Analysis and Design of Steel Deck - Concrete Composite Slabs

ANALYSIS AND DESIGN OF STEEL DECK CONCRETE COMPOSITE SLABS
Budi Ryanto Widjaja
Dissertation submitted to the Faculty of the Virginia Polytechnic Institute and State University in partial fulfillment of the requirements for the degree of
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 =0.90 for the SDI-M method and =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.......................................................................................................................... DEDICATIONS.................................................................................................................... ACKNOWLEDGMENTS .................................................................................................... TABLE OF CONTENTS...................................................................................................... LIST OF FIGURES............................................................................................................... LIST OF TABLES ................................................................................................................ LIST OF NOTATIONS ........................................................................................................ Chapter 1. Introduction 1.1. Motivation and Scope of the Research ......................................................................... 1.2. Organization of this Report ..........................................................................................
ii iv v vi ix xi xii
1 3
Chapter 2. Elemental Tests 2.1. General.......................................................................................................................... 2.2. Review of Research on Elemental Tests for Shear Bond and End Anchorages........... 2.3. Shear Bond Elemental Tests......................................................................................... 2.3.1. Specimen Description and Test Set Up............................................................. 2.3.2. Test Procedure................................................................................................... 2.3.3. Test Results ....................................................................................................... 2.4. End Anchorage Elemental Tests................................................................................... 2.4.1. Specimen Description and Test Set Up............................................................. 2.4.2. Test Procedure................................................................................................... 2.4.3. Test Results ....................................................................................................... 2.5. Concluding Remarks..................................................................................................... 4 4 6 6 10 10 13 13 15 15 17
Chapter 3. Strength and Stiffness Prediction of Composite Slabs by Simple Mechanical Model 3.1. General.......................................................................................................................... 18
vi
3.2. Review of Methods of Prediction of Composite Slab Strength by Means of Semi-Empirical Formulations and Simple Mechanical Models ................................... 3.3. SDI-M Method.............................................................................................................. 3.4. Iterative Method............................................................................................................ 3.5. Direct Method............................................................................................................... 3.6. Comparison of Calculated and Test Results................................................................. 3.7 Concluding Remarks..................................................................................................... 19 27 27 34 35 40
Chapter 4. Strength and Stiffness Prediction of Composite Slab by Finite Element Model 4.1. General.......................................................................................................................... 4.2. Review of Finite Element Method for Composite Slabs .............................................. 4.3. Finite Element Model ................................................................................................... 4.3.1. Structure Model................................................................................................. 4.3.2. Material Model.................................................................................................. 4.4. Method of Analysis....................................................................................................... 4.5. Results of Analysis and Discussion.............................................................................. 4.6 Concluding Remarks..................................................................................................... 41 42 43 43 45 48 49 52
Chapter 5. Long Span Composite Slab Systems 5.1. General.......................................................................................................................... 5.2. Construction Phase ....................................................................................................... 5.3. Service Phase ................................................................................................................ 5.4. Specimen Description and Instrumentation.................................................................. 5.5. Load Test Procedure ..................................................................................................... 5.6. Test vs. Analysis Results .............................................................................................. 5.7. Evaluation of the Floor Vibrations ............................................................................... 5.8. Proposed Detailed Connection ..................................................................................... 5.9. Concluding Remarks..................................................................................................... 53 57 60 60 64 65 67 69 70
vii
Chapter 6. Reduction Factor, 6.1. General.......................................................................................................................... 6.2. Review of Probabilistic Concepts of Load and Resistance Factor Design .................. 6.2.1. Reliability Index ................................................................................................ 6.2.2. AISC LRFD Approach for the Resistance Factor ............................................. 6.2.3. AISI LRFD Approach for the Resistance Factor .............................................. 6.3. Statistical Data.............................................................................................................. 6.3.1. Material Factor, M ............................................................................................ 6.3.2. Fabrication Factor F .......................................................................................... 6.3.3. Professional Factor, P........................................................................................ 6.3.4. Load Statistic..................................................................................................... 6.4. The Resistance Factor................................................................................................... 6.5. Concluding Remarks..................................................................................................... 71 71 72 74 75 76 77 78 79 79 80 82
Chapter 7. Conclusions and Recommendations..............................................................
83
References............................................................................................................................. VITA .....................................................................................................................................
85 96
viii
LIST OF FIGURES
2-1. 2-2. 2-3. 2-4. 2-5. 2-6. 2-7. 2-8. 2-9. 2-10.
Profile shapes .............................................................................................................. Embossment types....................................................................................................... Shear bond test ............................................................................................................ Shear bond specimen with frames for lateral force .................................................... Shear stress vs. slip of specimen SB2-2-A.................................................................. Shear stress vs. slip of specimen SB6-1-B.................................................................. Details of the end anchorage specimens ..................................................................... End anchorage test ...................................................................................................... Load vs. deck to concrete slip of specimen EA1-1-B ................................................. Load vs. deck to concrete slip of specimen EA2-1-A.................................................
8 8 9 9 12 12 13 14 16 16 20 22 22 24 25 26 26 27 28 29 31 32 34 37 38 39 44 44 45 46 47 47 48 49 50
3-1. m and k shear bond regression line ............................................................................. 3-2. Partial interaction theory (Stark and Brekelmans 1990)............................................. 3-3. Simplified relation between M p ' and N b (Stark and Brekelmans 1990) ................. 3-4. Boundary curve based on the partial interaction theory ............................................. 3-5. Free body diagram of the forces action in the composite slab section (Patrick 1990, Patrick and Bridge 1994)..................................................................... 3-6. Plot of M u vs. T (Patrick 1990, Patrick and Bridge 1994)........................................ 3-7. Boundary curve for the ultimate bending moment capacity (Patrick 1990, Patrick and Bridge 1994) ............................................................................................ 3-8. Reinforcing effects of some devices ........................................................................... 3-9. Forces acting on the cross section............................................................................... 3-10. Shear bond interaction ................................................................................................ 3-11. Concrete bottom fiber elongation, dL, and slip diagrams........................................... 3-12. Additional load carrying capacity from the deck........................................................ 3-13. Forces acting on the cross section for the direct method ............................................ 3-14. Test setup .................................................................................................................... 3-15. Test vs. predicted strength .......................................................................................... 3-16. Load vs. mid-span deflection: (a) slab-4, (b) slab-15, (c) slab-21 .............................. 4-1. 4-2. 4-3. 4-4. 4-5. 4-6. 4-7. Schematic model of steel deck to concrete slip .......................................................... Typical finite element model ...................................................................................... Von Mises yield surface in the principal stress space ................................................ Concrete failure surface in principal stress space....................................................... Concrete uniaxial compressive stress-strain relation.................................................. Typical shear bond shear stress vs. slip ...................................................................... (a) Shear stud to steel deck interaction, and (b) puddle weld to steel deck interaction.................................................................................................................... 4-8. General arc-length method.......................................................................................... 4-9. Slab-4: (a) Load vs. mid-span deflection. (b) Load vs. end-slip.................................
ix
4-10. Slab-15: (a) Load vs. mid-span deflection. (b) Load vs. end-slip............................... 4-11. Slab-21: (a) Load vs. mid-span deflection. (b) Load vs. end-slip............................... 4-12. Composite slab strength: FEM vs. experimental ........................................................ 5-1. Prototype 1 and prototype 2 of Ramsden (1987) deck profiles .................................. 5-2. Innovative light weight and long-span composite floor (Hillman 1990, Hillman and Murray 1994).......................................................................................... 5-3. Slimflor system (British Steel, Steel Construction Institute 1997)............................. 5-4. 6 in, 4.5 in and 3 in deep profiles................................................................................ 5-5. Yield strength and deflection limit states of the construction (non-composite) phase ................................................................................................ 5-6. Steel deck weight vs. span length of single span systems........................................... 5-7. Steel deck weight vs. span length of double span systems ......................................... 5-8. System configuration of LSS1 and LSS2.................................................................... 5-9. Strain gage and shear stud schedules of LSS1............................................................ 5-10. Strain gage and shear stud schedules of LSS2............................................................ 5-11. Test set-up ................................................................................................................... 5-12. Map of cracks in LSS1................................................................................................ 5-13. Map of cracks in LSS2................................................................................................ 5-14. Load vs. mid-span deflection of LSS1........................................................................ 5-15. Load vs. mid-span deflection of LSS2........................................................................ 5-16. Normalized relative power vs. frequency of LSS1 ..................................................... 5-17. Normalized relative power vs. frequency of LSS2 ..................................................... 5-18. Proposed beam to girder connection to reduce slab-beam height...............................
50 51 51 54 54 55 56 58 59 59 61 62 63 64 65 65 66 66 68 68 70
LIST OF TABLES
2-1. 2-2. 2-3. 2-4.
Test parameters ........................................................................................................... Summary of shear bond test results ............................................................................ Test parameters ........................................................................................................... Summary of the end anchorage test results.................................................................
7 11 14 15 36 37 49
3-1. Test parameters ........................................................................................................... 3-2. Prediction vs. test results............................................................................................. 4-1. Finite element vs. test results ...................................................................................... 5-1. Ratios of actual load capacities and permissible load based on allowable deflection to 50 and 150 psf design live loads ........................................... 5-2. Section properties of profiles 1, 2 and 3 ..................................................................... 5-3. Summary of ultimate load capacity and permissible load based on allowable deflection .................................................................................................... 6-1. vs. p f ....................................................................................................................... 6-2. Statistical data of f c ' , f y , f s,max and f s,min .............................................................. 6-3. 6-4. 6-5. 6-6. 6-7. 6-8. 6-9. Statistical data of t....................................................................................................... Statistical data of P...................................................................................................... Statistical data of dead and live loads ......................................................................... Calculated factors for SDI-M method (AISI-LRFD Approach) .............................. Calculated factors for direct method (AISI-LRFD Approach) ................................ Calculated factors for SDI-M method (AISC-LRFD Approach)............................. Calculated factors for direct method (AISC-LRFD Approach)...............................
57 58 67 73 77 78 79 79 81 81 81 81
xi
LIST OF NOTATIONS
A bf As A webs
= area of steel deck bottom flange / unit width of slab = steel deck cross sectional area = area of steel deck webs / unit width of slab = depth of concrete stress block = A sfy 0.85 f c ' b (Eqn.(3-6))
= b C c
Dn
Fs + Fst (Eqn.(3-24)) 0.85 f c ' b
= section width = resultant of concrete compressive force = depth of the neutral axis of composite section = nominal value of dead load = distance of the steel deck centroid to the top surface of the slab (effective depth) = length of each segment
dL, dL i dL c dc ds Es E o , E sc
= elongation of the bottom fiber of concrete slab of segment i = elongation of the segment at the mid-span = deflection of the partially composite section = deflection of the steel deck = elastic modulus of steel deck = initial and secant modulus of concrete
e1 , e 2 , e 3 = moment arms of T1 , T2 , T3 (Eqn.(3-9))
A = minimum anchorage force (Chapter 3) = f y A s webs A bf , (Eqn.(3-8)) 2 = fabrication factor (Chapter 6)
Fm
= mean of fabrication factor
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 anchorage = stress in the steel deck induced by end anchorages

f bond fc '
= stress in the steel deck induced by shear bond force, f b = concrete compressive strength = mean of concrete compressive strength = fc ' = stress in the steel deck induced by concrete casting = shear bond force per unit length = stress in the steel deck induced by shore removal
f c ' ,m
f cast fs f shore
f s,max , f s,min = maximum and minimum of f s

