DEEL STRUCTURES ar cr | iad ae || CHARLES G. SALMON JOHN E. JOHNSON EMPHASIZING LOAD AND RESISTANCE FACTOR DESIGN THIRD EDITIONC4 e1 ey G Cys Cos Coy @ Cour Ges Gory G & 4 bas mit yr Se fe = fillet weld leg dimension; ie., weld size, in. (Chap. 5); length (dimension parallel to load) of plate (Fig, 6.14.7); /EC,/GI = 1/% (Chap. 8); stiffener spacing (Chap. 11) = area of the cross-section; area = gross cross-sectional area of threaded fastener (bolt) ‘= gross area of threaded fastener based on major diameter = effective net area A, of tension member = gross area of one flange = gross area of cross-section = net area; “tensile stress area” through threaded portion of tension rod; net area through holes on tension member = net area subject to shear fracture (Sec. 3.6) = net area subject to tensile fracture (Sec. 3.6) area of stiffener = gross area subject to tensile yielding (Sec. 3.6) = gross area subject to shear yield (See. 3.6) = web area; df, for rolled I-shaped beams; hr,, for plate gitders = width; beam width; flange width; dimension of plate perpendicular to load direction (Fig. 6.14.7) = effective width of unstiffened compression element (Sec, 6.18); effective slab width (Chap. 16) = flange width, usually for steel W section; for beam (by,) = width of stiffener = magnification factor for member in braced frame (no lateral transtation at ends. of member), Eq. 12.103 = magnification factor for sway analysis of member in unbraced frame, Eqs, 12119 of L2ILIL tance from neutral axis to extreme fiber where stress is computed; measured. in x- or y-direction according to subscript = coefficient for columas such that F, = COF, = factor to account for moment gradient in beam strength, Eq. 96.11; with respect to bending about x- or y-axis, according to second subscript = slenderness ratio KL/r separating long and short columns for ASD, Eq, 6.7.11 = factor in moment magnification relating to moment gradient and end restraint (defined in Secs. 12.10 for LRED and 12.14 for ASD; also Table 123.1); with respect to bending about a- or y-axes, according to second subscript = Tal't = torsional ‘warping constant (Chap. 8}; compression force in web of steel section «Chap. 16) = overall depth of steel section; nominal bolt diameter = distance from outside of ange to PNA (Chap. 16) = diameter of stud = £112 — p*)] = flexural rigidity per unit length of plate (Chap. 6 Part IT and Eg, 11.11.14; factor to account for eccentric load on stiffeners, Eqs. 11.9.28 and 11.935; width of fastener hole to be deducted (Sec. 3.4); number of sixteenths of an inch in weld size = eccentricity of load = tension-compression modulus of elasticity . = modulus of elasticity of concrete = (w!®)/fi', where w is the density of conerete in pef (ie., 145 pef for normal-weight concrete) and is in ksi = modulus of elasticity of steel, 29,000 ksi = stress due to direct shear = P/A = stress. due to torsional moment = Tr/I, = service load axial tension stress; T/A, or T/A, (Chap. 3); service load axial stress = P/Ay service load bending stress = M/STow Soar toy fa Son Son tes i Fis Foxy Foy E F =a" Sa Eye, hes ty ey hyve K, Ky, KK, Ppp e = normal stress due to restraint of warping (lateral bending of flanges) = (Eph /4Xd¢/de*) = service Joad flexural stress based on primary moment, about the x-axis = M/S., or the y-axis = M/S, = specified 28-day compressive strength of concrete = factored normal. (compression or tension) stress = nominal stress in shear at factored load, Ry/Ay = service load shear stress, V/4,, oF R/Aj, (bolts) = service load shear flow (kips/im) (See. 11.11) = allowable service load axial stress in ASD, Eqs. 6.7.9 of 6.7.10 “= allowable bending stress M/S in ASD at service load, bending about the x- or ‘pmaxis according to the second subscript = critical stress in compression; buckling stress = Eq, 12.1411 = weld metal tensile strength = residual stress = factor of salety = tensile strength of structural stee]; tensile strength of base nietal (Chap. 5) = tensile strength of bolt material = design sivess\ in tension on bolt subject. to combined shear and tension (R;,/Ay), see Table 4.14.1 = allowable shear stress (for ASD) = allowable shear stress in the presence of tension (slip-critical connections) = allowable tension stress in bolt ~ yield stress forbear ange (F ” for column web (F,,); for flange (Fy); for login reinforcing bars (#,); for stiffeners (F,,,J; for web or weld metal = gage distance for fastener holes measured transverse to direction of load = shear modulus of elasticity £/{20 + p)) = distance between centers of flanges (Chap. 8); overall depth (Fig. 6.13.1); depth ‘of web plate (Chap. 11); story height (Chaps. 14 and 13) = unsupported web height (For unsymmetrical sections it relates to the compres: sion side of the neutral axis.) (see LRFD-B5.1) = factored lateral force causing sway deflection = moment of inertia of one flange about y~ axis (Chap. 8) = girder moment of inertia (Chaps. 14 and 15) = polar moment of inertia about shear center ‘= moment of inertia of the cross-sectional area of a transverse stiffener as defined. following Eq. ELL ‘= transformed cracked section moment of inertia of composite section = moment of inertia, about the x- or y- axis, respectively = product of inertia ~ /xyd4 (See, 7.10) = torsion constant - (ele (Becs, 62 and 12:2); distance from outer face of flange to web toe of let (See. 7.8); plate buckling coefficient, Eq. 6.14.28; spring constant (See. 9.13); spring constant = EP,/h, Eq. 1459 = coefficients relating to fillet weld strength; Eqs. 5.17.4 and $.17.7 effective length factor (Sec. 6.9), with respect to the x, y>, or z axes, respectively. = length; span; end distance measured in direction af line of force (Sec. 4.7) = laterally unbraced length = ASD; maximum laterally unbraced length for using F, ~ 0.66%, = LRFD: maximum laterally unbraced length for using M, = My, Bq. 9.63 = LRFD: maximum laterally unbraoed length for using plastic analysis, Eq. 9.62. = LRFD: maximum laterally unbraced length for using, M, > M,L, = ASD; maximum laterally unbraced length for using allowable stress Fi, ~ 0.60, when C, = 1 Ly ‘= length of fillet weld m = number of shear planes (Chap, 4); uniformly distributed service load torsional moment (Chap. 8) my = factored uniformly distributed torsional moment (Chap. 8) M = bending moment; service load moment (unfactored) for ASD MyM; = smaller moment (M,) and larger moment (iM,) at the ends of a laterally lunbraced segment Me = clastic laterat-torsional buckling moment strength Mp = equivalent constant moment = C,,iMf, for beam-column in braced frame subject toend moments M, and M, only (See. 12.4) My = lateral bending moment on one flange = (EC, /h\d*¢/d:") M, = primary bending moment from first-order analysis My, = first-order moment in sway analysis under H, force M, ‘= nominal moment strength My = nominal moment strength in the presence of shear My, = primary factored moment for the no translation case of beam-column Maes May = nominal moment strengths about x- and »- axes, respectively (Secs. 7.11 and 9.14) My. Myus Myy = plastic moment strength = ZF, with respect to the 4- and + axes, according to second subseript = moment sirength when extreme fiber reaches (F, ~ F,) = pure torsional moment (St, Venant torsion) (Chap. 8) + Myc Myy = factored service load moment, about the 3+ or y axes, according to second subscript Lace ~ equivalent factored bending moment on beam-column lp = factored moment on one flange Ia = warping torsional moment (lateral bending effect) (Chap. 8) ‘= moment about x-axis; bending moment from stress 0, (Sec. 6.14) = torsional moment from shear stress 1,, (Sec. 6.14) = bending moment from stress «, (Sec. 6.14); moment about y-axis when biaxial tending is being considered (Chaps. 