339 3 1 9 W-7 5.5 D Y 3900
20 2 3 0.0339 3 1 9 W-7 5.5 D N 3900
21 2 3 0.0558 3 1 12 W-7 5.5 D Y 5120
22 2 3 0.0558 3 1 12 W-7 5.5 D Y 4550
23 2 3 0.0558 3 1 12 W-7 5.5 D N 4550
24 4 6 0.0560 _ 1 20 S-6 8.5 D N 3070
25 4 6 0.0560 _ 1 20 S-6 8.5 D N 3070
26 5 4.5 0.0570 _ 1 20 S-6 7.0 C N 2330
27 5 4.5 0.0570 _ 1 20 S-6 7.0 C N 2330
Note
* End anchorages: S=stud, P=pour stop, W=puddle weld
* The number following S and W is the number of studs or welds installed
* Deck continuity: C=continuous over the support, D=discontinuous
* Deck profiles and embossment types: refer to Fig. 2-1 and 2-2, respectively
From Table 3-2, it can be observed that the iterative and direct methods predicted the
capacity of the slab reasonably well. The SDI-M method tends to give conservative predictions.
A graphical comparison of the test vs. predicted strengths using the iterative and direct methods
are shown in Fig. 3-15.
A comparison of the experimental and iterative method response histories for slab-4
(studded slab with trapezoidal deck profile), slab-15 (studded slab with re-entrant deck profile)
AIR BAG
1000 1000
TEST (psf)
TEST (psf)
800 800
600 600
400 400
-15%
1400
1200 15%
1000
TEST (psf)
800
600
400
200 non-studded
studded
0
0 200 400 600 800 1000 1200 1400
SDI-M (psf)
500 1000
400 800
LOAD (psf)
LOAD (psf)
300 600
200 400
test test
100 200
iterative analysis iterative analysis
0 0
0.0 1.0 2.0 3.0 4.0 5.0 0.0 1.0 2.0 3.0 4.0 5.0
(a) (b)
500
400
LOAD (psf)
300
200
100 test
iterative analysis
0
0.0 1.0 2.0 3.0 4.0 5.0
MID-SPAN DEFLECTION (in)
(c)
Figure 3-16. Load vs. mid-span deflection: (a) slab-4, (b) slab-15, (c) slab-21
4.1. General
Successful use of the finite element method in many studies involving complex
structures or interactions among structural members has been one of the motivations for applying
the method in this study. To compare with simple mechanical models discussed in the previous
chapter, finite element models may offer more accurate analyses because of the ability to model
the material and interaction of each part of the system in more detail. Further, the response
history of virtually any part of the model can be obtained. In this method, element and material
model types play an important role for the entire analysis. Selection of element and material
model types for the analysis is based on the structural system and any specific need or emphasis
of the study.
In this study, because the main concern is behavior of one-way composite slabs with a
large ratio of length to the cross sectional dimensions in a typical width of the slab, then the
choice of beam and spring elements for a finite element model is the most effective one. The
model is similar to the one proposed by An (1993) with modifications such as the inclusion of
end anchorages and a concrete fracture model for concrete in tension. With this concrete fracture
model, the mesh sensitivity problem in finite element analysis involving concrete (brittle)
material can be removed (Fracture 1992; Karihaloo 1995). Descending curves of end anchorages
and shear bond interaction are also included. ABAQUS is used to conduct the analyses.
y c = y d sin θ ≅ y d θ (4-1)
y s = u 1d − u1c + y d θ (4-2)
Plane
of
reference
c.g.c
ys c.g.s
u1d
Figure 4-1. Schematic model of steel deck to concrete relative slip
where y c = horizontal projection of y d , y d = depth of deck c.g. from concrete c.g., θ = rotation
of cross sectional plane, y s = horizontal slip of steel deck relative to the concrete, u1d = nodal
displacement of steel deck beam element in d.o.f.-1 direction (horizontal), u1c = nodal
displacement of concrete beam in d.o.f.-1 direction (horizontal).
For end anchorages, spring elements are placed at the supports to produce resistance to
horizontal movements of the concrete slab and steel deck relative to the support. The spring is
attached to the bottom surface of the deck. A schematic diagram of the model is shown in Fig. 4-
2.
STUD-CONCRETE
INTERACTION
CONCRETE
IMPOSED EQUATION
FOR HORIZONTAL SLIP
IMPOSED EQUATION
SHEAR BOND
FOR VERTICAL DSPL.
STEEL DECK
STUD-DECK
INTERACTION
WELD-DECK
INTERACTION
hydrostatic axis
σ3
π plane
failure surface
σ2
Figure 4-3. Von Mises yield surface in the principal stress space
The concrete material on the other hand, is pressure dependent. The general shape of
failure surface for concrete material is illustrated in Fig. 4-4. ABAQUS uses the Drucker-Prager
failure surface, a two-parameter model, for concrete material (Drucker and Prager 1952). This
model is valid only for problems with low confining pressures (Hibbitt 1987). For a high
confining pressure, many finer models of concrete failure surfaces are available, such as the
Ottosen four parameter model (Ottosen 1977), Hsieh-Ting-Chen four parameter model (Hsieh et
al. 1982), Willam-Warnke five parameter model (Chen 1982), etc. The Drucker-Prager model,
however, is sufficient for one-way composite slabs. Moreover, because of the conical shape of
the failure surface, singularity is only at the apex. Multi-vector return stress based on Koiter’s
(1953) approach is a common method to handle such singularity. Other methods such as a
multiple single vector return (Widjaja 1997b) may improve the accuracy of the former method.