ft fw
= concrete tensile strength = stress in the steel deck induced by puddle welds = steel deck yield stress = corrected steel deck yield stress due to concrete casting and shoring = remaining strength of the steel deck = mean of steel deck yield stress = f y = elastic concrete compressive and tensile stress at the extreme fiber = concrete depth above steel deck rib = depth of the concrete flange (concrete above steel deck rib) = effective cross sectional inertia of the slab = effective cross sectional inertia of a segment = sequence number of a segment = span length of the slab = shear span length = cantilever length
fy f yc
* fy
f y,m
f1 , f2
hb h1
I eff
Ii
i L L
Lc
xiii
Ln Ls
= nominal value of live load = shear bond length = bending moment, general (Chapter 3) = material factor (Chapter 6)
M et Mm
= first yield bending moment = mean of material factor
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
Mp
Mu , Mn
= steel deck plastic moment capacity = nominal bending moment = bending moment caused by a unit load = k f c ' h b b (Eqn.(3-3)) = number of shear studs / unit width of slab = number of segment from the support to the mid-span = professional factor = mean of professional factor = probability of failure = load effect, mean of load effect = nominal strength of single shear stud = load carrying capacity: total, phase-1, phase-2, respectively = reduction factor due to insufficient number of shear studs to provide anchorage = = support reaction NrQn F
m
Nb Nr
n P
Pm pf Qi , Q m Qn q, q c , q d
Rn, Rm
= nominal resistance, mean of resistance = steel deck section modulus = total slip at a section = resultant of tensile force in steel deck
S
si
xiv
T1 , T2 , T3 = forces acting in top flange, web and bottom flange of steel deck
t
tm
d u1 c u1
= steel deck thickness = mean of steel deck thickness = t = nodal displacement of steel deck beam element in d.o.f.-1 direction (horizontal) = nodal displacement of concrete beam element in d.o.f.-1 direction (horizontal)
V, VR , VQ = coefficients of variation: general, resistance, load effect VM , VF , VP , Vfc ' , Vfy , Vt = coefficients of variation of: material, fabrication, professional factors, concrete compressive strength, steel deck yield stress, steel deck thickness VM ,SDI , VM,Direct = coefficients of variation of material factor with regard to the SDI and Direct method, respectively
Vu x, x i yc yd ys y1 , y 2
= ultimate shear capacity = distance from the support to the section being investigated = horizontal projection of y d = depth of deck c.g. from concrete c.g. = horizontal slip of steel deck relative to the concrete = moment arm of Fs and Fst , respectively = reliability index = concrete strain, concrete strain at the peak compressive stress = steel deck strain = standard normal probability function = design resistance factor = dead and live load factors = design load factor = correction due to diagonal shear cracking = fraction of the support reaction, R in Eqn.(3-11)
, cu s

D, L i
xv
, R , Q = log-normal mean: general, resistance, load effect
= coefficient of friction between the deck and concrete = mean of concrete compressive strength = f c ' ,m = mean of steel deck yield stress = f y,m = mean of steel deck thickness = t m = rotation of cross sectional plane (Chapter 4) = central safety factor (Chapter 6)
fc ' fy
t
= reinforcement ratio = A s / bd
, P , t = standard deviations: general, professional factor, steel deck thickness shear bond = shear bond strength
D D = D n + L / 1.05 n + 1 , (Eqn.(6-18)) Ln Ln
, R , Q = log-normal standard of deviation: general, resistance, load effect M m ds

i
= i (Eqn.(3-22)) M m ds
L
xvi
CHAPTER 1 INTRODUCTION
1.1. Motivation and Scope of the Research Cold-formed steel decks have been widely used in composite slab systems in steelframed structures. The system has proven to be very attractive to structural designers because of many advantages it has over conventional systems of reinforced concrete slabs. These
advantages have been listed by Finzi (1968), Oudheusden (1971), Hogan (1976), Porter and Ekberg (1976), Fisher and Buettner (1979), Porter (1985), Wright et al. (1987), Evans and Wright (1988). Among them, elimination or significant reduction of the positive moment
reinforcement and form work for concrete casting are two of the most important ones. This is in contrast to the early use (before 1950) of the steel deck-concrete floor, where the concrete is used only as a filling material (Dallaire 1971). The knowledge of this composite interaction as well as elemental behavior involved in the system has progressed rapidly during the past two decades. Much effort has been put forth to better understand and model the behavior of the system. Research on the subject has been conducted worldwide (U.S., Canada, Europe and Australia). Motivations for the research can be summarized as follows: 1. To develop an efficient composite floor system that optimally utilizes the material and thus yields an economical design. 2. To avoid the dependency on many full-scale experiments which are expensive and time consuming.
Chapter 1 . Introduction
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 fullscale 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, , for flexure design of composite slabs are also sought. The current resistant factors, , for composite slab design (Standard for 1992) were taken from the
steel or concrete design specifications. Therefore, it is desired to obtain these factors based on test results and refined analytical studies of composite slabs.
1.2. Organization of this report This report is organized as follows. Following Chapter 1, elemental tests for shear bond and end anchorages that were performed are described in Chapter 2. Results of these elemental tests were used in the analytical methods of prediction for the composite slab strength and stiffness using simple mechanical and finite element models that are presented in Chapter 3 and Chapter 4, respectively. In these chapters, by using the afore-mentioned methods, predicted strength and stiffness of experimentally tested composite slabs were compared to test results. Chapter 5 discusses the investigation toward the long span composite slab systems. New deck profiles are introduced for these long span systems and the methods described in Chapter 3 were applied to predict the strength and behavior of the slab. In Chapter 6, factors for flexural design of composite slabs are derived and discussed. Finally, conclusions and recommendations for future research are presented in Chapter 7. Note that pertinent literature is reviewed in each chapter.
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).
Chapter 2 . Elemental Tests
This fact motivates research toward an alternative solution which can reduce the number of performance tests required or replace them with smaller elemental tests that are less expensive. Such elemental tests were set forth, such as the pull-out test (Daniels 1988; Daniels and Crisinel 1993; Sonoda et al. 1994), slip-block test (Patrick 1990; Patrick and Poh 1990), concrete-block bending test (An and Cederwall 1992; An 1993), push-test (Veljkovic 1993), push-out test (Tagawa et al. 1994). These elemental test results are to be used in the analysis for the slab strength and stiffness. One may argue that these elemental test results may not directly represent the actual behavior of the composite slabs because all the affecting parameters are inseparable with each other, such as clamping force and curvature of the slab. Analytical models using shear bond elemental test results, however, have shown good agreement with the full-scale test results, which indicates that those elemental tests are applicable. Another parameter that significantly affects the strength of composite slab systems is the end anchorage. The presence of end anchorages over the support has a favorable effect on the strength of the composite slabs because these end anchorages tend to block the relative slippage of the concrete to the steel sheeting (Stark 1978; Crisinel et al. 1986a, 1986b; OLeary et al. 1987; Jolly and Lawson 1992). End anchorages can be in one of the following forms: headed shear studs welded through the deck to the supporting beams, hot rolled angles welded to the beams, or cold formed members, such as pour stops. Porter and Greimann (1984) reported an increase of 8% to 33% in composite slab strength when stud end anchorages are used. The strength expression for the headed shear studs has been established by Ollgaard et al. (1971) and has been used in the AISC Specification. This expression, however, was derived in conjunction with composite beam design in which both the concrete and the steel deck slip toward the same direction relative to the supporting beam. This is not the case with composite slab action in which the concrete moves relatively to the steel sheeting. Therefore, elemental tests for this type of end anchorages are of interest. When a longitudinal slip occurs in the composite slabs, the steel deck is being pulled-out from between the supporting beam and the concrete. The strength of the anchorage for the steel sheeting is therefore a function of the sheeting strength and thickness, and also the clamping force provided by the concrete and the steel beam due to the support reaction. Hence, elemental tests for the end anchorage are needed to determine the force provided by the anchorage to the steel deck. Very detailed and extensive elemental tests for the end anchorages were carried out
by Daniels (1988). The results from both the pull-out (shear bond) tests and the push-out (end anchorage) tests were input to finite element analyses. The analytical results were reported to compare favorably to the results of full-scale slab tests.
2.3. Shear Bond Elemental Tests The shear bond interaction at the interface of steel deck and concrete can be separated into three components, namely, the chemical bonding, mechanical interlocking, and friction between the two materials. The first component is the type of bond that is developed through a chemical process as the concrete cured. This component of interaction is brittle in nature, and once it is broken it can not be restored. The mechanical interlocking gains its strength from the interlocking action between the concrete and the steel deck due to the embossments. This action is directly affected by the embossment shape and steel deck thickness. Finally, the presence of the friction between the concrete and the steel deck is due to the presence of internal pressure between the two materials. Unlike the first, the last two components are always present although they may change in magnitude. The shear bond elemental tests were designed to obtain as much information as possible about these three components.
2.3.1. Specimen Description and Test Set Up The specimens were cast in a horizontal position so as to simulate the actual casting position for a composite slab. The size of the specimen was made 1 ft wide by 2 ft long such that it has at least one complete typical shape of the deck profile. To prevent the deck from being bent during handling, which may result in the loss of the chemical bonding, each piece of deck was fastened to a steel plate. Concrete cover above the deck was at least 2 in. to provide enough bearing area for testing. After the concrete had cured, the specimens were coupled back to back. Finally, banding strips were used to keep the concrete from falling off from the deck during handling and storing. These strips were removed before the test. Test parameters considered were concrete compressive strength, steel deck strength, thickness, rib height, profile shape and embossment type as given in Table 2-1. The profile shapes and embossment types that were used are illustrated in Fig. 2-1 and Fig. 2-2, respectively. A single test frame was designed to handle both the shear bond and end anchorage tests. The test set up for the shear bond test as shown in Fig 2-3, is intended to apply axial force, i.e., to
pull the steel deck out from the concrete. This axial force was applied through a ram that was operated manually. The magnitude of the load applied was measured through a loading rod that was instrumented with strain gages and calibrated as a tensile load cell.
Table 2-1. Test Parameters

Concrete Steel Deck Internal fc' fy Thicknss Rib ht. Profile Emboss. Pressure (psi) (ksi) (in) (in) Shape type (psf) ** SB1-1 3850 50.3 0.031 2.00 1 3 500 SB1-2 3850 50.3 0.031 2.00 1 3 300 SB1-3 3850 50.3 0.031 2.00 1 3 100 SB2-1 3850 45.4 0.034 2.00 1 1 500 SB2-2 3850 45.4 0.034 2.00 1 1 300 SB2-3 3850 45.4 0.034 2.00 1 1 100 SB2-4 3850 45.4 0.034 2.00 1 1 300* SB2-5 3850 45.4 0.034 2.00 1 1 100* SB3-1 3850 46.5 0.047 2.00 1 2 500 SB3-2 3850 46.5 0.047 2.00 1 2 300 SB3-3 3850 46.5 0.047 2.00 1 2 100 SB3-4 3850 46.5 0.047 2.00 1 2 300* SB4-1 4710 55.5 0.034 3.00 2 3 500 SB4-2 4710 55.5 0.034 3.00 2 3 300 SB4-3 4710 55.5 0.034 3.00 2 3 100 SB5-1 4710 52.1 0.056 3.00 2 3 500 SB5-2 4710 52.1 0.056 3.00 2 3 300 SB5-3 4710 52.1 0.056 3.00 2 3 100 SB6-1 4710 50.8 0.034 2.00 3 _ 500 SB6-2 4710 50.8 0.034 2.00 3 _ 300 SB6-3 4710 50.8 0.034 2.00 3 _ 100 SB7-1 3840 48.2 0.056 6.00 4 _ 300 SB7-2 3840 48.2 0.056 6.00 4 _ 100 SB8-1 3840 49.6 0.057 4.50 5 _ 300 SB8-2 3840 49.6 0.057 4.50 5 _ 200 SB8-3 3840 49.6 0.057 4.50 5 _ 128 * initial pressure, no further adjustment ** internal pressure at the interface of steel deck-concrete F or profile s hapes and embos s ment types , refer to F ig. 2-1 and F ig. 2-2, res pectively ID#
For shear bond tests, a pair of additional frames is added to induced lateral force (Fig. 24). 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
(1)
5 7 12
2.5
(2)
4.75 7.25 12 1.5
(3)
0.5 6 12 9.25
2.5 2
(4)
6 1 3.75 7.125 12.875 9 2.5 4.5 1 1.5 9 12 1.125 0.375 1.5 0.5
(5)
Figure 2-1. Profile shapes (all dimensions are in inches)
type 1
type 2
type 3
Figure 2-2. Embossment types
hydraulic ram
tension rod instrumented with strain gages clevis
Load cells
specimen
Figure 2-3. Shear bond test
nuts to adjust internal pressure
specimen
load cell
Figure 2-4. Shear bond specimen with frames for lateral force
2.3.2. Test Procedure The test was performed by applying lateral and axial forces simultaneously. The lateral force is to produce internal pressure between the steel deck and concrete. The values of these internal pressures are listed in Table 2-1. These internal pressures were obtained by adjusting the nuts on the threaded rod of the lateral frame. The pressures were monitored from the load reading obtained from the load cells placed in the lateral frame as shown in Fig. 2-5. The axial force was applied by using a hydraulic ram. This axial force pulls the steel deck out of the specimen, and therefore produces shearing stress on the interface between the concrete and steel deck. At the bottom side of the specimen, slip transducers were placed to measure the slip between the concrete and the steel deck. The loads applied along with the corresponding slip were recorded. The tests were stopped after 1 in. of slip was reached where a relatively constant plateau was achieved.
2.3.3. Test Results A summary of the test results is listed in Table 2-2. Plots of shear stress vs. slip for specimen SB2-2-A and SB6-1-B are shown in Figs. 2-5 and 2-6. In specimen SB2-2-A, as shown in Fig. 2-5, chemical bond can be observed as the vertical line at zero slip. At a shear stress level of 6.86 psi, this chemical bond failed which caused a sharp drop in the shear stress value. Beyond this point, the strength was due to the mechanical interlocking and friction. Some specimens, however, did not show a clear chemical bond response. This has been caused by a loss of chemical bond during handling. The typical characteristic of this shear bond interaction is that after the loss of the chemical bond, the strength increases until the ultimate (peak) shear stress value and it is followed by a descending curve until it reaches a relatively long horizontal plateau at the end of the descending curve. The ascending and the descending curve represent the action of the mechanical interlocking, when the concrete tries to over-ride the embossment. In this action, the steel deck stiffness that is characterized by the thickness and rib height plays an important role. After the concrete completely over-rides the embossment of the deck, the resistant to the slip is relatively constant, in which case a horizontal plateau is resulted. In the case with un-embossed deck, the response is very brittle as shown in Fig. 2-6. A
10
horizontal plateau was obtained directly after the failure of chemical bond. This is because of the absence of mechanical interlocking due to the lack of embossments. The plateau is due to friction between the steel deck and concrete as observed from the test results that an increase in the lateral pressure results in a higher shear bond strength, in particular, in the plateau portion of the response. This fact is due to the friction between the two materials.
Table 2-2. Summary of shear bond test results