8 and 9; See 129); nominal moment strength M, when extreme fiber reaches F,,M,= FS, (except for biaxial ‘bending) = bending, moment or torsional moment at distance 2 along axis of member; for torsion, M, = M, + M, = modulus of elasticity ratio, E,/£,; number of fasteners (Chap. 4) = bearing length (Sec. 7.8) = pitch (spacing) of bolts; connector spacing (Chap. 16) = service axial load; factored load per bolt (Chap. 4) "7 = design strength of column web to resist a concentrated factored load Pa = critical buckling load; compression force at buckling PoP Py = Euler load = 9°EA,/(KE/r)? for axis of bending (using two subscripts for biaxial bending) EEEE ERE = x = poe ee Pag = equivalent column load for beam-column in ASD P, = nominal strength of an axially loaded compression member, £,4,; nominal strength of weld configuration (Fig. 5.17.2) z, = factored axial Toad (Sec. 1.9); factored reaction or load Prsg = equivalent factored column load for beam-column B, = yield load = FA, (Chap. 12) ¢ = distance from shear center to midethickness of web: on channel Q = form factor = Q,0, (Sec. 6.18); first moment of area (e., statical moment {ydA) about the neutral axis from extreme fiber to-section at which elastic shear Ssteess is computed (see Sec. 7.7) 2, = shape factor for stiffened compression element (Sec. 6.18)Steel StructuresSTEEL STRUCTURES DESIGN AND BEHAVIOR Emphasizing Load and Resistance Factor Design Third Edition Charles G. Salmon John E. Johnson ‘University of Wisconsin-Madison Biri epeiCollinsPablishersProject Coordination: Science Typographers, Inc, Cover Coordinator: Heather A. Ziegler Cover Design: Circa 86, Inc. Photo: Bill Ross, Woodfin Camp reduction: Beth Maglione STEEL STRUCTURES: Design and Behavior, Emphasizing Load and Resistance Factor Design, Third Edition Copyright © 1990 by HarperCollins Publishers Ine, All rights reserved. Primed in the United States of America. No part of ‘his book may be used of reproduced in any manner whatsoever without written permission, exeept in the case of brief quotations embodied in critical articles and reviews, For information address Harper CollinsPublishers Inc. 10 East $3rd Street, New York, NY 10022. Library of Congress Cataloging-in-Publication Data Salmon, Charles G. Steel structures: design and behavior: emphasizing load and resistance factor design /Charles G, Salmon, John E. Johnson. — 3rd ed, Pp cm Includes bibliographical references. ISBN 0-06-045703-1 L.Steel, Structural, I. Johnson, John Edwin, 1931- IL. Title TAG84.S24 1990 624,1°821—de20 89-6768 cp M 92 9876543Contents Preface xiii Conversion Factors xvii Chapter 1 - Introduction 1 1 12 13 14 15 16 wy 18 19 1.10 ll Structural Design = 1 Principles of Design — 1 Historical Background of Steel Structures 3 Loads 4 Types of Structural Steel Members 3 Steel Structures 18 Specifications and Building Codes 24 Philosophies of Design 25 Factors for Safety—ASD and LRFD Compared 31 Why Should LRFD be Used? 36 Analysis of the Structure 37 Selected References 38 Coa PS Steels and Properties 41 a 23 24 25 2.6 27 28 Structural Steels 4. Fastener Steels 47 Weld Electrode and Filler Material 49 Stress-Strain Behavior (Tension Test) at Atmospheric Temperatures 50 Material Toughness 52 Yield Strength for Multiaxial States of Stress 54 High Temperature Behavior 56 Cold Work and Strain Hardening 58vi CONTENTS: 29 2.10 211 212 Brittle Fracture 60 Lamellar Tearing 64 Fatigue Strength 66 Corrosion Resistance and Weathering Steels 67 Selected References 69 Chapier 3- Tension Members 72 7 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 Introduction = 72. Nominal Strength 72 ‘Net Area 15 Effect of Staggered Holes on Net Area 76 Effective Net Area 80 Tearing Failure at Bolt Holes 84 Stiffness as a Design Criterion 86 Load Transfer at Connections 86 Load and Resistance Factor Design-Tension Members 88 Tension Rods 95 Allowable Stress Design—Tension Members 98 Selected References 99 Problems 100 ‘Chapter 4 - Structural Fasteners 106 41 42 43 44 45 46 47 48 49 4.10 4.11 4.12 413 414 415 ‘Types of Fasteners 106 Historical Background of High-Strength Bolts 109 Causes of Rivet Obsolescence = 111 Details of High-Strength Bolts 111 Installation Procedures 11s Nominal Strength of Individual Fasteners 118 Load and Resistance Factor Design— Fasteners 123 Examples—Tension Member Bearing-Type Connections—LRFD 129 Slip-Critical Joints 135 Allowable Stress Design—Fasteners 140 Examples—Tension Members Using Allowable Stress Design 144 Eccentric Shear 146 Fasteners Acting in Axial Tension 167 ‘Combined Shear and Tension 171 Shear and Tension from Eccentric Loading 183 Selected References 190 Problems 191 Chapter 5- Welding 202 31 32 3.3 Introduction and Historical Development 202 Basic Processes 204 Weldability of Structural Steel 2105.4 3.5 5.6 5.7 5.8 5.9 5.10 3.1 5.12 5.13 5.14 5.15 5.16 5.17 5.18 5.19 CONTENTS vii Types of Joints 211 Types of Welds 213 Welding Symbols 216 Factors Affecting the Quality of Welded Connections 219 Possible Defects in Welds 223 Inspection and Control 225 Economics of Welded Built-up Members and Connections 226 Size and Length Limitations for Fillet Welds 228 ‘ Effective Areas of Welds 231 Nominal Strength of Welds 233 Load and Resistance Factor Design—Welds 237 Allowable Stress Design—Welds 244 Welds Connecting Members Subject to Direct Axial Load 248 Eccentric Shear Connections—Strength Analysis 261 Eocentric Shear Connections—Elastic (Vector) Analysis 272 Loads Applied Eccentric to the Plane of Welds 279 Selected References 285 Problems 287 Chapter 6- Compression Members. 298 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 69 6.10 6.11 6.12 6.13 6.14 6.15 6.16 6.17 PART I: COLUMNS General 298 Euler Elastic Buckling and Historical Background 298 Basic Column Strength 301 Inelastic Buckling 302 Residual Stress 307 Development of Column Strength Curves Including Residual Stress 309 Structural Stability Research Council (SSRC) Strength Curves 319 Load and Resistance Factor Design 324 Effective Length 328 Load and Resistance Factor Design of Rolled Shapes (W, S, and M) Subject to Axial Compression 335 Allowable Stress Design 342 Shear Effect 344 Design of Latticed Members 347 PART Il: PLATES Introduction to Stability of Plates 352 Strength of Plates Under Uniform Edge Compression 362 AISC Width/Thickness Limits 4, to Achieve Yield Stress Without Local Buckling | 365 AISC Width/Thickness Limits \,, to Achieve Significant Plastic Deformation 370viii CONTENTS 6.18 AISC Provisions to Account for the Buckling and Post-Buckling Strengths of Plate Elements 372 6.19 Design of Compression Members as Affected By Local Buckling Provisions 378 Selected References 385 Problems 389 ‘Chapter 7 - Beams: Laterally Supported — 396 7.1 Introduction —- 396. 7.2. Simple Bending of Symmetrical Shapes 397 7.3 Behavior of Laterally Stable Beams 398 7.4 Laterally Supported Beams—Load and Resistance Factor Design 401 7,5 Laterally Supported Beams— Allowable Stress Design 7.6 Deflection 410 7.7 Shear on Rolled Beams 418 7.8 Concentrated Loads Applied to Rolled Beams 424 7.9 Holes in Beams 429 7.10 General Flexural Theory 431 7.11 Biaxial Bending of Symmetrical Sections 438 ‘Selected References 443 Problems 445 Chapter 8- Torsion 451 8.1 Introduction 451 8.2 Pure Torsion of Homogeneous Sections 452 8.3 Shear Stresses Due to Bending of Thin-Walled Open Cross-Sections 455 8.4 Shear Center 457 8.5 Torsional Stresses in I-Shaped Steel Sections 460 8.6 Analogy Between Torsion and Plane Bending 472 8.7 Practical Situations of Torsional Loading 477 88 Load and Resistance Factor Design for Torsion— Laterally Stable Beams 481 89 Allowable Stress Design for Torsion— Laterally Stable Beams 487 8.10 Torsion in Closed Thin-Walled Sections 488 8.11 Torsion in Sections with Open and Closed Parts 494 8.12 Torsional Buckling 494 Selected References 502 Problems 504 Chapter 9- Lateral-Torsional Buckling of Beams 513 9.1 Rational Analogy to Pure Columns 513 9.2 Lateral Support 514 9.3. Strength of I-Shaped Beams Under Uniform Moment 9.4 Elastic Lateral-Torsional Buckling 51895 96 9.7 98 99 9.10 9.11 9.12 9.13 9.14 CONTENTS | ix Inelastic Lateral-Torsional Buckling 523 Load and Resistance Factor Design—I-Shaped Beams Subjected to Strong-Axis Bending 526 Allowable Stress Design—I-Shaped Beams Subjected to Strong-Axis Bending = 531 Effective Laterally Unbraced Length 543 Examples: Load and Resistance Factor Design 545 Examples: Allowable Stress Design 560 Weak-Axis Bending of I-Shaped Sections 569 Lateral Buckling of Channels, Zees, Monosymmetric I-Shaped Sections, and Tees 570 Lateral Bracing Design 579 Biaxial Bending of Doubly Symmetric Sections 592 Selected References 597 Problems 601. Chapter 10- Continuous Beams 609 10.1 10.2 10.3 10.4 10.5 10.6 Introduction 609 Plastic Strength of a Statically Indeterminate Beam 610 Plastic Analysis—Load and Resistance Factor Design Examples 620 Elastic Analysis—Load and Resistance Factor Design Example 636 Elastic Analysis—Allowable Stress Design Examples 638 Splices 644 Selected References 646 Problems 647 Chapter 11- Plate Girders 651 11 2 11.3 114 1L5 116 17 1L8 11.9 11,10 1.11 11.12 1L.13 Introduction and Historical Development 651 Difference Between