Recent developments in the application of fracture mechanics to concrete, in particular,
σ1
ε
θ hydrostatic axis
ρ σ3
deviatoric plane
π plane
σ2
failure surface
The uniaxial stress-strain relation for concrete in compression is modeled using the
Saenz equation up to the peak value (Saenz 1964). This model has been successfully used by
Razaqpur and Nofal (1990) to model a composite bridge. The expression of Saenz equation is
given by:
Eoε
σ = (4-3)
2
E ε ε
1 + o − 2 +
E sc ε cu ε cu
where σ and ε are the stress and the corresponding strain of the concrete respectively, E o and
E sc are the initial and the secant modulus of elasticity, respectively, ε cu = concrete strain at the
peak compressive stress. The descending branch of concrete-stress-strain relation is omitted in
this beam model configuration to preserve stability of the system when compressive strength of
concrete is approached. Figure 4-5 shows the concrete stress strain relation.
E sc
ε cu ε
The backward Euler integration scheme is used in the plastic analysis. The scheme
assumes that the return of the stress state to the yield surface is normal to the final yield surface
(note that the yield surface keeps changing to follow the work hardening rule when plastic flow
occurs).
Finally, a nonlinear elastic model is used to model end anchorages (welds or shear studs)
and shear bond interaction. The force-displacement relation of these end anchorages and shear
bond interactions were obtained from elemental tests as presented in Chapter 2. Typical shear
bond interaction is shown in Fig. 4-6 and typical shear stud to steel deck and puddle weld to steel
deck interactions, respectively, are shown in Figs. 4-7(a) and (b).
7.0
6.0
SHEAR STRESS (psi)
5.0
4.0
3.0
2.0
ACTUAL
1.0 SIMPLIFIED
0.0
0.0 0.2 0.4 0.6 0.8 1.0
SLIP (in)
10.0 3.0
2.5
FORCE (kips)
FORCE (kips)
8.0
2.0
6.0
1.5
4.0
1.0
2.0 0.5
0.0 0.0
0.0 0.1 0.2 0.3 0.4 0.0 0.2 0.4 0.6 0.8
u1 displacement, u
un
Table 4-1. Ultimate slab capacity: finite element vs. test results
SLAB FEM TEST RATIO SLAB FEM TEST RATIO
# TEST/ # TEST/
psf psf FEM psf psf FEM
1 627 730 1.16 15 985 1017 1.03
2 617 700 1.13 16 1037 1185 1.14
3 577 600 1.04 17 506 565 1.12
4 543 600 1.10 18 264 368 1.40
5 480 490 1.02 19 537 523 0.97
6 565 590 1.04 20 496 523 1.05
7 293 375 1.28 21 456 467 1.03
8 480 490 1.02 22 441 494 1.12
9 775 900 1.16 23 408 507 1.24
10 790 900 1.14 24 534 621 1.16
11 733 750 1.02 25 534 559 1.05
12 799 870 1.09 26 353 498 1.41
13 409 480 1.17 27 353 455 1.29
14 364 500 1.37
500 500
400 400
LOAD (psf)
LOAD (psf)
300 300
200 200
test
test
finite element
100 finite element 100
0 0
0.0 1.0 2.0 3.0 4.0 5.0 0.00 0.10 0.20 0.30 0.40
(a) (b)
Figure 4-9. Slab-4: (a) Load vs. mid-span deflection. (b) Load vs. end-slip
1200 1200
1000 1000
800 800
LOAD (psf)
LOAD (psf)
600 600
400 400
(a) (b)
Figure 4-10. Slab-15: (a) Load vs. mid-span deflection. (b) Load vs. end-slip
450
400 400
350
LOAD (psf)
LOAD (psf)
300 300
250
200 200
150
100 100
test test
finite element 50 finite element
0 0
0.0 1.0 2.0 3.0 4.0 5.0 0.00 0.02 0.04 0.06 0.08
(a) (b)
Figure 4-11. Slab-21: (a) Load vs. mid-span deflection. (b) Load vs. end-slip
Figure 4-12 shows graphical comparison of predicted vs. test values of slab strength. It
can be seen from the figure, the predicted values for studded slabs fall within ±15% margin.
For non-studded slabs, predicted values tend to be more conservative. This fact may be caused
by the exclusion of clamping force to the steel deck and friction at steel deck-concrete interface
at the supports.
-15%
1400
1200 15%
1000
TEST (psf)
800
600
400
200 non-studded
studded
0
0 200 400 600 800 1000 1200 1400
FEM (psf)
5.1. General
The maximum span length of unshored single span composite slabs used in the U.S.
based on available steel deck floor in the market is around 13 ft. The choice of unshored systems
is very common because these systems can save construction cost and time. If the span length
can be increased by a factor of, for example, 1.5 or 2, significant cost savings can be expected
from elimination of some intermediate beams and their connections to the girders. These
potential advantages have motivated research in the area of long span slab systems. In this case,
long span slab systems that do not cause any significant increase in the depth and weight of the
slabs compared to regular span slabs are particularly attractive. This has been one of the main
objectives of this part of the study.