Max. Shear Constant Slip at Max. Shear before slip (psi) after Shear (chemical bond) slip (plateau) (in) A B (psi) (psi) A B SB1-1 6.87 8.10 8.25 5.95 0.160 0.102 SB1-2 6.03 5.81 6.67 4.32 0.124 0.154 SB1-3 7.07 7.07 6.23 2.25 0.208 0.041 SB2-1 4.44 5.75 7.19 3.92 0.094 0.093 SB2-2 6.86 5.09 6.49 3.60 0.119 0.167 SB2-3 2.85 3.22 5.02 2.03 0.159 0.067 SB2-4 4.40 4.15 5.60 3.19 0.151 0.129 SB2-5 4.87 2.87 4.80 2.73 0.085 0.142 SB3-1 9.78 9.36 12.17 8.78 0.064 0.104 SB3-2 9.88 6.81 11.92 6.10 0.062 0.166 SB3-3 7.97 9.83 8.67 3.27 0.126 0.226 SB3-4 7.07 6.02 10.45 6.21 0.094 0.103 SB4-1 7.05 9.18 10.59 6.20 0.031 0.072 SB4-2 6.91 7.03 8.79 3.70 0.057 0.057 SB4-3 3.38 5.53 6.93 1.50 0.061 0.088 SB5-1 4.46 5.43 15.17 6.00 0.064 0.067 SB5-2 3.88 8.06 15.44 4.50 0.078 0.095 SB5-3 3.73 2.89 12.76 2.00 0.112 0.062 SB6-1 11.59 10.54 10.54 6.50 0.102 0.004 SB6-2 9.50 8.91 10.00 5.00 0.699 1.001 SB6-3 7.48 7.38 7.38 5.80 0.005 0.216 SB7-1 10.78 17.66 21.12 7.66 0.431 0.235 SB7-2 17.99 12.49 17.99 8.70 0.195 0.476 SB8-1 11.34 13.18 11.34 5.27 0.108 0.082 SB8-2 11.65 11.40 11.40 4.06 0.065 0.106 SB8-3 10.30 14.53 15.55 3.99 0.004 0.204 A and B indicate the two s pecimen halves ID#
11
8.0 7.0 SHEAR STRESS (psi) 6.0 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)
Figure 2-5. Shear stress vs. slip of specimen SB2-2-A
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)
Figure 2-6. Shear stress vs. slip of specimen SB6-1-B
12
2.4. End Anchorage Elemental Tests Three types of end anchorages were tested: headed shear studs, pour stops, and a combination of the two.
2.4.1. Specimen Description and Test Set Up Similar to the shear bond specimens, the end anchorage specimens were cast in a horizontal position. The width of the specimens was 3 ft and the concrete cover above the deck was at least 2 in. to provide enough bearing area for testing. Details of end anchorage tested are illustrated in Fig. 2-7. In specimens EA2 and EA3, the deck was puddle welded to the beam and fillet welds were used for the pour stop. After the concrete had cured, the specimens were coupled back to back. Parameters of the tests are listed in Table 2-3. For end anchorage tests, the shear bond test frame was used with a slight modification. A pair of rods was used to pull the deck out from the specimens. Figure 2-8 shows the test set up. A hydraulic ram, operated by an electric powered hydraulic pump, was put on top of the load cells and an additional frame, as shown in Fig. 2-8, was added to hold the ram. A load beam, made from a box section, was placed on top of the ram. In the space between the two specimens, several displacement transducers were placed to measure the relative slip of the concrete to the deck, and the deck to the beam.
puddle welds
EA1
EA2
EA3
Figure 2-7. Details of the end anchorage specimens
13
load beam
hydraulic ram
load cell tension rod
specimen
Figure 2-8. End anchorage test
Table 2-3. Test Parameters

Concrete Deck End fc' fy Thicknss Emboss. Rib ht. Profile Anchor. (psi) (ksi) (in) Type (in) Type Type EA1-1 4050 45.4 0.034 2 2.00 1 S EA1-2 4050 45.4 0.034 2 2.00 1 S EA2-1 4050 45.4 0.034 2 2.00 1 PS EA2-2 4050 45.4 0.034 2 2.00 1 PS EA3-1 4050 45.4 0.034 2 2.00 1 P EA3-2 4050 45.4 0.034 2 2.00 1 P End anchorage types: S=shear studs, P=pour stop, PS=pour stop and shear studs Embossment and profile type, refer to Fig. 2-2 and 2-1, respectively * Puddle weld: 3/4" visible diameter ID# No. of Studs /side 2 2 2 2 _ _ No. of Puddle Welds on deck* _ _ 4 4 4 4 Fillet Weld on pour stop _ _ 1" - 12" 1" - 12" 1" - 12" 1" - 12"
14
2.4.2. Test Procedure In this test, there was no lateral force applied to the specimens. The axial force from the ram was incremented with an interval of 5 minutes to allow the system to settle. The load and the corresponding slips were recorded and the test was stopped when failure occurred as indicated by a consistently decreasing resistance to load. As shown in Fig. 2-8, the ram pushes the load beam upward during the load test and the two rods held by this beam will pull the steel deck out of the specimens. The concrete part of the specimen is sustained by the frame.
2.4.3. Test Results A summary of the test results is given in Table 2-4. Figure 2-9 and 2-10 show load vs. deck to concrete slip for specimen EA1-1-B (shear stud end anchorages) and specimen EA2-1-A (shear stud and pour stop). The failure mode in the later specimen is deck tearing around the weld, which is typical for other specimens with deck welded to the beam. The shear studs in this case do not give significant contribution to the strength because they were not welded through the deck.
Table 2-4. Summary of the end anchorage test results

ID# Max. Load per Stud or Weld (k) 10.45 9.90 6.87 7.16 5.86 5.70 Computed Strength Stud Weld (k) (k) 26.59 _ 26.59 _ 26.59 3.03 26.59 3.03 _ 3.03 _ 3.03
EA1-1 EA1-2 EA2-1 EA2-2 EA3-1 EA3-2
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
Chapter 2 . Elemental Tests 15
propagates to the edge.
12.0 10.0 8.0 6.0 4.0 2.0 0.0 0.00 0.05 0.10 0.15 SLIP (in) 0.20 0.25 0.30
Figure 2-9. Load vs. deck to concrete slip of specimen EA1-1-B
LOAD (kips) LOAD (kips)
7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 0.00 0.05 0.10 0.15 SLIP (in) 0.20 0.25 0.30
Figure 2-10. Load vs. deck to concrete slip of specimen EA2-1-A
16
2.5. Concluding Remarks Based on the results of the shear bond test, it can be concluded that the shear bond strength is influenced by the internal pressure developed between the deck and the concrete. A more accurate determination of the internal pressure will lead to a more accurate shear bond strength prediction. This raises new issues on the relation of the internal pressure to the shear bond strength as well as the determination of the internal pressure. From the comparison shown in Table 2-4, it can be noted that the strength of the puddle welds that were resulted from the tests are approximately double to the computed single weld strength values (LRFD Cold-Formed, 1991). The strength of the anchorage by the shear stud, however, is less than half of the single stud strength computed by using the AISC (1993) specifications. In the first case, the higher strength was suspected due to the clamping effect on the deck between the concrete and the steel beam. In the later case, the lower strength was caused by the deck tearing rather than the stud shearing.
17
CHAPTER 3 STRENGTH AND STIFFNESS PREDICTIONS OF COMPOSITE SLABS BY SIMPLE MECHANICAL MODEL
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 nonstudded 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
Chapter 3 . Strength and Stiffness Prediction - Simple Mechanical Model
18
history of the slab can be obtained. A computer program is required to perform the iterations. The method is referred to as the iterative method. The second method is one in which simple expressions are used in the formulation. Thus, it is suitable for hand computation. The method is referred to as the direct method. Along with these two new methods, a modified version of the SDI method is presented. This method is referred to as the SDI-M method. The modifications include a corrected yield stress due to concrete casting and omission of the shoring effect to the steel deck yield stress. Modifications were introduced because the SDI method often yields unconservative results if the casting stresses are not introduced and it may give very unconservative results if the shoring stresses are included using the simple approach.
3.2. Review of Methods for Prediction of Composite Slab Strength by Means of Semi-Empirical Formulations and Simple Mechanical Models Although the use of cold-formed steel decks in the U.S. began as early as the 1920s, the standard design procedures for composite steel deck-concrete slabs were not formulated until much later. A landmark research program that led to a design specification for composite slabs was initiated in 1966 at Iowa State University (ISU) under the sponsorship of the American Iron and Steel Institute (Ekberg and Schuster 1968; Porter and Ekberg 1971, 1972). The results of the research led to design recommendations for composite slabs, which later became the basis for an American Society of Civil Engineers design standard for composite slabs (Standard for 1992). These design recommendations were based on two limit states, namely, the flexural and the shear bond limit states. Determination of the slab strength based on shear bond requires a series of full scale-tests. The flexure limit state is characterized by the achievement of the flexural capacity, M u (ASCE nomenclature), of the cross section at the maximum positive bending moment location, although slip between the steel deck and concrete may occur anywhere in the slab including at the end of the slab. The shear bond limit state is characterized by the occurrence of slip such that it limits the capability of a section to reach its flexural capacity. Yielding of the steel deck section, however, may occur prior to the failure. The shear bond limit state was found to be the governing limit state in most composite slab tests conducted at ISU, as well as in other research programs. The formulation of the design
19
method, which is commonly referred to as the m-k method, was chosen to follow the shear equation from the ACI Building Code (Building Code 1995). The expression was developed by Schuster (1970) and refined by Porter and Ekberg (1971). The equation for the limit state is given by:
md + k fc ' Vu = bd L'
(3-1)
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,
= 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
Vu bd f c '
k k
d L fc '
Figure 3-1. m and k shear bond regression line
20
Because shear bond was found to be the predominant failure mode of composite slabs, the focus of recent research in this area has been to study more closely the behavior of this shear bond action and to improve the performance of this action with or without adding other devices such as end anchorages. Three components were identified in the shear bond action: chemical or adhesion bonding, mechanical interlocking, and surface friction. The afore-mentioned m-k
method does not explicitly reflect the action of these components. To substantiate the effects of these actions, tests have been performed and semi-empirical formulations have been developed separately by Schuster and Ling (1980), Luttrell and Prasanan (1984), and Luttrell (1987a, 1987b). The natures of those design procedures previously described are semi-empirical which rely heavily on full-scale tests. This fact raises some problems as to how to incorporate more parameters without significantly increase the number of full-scale test required and how to crossexamine the design calculations analytically. In 1978, Stark introduced a partial interaction theory similar to that used for composite beam design. The method was developed further by Stark and Brekelmans (1990), in which they view the ultimate bending moment capacity of the slab as built up from two components: (1) the contribution of the normal force of the steel sheet and (2) the contribution of the reduced plastic moment M p ' of the deck. The formulation is given by: Mu = N b.d + Mp '
(3-2)
N b = k. f c '. h b . b
(3-3)
Nb M p ' = 1.25M p 1 A .f Mp s y
(3-4)
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
21
N b illustrated in Fig. 3-3.