Beam and Plate Girder 654 Vertical Flange Buckling Limit State 656 Nominal Moment Strength—Load and Resistance Factor Design 660 Moment Strength—Allowable Stress Design 664 Moment Strength Reduction Due to Bend-Buckling of the Web 665 Nominal Moment Strength—Hybrid Girders 673 Nominal Shear Strength—Elastic and Inclastic Buckling 675 Nominal Shear Strength—Including Tension-Field Action 683 Strength in Combined Bending and Shear 694 Intermediate Transverse Stiffeners 699 Bearing Stiffener Design 706 Longitudinal Web Stiffeners 709x CONTENTS 11.14 11.15 Proportioning the Section 711 Plate Girder Design Example—LRFD 718 Selected References 742 Problems 744 ‘Chapter 12-Beam-Columns 7511 121 122 12.3 12.4 12.5 12.6 12.7 128 12.9 12.10 12,11 12.12 12,13 12.14 12.15 Introduction 751 Differential Equation for Axial Compression and Bending 752 Moment Magnification—Simplified Treatment for Members in Single Curvature Without End Translation 757 Moment Magnification—Members Subject to End Moments Only; No Joint Translation 767 Moment Magnification—Members with Sidesway Possible 764 Nominal Strength—Instability in the Plane of Bending 765 Nominal Strength—Failure by Combined Bending and Torsion 767 Nominal Strength—Interaction Equations 768 Biaxial Bending 771 Load and Resistance Factor Design Criteria 773 Unbraced Frame—Load and Resistance Factor Design 778 Design Procedures—Load and Resistance Factor Design 786 Examples—Load and Resistance Factor Design 788 Allowable Stress Design Criteria. 812 Design Procedures—Allowable Stress Design 818 Selected References 829 Problems 832 Chapter 13-Connections 843 13.1 13.2 13.3 13.4 13.5 13.6 137 13.8 13.9 13.10 Types of Connections 843 Framed Beam Connections 847 Seated Beam Connections— Unstiffened 864 Stiffened Seat Connections —_ 871 Triangular Bracket Plates 878 Continuous Beam-to-Column Connections 884 Continuous Beam-to-Beam Connections 918 Rigid-Frame Knees 919 ‘Column Base Plates 927 Beam Splices 933 Selected References 938 Problems 943 Chapter 14 - Frames—Braced and Unbraced = 950 14.1 14.2 143 General 950 Elastic Buckling of Frames 954 General Procedure for Effective Length 96314.4 14.5 14.6 CONTENTS. xi Stability of Frames Under Primary Bending Moments 963 Bracing Requirements—Braced Frame 969 Overall Stability when Plastic Hinges Form 974 Selected References 975 Chapter 15 - Design of Rigid Frames 978 15.1 15.2 15.3 15.4 Introduction = —- 978 Plastic Analysis of One-Story Frames 978 Load and Resistance Factor Design— One-Story Frames 991 Multistory Frames 1006 Selected References 1007 Problems 1008 Chapter 16 - Composite Steel-Concrete Construction 1010 16.1 16.2 16.3 16.4 16.5 16.6 16.7 16.8 16.9 16.10 16.11 16.12 16.13 16.14 16.15 16.16 16.17 Appendix Table Al Table A2 Index Historical Background = 1010 ‘Composite Action 1013 Advantages and Disadvantages 1015 Effective Width 1016 ‘Computation of Elastic Section Properties 1018 Service Load Stresses With and Without Shoring 1021 Nominal Moment Strength of Fully Composite Sections 1023 Shear Connectors 1029 Hybrid Composite Girders 1040 Composite Flexural Members Containing Formed Steel Deck 1041 Design Procedure—Load and Resistance Factor Design 1041 Design Procedure—Allowable Stress Design 1042 LRFD Examples—Simply Supported Beams 1043 ASD Example—Simply Supported Beam 1051 Deflections = 1054 Continuous Beams 1056 Composite Columns —-1059 Selected References 1061 Problems 1063 1067 Approximate Radius of Gyration 1068 Torsional Properties 1069 1070Preface ‘The publication of this third edition reflects the continuing changes occurring in design requirements for structural steel, particularly the formal adoption by the American Institute of Steel Construction (AISC) of a specification for Load and Resistance Factor Design (LRFD). Design of structural steel members has developed over the past 90 years from a simple approach involving a few basic properties of steel and elemen- tary mathematics to a sophisticated treatment demanding a thorough knowl- edge of structural and material behavior. Present design practice utilizes knowledge of mechanics of materials, structural analysis, and particularly, structural stability, in combination with nationally recognized design rules for safety. The most widely used design rules are those of the American Institute of Steel Construction (AISC), given in Load and Resistance Factor Design Specification for Structural Steel Buildings and Specification for Structural Steel Buildings—Allowable Stress Design and Plastic Design, referred to hereafter as LRED Specification and ASD Specification, respectively. ‘The specific occurrences dictating this third edition are the publication of the 1986 LRFD Specification (effective September 1, 1986) with Commentary; the corresponding LRFD handbook, Manual of Steel Consiruction—Load and Resistance Factor Design, 1st edition, 1986; the 1989 ASD Specification (effective June 1, 1989) with Commentary; and the corresponding ASD handbook, Manual of Steel Construction—Allowable Stress Design , 9th edition, 1989. Steel members and components are selected from these handbooks, referred to hereafter as LRFD Manual and ASD Manual. ‘The third edition follows the same philosophical approach that has gained wide acceptance of users since the first edition was published in 1971. This edition continues to strive to present in a. logical manner the theoretical xilixiv PREFACE background needed for developing and explaining design requirements, partic- ularly those of the 1986 LRFD Specification. Beginning with coverage of background material including references to pertinent research, the develop- ment of specific formulas used in the AISC Specifications is followed by a generous number of design examples explaining in detail the process of selecting minimum weight members to satisfy given conditions. Emphasis throughout this third edition is on the 1986 LRFD Specifica- tion. That specification is based on statistical studies of loads and the resistances of steel structures subject to various types of load effects, such as bending moment, shear, axial force, and torsional moment. The rational treatment of both loads and resistances provides results in steel structures having more uniform safety throughout. This new philosophy of design, discussed only briefly in one section of the second edition, is expected in time to become the predominant approach to design. Considerable emphasis has been placed on presenting for the beginner, as well as the advanced student, the necessary elastic and inelastic stability concepts, the understanding of which is essential to properly apply steel design rules. The same concepts are applicable whether design is according to the LRED Specification or the ASD Specification. The explanation of stability concepts is incorporated into the chapters in such a way that the reader may either study in detail the stability concepts in logical sequence, or omit or postpone study of sections containing detailed development, merely accepting qualitative explanation and proceeding directly to design. As this third edition is prepared, most design is still done according to the traditional Allowable Stress Design (ie., Working Stress Design). This method focuses on service (working) loads and elastically computed stresses, compar- ing these stresses with allowable limiting values. However, the logical trend over the next several years should be toward the more rational Load and Resistance Factor Design Specification. Strength design philosophy (reflected in the 1986 LRFD Specification) uses factored service loads and compares the strength provided with such factored loads. The strength in any given case depends on the “limit state,” or mode of failure, such as yielding, fracture, or buckling. The traditional “plastic design,” included as Chapter N of the 1989 ASD Specification, is an option integrally included as part of the LRFD Specification. Throughout the text, the theory and background material, being common to both the LRFD and ASD philosophies of design, have been integrated. The specific design provisions and illustrative examples are, however, treated generally in separate sections within the chapters so that the user may study either the Allowable Stress Design or the Load and Resistance Factor Design portions separately. The third edition continues the use of SI units as an addition to the primary use of Inch-Pound units. Although neither the LRFD nor the ASD Specification