Research in this area has been carried out by other researchers. Notable among these
are, the investigations by Ramsden (1987), the innovative lightweight floor system by Hillman
and Murray (Hillman 1990, Hillman and Murray 1990, 1994) and the slimflor system (British
Steel, Lawson et al. 1997). Ramsden (1987) conducted a study on two new prototypes of deck
profiles that can span a distance up to about 24 ft (7.5 m). The prototypes have holes in the web
to ensure the composite action between the deck and the concrete. The second prototype is an
improved version of the first one. These two prototypes are shown in Fig. 5-1. Because of the
shape of the profile, the concrete slab is virtually a solid slab with a thickness of 5 in. to 6 in.,
which is disadvantageous because of it selfweight. There is no mention in the paper whether
shoring of the slab during the construction was provided.
Prototype 2
An innovative composite slab floor system design was developed and reported by
Hillman (1990) and Hillman and Murray (1990, 1994). The floor system developed was not only
lightweight but also able to span up to 30 ft without any intermediate beams. Figure 5-2 shows
schematically the design of the composite slab.
Concrete slab
Perpendicular
steel decks
Concrete
Support beam
Steel deck
Figure 5-3. Slimflor system (British Steel, Steel Construction Institute 1997)
In the current study, two 16 ga, deep steel deck profiles are investigated. The first
profile, referred to as profile 1, has a 6 in. rib height. The profile is currently not available in the
market so it was designed and manufactured by a press-brake process for this project. Because
of this, the length of the deck was limited to 25 ft. For long span slab specimens, the length is
only enough for a single span configuration. The second profile, i.e. profile 2, is a currently
available roof deck profile whose stiffness, as discussed later in this chapter, satisfies the
requirements for a long span slab in a double span configuration. This section was manufactured
through a cold-rolling process. Profile shapes of these sections are shown in Fig. 5-4. Note that
neither of these shapes incorporated embossments. This is because neither are currently
available composite deck profiles. For comparison, a 3 in. deep trapezoidal section is also
included in the figure.
Two design phases have to be considered in the development of these new deck profiles
for long span composite slab systems, namely the construction (non-composite) phase and
service (composite) phase. The construction phase considers the strength and stiffness of the
steel deck as a working platform that is subject to concrete self-weight and construction loads.
This phase is important in the determination of the required deck stiffness. It is shown later that
9.25
6
profile 1
1
3.75 7.125 1.5 0.5
12.875
profile 2 4.5
1
profile 3
3
4.75
7.25
12
The service phase deals with a composite section of steel deck-concrete slab that is
subject to occupancy loads. Studies on composite slabs with typical span lengths (Terry and
Easterling 1994, Widjaja and Easterling 1995, 1996, 1997) revealed that the actual load capacity
of the slabs are very high compared to the standard design live loads (50 to 150 psf). Table 5-1
shows that the ratios of actual load capacities (from the tests) to a 150 psf design live load range
from 2.45 to 7.90. At these (ultimate) load capacities, however, the slabs have undergone
excessive deflections. If the allowable deflection is limited to L/360 (SDI 1992), then, the
permissible loads based on this allowable deflection will be much lower than the ultimate load
capacities. The ratios of these permissible loads to a 150 psf design live load, as shown in Table
5-1, range from 1.37 to 3.11.
These ratios suggest that the service phase rarely governs the design of composite slabs.
However, this is not always the case for long span composite slabs as latter shown by the
analysis and test results. For long span slabs, both the construction and service phase have an
equal change to govern the design.
slab ultimate load load at allow. test load / 50 test load / 150
# capacity deflection *) ultimate load load at allow. ultimate load load at allow.
(psf) (psf) capacity deflection capacity deflection
1 730 345 14.60 6.91 4.87 2.30
2 700 326 14.00 6.52 4.67 2.17
3 600 238 12.00 4.76 4.00 1.59
4 600 223 12.00 4.47 4.00 1.49
5 490 310 9.80 6.20 3.27 2.07
6 590 316 11.80 6.32 3.93 2.11
7 375 301 7.50 6.01 2.50 2.00
8 490 320 9.80 6.41 3.27 2.14
9 900 374 18.00 7.49 6.00 2.50
10 900 388 18.00 7.76 6.00 2.59
11 750 352 15.00 7.04 5.00 2.35
12 870 418 17.40 8.36 5.80 2.79
13 480 399 9.60 7.98 3.20 2.66
14 500 389 10.00 7.78 3.33 2.59
15 1017 407 20.34 8.14 6.78 2.71
16 1185 466 23.70 9.32 7.90 3.11
17 565 301 11.30 6.02 3.77 2.01
18 368 303 7.36 6.07 2.45 2.02
19 523 396 10.46 7.93 3.49 2.64
20 523 445 10.46 8.91 3.49 2.97
21 467 229 9.34 4.57 3.11 1.52
22 494 206 9.88 4.11 3.29 1.37
23 507 246 10.14 4.92 3.38 1.64
*) based on L/360
30
YIELD STRENGTH
LIMIT STATE
25 DEFLECTION
LIMIT STATE
STEEL DECK WEIGHT (lb/ft2)
20
15
10
0
6 9 12 15 18 21 24
SPAN LENGTH (ft)
It can be observed from Figs. 5-6 and 5-7, that for a same weight of steel deck, profiles 1
and 2 allow one to have a longer span than that of profile 3. This indicates that profiles 1 and 2
are more efficient than profile 3. Therefore, for a long span slab system of 20 ft, only the 4.5 in.