k. f c '
hb
fy
Nb
z
Na
Mp '
fy
Figure 3-2. Partial interaction theory (Stark and Brekelmans 1990)
Mp ' Mp 1.0 0.8 0.6 0.4 0.2
Nb M p ' = 1.25 M p 1 - A f M p s y
0.0
0.2
0.4
0.6
0.8
1.0
Nb A sfy
Figure 3-3. Simplified relation between M p ' and N b (Stark and Brekelmans 1990)
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.
22
The advantage of the SDI procedure is that the effect of end anchorages can be taken into account in a simple manner. In the procedure, there is no distinction between ductile and brittle behavior of the slab, however, it recognizes the studded and non-studded slab condition in which generally, the studded shows ductile behavior and the non-studded sometimes has brittle behavior. The nominal moment capacity is calculated based on the expression for a singly reinforced concrete section, given by:
a M n = R.A s . f y d - 2
(3-5)
where
a =
A sf y
' 0.85f c b
(3-6)
R =
N rQn F
(3-7)
A F = f y A s webs A bf 2
(3-8)
with M n = nominal moment, A s , f c ' , f y , b, and d are previously defined, N r = number of 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 nonstudded 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,
23
respectively, and e1 , e 2 , e 3 are the corresponding moment arms of Ti s to the centroid of the compression side of concrete. Linear interpolation between the full nominal moment capacity and the first yield moment for slabs that do not have sufficient number of shear studs to provide full anchorage was introduced to the method based on the research by Terry and Easterling (1994). With this interpolation, the studded and non-studded cases can be unified. Following the development by Stark and Brekelmans (1990), Bode and Sauerborn (1992) developed a method based on the same partial interaction theory that can include the shear bond effect explicitly. To determine the strength of a composite slab, a boundary curve of the slab nominal bending moment resistance vs. the shear bond length for the particular slab for various degree of partial interaction need to be generated (see Fig. 3-4). The expression for the shear bond length is given by:
Ls =
Nb b. shear bond
(3-10)
where L s = shear bond length, N b = normal force developed in the concrete slab (see Fig. 3-4), b = slab unit width, 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.
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)
24
Patrick and Bridge (1990, 1994) developed a partial shear connection method, which is also based on partial interaction theory. In this method, the effect of the end anchorages and clamping forces over the support as well as the shear bond strength can be taken into account. Similar to the ASCE procedure, the principle of a singly reinforced rectangular concrete section is used to obtain the nominal bending moment, M n . The normal force, T, in the steel deck, which can be viewed as the reinforcing force in a concrete section, can be determined from the free body diagram shown in Fig. 3-5:
T = f s (x + L c D) + R f y A s
(3-11)
where f s = shear bond force per unit length, x = distance from the support to the section being investigated, L c = cantilever length, D = correction due to diagonal shear cracking, = coefficient of friction between the deck and concrete, R = support reaction, and = fraction of R that has some contribution in T through a frictional action. With the T value calculated from Eqn. (3-11), the corresponding M n value can be determined. However, because the shear bond 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
M T
C M T
R
x
Figure 3-5. Free body diagram of the forces acting in the composite slab section (Patrick 1990, Patrick and Bridge 1994)
25
= 0.0
Mn
= 10 .
Figure 3-6. Plot of M n vs. T (Patrick 1990, Patrick and Bridge 1994)
B A M A
Distance from the support, x
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
Chapter 3 . Strength and Stiffness Prediction - Simple Mechanical Model 26
section. In the procedure, however, the remaining strength in the steel deck, i.e. the reduced plastic moment of the deck in the method by Stark and Brekelmans (1990), is omitted and on the other hand, as can be noted from Eqn. (3-11), the clamping force at the support due to support reaction is accounted for. In their research, the shear bond strength used for the procedure was obtained from the slip block test instead of the full-scale tests.
3.3. SDI-M Method The SDI-M method is a modified version of the SDI design procedure. All the equations given by Eqns. (3-5) to (3-9) apply. The modifications are introduced by: (1) replacing f y (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.
3.4. Iterative Method The method utilizes a singly reinforced concrete beam section as the basis for the approach. All effects that help the concrete resists cracking in the positive moment regime are considered as reinforcement as indicated in Fig. 3-8. Such effects come from shear bond action ( f s ), end anchorages ( Fst ), reinforcing bars, etc. qc
Fst
fs
Figure 3-8. Reinforcing effects of some devices 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
27
as the non-composite action of the system. In phase-1, analysis is performed exactly in the same manner as one treats a singly reinforced concrete section. Two equilibrium equations are considered: equilibrium of forces and equilibrium of moments on the cross section. Assumptions used in the procedure therefore follow directly from the concrete beam section procedure, with one exception. Because in this procedure one wants to obtain the response of the slab through the entire loading history, the Whitney stress block (equivalent rectangular stress block) for the concrete is replaced by an elasto-plastic model of the stress distribution. This is illustrated in Fig. 3-9 in which, Fs and Fst are forces resulting from the effect of shear bond and end anchorages, respectively. Additional effects of welds or pour stop can be added in a way similar to Fs and Fst .
f1 fc ' h1 c C(c, f1) T(c, f1) Fs (c, f1) Fst (c, f1) M
ft f2 (c, f1)
Figure 3-9. Forces acting on the cross section
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.
The afore-mentioned shear bond force, Fs , is computed as follows.
Consider the
schematic illustration of the shear bond interaction in Fig. 3-10. Figure 3-10a shows a typical relation between shear bond force per-unit length, f s , versus slip at the interface of steel deckconcrete. This relationship is obtained from elemental tests. In general, at a certain load level, the distribution of f s along the slab is not uniform due to the difference in the amount of slip at 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 greater than the strength of the steel deck, f yc A s .
fs (a)
f s,A
fs,B
slip fs,B fs diagram
(b)
Fs
fs,A
(c)
f yc . A s
Fs limit
Figure 3-10. Shear bond interaction
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:
29
Fs = s . E s . A s Fs,limit
(3-12)
where Fs = shear bond force, 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 deck, and hence we can state: Fs,lim it f yc .A s
(3-13)
with f yc as the corrected steel deck yield stress. 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-
30
11d is used. In this case, the elongation of the bottom fiber of a segment located at x i from the support can be written as:
dL i =
xi dL c L/2
(3-14)
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
(c) d (d)
slip diagram
xi
L/2
simplified dL diagram dLi dLc
Figure 3-11. Concrete bottom fiber elongation, dL, and slip diagrams
s i = dL i = (x i + x i +1 +...+ x n )
i=1
dL c dL c = (i + (i + 1)+...+n) d L/2 L/2
(3-15)
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-
31
15) by (1 + 2 +...+ n) (1 + 2 +...+(i 1)) , the slip at a cross section can be expressed in terms of elongation of that particular segment as follow:
n(n + 1) i(i 1) 1 si = dL i 2 i 2
(3-16)
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 (a) composite action qd ds (b) non-composite action
Figure 3-12. Additional load carrying capacity from the deck
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:
* fy = f y f cast f shore f bond - f anchorage f w
(3-18)
32
where f cast , f shore , f bond , f anchorage , f w = stress in the steel deck induced by concrete casting, 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:
M m M m M m Mm = L ds = 1 1 1 ds + 2 2 2 ds + ... + n n n ds EI eff EI 1 EI 2 EI n
(3-20)
where 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:
33
1 = 1 + 2 +...+ n I eff I1 I2 In
(3-21)
where M m ds i = i M m ds
L
(3-22)
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).
3.5. Direct Method The direct method shares the same basic concept as the iterative method. In fact, the direct method is just one point, namely the ultimate load point of the iterative analysis, therefore, all assumptions of the iterative method are applicable. In this case, a fully plastic condition of the cross section is assumed and the Whitney stress block for the concrete is utilized. The stress distribution is illustrated in Fig. 3-13.
0.85 fc ' C y1 y2 Fst Fs M
Figure 3-13. Forces acting on the cross section for the direct method
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
34
still be taken into account. Partial interaction between the deck and concrete is also considered as in the iterative procedure. The nominal moment capacity provided by the composite action of the steel deck and the concrete is expressed as:
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
' 0.85f c b
a =
(3-24)
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.
3.6. Comparison of Calculated vs. Test Results Predicted values of the slab strength were made by using the iterative, direct and SDI-M methods. They were compared to experimental results. The tests were performed using several different deck profiles, embossment patterns and steel thicknesses. Different span lengths, slab
35
depths, end anchorages and concrete strengths were also incorporated in the tests. The width of the specimens was 6 ft. Loading was applied through an air bag to the top surface of the concrete slab to produce a uniformly distributed load. The test setup is shown in Fig. 3-14. Table 3-1 lists main parameters of the specimens and computed values using previously described methods are listed in Table 3-2. Test data are taken from Terry and Easterling (1994), and Widjaja and Easterling (1995, 1996). Table 3-1. Test parameters
SLAB DECK RIB STEEL EMBSM. OVERSPAN 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)
and slab-21 (non-studded slab) are shown in Fig. 3-16.
AIR BAG
Figure 3-14. Test Setup
Table 3-2. Prediction vs. test results