contains SI units, some use of SI units is made in the text. LRFD and ASD formulas have their SI equivalent (conversions are the practical ones made by the authors) given as a footnote on the text page containing thePREFACE XW AISC-specified Inch-Pound units, Tables and diagrams generally contain both Inch-Pound and SI units. Depending on the proficiency required of the student, this textbook may provide material for two courses of three or four semester-credit hours each, It is suggested that the beginning course in steel structures for undergraduate students might contain the material of Chapters 1 through 7, 9, 10, 12 and 16, except Sections 6.4, 6.6, 6.12 to 6.19, 7.9 to 7.11, 9.3 to 9.5, 9.8, 9.12 to 9.14, and 12.6 to 12.7. The second course would review some of the same topics of the first course, but more rapidly, emphasizing items omitted in the first course. In addition, the remaining chapters—namely Chapter 8 on torsion, Chapter 11 on plate girders, Chapter 13 on connections, Chapter 14 on frames, and Chapter 15 on frame design—are suggested for inclusion. In both courses, emphasis should be on Load and Resistance Factor Design. The reader will need ready access to the LRFD Manual* throughout the study of the text, particularly when working with the examples. However, it is not the objective of this text that the reader become proficient in the routine use of tables; the tables serve only as a guide to obtaining experience with variation of design parameters and as an aid in arriving at good design. The LRFD Specification and Commentary are contained in the LRFD Manual and are therefore not included in this book, except for various individual provisions quoted where they are explained. ‘The direct use of the computer is not specifically employed anywhere in the text. The authors believe the study of basic principles in the classroom is of the highest priority. However, the reader may find that acquiring the data base of standard section properties, available for purchase from AISC, will be helpful. The authors recommend the use of a spreadsheet software, such as Lotus 1-2-3," along with the use of the database properties. New features of this third edition are: (1) detailed presentation of strength-related background and design rules for Load and Resistance Factor Design; (2) an integrated treatment of both the 1986 Load and Resistance Factor Design Specification along with the 1989 Allowable Stress Design Specification. ‘Other special features of this text are (3) comprehensive treatment of design of J-shaped members subject to torsion (Chapter 8), including a simplified practical method; (4) detailed treatment of plate girder theory as it relates to Load and Resistance Factor Design (Chapter 11) and a comprehen- sive design example of a two-span continuous girder using two different grades “Manual of Steel Construction—Load and Resistance Factor Design, 1st edition, 1986. Since nearly continuous reference is made to the LRFD Manual (which also contains the 19% LRFD Specification and Commentary), the reader will find it desirable to secure a copy of it from ‘American Institute of Steel Construction, Inc., One East Wacker Drive, Suite 3100, Chicago, IL 6060-200. The LRED Specification and Commentary is also available as a separate paper cover document. The ASD Manual and the paper cover ASD Specification and Commentary are also available from AISC. + Lotus ‘1-2-3 is a trademark of Lotus Development Co,xvi PREFACE of steel; (5) extensive treatment of connections (Chapter 13), including signifi- cant discussion and illustration of the design of components. All chapters have been almost completely rewritten to reflect LRFD approaches. The authors are indebted to students, colleagues, and other users of the first two editions who have suggested improvements of wording, identified errors, and recommended items for inclusion or deletion. The suggestions have been carefully considered and included in this complete revision where possi- ble. Particularly, the authors are appreciative of the suggestions of Professor Abdul H. Zureick of Georgia Institute of Technology regarding several as- pects, and Professor Geoffrey L. Kulak of the University of Alberta at Edmonton regarding bolted connections. The help of Professor Donald R. Sherman of the University of Wisconsin-Milwaukee in providing information regarding specification-related developments is appreciated. The cooperation and help of AISC through Nestor Iwankiw and Robert F. Lorenz are also acknowledged. The authors also thank University of Wisconsin Teaching Assistants Minhah Chun and Jae-Hoon Lee for their help in locating manuscript errors. Users of the third edition are urged to communicate with the authors regarding all aspects of this book, particularly on identification of errors and suggestions for improvement. ‘The senior author continues to give special credit to his wife Bette Salmon for her continued patience and encouragement, without which this extensive revision would not have been possible. Charles G. Salmon John E. JohnsonConversion Factors Some Conversion Factors, between Inch-Pound Units and SI Units, Useful in Structural Stee! Design To Convert To Multiply by Forces kip force kN 4.448 Ib N 4448 KN kip 0.2248 Stresses: ksi MPa (ie, N/mm?) 6.895 psi MPa 0.006895 MPa ksi 01450 MPa psi 145.0 ‘Moments ft-kip kN-m 1.356 kN-m ft-kip 0.7376 Uniform Loading kip/ft kN/m 14:59 kN/m kip ft 0.06852 kip /ft? KN/o? 47.88 pst N/a? 41.88 KN/o? kip/ft? 0.02089 For proper use of SI, sce Standard for Metric Practice (ASTM 380-89), American Society for Testing and Materials, Philadelphia, 1989, Also see Standard Practice for the Use of Metric (SI) Units in Buildings Design and Construction (Committee E-6 Supple- ment to E380) (ANSI/ASTM E621-78), American Society for Testing and Materials, Formula kg-m/s* N/a? Nem Philadelphia, 1978. Basis of Conversions (ASTM E380): 1 in. = 25.4 mm; 1 Ib force = 4.448 221 615 260 5 newtons. Basic SI units relating to structural steel design: Quantity Unit Symbol Tength metre m mass kilogram kg time second 8 Derived SI units relating to structural design: Quantity Unit Symbol force newton N pressure, stress pascal Pa energy, or work joule I xvilChapter 1 Introduction 1.1 STRUCTURAL DESIGN Structural design may be defined as a mixture of art and science, combining the experienced engineer's intuitive feeling for the behavior of a structure with a sound knowledge of the principles of statics, dynamics, mechanics of materials, and structural analysis, to produce a safe economical structure which will serve its intended purpose. Until about 1850, structural design was largely an art relying on intuition to determine the size and arrangement of the structural elements. Early man-made structures essentially conformed to those which could also be observed in nature, such as beams and arches. As the principles governing the behavior of structures and structural materials have become better understood, design procedures have become more scientific. Computations involving scientific principles should serve as a guide to decision making and not be followed blindly. The art or intuitive ability of the experienced engineer is utilized to make the decisions, guided by the computa- tional results. 