(profile 2) and 6 in. (profile 1) sections are considered in this study.
0
6 9 12 15 18 21 24
SPAN LENGTH (ft)
10
profile 1
profile 2
profile 3
8
STEEL DECK WEIGHT (lb/ft2)
6
16 ga
16 ga
18 ga
18 ga
4
20 ga
16ga 20 ga
18 ga
20 ga
2
0
6 9 12 15 18 21 24
SPAN LENGTH (ft)
corresponds to an ability to span 1.6 (= 4 7 ) times further. The steel deck self-weight is almost
double the one of profile 3. For a 20 ft long piece of deck with only one typical rib of profile 1,
the piece weighs about 116 lb and it can be handled by two people in the construction site.
For profile 2, the increase of the moment of inertia is about 3 times of that of profile 3,
and it corresponds to an ability to span 1.3 times further. The total weight of the slab, for the
same 2.5 in concrete cover above the rib, is slightly lighter than the slab with profile 3 as the
steel deck.
permissible loads based on deflection limit state. The analysis was performed in the same ways
as those with typical span length.
SLAB 2
concrete slab
steel deck
1′ 20′ 20′ 1′
Strain gages were placed at the bottom surface of the deck to measure the steel deck
strain during concrete casting and the load test. Three cross sections were monitored in each
span of the slab: the exterior support, interior support and mid-span. A set of six strain gages
was used at each of those cross sections. The schedules of these strain gage and shear stud
locations are shown in Figs. 5-9 and 5-10 for LSS1 and LSS2, respectively. In addition to these
strain gages, potentiometers were also placed at each end of the slab to measure the slip between
the concrete and the deck. Several displacement transducers were also used to measure vertical
displacements.
No shoring was provided during the construction of the slabs. The measured mid-span
deflections of the steel deck during concrete casting were 0.695 in. and 0.685 in. for LSS1 and
LSS2, respectively. Concrete compressive strength at 28 days were 3060 psi and 2330 psi for
LSS1 and LSS2, respectively.
Section A-A
A A A A A A
A A A A A A
8.5
Section B-B
B BB B
B BB B
12 240 240 12
Section A-A
A A A A A A
A A A A A A
Section B-B
B B B
B B B
12 240 240 12
air bag
At the beginning of each load test, the tested span was preloaded with approximately
0.35 psi (50 psf) to settle the system and check the instrumentations. The slab was unloaded
afterward and the loading was restarted and continued until a permanent set in the system was
obtained. This permanent set can be observed from the presence of the nonlinear relation of the
load versus mid-span displacement. Load increments of approximately 0.25 psi (36 psf) was
applied with a pause, of approximately two minutes before any data recording, to allow the
system to settle. When a permanent set had been noted, the system was once again unloaded
completely. The loading was then restarted until failure or excessive deflection was obtained.
In the inelastic region where the stiffness of the slab had decreased considerably,
displacement control loading was used with a displacement increment of approximately 0.5 in.
x x x x
x x x x
700 700
Direct Direct
600 600
Test Iterative
Iterative
500 500
SDI-M SDI-M
LOAD (psf)
LOAD (psf)
400 400 Test
300 300
200 200
100 100
0 0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
500 500
LOAD (psf)
250 250
200 200
150 150
100 100
50 50
0 0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
slab ultimate load load at allow. test load / 50 test load / 150
# capacity deflection *) ultimate load load at allow. ultimate load load at allow.
(psf) (psf) capacity deflection capacity deflection
LSS1a 621 245 12.42 4.90 4.14 1.63
LSS1b 559 210 11.18 4.20 3.73 1.40
LSS2a 498 163 9.96 3.26 3.32 1.09
LSS2b 455 121 9.10 2.43 3.03 0.81
*) based on L/360
From the above table, it can be noted that for LSS2, the permissible loads based on the
allowable deflection are relatively low compared to those of typical span slabs and LSS1.
Therefore, in the case of long span composite slab, it is important to check the deflection limit
state.
1.0
0.6
0.4
0.2
0.0
4 6 8 10 12 14 16
FREQUENCY (Hz)
1.2
1.0
NORMALIZED RELATIVE POWER
0.8
0.6
0.4
0.2
0.0
4 6 8 10 12 14 16
FREQUENCY (Hz)
6.1. General
Probability-based design criteria in the form of load and resistance factor design (LRFD)
are now applied for most construction materials. The design requirements have to insure
satisfactory performance of structures. The main advantage of the approach is the ability to
achieve a uniform level of reliability for structural members, or to impose a certain level of
reliability (higher or lower) of some certain parts of the structures. This gives a strong rationale
to the load and resistance factors as compared to the design safety factors of the allowable stress
design. Additionally, a unified design strategy as to setting up common load combinations and
load factors can be obtained.
In this part of the study, resistance factors, φ, for the flexural design of composite slabs
were evaluated based on test data of 39 full scale composite slab specimens. The tests were
performed at the Structures and Materials Laboratory of Virginia Polytechnic Institute and State
University, Blacksburg, Virginia. The φ factors evaluated correspond to the SDI-M method and
direct method described in Section 3.