SLAB SDI-M # psf 1 503 2 503 3 521 4 515 5 351 6 348 7 297 8 293 9 751 10 753 11 595 12 595 13 357 14 461 ITER. DIRECT TEST psf 673 637 633 507 366 510 346 487 766 1008 672 876 443 457 psf 755 657 657 519 337 534 321 519 802 1060 658 881 388 528 psf 730 700 600 600 490 590 375 490 900 900 750 870 480 500 RATIO OF TEST/PREDICTED SDI-M ITER. DIRECT 1.45 1.08 0.97 1.39 1.10 1.07 1.15 0.95 0.91 1.16 1.18 1.16 1.40 1.34 1.45 1.69 1.16 1.11 1.26 1.08 1.17 1.67 1.01 0.94 1.20 1.18 1.12 1.20 0.89 0.85 1.26 1.12 1.14 1.46 0.99 0.99 1.34 1.08 1.24 1.08 1.09 0.95 SLAB SDI-M # psf 15 1153 16 1158 17 627 18 353 19 507 20 507 21 425 22 422 23 422 24 476 25 476 26 294 27 294 ITER. DIRECT TEST psf 798 1218 411 251 515 495 510 421 421 638 638 445 445 psf 908 1256 572 352 533 478 486 485 431 654 654 464 464 psf 1017 1185 565 368 523 523 467 494 507 621 559 498 455 RATIO OF TEST/PREDICTED SDI-M ITER. DIRECT 0.88 1.27 1.12 1.02 0.97 0.94 0.90 1.37 0.99 1.04 1.47 1.05 1.03 1.01 0.98 1.03 1.06 1.09 1.10 0.91 0.96 1.17 1.17 1.02 1.20 1.20 1.18 1.30 0.97 0.95 1.17 0.88 0.85 1.69 1.12 1.07 1.54 1.02 0.98
37
1400
-15%
1400 15%
-15%
1200
1200
15%
1000 TEST (psf) TEST (psf)
1000
800
800
600
600
400
400
200
non-studded studded
200
non-studded studded
0 200 400 600 800 1000 1200 1400
0 0 200 400 600 800 1000 1200 1400
ITERATIVE (psf)
DIRECT (psf)
1400
-15%
1200
15%
1000 TEST (psf)
800
600
400
200
non-studded studded
0 200 400 600 800 1000 1200 1400
SDI-M (psf)
Figure 3-15. Test vs. predicted strength
38
600 500 400 LOAD (psf) 300 200 100 0 0.0 1.0 2.0 3.0 4.0 5.0 test iterative analysis LOAD (psf)
1200 1000 800 600 400 200 0 0.0 1.0 2.0 3.0 4.0 5.0 test iterative analysis
MID-SPAN DEFLECTION (in)
(a)
500
(b)
400
LOAD (psf)
300
200
100
test iterative analysis
0 0.0 1.0 2.0 3.0 4.0 5.0
(c)
Figure 3-16. Load vs. mid-span deflection: (a) slab-4, (b) slab-15, (c) slab-21
39
3.7.
Concluding Remarks From the comparison and discussion presented, it can be concluded that the iterative and
direct methods generally predict the slab strength reasonably well. The methods are simple to perform and because they are based on a mechanical model rather than an empirical one, they are able to take into account parameters such as shear bond interaction and end-anchorages. Therefore, the methods can offer an alternate solution to performing full size slab tests. The SDI-M method, while not as accurate, provides a conservative design approach that is very simple to apply.
40
CHAPTER 4 STRENGTH AND STIFFNESS PREDICTIONS OF COMPOSITE SLABS BY FINITE ELEMENT MODEL
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.
Chapter 4 . Strength and Stiffness Prediction - Finite Element Model
41
4.2. Review of Finite Element Method for Composite Slabs Due to the complex nature of interactions within composite slab systems, finite element modeling has become a powerful tool in predicting the slab strength and stiffness. The power rests in the ability to locally model each different part or interaction of the system and systematically integrate contributions of those parts or interactions to represent the whole system. For composite slabs, various models have been proposed. The selection of model types depends on the physical system of the slabs and specific need of the study. Daniels et al. (1989, 1990), Ren and Crisinel (1992) and Ren et al. (1993) used planebeam elements to model one-way composite slabs. Special ten-degree of freedom beam elements that can take into account nonlinear slip behavior between the steel deck and concrete slab was used. For this purpose, a special finite element code was developed. By using ABAQUS, a commercial general-purpose finite element code, two-node plane Timoshenko beam elements were used by An (1993) for one-way slab systems. Two series of beam elements were generated, one for the concrete slab and the other for the steel deck. Shear bond interaction was modeled by using series of spring elements and additional set of equations to the stiffness equations to prescribe imposed relations among the degree of freedoms of the spring, concrete beam and steel deck beam nodes. Three dimensional brick elements were used for a two-way composite slab system. Some difficulties concerning numerical convergence was reported in the 3D model (An 1993). The problem was thought to be due to mesh sensitivity in relation to the tension-stiffening model of the concrete material. Because of this problem, the concrete material model was replaced by two different J 2 (metal) plasticity models, each of which is representing the tension and compression parts of the concrete separately. Other 3D models using brick elements were proposed by Veljkovic (1993, 1994, 1996) and Oloffsson et al. (1994). In their study, DIANA, a general-purpose finite element code was used. It was reported that some trials for concrete tension stiffening functions were needed in some cases before a stable numerical result can be obtained.
42
4.3. Finite Element Model 4.3.1. Structure Model A simply supported beam configuration is chosen as a typical model of the system. In the case with continuous deck over an interior support, a rotational spring element is added to the continuous end. The stiffness of this spring represents an elastic rotational stiffness of the adjacent span at the common support. This type of configuration (simply supported beam) is based on observations during experimental tests. Because of the absence of negative
reinforcement over interior supports, negative cracks along these supports were developed at a relatively low load level. Therefore, the assumption that the concrete slab is discontinuous over the interior support will not have any significant effect to the analysis. Two series of Euler-Bernoulli beam elements with 12 in. typical length were generated. One series is for the concrete slab and the other is for the steel deck. Only a single typical longitudinal slice of the slab is considered in the model. Vertical nodal displacements of these two series of beam elements are forced to be the same. This is based on previous study (An and Cederwall 1994) which concluded that the effect of vertical separation between the two parts is minor. End anchorages and shear bond interactions at the interface of the concrete and steel deck are modeled by using horizontal spring elements. In the case of the shear bond interaction, the spring elements are placed along the slab. One end of each spring element is attached to the steel deck beam element and the other to the concrete beam element. Both are at the steel deck centroid elevation. This means that the attachment of the spring elements to the concrete beam element is not at the centroid of concrete beam elements. In ABAQUS, this can be assigned by using the EQUATION option in which the magnitude of a certain degree of freedom can be made equal to scalar multiplications of any other degree of freedoms. This compatibility condition is shown schematically in Fig. 4-1 and can be expressed as:
y c = y d sin y d
(4-1)
d c u1 y s = u1 + yd
(4-2)
43
c u1
yc
yd
Plane of reference
c.g.c
ys
d u1
c.g.s
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., = rotation
d of cross sectional plane, y s = horizontal slip of steel deck relative to the concrete, u1 = nodal c displacement of steel deck beam element in d.o.f.-1 direction (horizontal), u1 = 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. 42.
STUD-CONCRETE INTERACTION
CONCRETE IMPOSED EQUATION FOR HORIZONTAL SLIP SHEAR BOND IMPOSED EQUATION FOR VERTICAL DSPL.
STEEL DECK STUD-DECK INTERACTION WELD-DECK INTERACTION
Figure 4-2. Typical finite element model
44
4.3.2. Material Model Incremental plastic flow theory is applied for the steel and concrete materials whereas nonlinear elasticity theory is applied for end anchorages and shear bond interaction.
J2 -
plasticity (von Mises) with associative flow rule is used for the steel material of the steel deck. In this case, the yield surface is independent of the hydrostatic component of the stress vector as shown in Fig. 4-3. Although top flange buckling at the maximum positive moment region was observed in some specimens, no buckling is assumed in the model.
hydrostatic axis 3 plane failure surface 2
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 Koiters (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,
45
concrete in tension, enabled a fracture mechanics model to be used for the tensile portion of the concrete. This model can avoid the mesh sensitivity effect of a tension-stiffening model.
Further, in this study, an associative flow and isotropic work hardening rule is assumed.
1
hydrostatic axis 3 deviatoric plane plane 2 failure surface
Figure 4-4. Concrete failure surface in principal stress space
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: Eo E + 1 + o 2 E sc cu cu
2
(4-3)
where and 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, 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.
46
Eo
E sc
cu
Figure 4-5. Concrete uniaxial compressive stress-strain relation 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 0.0 0.0 0.2
SIMPLIFIED
0.4
0.6
0.8
1.0
SLIP (in)
Figure 4-6. Typical shear bond shear stress vs. slip
47
12.0 10.0 FORCE (kips) 8.0 6.0 4.0 2.0 0.0 0.0 0.1 0.2 SLIP (in) 0.3 0.4 FORCE (kips)
3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.0 0.2 0.4 SLIP (in) 0.6 0.8
(a)
(b)
Figure 4-7. (a) Shear stud to steel deck interaction, and (b) puddle weld to steel deck interaction
4.4. Method of Analysis Among the three sources of nonlinearity: material, geometrical and boundary, only the first two are applicable to composite slab problems in this study. Both the material and
geometrical nonlinearity were applied in the analyses. The integration scheme used to trace the equilibrium path was the arclength method with a cylindrical constraint surface as suggested by Crisfield (1981). The cylindrical constraint surface converges much faster than the general spherical one. Despite the problem with inconsistency in the physical units used in its constraint equations (Yang and McGuire 1985; Chen and Blandford 1993), no serious problem related to this inconsistency was reported. However, Widjaja (1997a) shows that the method is sensitive to the selection of physical units used. A choice of units that make the order of magnitude of each d.o.f. type (rotation, translational, etc.) comparable may improve the performance of the method. Other problems were indicated by Carrera (1994), such as no convergence due to a relatively large load step, very slow or no convergence due to oscillation near the equilibrium path, or no real root that satisfies the constraint surface. These later problems can be overcome by avoiding the use of large step sizes. Figure 4-8 illustrates the method with a general constraint surface.
48
load, F
1F
nF
equilibrium path
= load proportionality factor
u1 un
displacement, u
Figure 4-8. General arclength method
4.5. Results of Analysis and Discussion Finite element analyses have been performed to simulate composite slab tests and results are listed in Table 4-1. Parameters of each slab specimens are listed in Table 3-1. Among the analysis results, load vs. mid-span deflection and load vs. end-slip response histories of slab-4 (studded slab with trapezoidal deck profile), slab-15 (studded slab with re-entrant deck profile) and slab-21 (welded slab, shored during the construction) are shown in Figs. 4-9, 4-10 and 4-11, respectively. Table 4-1. Ultimate slab capacity: finite element vs. test results
SLAB # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 FEM psf 627 617 577 543 480 565 293 480 775 790 733 799 409 364 TEST psf 730 700 600 600 490 590 375 490 900 900 750 870 480 500 RATIO TEST/ FEM 1.16 1.13 1.04 1.10 1.02 1.04 1.28 1.02 1.16 1.14 1.02 1.09 1.17 1.37 SLAB # 15 16 17 18 19 20 21 22 23 24 25 26 27 FEM psf 985 1037 506 264 537 496 456 441 408 534 534 353 353 TEST psf 1017 1185 565 368 523 523 467 494 507 621 559 498 455 RATIO TEST/ FEM 1.03 1.14 1.12 1.40 0.97 1.05 1.03 1.12 1.24 1.16 1.05 1.41 1.29
49
600 500 400 LOAD (psf) 300 200 100 0 0.0 1.0 2.0 3.0 4.0 5.0 test finite element LOAD (psf)
600 500 400 300 200 100 0 0.00 0.10 0.20 END SLIP (in) 0.30 0.40
test finite element
(a)
(b)
Figure 4-9. Slab-4: (a) Load vs. mid-span deflection. (b) Load vs. end-slip
1200 1000 800 LOAD (psf) 600 400 200 0 0.0 1.0 2.0 3.0 4.0 5.0 test finite element LOAD (psf)
1200 1000 800 600 400 200 0 0.00 0.10 0.20 0.30 0.40 0.50
test finite element
END SLIP (in)
(a)
(b)
Figure 4-10. Slab-15: (a) Load vs. mid-span deflection. (b) Load vs. end-slip
50
500
500 450
400
400 350
LOAD (psf)
LOAD (psf) test finite element
300
300 250 200 150
200
100
100 50 0
test finite element 0.02 0.04 END SLIP (in) 0.06 0.08
0 0.0 1.0 2.0 3.0 4.0 5.0
0.00
(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.
1400 -15%
1200
15%
1000 TEST (psf)
800
600
400
200
non-studded studded
0 200 400 600 800 1000 1200 1400
FEM (psf)
Figure 4-12. Composite slab strength: FEM vs. experimental
51
4.6. Concluding Remarks Comparison of the finite element results to those of the tests for a relatively wide range of parameters demonstrates the ability of the method in predicting composite slab strength and behavior. This ability may reduce the number of expensive full-scale experimental tests for composite slabs. Further, the stress-strain response history of virtually any point in the system can be obtained. In comparison to the iterative method, the nonlinear finite element method offers some advantages, such as the possibility to obtain stresses and strains at virtually any location of the slab. The method, however, requires more advanced users knowledge than the iterative method. Therefore, the iterative method is more suitable for design purposes.
52
CHAPTER 5 LONG SPAN COMPOSITE SLAB SYSTEMS
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.
Chapter 5 . Long Span Composite Slab Systems
53
Prototype 1
Prototype 2
Figure 5-1. Prototype 1 and prototype 2 of Ramsden (1987) deck profiles
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
Figure 5-2. Innovative light weight and long-span composite floor (Hillman 1990, Hillman and Murray 1990, 1994)
54
The slimflor system, marketed by the British Steel, utilizes deep deck sections of ComFlor 210 (210 mm, approximately 8.25 in. deep) and SlimDek 225 (225 mm, approximately 8.86 in. deep) sections. With lightweight concrete, the ComFlor 210 deck section can span up to 6 m (approximately 19.7 ft). Figure 5-3 shows schematic view of this slimflor system.
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
Chapter 5 . Long Span Composite Slab Systems 55
when a long span system is involved, the deflection (stiffness) limit state becomes very crucial.
9.25
profile 1
1 3.75 7.125 12.875 9 1.5 0.5
profile 2
1 1.5 9 12 1.125 0.375
4.5
profile 3
4.75 7.25 12
Figure 5-4. 6 in., 4.5 in. and 3 in. deep profiles
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.
56
Table 5-1. Ratios of actual load capacities and permissible load based on allowable deflection to 50 and 150 psf design live loads.
ultimate load capacity (psf) 1 730 2 700 3 600 4 600 5 490 6 590 7 375 8 490 9 900 10 900 11 750 12 870 13 480 14 500 15 1017 16 1185 17 565 18 368 19 523 20 523 21 467 22 494 23 507 *) based on L/360 slab # load at allow. deflection *) (psf) 345 326 238 223 310 316 301 320 374 388 352 418 399 389 407 466 301 303 396 445 229 206 246 test load / 50 ultimate load load at allow. capacity deflection 14.60 6.91 14.00 6.52 12.00 4.76 12.00 4.47 9.80 6.20 11.80 6.32 7.50 6.01 9.80 6.41 18.00 7.49 18.00 7.76 15.00 7.04 17.40 8.36 9.60 7.98 10.00 7.78 20.34 8.14 23.70 9.32 11.30 6.02 7.36 6.07 10.46 7.93 10.46 8.91 9.34 4.57 9.88 4.11 10.14 4.92 test load / 150 ultimate load load at allow. capacity deflection 4.87 2.30 4.67 2.17 4.00 1.59 4.00 1.49 3.27 2.07 3.93 2.11 2.50 2.00 3.27 2.14 6.00 2.50 6.00 2.59 5.00 2.35 5.80 2.79 3.20 2.66 3.33 2.59 6.78 2.71 7.90 3.11 3.77 2.01 2.45 2.02 3.49 2.64 3.49 2.97 3.11 1.52 3.29 1.37 3.38 1.64
5.2. Construction Phase As previously mentioned, this design phase considers the strength and stiffness of steel deck due to the fresh concrete weight. For typical span lengths, the flexural strength limit state is generally the governing condition in the design. For a longer span length, the governing
condition is shifted toward the stiffness or deflection limit state. This condition is schematically shown in Fig. 5-5. The deflection is limited to l/180, as required in the SDI Composite Deck Design Handbook (Heagler et al 1997). The 0.75 in. maximum deflection limitation was not used because it is considered to be to restrictive for long span slabs. For the purpose of this study, the construction phase is utilized to determine the profile shape of the steel deck that can be used for a desired span length. This was performed by generating charts of steel deck weight vs. span length as shown in Fig. 5-4, for various types of
57
profile shapes. For the profiles shown in Fig. 5-4 with a 2.5 in. concrete cover in a single span system, Fig. 5-6 gives the plots of the steel deck weight vs. span length. Figure 5-7 shows similar plots for double span (continuous) condition. The steel deck weight was calculated based on the deck thickness that corresponds to the required moment of inertia for a certain span length with a given concrete self-weight plus the construction load.
30
YIELD STRENGTH LIMIT STATE
25 STEEL DECK WEIGHT (lb/ft2)
DEFLECTION LIMIT STATE
20
15
10
0 6 9 12 15 18 21 24
SPAN LENGTH (ft)
Figure 5-5. Yield strength and deflection limit states of the construction (non-composite) phase 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.
Table 5-2. Section properties of profiles 1, 2 and 3