1.2 PRINCIPLES OF DESIGN Design is a process by which an optimum solution is obtained. In this text the concern is with the design of structures—in particular, sree! structures. In any design, certain criteria must be established to evaluate whether or not an optimum has been achieved. For a structure, typical criteria may be (a) minimum cost; (b) minimum weight; (c) minimum construction time; (d) 12 1 /INFROOUCTION 1350 ft World Trade Center, New York (Photo by C. G. Salmon) minimum labor; (c) minimum cost of manufacture of owner’s products; (f) maximum efficiency of operation to owner. Usually several criteria are in- volved, each of which may require weighting. Observing the above possible criteria, it may be apparent that setting clearly measurable criteria (such as weight and cost) for establishing an optimum frequently will be difficult, and perhaps impossible. In most_practical situations the evaluation must_be qualitative. : “Tf a specific objective criterion can be expressed mathematically, then optimization techniques may be employed to obtain a maximum or minimum for the objective function. Optimization procedures and techniques comprise an entire subject that is outside the scope of this text. The criterion of minimum weight is emphasized throughout, under the general assumption that minimum material represents minimum cost. Other subjective criteria must be kept in mind, even though the integration of behavioral principles with design of structural steel elements in this text utilizes only simple objective criteria, such as weight or cost.1.3 /HISTORICAL BACKGROUND OF STEEL STRUCTURES = 3 Design Procedure ‘The design procedure may be considered to be composed of two parts—func- tional design and structural framework design. Functional design ensures that intended results are achievéd, such as (a) adequate working areas and clear- ances; (b) proper ventilation and/or air conditioning; (c) adequate transporta- tion facilities, such as elevators, stairways, and cranes or materials handling equipment; (d) adequate lighting; and (e) aesthetics. ‘The structural framework design is the selection of the arrangement and sizes of structural elements so that service loads may be safely carried. The iterative design procedure may be: outlined as follows: 1. Planning. Establishment of the functions for which the structure must serve. Set criteria against which to measure the resulting design for being an optimum. Preliminary structural configuration. Arrangement of the elements to serve the functions in step 1. . 3. Establishment of the loads to be carried, 4. Preliminary member selection. Based on the decisions of steps 1, 2, and 3 selection of the member sizes to satisfy an objective criterion, such as least weight or cost. 5. Analysis. Structural analysis involving modeling the loads and the structural framework to obtain internal forces and any desired deflec- tions. 6. Evaluation. Are all strength and serviceability requirements satisfied and is the result optimum? Compare the result with predetermined criteria. 7. Redesign. Repetition of any part of the sequence 1 through 6 found necessary or desirable as a result of evaluation. Steps 1 through 6 represent an iterative process. Usually in this text only steps 3 through 6 will be subject to this iteration since the structural configuration and external loading will be prescribed. 8. Final decision. The determination of whether or not an optimum design has been achieved. n 1.3 HISTORICAL BACKGROUND OF STEEL STRUCTURES Metal as a structural material began with cast iron, used on a 100-ft (30-m) arch span which was built in England in 1777-1779 [1.1].* A number of cast-iron bridges were built during the period 1780-1820, mostly arch-shaped with main girders consisting of individual cast-iron pieces forming bars or ‘Numbers in brackets refer to the Selected References at the end of the chapter.41, /INTRODUCTION trusses. Cast iron was also used for chain links on suspension bridges until about 1840. Wrought iron began replacing cast iron soon after 1840, the earliest important example being the Brittania Bridge over Menai Straits in Wales which was built in 1846-1850. This was a tubular girder bridge having spans 230-460-460-230 ft (70-140-140-70 m), which was made from wrought-iron plates and angles. The process of rolling various shapes was developing as cast iron and wrought iron received wider usage. Bars were rolled on an industrial scale beginning about 1780. The rolling of rails began about 1820 and was extended to I-shapes by the 1870s. The developments of the Bessemer process (1855), the introduction of a basic liner in the Bessemer converter (1870), and the open-hearth furnace brought widespread use of iron ore products in building materials. Since 1890, steel has replaced wrought iron as the principal metallic building material Currently (1989), stecls having yield stresses varying from 24,000 to 100,000 pounds per square inch, psi (165 to 690 megapascals,’ MPa), and available for structural uses. The various stecls, their uses and their properties are discussed in Chapter 2. 1.4 LOADS The accurate determination of the loads to which a structure or structural element will be subjected is not always predictable. Even if the loads are well known at one location in a structure, the distribution of load from element to element throughout the structure usually requires assumptions and approxima- tions. Some of the most common kinds of loads are discussed in the following sections. Dead Load Dead load is a fixed position gravity service load, so called because it acts continuously toward the earth when the structure is in service. The weight of the structure is considered dead load, as well as attachments to the structure such as pipes, electrical conduit, air-conditioning and heating ducts, lighting fixtures, floor covering, roof covering, and suspended ceilings: that is, all items that remain throughout the life of the structure, Dead loads are usually known accurately but not until the design has been completed. Under steps 3 through 6 of the design procedure discussed in Sec. 1.2, the weight of the structure or structural element must be estimated, preliminary section selected, weight recomputed, and member selection revised ‘MPa, megapaseals. are equivalent to Newtons per square millimeter, N/mm, in SI units.14/LOADS 5 if necessary. The dead load of attachments is usually known with reasonable accuracy prior to the design. Live Load Gravity loads acting when the structure is in service, but varying in magnitude and location, are termed Jive loads. Examples of live loads are human occu- pants, furniture, movable equipment, vehicles, and stored goods. Some live loads may be practically permanent, others may be highly transient. Because of the unknown nature of the magnitude, location, and density of live load items, realistic magnitudes and the positions of such loads are very difficult to determine. Because of the public concern for adequate safety, live loads to be taken as service loads in design are usually prescribed by state and local building codes. These loads are generally empirical and conservative, based on experi- ence and accepted practice rather than accurately computed values. Wherever TABLE 1.4.1 TYPICAL MINIMUM UNIFORMLY DISTRIBUTED LIVE LOADS (Adapted from Ret. 1.2) Live load Occupancy or Use pst Pa® 1. Hotel guest rooms: 40 1900 School classrooms Private apartments Hospital private rooms 2. Offices 50 2400 3. Asscmbly halls, fixed seat 60 2900 Library reading rooms 4, Corridors, above first floor in schools, libraries, and hospitals 803800 5. Assembly areas; theater lobbies 100 4800 Dining rooms and restaurants Office building lobbies Main floor, retail stores ‘Assembly hall, movable seats 6. Wholesale stores, all floors 125 6000. Manufacturing, light Storage warehouses, light 7. Armories and drill halls 150 7200 Stage floors Library stack rooms ; & Manufacturing, heavy 250 12000 Sidewalks and driveways subject to trucking. Storage warehouses, heavy *SI values are approximate conversions. 1 psf (Ib/sq ft) = 47.9 Pa.6 1 /INTRODUCTION local codes do not apply, or do not exist, the provisions from one of several regional and national building codes may be used. One such widely recognized code is the American National Standard Minimum Design Loads for Buildings and Other Structures ANSI A58.1 of the American National Standards Insti- tute (ANSI) [1.2], from which some typical live loads are presented in Table 1:4.1. The code will henceforth be referred to as the ANSI Standard. This Standard is updated from time to time, most recently in 1982. Live load when applied to a structure should be positioned to give the maximum effect, including partial loading, alternate span loading, or full span loading as may be necessary. The simplified assumption of full uniform loading everywhere should be used only when it agrees with reality or is an appropriate approximation. The probability of having the prescribed loading applied uniformly over an entire floor, or over all floors of a building simultaneously, is almost nonexistent. Most codes recognize this by allowing for some percentage reduction from full loading. For instance, for live loads of 100 psf or(thore ANSI Standard [1.2] allows members having an influence area of 400 sq frormore to be designed for a reduced live load according to Eq, 1.4.1, as follows: L= toa ee (14.1) where L = reduced live load per sq ft of area supported by the member Lo = unreduced live load per sq ft of area supported by the member (from Table 1.4.1) influence area, sq ft A,= The influence area A, is four times the tributary area for a column, two