φ R n ≥ ∑ γ i Qi (6-1)
i
in which, R n = nominal resistance, Q i = load effect, φ = resistance factor and γ i = load factor.
p f = 1 − Φ(β) (6-2)
where Φ is the standard normal probability function, and β is the reliability index.
R
ln m
λR − λQ Qm
β = ≈ (6-3)
ζ 2R + ζ 2Q VR2 + VQ2
where λ, ζ and V, respectively, denote the log-normal mean, log-normal standard deviation and
the coefficient of variation. Subscript R and Q denote the resistance and the load effect,
respectively. R m and Q m are the means of resistance and load, respectively. Introduce a
linearization given by:
with ±6% maximum error. Equation (6-5) forms the basis equation for the AISC and AISI load
and resistance factor design specification for structural steel and cold-formed steel. From Eqn.
(6-5), the central safety factor can be expressed as:
Rm αβ (VR + VQ )
θ = = e (6-6)
Qm
By minimizing the error of this central safety factor, Galambos and Ravindra (1977) suggested a
value of α = 0.55 which was later adopted in AISC LRFD. The reliability index, β , can be
determined from Eqn. (6-5). As an illustration, the following table shows some β values and
β pf
5.0 2.9 x 10 -7
4.0 3.2 x 10-5
3.0 1.4 x 10-3
2.0 2.3 x 10-2
R m ≥ θ Qm (6-7)
αβVQ
Rm e− αβVR ≥ Qm e (6-8)
or,
φ R n ≥ γ Qn (6-9)
R m − αβVR
φ = e (6-10)
Rn
Qm αβVQ
γ = e (6-11)
Qn
Further, the mean resistance, R m , can be expressed in terms of the nominal resistance and
statistical parameters that represent the variability of material strength and stiffness, M,
R m = R n (M m Fm Pm ) (6-12)
where M m , Fm and Pm are the means of M, F, and P, respectively. Accordingly, the coefficient
of variation of the resistance can be approximated by using:
VR ≈ ( VM ) 2 + ( VF ) 2 + ( VP ) 2 (6-13)
φ = (M m Fm Pm ) e − αβVR (6-14)
D
φ R n = c (γ D D n + γ L L n ) = γ D n + γ L c L n (6-15)
Ln
where γ D and γ L are the dead and live load factors, and D n and L n are the nominal values of
the dead and live load. Similarly, the mean of the load effect can be expressed as:
Rm ψ Rm
= (6-17)
Qm φ Rn
with,
D D
ψ = γ D n + γ L / 1.05 n + 1 (6-18)
Ln Ln
By combining Eqns. (6-3), (6-12) and (6-17), an expression of the resistance factor can be
obtained:
2 + V2
-β VR
φ = ψ (M m Fm Pm ) e Q
(6-19)
Using this equation, determination of the α coefficient can be avoided. However, the coefficient
of variation of the load has to be known.
steel members by Hsiao et al. (1990). For these two parameters, the data obtained from the lab
tests from the composite slab specimens were used as a comparison only.
Data obtained from the lab tests, which are not available elsewhere from larger sets of
database, were used for the determination of the resistance factor. These data are the statistical
data of deck thickness, t, maximum and minimum shear bond strength at the interface of steel
deck - concrete, f s,max and f s,min , respectively.
µ σ V
fc' (MacGregor, 1997) 3940 psi 615 psi 0.156
fc' (test) 3867 psi 878 psi 0.227
fy (Hsiao et al. 1990) 1.100 fy 0.121 fy 0.110
fy (test) 1.002 fy 0.058 fy 0.058
fs,max 0.999 fs,max 0.035 fs,max 0.035
fs,min 1.001 fs,min 0.073 fs,min 0.073
Note: µ = mean, σ = standard of deviation, V = coefficient of variation
f c ' ,m f y,m
( )
m
(
f s,max f s,min
m
)
M m,Direct = f
fc ' y f s,max f s,min
VF = Vt = 0.313 (6-24)
t ,m µt
Fm = = = 0.966 (6-25)
t t
The prediction is the resistance of the slab as predicted by the design equation based on the
measured (actual) values of its parameters. Based on the lab tests performed on the afore-
mentioned full-scale composite slab specimens, the following statistical data is obtained:
dead and live loads were taken from a special publication of the National Bureau of Standards
(Ellingwood et al. 1980). These data are summarized in Table 6-5. D n and L n denote the
nominal dead and live loads.
For the combination of the dead and live loads given by:
Q = γD D + γL L (6-26)
the mean and standard of deviation of this combination can be expressed by:
µQ = γ D µD + γ L µ L (6-27)
assuming that the distribution of D and L are statistically independent. In Eqn. (6-28), Var(Q)
denotes the variance of Q. By substituting values from Table 6-5 into Eqn. (6-27) and Eqn. (6-
28), the coefficient of variation of Q can be obtained as:
2
D D
VQ = 0.011 γ 2D n + 0.063 γ 2L / 1.05 γ D n + γ L (6-29)
Ln Ln
for composite slabs. A range of dead to live load ratios between 0.5 (short to normal span slabs
with relatively heavy live load, approximately 100 psf) and 1.5 (long span slabs up to 20 ft with
relatively light live load, approximately 50 psf) is considered.