Profile # 1 2 3 Thickness (in) 0.056 0.056 0.056 Area (in2/ft) 1.694 1.380 0.895 Inertia (in4/ft) 10.54 4.70 1.49 yp (in) 3.197 2.610 1.500 weight (lb/ft2) 5.8 4.8 3.1 slab weight (lb/ft2) 61.8 48.6 51.4
58
10 profile 1 profile 2 8 STEEL DECK WEIGHT (lb/ft2) profile 3
6 16 ga 18 ga 4 16ga 18 ga 2 20 ga 20 ga 20 ga
16 ga 18 ga
0 6 9 12 15 SPAN LENGTH (ft) 18 21 24
Figure 5-6. Steel deck weight vs. span length of single span systems
10 profile 1 profile 2 8 STEEL DECK WEIGHT (lb/ft2) profile 3
6 16 ga 16 ga 18 ga 4 16ga 18 ga 2 20 ga 20 ga 20 ga 18 ga
0 6 9 12 15 SPAN LENGTH (ft) 18 21 24
Figure 5-7. Steel Deck weight vs. span length of double span systems
59
From Table 5-2, by comparing values of profiles 1 and 3, it can be seen that the steel deck moment of inertia of profile 1 is approximately 7 times higher than that of profile 3, which 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.
5.3. Service Phase In the service phase, predicted maximum test loads of composite slabs can be calculated using various ways. The iterative, direct, SDI-M and finite element methods were used to predict the capacities of the specimens built using profiles 1 and 2 in this study. However, only the iterative and finite element methods can provide response histories of load vs. deflection of the slabs. In the SDI-M and direct methods, I avg is used with an elastic analysis to obtain
permissible loads based on deflection limit state. The analysis was performed in the same ways as those with typical span length.
5.4. Specimen Descriptions and Instrumentation Long span slab 1 (LSS1) has a configuration of two single deck spans of 20 ft each and 1 ft cantilever as shown in Fig. 5-8 (a). The total slab depth was 8.5 in. (2.5 in. concrete cover above the 6 in. deck rib height). Six 3/4 in. diameter, 8-3/16 in. tall shear studs were used at each end of the slab, spaced at 1 ft on center as shown in Fig. 5-9. For long span slab 2 (LSS2), a two-span system was used with 20 ft span lengths. The configuration is shown in Fig. 5-8 (b). The total depth of the slab was 7 in. (2.5 in. concrete cover above 4.5 in. rib height). Six 3/4 in. diameter, 6-3/16 in. tall shear studs were used at the support, spaced at 1 ft on center as shown in Fig. 5-10.
60
SLAB 1 concrete slab steel deck steel deck
SLAB 2 concrete slab steel deck
20
20
Figure 5-8. System configurations of LSS1 and LSS2
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.
61
strain gage locations Section A-A A A A A A A
12 9
111
111
9 9
111
111
9 12
Strain gage locations
8.5 Section B-B B BB B
BB
12
240
240
12
Shear stud locations Figure 5-9. Strain gage and shear stud schedules of LSS1
62
strain gage locations Section A-A A A A A A A
12 9
111
111
9 9
111
111
9 12
Strain gage locations
7 Section B-B B B B
12
240
240
12
Shear stud locations Figure 5-10. Strain gage and shear stud schedules of LSS2
63
5.5. Load Test Procedure A uniform load configuration was used for the load tests. An air bag, placed on the top surface of the slab, was used for this purpose, and the load was applied by gradually increasing the pressure in the air bag. The air bag has a capacity of 20 psi in a fully constrained condition. The view of the test set-up is shown in Fig. 5-11. Each span was tested separately and in an attempt to prevent development of negative cracks into the adjacent span, crack inducers were placed along the interior supports of LSS1 and LSS2. The crack inducers were groves,
approximated 0.5 in. deep, and made along the interior supports on the top surface of the concrete when it was still wet after the casting.
air bag
Figure 5-11. Test set-up 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.
Chapter 5 . Long Span Composite Slab Systems 64
The test was terminated after 7 in. (LSS1) or 8 in. (LSS2) deflection was obtained.
5.6. Test vs. Analysis Results Before the load tests, fine cracks through the depth of the concrete were observed on the sides of the slabs over the interior supports. During the load test, as the load was increased, flexural cracks developed within the tested span. In LSS2, because of the continuity of the steel deck over the interior support, cracks appeared in the adjacent span on the top surface of the slab. Maps of the cracks of LSS1 and LSS2 after the test are shown in Figs. 5-12 and 5-13. In Fig. 513, cracks indicated by x are cracks that were developed during the test of the adjacent span.
1st test
2nd test
Figure 5-12. Map of cracks in LSS1

x x x x
2nd test
1st test
Figure 5-13. Map of cracks in LSS2
65
Flexural cracks in the positive moment regime appeared on the side of the slabs tend to turn horizontally approximately at the level of the top flange of the steel deck. This may indicate some separation of the slab portion (concrete cover) from the beam portion (concrete rib) of the concrete. Load vs. mid-span deflection response from the tests and analyses of LSS1 and LSS2 are compared in Figs. 5-14 and 5-15. It can be observed from these figures, that the response of the second test of each LSS was relatively weaker and softer compared to the first. This may be caused by damage that occurred in the adjacent span (first test), so that less (horizontal) restraint was resulted. In LSS2, the occurrence of the negative cracks before the test on the second span may have increased this effect.
700 Direct 600 Test 500 LOAD (psf) Iterative SDI-M 400 LOAD (psf) 400 Test 500 SDI-M 600 Iterative 700 Direct
300
300
200
200
100
100
0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
(a) 1st test
(b) 2nd test
Figure 5-14. Load vs. mid-span deflection of LSS1