times the tributary area for a beam, and is equal to the panel area for a two-way slab. The reduced live load L shall not be less than 50% of the live load Lo for members supporting one floor, nor less than 40% of the live load Ly otherwise. The live load reduction referred to above is not permitted in areas to be occupied as“places of public assembly and for one-way slabs, when the live load L is 100 psf or less.)Reductions are permitted for occupancies where Ly is greater than 100 psf and for garages and roofs only under special circum- stances (ANSI-4,7.2) [1.2]. Highway Live Loads Highway vehicle loading in the United States has been standardized by the American Association of State Highway and Transportation Officials (AASHTO) [1.3] into standard truck loads and lane loads that approximate a series of trucks. There are two systems, designated H and HS, that are identified by the number of axles per truck. The H system has two axles, whereas the HS’system has three axles per truck. There are several classes of14/LOADS 7 bal feo ftw reao-| 6 orf] bz a" 8 ZR 2 ie loads hips) H20-Truck HS20-Truck {i for wore 1a be positioned: 26k for shoar for maximum effect yw = 0.64 kipsilinaar ft of kant HZ0 and H520 Lane Figure 1.41 AASHTO Highway 120 and S20 loadings (1.3). (1 kip = 4.45 KN). loading; however, the usual ones are known as H20 and HS20, shown in Fig. 14.1, In designing a given bridge, either one truck loading is applied to the entire structure, or the lane loading is applied. When the lane loading is used, the uniform portion is distributed over as much of the span or spans as will cause the maximum effect. In addition, the one concentrated load (for maximum negative moment on continuous spans a second concentrated load is also used) is positioned for the most severe loading effect. The load distribution across the width of a bridge to its various supporting members is taken in accordance with semiempirical rules that depend on the type of bridge deck and supporting structure. The single truck loading provides the effect of a heavy concentrated load and usually governs on relatively short spans. The uniform lane load is to simulate a line of traffic, and the added concentrated load is to account for the possibility of one extra heavy vehicle in the line of traffic. These loads have been used with no apparent difficulty since 1944, before which time a line of trucks was actually used for the loading. On the interstate system of highways a military loading is also used that consists of two 24-kip (107-KN) axle loads spaced 4 ft (1.2 m) apart. Railroad bridges are designed to carry a similar semiempirical loading known as the Cooper E72 train, consisting of a series of concentrated loads a fixed distance apart followed by uniform loading, This loading is prescribed by the American Railway Engineering Association (AREA) [1.4]. Impact The term impact as ordinarily used in structural design refers to the dynamic effect of a suddenly applied load. In the building of a structure the materials are added slowly; people entering a building are also considered a gradual8 — 1 /INTRODUCTION D Spring force ‘j oe xeo Spring force = ke (wy Lose suddenty ee " released from hare d oo # _ TD ones wo w = ‘Oscillation above aol and below here (@) Novvibration: (0) Free vibration rma spring Yoree = ¥F tne spring force = 20 Figure 1.4.2 Comparison of static and dynamic loading. loading. Dead loads are static loads; i.e., they have no effect other than weight. Live loads may be either static or they may have a dynamic effect. Persons and furniture would be treated as static live load, but cranes and various types of machinery also have dynamic effects. Consider the spring-mass system of Fig. 1.4.2a where the spring may be thought of as analogous to an clastic beam. When load is gradually applied (ie, static loading) the mass (weight) deflects an amount x and the load on the spring (beam) is equal to the weight I. In Fig. 1.4.2b the load is suddenly applied (dynamic loading) and the maximum deflection is 2x; ie., the maxi- mum load on the spring (beam) is 21”. In this case the mass vibrates in simple harmonic motion with its neutral position equal to its static deflected position. In real structures the harmonic (vibratory) motion is damped out (reduced to zero) very rapidly. Once the motion has stopped the force remaining in the spring is the weight W. To account for the increased force during the time the member is in motion a load equal to twice the static load W’ should be used—add 100% of the static load to represent the dynamic effect. This is called a 100% impact factor. ‘Any live load that can have a dynamic effect should be increased by an impact factor, While a dynamic analysis of a structure could be made, such a procedure is unnecessary in ordinary design, Thus empirical formulas and impact factors are usually used, In cases where the dynamic effect is small (say where impact would be less than about 20%) it is ordinarily accounted for by using a conservative (higher) value for the specified live load. The dynamic effects of persons in buildings and of slow-moving vehicles in parking garages are examples where ordinary design live load is conservative and no explicit impact factor is usually added. For highway bridge design, however, impact is always to be considered. AASHTO [1.3] prescribes empirically that the impact factor expressed as a portion of live load is ——_ <= 0.30 z T+ * (42) In Eq. 1.4.2, L (expressed in feet) is the length of the portion of the span that14/LOADS 9 is loaded to give the maximum effect on the member. Since vehicles travel directly on the superstructure, all parts of it are subjected to vibration and must be designed to include impact. The substructure, including all portions not rigidly a attached to the superstructure suc! such as abutments, retaining walls, and piers, are _assumed to have adequate dam; from | the application poi of the dynamic | ered, Again, conservative: static loads may account for the effects, In buildings, it is principally in the design of supports for cranes and heavy machinery that impact is explicitly considered. The American Institute of Steel Construction (AISC) Allowable Stress Design (ASD) and Load and Resistance Factor Design (LRFD) Specifications* (1.5, 1.15] (ASD and LRFD-A4.2) state that if not otherwise specified, the impact percentage shall be: For supports of elevators and elevator machinery 100% For cab-operated traveling crane support girders and their connections 25% For pendant-operated traveling crane support girders and their connections 10% For supports of light machinery, shaft or motor driven, not less than 20% For supports of reciprocating machinery or power driven units, not less than 50% For hangers supporting floors and balconies 33% In the design of crane runway beams (see Fig. 1.4.3) and their connec- tions, the horizontal forces caused by moving crane trolleys must be consid- ered. Both LRFD and ASD-A4.3 prescribe using a minimum of “20% of the sum of the lifted load and the crane trolley (but exclusive of other parts of the crane), The force shall be assumed to be applied at the top of the rails, acting in either direction normal to the runway rails, and shall be distributed with due regard for lateral stiffness of the structure supporting the rails.” In addition, due to acceleration and deceleration of the entire crane, a longitudinal tractive force is transmitted to the runway. girder through friction of the end truck wheels with the crane rail. LRFD and ASD-A4.3 require this force, if not otherwise specified, to be taken not less than 10% of the maximum wheel loads of the crane applied at the top of the rail. The reader will find continued reference to the AISC Specifications (ASD and LRFD) which are contained, respectively, in the AFSC ASD Manual [1.7] and AISC LRFD Manual (1.17). These two books may be purchased from AISC, 400 North Michigan Avenue, Chicago, IL 60611-4185.10 1 /INTRODUCTION Trolley movaront to Trolley Section AA End truck Plan View Runway girder Section B-B Figure 1.43 Crane arrangement, showing movements that contribute impact loading, Snow Load ‘The live loading for which roofs are designed is either totally or primarily a snow load. Since snow has a variable specific gravity, even if one knows the depth of snow for which design is to be made, the load per unit area of roof is at best only a guess. The best procedure for establishing snow load for design is