Based on the statistical data presented in section 6.3 and equations given in section 6.2, φ
factors for several values of D/L (0.5, 1.0 and 1.5) were computed and the results are listed in
Table 6-6 and Table 6-7 for the SDI-M and direct design procedures, respectively. Again, these
results are based on the AISI-LRFD approach presented in section 6.2.3.
Based on the results in Tables 6-6 and 6-7, φ = 0.90 is chosen for the SDI-M method and φ = 0.85
is selected for the direct method. For comparison, φ factors computed by using the AISC-LRFD
approach are listed in Table 6-8 for the SDI-M method and Table 6-9 for the direct method for
several combinations of α and β values. This later approach is not influenced by the ratio of the
dead to live load (D/L). As shown in these tables, the choice of α between 0.65 to 0.75 show
relatively close results to the AISI approach.
A study of the strength and behavior of composite slabs in general, with a particular
investigation of the use of long span composite slab systems, has been carried out analytically
and experimentally. Two new methods of predicting composite slab strength and stiffness based
on simple mechanical models have been developed. The methods, which are supported by
experimental data obtained from elemental tests of shear bond and end anchorages, offer an
alternative solution to the m-k method, which requires a number of full-scale tests. Experimental
test results conducted on full-scale composite slab specimens reveal that the methods predict the
slab strength more accurately than the SDI-M method. This is due to the ability of these methods
to include the effects of shear bond strength, weld strength, end anchorage strength and any
remaining strength of the deck.
The nonlinear finite element method was used to model the complex nature of composite
slabs. From this analysis, a response history of virtually any point of the system can be obtained.
The development and use of a special purpose finite element code, which is particularly designed
for composite slabs and incorporates a concrete plasticity model with three or higher number of
parameters for the concrete failure surface and an energy based path following technique is
recommended. This is based on the fact that the concrete material is one of the most sensitive
aspects of the composite slab analysis, particularly when the concrete is in tension. The
suggested energy based path following technique is due to the inconsistency of the physical units
in the arc length method, which may lead to numerical problems.
An, L. (1993). “Load Bearing Capacity and Behavior of Composite Slabs with Profiled Steel
Sheet.” Ph.D.-Dissertation, Chalmers University of Technology, Division of Concrete
Structures, Goteborg, Sweden.
An, L. and Cederwall, K. (1992). “Composite slabs analyzed by block bending test.” Proc.
11th. International Specialty Conference on Cold-Formed Steel Structures. Ed.: W. W. Yu, 268-
282.
An, L. and Cederwall, K. (1994). “Slip and separation at interface of composite slabs.”
Proceedings of the 12th. International Specialty Conference on Cold-Formed Steel Structures.
Ed.: W. W. Yu. 385-397.
Ang, A. H. S. and Cornell, C. A. (1974). “Reliability bases of structural safety and design.”
ASCE Journal of the Structural Division, 100, ST9, 1755-1769.
ASCE Task Committee on Structural Safety of the Administrative Committee on Analysis and
Design of the Structural Division. (1972). “Structural safety - a literature review.” ASCE
Journal of the Structural Division, 98, ST4, 845-885.
85
Bode, H. and Sauerborn, J. (1992). “Modern design concept for composite slabs with ductile
behavior.” Composite Construction in Steel and Concrete II, Proceedings of an Engineering
Foundation Conference, ASCE. Ed.: W. S. Easterling, W. K. M. Roddis, 125-141.
Building Code Requirements for Structural Concrete (ACI 318-95) and Commentary (ACI 318R-
95). (1995). American Concrete Institute, Michigan.
Carrera, E. (1994). “A study on arc-length -type methods and their operation failures illustrated
by simple model.” Computers and Structures, 50, 2, 217-229.
Chen, W. F. (1982). “Plasticity in Reinforced Concrete.” McGraw-Hill Book Co., New York.
“Commentary on the 1996 specification for the design of cold-formed steel structural members.”
(1996). American Iron and Steel Institute, Washington, D. C.
Cornell, A. C. (1969). “A probability based structural code.” ACI Journal, 66, 12, 974-985.
Crisinel, M., Fidler, M. J. and Daniels, B. J. (1986a). “Behavior of steel deck reinforced
composite floors.” IABSE Colloquium, Stockholm.
Crisinel, M., Fidler, M. J. and Daniels, B. J. (1986b). “Flexure Tests on Composite Floors with
Profiled Steel Sheeting.” ICOM 158, Ecole Polytechnique Federale de Lausanne.
86
Dallaire, E. E. (1971). “Cellular steel floors mature.” Civil Engineering, ASCE, 41, 7, 70-74.
Daniels, B. J. (1988). “Shear Bond Pull-Out Tests for Cold Formed Steel Composite Slabs.”
ICOM 194, Ecole Polytechnique Federale de Lausanne.
Daniels, B. J. and Crisinel, M. (1987). “Composite slabs with profiled sheeting.” Composite
Construction in Steel and Concrete, Proceedings of an Engineering Foundation Conference,
ASCE. Eds.: C. D. Buckner and I. M. Viest, 656-662.
Daniels, B. J. and Crisinel, M. (1993). “Composite slab behavior and strength analysis. Part I:
calculation procedure.” Journal of Structural Engineering, ASCE, 119, 1, 16-35.
Daniels, B. J., Isler, A. and Crisinel, M. (1990). “Modeling of Composite Slab with Thin Walled
Cold-Formed Decking.” ICOM 235, Ecole Polytechnique Federale de Lausanne.