500 450 400 350 SDI-M LOAD (psf) LOAD (psf) 300 250 200 150 100 50 0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 300 250 200 150 100 50 0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 Direct 500 450 400 Iterative 350 SDI-M Direct Test
Test
Iterative
(a) 1st test
(b) 2nd test
Figure 5-15. Load vs. mid-span deflection of LSS2
66
Predicted responses using the iterative method, as shown in Figs. 5-14 and 5-15, show reasonable agreement to those of the tests, particularly the first test of each slab. In terms of the slab strength, the direct method also shows relatively good agreement to the test results. The SDI-M method, however, predicted rather low strength (very conservative). This is due to the very low values of the reduction factor, R, based on the required anchorage forces. These were 0.545 and 0.447 for LSS1 and LSS2, respectively. Finally, a summary of the maximum test load capacity and permissible load based on the allowable deflection is given in Table 5-3.
Table 5-3. Summary of maximum test load and permissible load based on allowable deflection
ultimate load capacity (psf) LSS1a 621 LSS1b 559 LSS2a 498 LSS2b 455 *) based on L/360 slab # load at allow. deflection *) (psf) 245 210 163 121 test load / 50 ultimate load load at allow. capacity deflection 12.42 4.90 11.18 4.20 9.96 3.26 9.10 2.43 test load / 150 ultimate load load at allow. capacity deflection 4.14 1.63 3.73 1.40 3.32 1.09 3.03 0.81
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.
5.7. Evaluation of the Floor Vibrations Vibration tests on LSS1 and LSS2 were conducted prior to the load tests to determine the frequency of the fundamental mode of these slabs. For LSS1, the frequency of the fundamental mode was 10.63 Hz., and it was 8.13 Hz. for LSS2. Plots of the frequency spectra in terms of the normalized relative power vs. the frequency resulting from the tests are shown in Figs. 5-16 and 5-17.
67
1.2
1.0 NORMALIZED RELATIVE POWER
0.8
0.6
0.4
0.2
0.0 4 6 8 10 FREQUENCY (Hz) 12 14 16
Figure 5-16. Normalized relative power vs. frequency of LSS1
1.2
1.0 NORMALIZED RELATIVE POWER
0.8
0.6
0.4
0.2
0.0 4 6 8 10 FREQUENCY (Hz) 12 14 16
Figure 5-17. Normalized relative power vs. frequency of LSS2
68
Analytical calculations were made to determine if the frequencies satisfy the acceptance criteria for human comfort (Murray et al. 1997). Two criteria were considered in this case. LSS1 was classified as a footbridge with 6 ft effective width. The estimated peak acceleration was 3.13% g. This estimated peak acceleration is higher than the specified value of 1.5% g and thus the slab can not be considered satisfactory. The effective width and the occupational load of the slab influence the vibration performance of the slab. For an effective width of 15 ft and for office and residential use of the same slab, the estimated peak acceleration becomes 0.31% g, which is lower than the maximum peak acceleration limit of 0.50% g. The slab stiffness
requirement was also satisfactory (5.89 k/in, experimental, compared to the minimum requirement of 5.70 k/in). Therefore, in the later case, the slab can be considered satisfactory. These estimations, however, are rather approximate, and further investigation is necessary. The vibration response of LSS2 was not as good as those of LSS1. The estimated peak acceleration for a footbridge condition is 10.2% g compared to the maximum peak acceleration limit of 1.5% g and the experimental slab stiffness was 2.48 k/in which is below the minimum required stiffness of 5.7 k/in. For the condition with an effective width of 15 ft for office and residential purpose, the estimated peak acceleration is 0.95%, and again is greater than the specified value of 0.50% g. Further evaluations are necessary based on these preliminary
evaluations of the composite slabs.
5.8. Proposed Detail Connection The total depth of composite floor system using steel deck profiles as described in this study is relatively shallow. In comparison with the 3 in. trapezoidal deck profile using a same thickness of concrete cover, profiles 1 and 2 will result in 3 in. and 1.5 in., respectively, of additional slab depth. Therefore, typical beam to girder connection for composite slabs with regular span length can be used without adding any significant height to most structures. However, should this additional structure height be objectionable, it can be reduced or eliminated by using a beam to girder connection as shown in Fig. 5-18.
69
beam
girder
Figure 5-18. Proposed beam to girder connection to reduce slab-beam height.
5.9. Concluding Remarks A study on long span composite slab systems has been presented and two steel deck profiles have been investigated. The study, verified by experimental tests, shows very promising results on the use of relatively slim slabs (8.5 in. and 7 in. total slab depth), with almost the same concrete volume or weight as of the typical span slabs. With the proposed beam to girder connection, the slab-beam depth may be reduced to a total floor depth comparable to currently used floors. This feature of slab depth and weight promise potential advantages over the slimflor systems that are now used in European countries. The design method for the development of the deck profile by generating charts of the steel deck weight vs. the span length, and the analytical methods for the prediction of the composite strength and stiffness of the slab were shown to be good tools. These methods of analyses are very promising for the development of new deck profiles before any experimental tests. They can also reduce the number of full-scale tests needed. Permissible loads based on the deflection limit state of the service phase may become the governing limit state in the case of long span composite slabs. This limit state rarely governs the design in typical span slab systems. Therefore, in the case of long span slab systems, both the construction (non-composite) and service (composite) phases have to be evaluated carefully. Results of the evaluation of floor vibrations suggest further study be required to improve the performance of the slabs with respect to the floor vibration criteria. A deeper slab thickness with a little sacrifice in span length could be considered to give higher slab stiffness, which may improve the vibration characteristics.
70
CHAPTER 6 RESISTANCE FACTOR FOR THE DESIGN OF COMPOSITE SLABS
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, , 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 factors evaluated correspond to the SDI-M method and direct method described in Section 3.
6.2. Review of Probabilistic Concepts of Load and Resistance Factor Design Discussions on the probabilistic concepts of the LRFD approach are given in detail by many sources (Cornell 1969, Lind 1971, Ang and Cornell 1974, Galambos and Ravindra 1977, Ravindra and Galambos 1978, Ellingwood et al. 1980, Load and 1986, Hsiao et al. 1990,
Chapter 6 . Resistance Factor
71
Geschwindner et al. 1994, Commentary on 1996, Barker and Puckett 1997). In principle the following inequality applies: R n i Qi
i
(6-1)
in which, R n = nominal resistance, Q i = load effect, = resistance factor and i = load factor. The probability of failure can be expressed by: p f = 1 () where is the standard normal probability function, and is the reliability index.
(6-2)
6.2.1. Reliability Index The reliability index, , used in Eqn. (6-2), in a log-normal format, can be expressed by: R ln m Qm
2 VR 2 + VQ
R Q 2 R + 2 Q
(6-3)
where , and V, respectively, denote the log-normal mean, log-normal standard deviation and the coefficient of variation. respectively. Subscript R and Q denote the resistance and the load effect,
R m and Q m are the means of resistance and load, respectively. Introduce a
linearization given by:
2 2 VR + VQ = VR + VQ
(6-4)
then Eqn. (6-3) can further be approximated as:
72
R ln m Qm VR + VQ
(6-5)
where according to Lind (1971), for 1 / 3 VQ / VR 3 , = 0.75 gives a good approximation 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:
(VR + VQ ) Rm = e Qm
(6-6)
By minimizing the error of this central safety factor, Galambos and Ravindra (1977) suggested a value of = 0.55 which was later adopted in AISC LRFD. The reliability index, , can be determined from Eqn. (6-5). As an illustration, the following table shows some values and
the corresponding probability of failure, pf . Table 6-1. vs. pf

pf
5.0 4.0 3.0 2.0
2.9 x 10 -7
3.2 x 10-5 1.4 x 10-3 2.3 x 10-2
AISC-LRFD uses the following values: = 3.0, for members, under DL + LL or Snow = 4.5, for connections, under DL + LL or Snow = 2.5, for members, under DL + LL + Wind = 1.75, for members, under DL + Earthquake
Chapter 6 . Resistance Factor 73
whereas the AISI-LRFD uses the following values: = 2.5, for members = 3.5, for connections Galambos et al. (1982) give values for various structural members under conditions of ratio of basic specific live load to normal value of dead load equal to 1, 2 and 5. Ranges of values were also given by Ellingwood et al. (1980). These values of range between 1.9 - 3.5 for reinforced concrete members and 3.0 - 4.5 for steel members.
6.2.2. AISC LRFD Approach for the Resistance Factor Using the central safety factor given in Eqn. (6-6), the following inequality can be written:
R m Qm
(6-7)
which leads to: Rm e VR Qm e

VQ
(6-8)
or,
R n Qn
(6-9)
in which, R n and Q n are nominal values of resistance and load,
= =
R m VR e Rn Qm VQ e Qn
(6-10)
(6-11)
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,
74
fabrication, F, and the uncertainties involved in the assumptions of the engineering design equation, P (Ravindra and Galambos 1978): 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)
in which, VM , VF and VP are, respectively, the coefficients of variation of M, F and P. Here, Eqn. (6-13) assumes independent relations among M, F and P variables. By using Eqn. (6-12), the resistance factor given by Eqn. (6-10) can be modified to: = (M m Fm Pm ) e VR
(6-14)
6.2.3. AISI LRFD Approach for the Resistance Factor The AISI specification for cold-formed steel structures follows a different approach in determining the resistance factor, . The approach is based on the research by Hsiao et al. (1990). Instead of using Eqn. (6-10), it starts by expressing the effective resistance in terms of the nominal loads and load factors multiplied by a deterministic coefficient, c, that relates the load intensities to the load effect and is given by: D R n = c ( D D n + L L n ) = D n + L c L n Ln
(6-15)
where D and 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:
75
D Q m = 1.05 n + 1 c L n Ln
(6-16)
Notice that in the last equation, D m = 105 . D n and L m = L n were used (based on load statistic by Ellingwood et al. 1980). From Eqns. (6-15) and (6-16), one obtains:
Rm Rm = Rn Qm
(6-17)
with, D D = D n + L / 1.05 n + 1 Ln Ln
(6-18)
By combining Eqns. (6-3), (6-12) and (6-17), an expression of the resistance factor can be obtained:
= (M m Fm Pm ) e
2 + V2 - VR Q
(6-19)
Using this equation, determination of the coefficient can be avoided. However, the coefficient of variation of the load has to be known.
6.3. Statistical Data Evaluation of the resistance factor, , as given by Eqn. (6-14) or (6-19) requires statistical values of the parameters involved. These data are available from the lab tests
conducted on the composite slab specimens previously mentioned. However, larger sets of database are preferred to give more representative values of means, standard deviations, and coefficients of variation of the afore-mentioned parameters. Therefore, statistical values from other sources that were based on larger sets of database were used. These values are the statistical values of the concrete compressive strength, f c ' , which was based on the study by
76
MacGregor (1997), and steel deck yield stress, f y , which was based on the study on cold-formed 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.
6.3.1. Material Factor, M The material factor, M, represents the variability of the strength and stiffness of the material. In this case, M is affected by the variability of f c ' , f y , f s,max and f s,min . Statistical data of these parameters are listed in Table 6-2.
Table 6-2. Statistical data of f c ' , f y , f s,max and f s,min

fc' (MacGregor, 1997) 3940 psi 615 psi fc' (test) 3867 psi 878 psi fy (Hsiao et al. 1990) 1.100 fy 0.121 fy fy (test) 1.002 fy 0.058 fy fs,max 0.999 fs,max 0.035 fs,max fs,min 1.001 fs,min 0.073 fs,min Note: = mean, = standard of deviation, V = coefficient of variation V 0.156 0.227 0.110 0.058 0.035 0.073
Assuming that those parameters are statistically independent, coefficients of variation of the material factor can be approximated by:
VM ,SDI =
Vf2 ' + Vf2 = 0.191 c y
(6-20)
for the SDI-M method while for the Direct method:
VM ,Direct =
+ Vf2 Vf2 ' + Vf2 + Vf2 = 0.208 c y s, max s,min
(6-21)
77
for the SDI-M or direct design procedure, respectively. The mean values of M for the SDI-M and direct design procedures can be evaluated from: fy = 1.445 f y
f c ' ,m M m,SDI = fc '
f y,m f ' = c f fc ' y f y,m f y
(6-22)
f c ' ,m M m,Direct = fc ' f ' M m,Direct = c fc '
f s,max f s,min m m f s,max f s,min fs,min = 1.397 f s,min (6-23)
fy f y
fs,max f s,max '
6.3.2. Fabrication Factor, F The fabrication factor, F, represents the variability of the manufacturing process. In this case, the variability of the steel deck thickness, t, is considered. The statistical data for this steel deck thickness are listed in Table 6-3. These data were based on the measurement conducted on the steel decks that were used for the composite slab specimen tests.
Table 6-3. Statistical data of t

t 0.966 t t 0.030 t Vt 0.313
Based on the above statistical values, VF and Fm can be computed as follow:
VF = Vt = 0.313
(6-24)
Fm =
t ,m t
t = 0.966 t
(6-25)
6.3.3. Professional Factor, P The professional factor, P, takes into account the uncertainties of the design equation. This professional factor is defined as (Ravindra and Galambos 1978, Geschwindner et al. 1994):
78
P =
test prediction
(6-25)
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 aforementioned full-scale composite slab specimens, the following statistical data is obtained:
Table 6-4. Statistical data of P

SDI Direct p = Pm 1.193 1.071 p 0.244 0.183 Vp 0.205 0.172
6.3.4. Load Statistic Information regarding statistical data of the load in terms of the coefficient of variation, VQ , is needed for the AISI approach as shown in Eqn. (6-19). For this reason, statistical data of 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.
Table 6-5. Statistical data of dead and live loads

D L 1.05 Dn 1.00 Ln 0.105 Dn 0.250 Ln V 0.10 0.25
For the combination of the dead and live loads given by:
Q = D D + L L
(6-26)
the mean and standard of deviation of this combination can be expressed by: Q = D D + L L
(6-27)
79
Q =
Var(Q) =
2 2 2 2 D D + L L
(6-28)
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. (628), the coefficient of variation of Q can be obtained as:
VQ =
Dn 0.011 2 D Ln
Dn + 0.063 2 + L L / 1.05 D Ln
(6-29)
6.4. The Resistance Factor In this study, the AISI-LRFD approach is adopted. The AISC-LRFD approach is used to give a comparison. Considering the fact that composite slabs are generally used in steel framed structures, a (reliability index) value greater than 3.0 is not considered necessary ( =3.0 for steel members). Therefore, =3.0 is chosen as the target reliability index (AISI uses =2.5 as the basic case). The final result of factors, however, is rounded to the closest 0.05 and hence, the actual values used will not be exactly 3.0. A minimum limit of =2.5 is used. A load combination with D = 1.2 and L = 1.6 as given in the SDI Composite Deck Design Handbook (Heagler et al. 1997) is used as the basic load case. The combination using
D = 1.4 and L = 1.0 is not considered because the ratio of dead to live load is typically < 3.0
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, 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.
80
Table 6-6. Calculated factors for SDI-M method (AISI-LRFD Approach)

D/L 0.5 1.0 1.5 3.00 0.8790 0.8773 0.8704 2.75 0.9558 0.9497 0.9403 2.50 1.0393 1.0282 1.0158
Table 6-7. Calculated factors for direct method (AISI-LRFD Approach)