to follow the ANSI Standard [1.2]. This Code uses a map of the United States giving isolines of ground snow corresponding to a 50-year mean recurrence interval for use in designing most permanent structures, The ground snow is then multiplied by a cocfficient that includes the effect of roof slope, wind expo- sure, nonuniform accumulation on pitched or curved roofs, multiple series roofs, and multilevel roofs and roof areas adjacent to projections on a roof level. It is apparent that the steeper the roof the less snow can accumulate. Also partial snow loading should be considered, in addition to full loading, if it is believed such loading can occur and would cause maximum effects. Wind may also act on a structure that is carrying snow load. It is unlikely, however, that ‘maximum snow and wind loads would act simultaneously. Tn general, the basic snow load used in design varies from 30 to 40 psf (1400 to 1900 MPa) in the northern and eastern states to 20 psf (960 MPa) or less in the southern states. Flat roofs in normally warm climates should be designed for 20 psf (960 MPa) even when such accumulation of snow may seem doubtful. This loading may be thought of as due to people gathered on such a roof, Furthermore, though wind is frequently ignored as a vertical force on a roof, nevertheless it may cause such an effect. For these reasons a 20 psf (960 MPa) minimum loading, even though it may not always be snow, is reasonable. Local codes, actual weather conditions, ANSI [1.2], or the Cana- dian Structural Design Manual [1.9], should be used when designing for snow.1asvoaps if Other snow load information has been provided by’ Lew, Simiu, and Elling- wood in the Building Structural Design Handbook (Chapter 2) (1.10), and in the works of O'Rourke and Stiefel [1.38], Templin and Schriever [1.39), O'Rourke, Tobiasson, and Wood [1.40], O'Rourke, Redfield, and von Bradsky [1.41], and O'Rourke, Speck, and Stiefel [1.42]. ‘Wind Load All structures are subject to wind load, but it is usually only those more. than three or four stories high, other than long bridges, for which special considera tion of wind is required. ~ On any typical building of rectangular plan and elevation, wind exerts pressure on the windward side and suction on the leeward side, as well as either uplift or downward pressure on the roof, For most ordinary situations vertical roof loading from wind is neglected on the assumption that snow loading will require a greater strength than wind loading. This assumption is not true for southern climates where the vertical loading due to wind must be included. Furthermore, the total lateral wind load, windward and leeward effect, is commonly assumed to be applied to the windward face of the building, In accordance with Bernoulli's theorem for an ideal fluid striking an object, the increase in static pressure equals the decrease in dynamic pressure, or q = 3p¥? (1.4.3) where q is the dynamic pressure on the object, p is the mass density of air (specific weight w = 0.07651 pef at sea level and 15°C), and V is the wind velocity. In terms of velocity V in miles per hour, the dynamic pressure q (psf) would be 1 / 0.07651 \ / 52807 32.2 i 3600 2 ) = 0.0026? (14.4)* In design of usual types of buildings the dynamic pressure q is commonly converted into equivalent static. pressure p, which may be expressed [1.9] p= aC, (1.45) where C, is an exposure factor that varies from 1.0 (for 0-40-ft height) to 2.0 (for 740-1200-ft height); C, is a gust factor, such as 20 for structural members and 2.5 for small elements including cladding; and C, is a shape factor for the building as a whole. Excellent details of application of wind loading to structures are available in the ANSI Standard [1.2] and in the National Building Code of Canada [1.9]. *in SI units, 4= 063%, for g in MPa and V in m/sec 4a)12 1,/INTRODUCTION The commonly used wind pressure of 20 psf, as specified by many building ‘codes, corresponds to a velocity of 88 miles per hour (mph) from Eq. 1.4.4, An exposure factor C, of 1.0, a gust factor C, of 2.0, and a shape factor C, of 1.3 for an airtight building, along with a 20 psf equivalent static pressure p, will give from Eq. 1.4.5 a dynamic pressure g of 7.7 psf, which corresponds, using Eg. 1.4.4, to a wind velocity of 55 mph. For all buildings having nonplanar surfaces, plane surfaces inclined to the wind direction, or surfaces having significant openings, special determination of the wind forces should be made using such sources as the ANSI Standard [1.2], or the National Building Code of Canada (1.9]. For more extensive treatment of wind loads, the reader is referred to the Task Committee on Wind Forces [1.35], Lew, Simiu, and Ellingwood in the Building Structural Design Handbook (1.10), Mehta [1.36], and Stathopoulos, Surry, and Davenport [1.37]. Earthquake Load . An earthquake consists of horizontal and vertical ground motions, with the vertical motion usually having much the smaller magnitude, Since the horizon- tal motion of the ground causes the most significant effect it is that effect which is usually thought of as carthquake load, When the ground under an object (structure) having a certain mass suddenly moves, the inertia of the mass tends to resist the movement, as shown in Fig. 1.4.4. A shear force is developed between the ground and the mass. Most building codes having earthquake provisions require the designer to consider a lateral force CW that is usually empirically prescribed. The dynamics of earthquake action on structures is outside the scope of this text, and the reader is referred to Chopra [1.45] and Clough and Penzien [1.46]. In order to simplify the design process, most building codes contain an equivalent lateral force procedure for designing to resist earthquakes, One of the most widely used design recommendations is that of the Structural Engineers Association of California (SEAOC), the latest version of which is 1974 [1.43]. Since that time, the Applied Technology Council (ATC) prepared a set of design provisions [1.47]. Some recent rules for the equivalent lateral force procedure are those given by the ANSI Standard [1.2]. In ANSI the a w CW = inertia reaetion ar Earth motion (a) At rest (6) Under horizontal motion from earthquake Figure 1.4.4 Force developed by earthquake.1.5/ TYPES OF STRUCTURAL STEEL MEMBERS = 13 lateral seismic forces V, expressed as follows, are assumed to act nonconcur- rently in the direction of each of the main axes of the structure, V = ZIKCSW (4.6) where Z = seismic zone coefficient, varying from 4 for the zone of lowest seismicity, to 1 for the zone of highest seismicity I= occupancy importance factor, varying from 1.5 for buildings designated as “essential facilities,” 1.25 for buildings where the primary occupancy is for assembly for greater than 300 persons, to 1.0 for usual buildings K=horizontal force factor, varying from 0.67 to 2.5, indicating capacity of the structure to absorb plastic deformation (low values indicate high ductility) c= <012 (47) ar ~ the seismic coefficient, equivalent to the maximum acceleration in terms of a¢celeration due to gravity T = fundamental natural period, i.c., time for one cycle of vibration, of the building in the direction of motion S = soil profile coefficient, varying from 1.0 for rock to 1.5 for soft- to medium-stiff clays and sands W = total dead load of the building, including interior partitions When the natural period T cannot be determined by a rational means from technical data, it may be obtained as follows for shear walls or exterior concrete frames utilizing deep beams or wide piers, or both: (1.4.8) where D is the dimension of the structure in the direction of the applied forces, in feet, and h, is the height of the building. Once the base shear V has been determined, the lateral force must be distributed over the height of the building. More details of the ANSI Standard [1.2] procedure are available in the Building Structural Design Handbook [1.10]. Various building code formulas for earthquake-resistant design are compared by Chopra and Cruz [1.44] Many states have adopted the Uniform Building Code (UBC) [1.48}, the most recent version of which is 1985, which contains provisions for design to resist earthquakes generally based on the ANSI Standard [1.2]. 