Drucker, D. C. and Prager, W. (1952). “Soil mechanics and plastic analysis or limit design.”
Quarterly of Applied Mathematics, 10, 2, 157-165.
Ekberg, C. E. and Schuster, R. M. (1968). “Floor Systems with Composite Form Reinforced
Concrete Slabs.” IABSE New York, Final Report, 385-394.
87
Ellingwood, B., Galambos, T. V., MacGregor, J. G. and Cornell, C. A. (1980). “Development of
a probability based load criterion for American National Standard A58: Building code
requirements for minimum design loads in buildings and other structures.” National Bureau of
Standards, U.S. Department of Commerce.
Finzi, L. (1968). “Light gage steel floor systems provided to include utilities - proposal and
experiments.” IABSE New York, Final Report, 367-374.
Galambos, T. V. and Ravindra, M. K. (1977). “The basis for load and resistance factor design
criteria of steel building structures.” Canadian Journal of Civil Engineering, 4, 178-189.
Geschwindner, L. F., Disque, R. O. and Bjorhovde, R. (1994). “Load and resistance factor
design of steel structures.” Prentice Hall, New Jersey.
88
Heagler, R. B., Luttrell, L. D. and Easterling, W. S. (1992). “The Steel Deck Institute method
for composite slab design.” Composite Construction in Steel and Concrete II, Proceedings of an
Engineering Foundation Conference, ASCE. Ed.: W. S. Easterling, W. K. M. Roddis.
Hibbitt, H. D. (1987). “A simplified model for concrete at low confining pressures.” Nuclear
Engineering and Design, 104, 313-320.
Hillman, J. R. (1990). “Innovative Lightweight Floor System for Steel Framed Buildings.” M.S.
Thesis, Virginia Polytechnic Institute and State University.
Hillman, J. R. and Murray, T. M. (1990). “Innovative floor systems.” Proc. 1990 National Steel
Construction Conference, Kansas City, Missouri, American Institute of Steel Constructions,
12/1-12/21.
Hillman, J. R. and Murray, T. M. (1994). “An innovative cold-formed floor system.” Proc.
12th. International Specialty Conference on Cold-Formed Steel Structures, ed.:W. W. Yu, 513-
521.
Hsiao, L. E., Yu, W. W. and Galambos, T. V. (1990). “AISI LRFD method for cold-formed
steel structural members.” ASCE Journal of Structural Engineering, 116, 2, 500-517.
Hsieh, S. S., Ting, E. C., and Chen, W. F. (1982). “A plastic-fracture model for concrete.”
International Journal of Solids and Structures, 18, 3, 181-197.
Jolly, C. K. and Lawson, R. M. (1992). “End anchorage in composite slabs: An increased load
89
carrying capacity.” The Structural Engineer, 70, 11/2, 202-205.
Koiter, W. T. (1953). “Stress-strain relations, uniqueness and variational theorems for elastic-
plastic materials with a singular yield surface.” Quarterly of Applied Mathematics, 11, 3, 350-
354.
“Load and resistance factor design specification for structural steel buildings.” (1986).
American Institute of Steel Constructions, Chicago, Illinois.
Lind, N. C. (1971). “Consistent partial safety factors.” ASCE Journal of the Structural Division,
97, ST6, 1651-1669.
LRFD Cold-Formed Steel Design Manual. (1991). American Iron and Steel Institute,
Washington, D. C.
Luttrell, L. D. and Prasannan, S. (1984). “Strength formulations for composite slabs.” Proc.
7th. International Specialty Conference on Cold-Formed Steel. Ed.: W. W. Yu, 307-326.
MacGregor, J. G. (1997). “Reinforced concrete mechanics and design.” Prentice Hall, New
90
Jersey.
Murray, T. M., Allen, D. E. and Ungar, E. E. (1997). “Floor Vibrations due to Human Activity.”
American Institute of Steel Construction and Canadian Institute of Steel Construction, Chicago,
Illinois.
O’Leary, D., El-Dharat, A. and Duffy, C. (1987). “Composite reinforced end-anchored concrete
floors.” Composite Steel Structures. Ed.: R. Narayanan, 117-126.
Ollgaard, J. G., Slutter, R. G. and Fisher, J. W. (1971). “Shear strength of stud connectors in
light-weight and normal-weight concrete.” AISC Engineering Journal, 8, 2, 55-64.
Ottosen, N. S. (1977). “A failure criterion for concrete.” Journal of the Engineering Mechanics
Division, 103, EM4, 527-535.
Patrick, M. (1990). “A new partial shear connection strength model for composite slabs.” Steel
Construction, 24, 3, 2-17.
Patrick, M. and Bridge, R. Q. (1990). “Parameters affecting the design and behavior of
composite slabs.” IABSE Symposium, Brussels, 221-225.
Patrick, M. and Bridge, R. Q. (1994). “Partial shear connection design of composite slabs.”
Patrick, M. and Poh, K. W. (1990). “Controlled test for composite slab design parameters.”
IABSE Symposium, Brussels, 227-231.
91
Porter, M. L. (1985). “Proposed design criteria for composite steel deck slabs.” Composite and
Mixed Construction. Ed.: C. W. Roeder, 28-41.