D/L 0.5 1.0 1.5 3.00 0.8112 0.8109 0.8052 2.75 0.8801 0.8757 0.8677 2.50 0.9548 0.9458 0.9350
Based on the results in Tables 6-6 and 6-7, = 0.90 is chosen for the SDI-M method and = 0.85 is selected for the direct method. For comparison, 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 and 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 between 0.65 to 0.75 show relatively close results to the AISI approach. Table 6-8. Calculated factors for SDI-M method (AISC-LRFD Approach)
3.00 2.75 2.50 0.55 1.046 1.087 1.130 0.65 0.961 1.006 1.053 0.75 0.883 0.931 0.982
Table 6-9. Calculated factors for direct method (AISC-LRFD Approach)

3.00 2.75 2.50 0.55 0.957 0.993 1.031 0.65 0.882 0.922 0.963 0.75 0.813 0.856 0.900
81
6.5. Concluding Remarks Resistance factors for the flexural design of composite slabs based on the SDI-M and direct methods have been presented. The AISI LRFD approach for evaluating the resistance factor was adopted. By this approach, the determination of the coefficient is not necessary. A target reliability index =3.0 and minimum limit of =2.5 were used. This choice was based on the target reliability =3.0 for steel members (AISC-LRFD), and the lower bound =2.5 used in AISI-LRFD for the basic load case. The resulting resistance factors are =0.90 for the SDI-M method and =0.85 for the direct method. These values were based on a range of dead to live load ratios between 0.5 and 1.5, which is representative of typical composite slab designs.
82
CHAPTER 7 CONCLUSIONS AND RECOMMENDATIONS
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.
Chapter 7 . Conclusions and Recommendations
83
Application of the methods of analysis described earlier shows a promising ability in providing analytical tools and an alternate solution to performing a large number of full-scale tests. These later tests can be replaced by elemental tests of shear bond and end anchorages, which are less expensive. The study on the long span composite slab systems indicates that the system can be used without significantly increasing the depth or weight of a floor system. This promises a potential advantage over the slimflor systems that are now used in European countries. With the long span systems, more efficient use of the material can be expected as some of the filler beams and their connection to the girders can be eliminated, and therefore, less construction work is required. More detailed study regarding the economy of the system is recommended for future research. Further study on the vibration behavior of the long span system is also recommended for future research. Finally, the study on the resistance factor, , for flexure design of composite slabs concluded that =0.90 and =0.85 can be used for the SDI-M and direct method, respectively. As more databases on the shear bond strength become available, further study of these factors is suggested. It is also recommended to extend the study to factors for other limit states of the design used for composite slab.
Chapter 7 . Conclusions and Recommendations
84
REFERENCES
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, 268282.
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.
Barker, R. M. and Puckett, J.A. (1997). Design of highway bridges. J. Wiley.
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.
British Steel. Design in Steel 3 Fast Track Slimflor.
Building Code Requirements for Structural Concrete (ACI 318-95) and Commentary (ACI 318R95). (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, H. and Blandford, G. (1993). Work-increment-control method for non-linear analysis. International Journal for Numerical Methods in Engineering, 36, 909-930.
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.
Crisfield, M. A. (1981). A fast incremental/iterative solution procedure that handles "snapthrough". Computers and Structures, 13, 55-62.
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., Nussbaumer, A. and Crisinel, M. (1989). Nonlinear Analysis of Composite Members in Bending and Shear. ICOM 223, Ecole Polytechnique Federale de Lausanne.
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.
Easterling, W. S. and Young, C. S.
(1992).
Strength of Composite Slabs. Journal of
Structural Engineering, ASCE, 118, 9, 2370-2389.
Ekberg, C. E. and Schuster, R. M. (1968). Floor Systems with Composite Form Reinforced Concrete Slabs. IABSE New York, Final Report, 385-394.
Ellingwood, B. and Ang, A. H. S. (1974). Risk-based evaluation of design criteria. ASCE Journal of the Structural Division, 100, ST9, 1771-1789.
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.
Evans, H. R. and Wright, H. D. (1988). Steel-concrete composite flooring deck structures. Steel-Concrete Composite Structures: Stability and Strength. Ed.: Narayanan, 21-52.
Finzi, L. (1968). Light gage steel floor systems provided to include utilities - proposal and experiments. IABSE New York, Final Report, 367-374.
Fisher, J. M. and Buettner, D. R. (1979). Application of light gauge steel in composite construction. Handbook of Composite Construction Engineering. Ed.: G. M. Sabnis, 80-96.
Fracture mechanics of concrete: concepts, models and determination of material properties. state of the art report by ACI Committee 446. (1992). Proceedings of the First International Conference on Fracture Mechanics of Concrete Structures. Ed.: Zdenek P. Bazant.
Galambos, T. V., Ellingwood, B., MacGregor, J. G. and Cornell, C. A. (1982). Probability based load criteria: Assessment of current design practice. ASCE Journal of the Structural Division, 108, ST5, 959-977.
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.
Heagler, R. B., Luttrell, L. D. and Easterling, W. S. Handbook. Steel Deck Institute, Canton, Ohio.
(1991).
Composite Deck Design
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.
Heagler, R. B., Luttrell, L. D. and Easterling, W. S. Handbook. Steel Deck Institute, Canton, Ohio.
(1997).
Composite Deck Design
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, 513521.
Hogan, T. J. (1976). Composite Steel Beams in Buildings. Australian Institute of Steel Construction.
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.
Karihaloo, B. L. (1995). Fracture Mechanics and Structural Concrete. Longman Scientific and Technical, England.
Koiter, W. T. (1953). Stress-strain relations, uniqueness and variational theorems for elasticplastic materials with a singular yield surface. Quarterly of Applied Mathematics, 11, 3, 350354.
Lawson, R. M., Mullett, D. L. and Rackham, J. W. (1997). Design of Asymmetric Slimflor Beams Using Deep Composite Decking. Steel Construction Institute, Berkshire.
Load and resistance factor design specification for structural steel buildings. American Institute of Steel Constructions, Chicago, Illinois.
(1986).
Lind, N. C. (1971). Consistent partial safety factors. ASCE Journal of the Structural Division, 97, ST6, 1651-1669.
LRFD Cold-Formed Steel Design Manual. Washington, D. C.
(1991).
American Iron and Steel Institute,
Luttrell, L. D. (1987a). Flexural strength of composite slabs. Composite Steel Structures. Ed.: R. Narayanan, 106-116.
Luttrell, L. D.
(1987b).
Methods for predicting strength in composite slabs.
Building
Structures. Ed. : D. R. Sherman, 425-437.
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.
OLeary, 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.
Olofsson, T., Noghabai, K. and Veljkovic, M. (1994). Computational modeling of interaction between concrete and steel. Diana World, 1, Diana Analysis bv.
Ottosen, N. S. (1977). A failure criterion for concrete. Journal of the Engineering Mechanics Division, 103, EM4, 527-535.
Oudheusden, A. J. (1971). Composite construction-applications. Proc. 1st. International Specialty Conference on Cold-Formed Steel Structures. Ed.: W. W. Yu, 173-178.
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. (1971). Investigation of cold-formed steel-deck-reinforcedconcrete floor slabs. Proc. 1st. International Specialty Conference on Cold-Formed Steel Structures. Ed.: W. W. Yu, 179-185.
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.
Razaqpur, A. G. and Nofal, M.
(1990).
Analytical modeling of nonlinear behavior of
composite bridges. Journal of Structural Engineering, 116, 6, 1715-1733.
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.
Schuster, R. M.
(1970).
Strength and Behavior of Cold-Rolled Steel-Deck-Reinforced
Concrete Floor Slabs. PhD Dissertation, Iowa State University.
Schuster, R. M. and Ling, W. C. (1980). Mechanical interlocking capacity of composite slabs. Proc. 5th. International Specialty Conference on Cold-Formed Steel Structures. Ed.: W. W. Yu, 387-407.
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.
Stark, J. W. B. and Brekelmans, J. W. P. M. (1990). Plastic design of continuous composite slabs. Journal of Construction Steel Research, 15, 23-47.
93
Tagawa, Y., Suko, M. and Tsushima, A. (1994). Push-out test of composite slab. SteelConcrete Composite Structures, Proceedings of the 4th International Conference by ACCS, 316319.
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.
(1993).
Development of a New Sheeting Profile for Composite Floors.
Research Report Tulea 1993:47, Division of Steel Structures, Lulea University of Technology.
Veljkovic, M.
(1994).
3D nonlinear analysis of composite slabs. Diana Computational
Mechanics 94. Eds.: G. M. A. Kusters, M. A. N. Hendriks, 395-404.
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. (1997a). Path-following technique based on residual energy suppression for nonlinear finite element analysis. To appear in Computers and Structures.
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. (1995). Evaluation of Composite Slabs Using Versa-Dek Profile. Report No. CE/VPI-ST 95/02, Department of Civil Engineering, Virginia Polytechnic Institute and State University.
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 NUMETA85 Conference. Eds.: J. Middleton, G. N. Pande, 913921.
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

Analysis and Design of Steel Deck - Concrete Composite Slabs

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Analysis and Design of Steel Deck - Concrete Composite Slabs

Uploaded by

Copyright:

Available Formats

ANALYSIS AND DESIGN OF STEEL DECK CONCRETE COMPOSITE SLABS

Budi Ryanto Widjaja

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

ANALYSIS AND DESIGN OF STEEL DECK CONCRETE COMPOSITE SLABS

Budi R. Widjaja Dr. W. S. Easterling, Chairman Department of Civil Engineering (ABSTRACT)

In Memory of my Father and In Love of my Mother

Chapter 7. Conclusions and Recommendations..............................................................

References............................................................................................................................. VITA .....................................................................................................................................

2-1. 2-2. 2-3. 2-4.

Fs + Fst (Eqn.(3-24)) 0.85 f c ' b

e1 , e 2 , e 3 = moment arms of T1 , T2 , T3 (Eqn.(3-9))

A = minimum anchorage force (Chapter 3) = f y A s webs A bf , (Eqn.(3-8)) 2 = fabrication factor (Chapter 6)

= mean of fabrication factor

f anchorage = stress in the steel deck induced by end anchorages

f s,max , f s,min = maximum and minimum of f s

= first yield bending moment = mean of material factor

, R , Q = log-normal mean: general, resistance, load effect

, R , Q = log-normal standard of deviation: general, resistance, load effect M m ds

CHAPTER 2 ELEMENTAL TESTS

Chapter 2 . Elemental Tests

Chapter 2 . Elemental Tests

Chapter 2 . Elemental Tests

Table 2-1. Test Parameters

Chapter 2 . Elemental Tests

Figure 2-1. Profile shapes (all dimensions are in inches)

Figure 2-2. Embossment types

Chapter 2 . Elemental Tests

tension rod instrumented with strain gages clevis

Figure 2-3. Shear bond test

nuts to adjust internal pressure

Chapter 2 . Elemental Tests

Chapter 2 . Elemental Tests

Table 2-2. Summary of shear bond test results

Chapter 2 . Elemental Tests

Figure 2-5. Shear stress vs. slip of specimen SB2-2-A

Figure 2-6. Shear stress vs. slip of specimen SB6-1-B

Chapter 2 . Elemental Tests

Figure 2-7. Details of the end anchorage specimens

Chapter 2 . Elemental Tests

load cell tension rod

Figure 2-8. End anchorage test

Table 2-3. Test Parameters

Chapter 2 . Elemental Tests

Table 2-4. Summary of the end anchorage test results

EA1-1 EA1-2 EA2-1 EA2-2 EA3-1 EA3-2

propagates to the edge.

Figure 2-9. Load vs. deck to concrete slip of specimen EA1-1-B

LOAD (kips) LOAD (kips)

Figure 2-10. Load vs. deck to concrete slip of specimen EA2-1-A

Chapter 2 . Elemental Tests

Chapter 2 . Elemental Tests

Chapter 3 . Strength and Stiffness Prediction - Simple Mechanical Model

Chapter 3 . Strength and Stiffness Prediction - Simple Mechanical Model

Figure 3-1. m and k shear bond regression line

Chapter 3 . Strength and Stiffness Prediction - Simple Mechanical Model

Chapter 3 . Strength and Stiffness Prediction - Simple Mechanical Model

N b illustrated in Fig. 3-3.

Figure 3-2. Partial interaction theory (Stark and Brekelmans 1990)

Mp ' Mp 1.0 0.8 0.6 0.4 0.2

Chapter 3 . Strength and Stiffness Prediction - Simple Mechanical Model

Chapter 3 . Strength and Stiffness Prediction - Simple Mechanical Model