Porter, M. L. and Ekberg, C. E. (1972). “Summary of full scale laboratory tests of concrete
slabs reinforced with cold-formed steel decking.” International Association for Bridge and
Structural Engineering (IABSE), Amsterdam, Preliminary Report, 173-183.
Porter, M. L. and Ekberg, C. E. (1976). “Design recommendations for steel deck floor slabs.”
Journal of the Structural Division, 102, ST11, 2121-2136.
Porter, M. L., Ekberg, C. E., Greimann, F. and Elleby, H. A. (1976). “Shear bond analysis of
steel deck reinforced slabs.” Journal of the Structural Division, 102, ST12, 2255-2267.
Porter, M. L. and Greimann, L. F. (1984). “Shear bond strength of studded steel deck slabs.”
Proc. 7th. International Specialty Conference on Cold-Formed Steel Structures. Ed.: W. W. Yu,
285-306.
Ramsden, J. A. (1987). “A light gauge structural panel for composite flooring.” Composite
Construction in Steel and Concrete, Proceedings of an Engineering Foundation Conference,
ASCE, Eds.:C. D. Buckner, I.M. Viest, 374-386.
Ravindra, M. K. and Galambos, T. V. (1978). “Load and resistance factor design for steel.”
ASCE Journal of the Structural Division, 104, ST9, 1337-1353.
92
Ren, P. and Crisinel, M. (1992). “Nonlinear Analysis of Composite Members.” ICOM 280,
Ecole Polytechnique Federale de Lausanne.
Ren, P., Couchman, G. and Crisinel, M. (1993). “Comparison Between Theoretical Model
COMPCAL and Test Results.” ICOM 286, Ecole Polytechnique Federale de Lausanne.
Saenz, L. P. (1964). “Discussion of the paper: equation for the stress-strain curve of concrete.”
ACI Journal, 61, 1229-1235.
SDI Specifications and Commentary for Composite Steel Floor Deck. (1992). “Steel Deck
Institute Design Manual.” Publication no. 28, SDI, Canton, Ohio.
Sonoda, K., Kitoh, H. and Horikawa, T. (1994). “Shear bond characteristic of embossed steel
plate with or without stud connectors under a constant lateral pressure.” Steel-Concrete
Composite Structures, Proceedings of the 4th International Conference by ACCS, 307-310.
“Standard for the Structural Design of Composite Slabs.” (1992). ANSI/ASCE 3-91, American
Society of Civil Engineers, New York.
Stark, J. (1978). “Design of composite floors with profiled steel sheet.” Proc. 4th. International
Specialty Conference on Cold-Formed Steel Structures. Ed.:W. W. Yu, 893-922.
93
Tagawa, Y., Suko, M. and Tsushima, A. (1994). “Push-out test of composite slab.” Steel-
Concrete Composite Structures, Proceedings of the 4th International Conference by ACCS, 316-
319.
Terry, A. S. and Easterling, W. S. (1994). “The Effects of Typical Construction Details on The
Strength of Composite Slabs.” Report No. CE/VPI-ST 94/05, Department of Civil Engineering,
Virginia Polytechnic Institute and State University.
Veljkovic, M. (1996). “Behavior and Resistance of Composite Slabs.” Doctoral Thesis, Lulea
University of Technology, Department of Civil and Mining Engineering, Division of Steel
Structures, Sweden.
Widjaja, B. R. (1997b). “Numerical integration of stress return for singular loading surface of
concrete.” Submitted for publication.
Widjaja, B. R. and Easterling, W. S. (1996). “Strength and stiffness calculation procedures for
composite slabs.” Proc. 13th. International Specialty Conference on Cold-Formed Steel
Structures. Ed.:W. W. Yu, 389-401.
94
Widjaja, B. R. and Easterling, W. S. (1996). “Vulcraft 3VLI Composite Deck Tests.” Report
No. CE/VPI-ST 96/18, Department of Civil Engineering, Virginia Polytechnic Institute and State
University, Blacksburg, Virginia.
Widjaja, B. R. and Easterling, W. S. (1997). “Vulcraft 2VLI Composite Deck Tests.” Report
No. CE/VPI-ST 97/09, Department of Civil Engineering, Virginia Polytechnic Institute and State
University, Blacksburg, Virginia.
Wright, H. D., Evans, H. R. and Harding, P. W. (1987). “The used of profiled steel sheeting in
floor construction.” Journal of Construction Steel Research, 7, 279-295.
Yang, Y. B. and McGuire, W. (1985). “A work control method for geometrically nonlinear
analysis.” Proceedings of the NUMETA’85 Conference. Eds.: J. Middleton, G. N. Pande, 913-
921.
95
VITA
Budi R. Widjaja was born on May 9, 1961 in Semarang, Indonesia. He obtained his B. S. degree
in civil engineering from Parahyangan Catholic University in Bandung, Indonesia, in April 1985.
After a year working in a contracting company, he joined with Parahyangan Catholic University
as a part time teaching assistant. Simultaneously he worked as a structural engineer in a
consulting firm in Bandung, Indonesia, until Fall of 1990, at which time he entered the graduate
program in civil engineering at Virginia Polytechnic Institute and State University. He
completed his Master of Science degree in May 1993 and continued pursuing a Ph.D. degree in
civil engineering. He worked as a research assistant at the Structures and Materials Laboratory
during his doctoral study.
96