1.5 TYPES OF STRUCTURAL STEEL MEMBERS As discussed in Sec. 1.2 the function of a structure is the principal factor determining the structural configuration. Using the structural configuration along with the design loads, individual components are: selected to properly44 1,/INTRODUCTION tite fF WT of ST Wide- American American Angle Structural flange standard standard we shape beam channel @) (by o @) «) © a Pipe Structure! Bars Plates section tubing oO ® (hy wo Figure 1.5.1 Standard rolled shapes. support and transmit loads throughout the structure. Steel members are selected from among the standard rolled shapes adopted by the American Institute of Steel Construction (AISC) (also given by American Society for ‘Testing and Materials (ASTM) AG Specification). Of course, welding permits combining plates and/or other rolled shapes to obtain any shape the designer may require. Typical rolled shapes, the dimensions for which are found in the AISC Manual* (1.7, 1.17], are shown in Fig. 1.5.1. The most commonly used section is the wide-flange shape (Fig. 1.5.1a) which is formed by hot rolling in the steel mill. The wide-flange shape is designated by the nominal depth and the weight per foot, such as a W18X97 which is nominally 18 in. deep (actual depth = 18.59 in. according to AISC Manual) and weighs 97 pounds per foot. (In SI units the W18X97 section could be designated W460%142, meaning nominally 460 mm deep and having a mass of 142 kg/m.) Two sets of dimensions are found in the AISC Manual, one set stated in decimals for the designer to use in computations, and another set expressed in fractions (4 in. as the smallest increment) for the detailer to use on plans and shop drawings. Rolled W shapes are also designated by ANSI/ASTM A6 [18] in accordance with web thickness as Groups I through V, with the thinnest web sections in Group I. The American Standard beam (Fig. 1.5.1b), commonly called the I-beam, has relatively narrow and sloping flanges and a thick web compared to the wide-flange shape. Use of most I-beams has become relatively uncommon because of exeessive material in the web and relative lack of lateral stiffness due to the narrow flanges. “When reference is made to AISC Manual, the material is available in both Refs. 1.7 and 1.17.1.5/ TYPES OF STRUCTURAL STEEL MEMBERS = 15, LC LL IL (@) Channels 0) Zoes (@) [shaped double ‘channels (@) Angie (e) Hat sections Figure 1.5.2 Some cold-formed shapes. ‘The channel (Fig. 1.5.1¢) and angle (Fig. 1.5.1d) are commonly used either alone or in combination with other sections. The channel is designated, for example, as C12X20.7, a nominal 12-in. deep channel having a weight of 20.7 pounds per foot. Angles are designated by their leg length (long leg first) and thickness, such as, L6x4x}. The structural tee (Fig. 1.5.le) is made by cutting wide-flange or I-beams in half and is commonly used for chord members in trusses. The tee is designated, for example, as WT544, where the 5 is the nominal depth and 44 is the weight in pounds per foot; this tee being cut from a W10x88. Pipe sections (Fig. 1.5.1f) are designated “standard,” “extra strong,” and “double-extra strong” in accordance with the thickness and are also nominally prescribed by diameter; thus 10-in.-diam double-extra strong is an example of a particular pipe size. Structural tubing (Fig. 1.5.1g) is used where pleasing architectural appear- ance is desired with exposed steel. Tubing is designated by outside dimensions and thickness, such as structural tubing, 8X64. ‘The sections shown in Fig. 1.5.1 are all hot-rolled; that is, they are formed from hot billet steel (blocks of steel) by passing through rolls numerous times to obtain the final shapes. Many other shapes are cold-formed from plate material having a thickness not exceeding 1 in., as shown in Fig. 1.5.2, Regarding size and designation of cold-formed steel members, there are no truly standard shapes even though the properties of many common shapes are given in the Cold-Formed Steel Design Manual (1.12). Various manufactur- ers produce many proprietary shapes. Tension Members The tension member occurs commonly as a chord member in a truss, as diagonal bracing in many types of structures, as direct support for balconies, as cables in suspended roof systems, and as suspension bridge main cables and16 1 /INTRODUCTION oo 3 J J (a), Flothd and rectangular (b) Cables composed (c) Single and bars, including eve ‘of meny small double bars and upset bers wires angles | Perforated (4) | Rolled Wand (©) Structural (1) Built-up box ‘S-sections tee sections Figure 1.5.3 Typical tension members. suspenders that support the roadway. Typical cross-sections of tension mem- bers are shown in Fig. 1.5.3, and their design (except for special factors relating to suspension-type cable supported structures) is treated in Chapter 3. ‘Compression Members Since compression member strength is a function of the cross-sectional shape (radius of gyration), the area is generally spread out as much as is practical. Chord members in trusses, and many interior columns in buildings are te J (a) Rolled W- and (b) Double (e) Structural (@)_ Structural S-shapes angie toe O LL] (©) Pipe section (®) Built-up sections Figure 1.5.4 ‘Typical compression members.1.5/ TYPES OF STRUCTURAL STEEL MEMBERS 17 examples of members subject to axial compression. Even under the most ideal condition, pure axial compression is not attainable; so, design for “axial” loading. assumes the effect of any small simultaneous bending may be ne- glected. Typical cross sections of compression members are shown in Fig. 1.5.4 and their behavior and design are treated in Chapter 6. Beams Beams are members subjected to transverse loading and are most efficient when their area is distributed so as to be located at the greatest practical distance from the neutral axis. The most common beam sections are the wide-flange (W) and I-beams (S) (Fig. 1.5.5a), as well as smaller rolled L-shaped sections designated as “miscellaneous shapes” (M). For deeper and thinner-webbed sections than can economically be rolled, welded I-shaped sections (Fig. 1.5.5b) are used, including stiffened plate girders, For moderate spans carrying light loads, open-web “joists” are often used (Fig. 1.5.5c). These are parallel chord truss-type members used for the support of floors and roofs. The steel may be hot-rolled or cold-formed. Such joists are designated “K-Series,” “LH-Series,” and “DLH-Series.” The K-Series is suitable for members having the direct support of floors and roof decks in buildings. The LH-Series and DLH-Series are known as Longspan and Deep Longspan, respectively. Longspan Steel Joists are shop-fabricated trusses used “for the direct support of floor or roof slabs or decks between walls, beams and main structural members” [1.13]. Deep Longspan Joists are used “for the direct support of roof slabs or decks between wall, beams and main structural members” [1.13]. The design of the chords for K-Series trusses is based on a yield strength* of 50 ksi (343 MPa), while the web sections may use either 36 (248 MPa) or 50 ksi (345 MPa), For the LH- and DLH-Scries the chord and web sections design must be based on a yield strength of at least 36 ksi (248 MPa) but not greater than 50 ksi (345 MPa). The K-Series joists have depths from 8 to 30 in, for clear spans to 60 ft. The Longspan joists (LH-Series) have depths from 18 to 48 in. for clear spans to 96 ft. The Deep Longspan joists (DLH-Series) have depths from 52 to 72 in. for clear spans to 144 ft. Alll of these joists are designed according to Specifications adopted by the Steel Joist Institute (SJI) [1.13], which generally are in agreement with AISC Specification [1.5] for hot-rolled steels and AISI Specification [1.11] for cold-formed steels, “Refer to Sec, 2.4 for definition.18 = 1,/INTRODUCTION TTX | (a) Rolled W- (b) Welded {c) Open web joists ‘nd other Tahope shaped (plate girder) sections L | [ee (qd) Angie — {e) Channel (f) Built-up ames: (8) Composite steel- conerete Figure 1.5.5 Typical beam members. For beams (known. as lintels) carrying loads across window and door openings, angles are frequently used; and for beams (known as girts) in wall panels, channels are frequently used. Bending and Axial Load When simultaneous action of tension or compression along with bending occurs, a combined stress problem arises and the type of member used will be dependent on the type of stress that predominates. A member subjected to axial compression and bending is usually referred to as a beam-column, the behavior and design of which is dealt with in Chapter 12. The aforementioned illustration of types of members to resist various kinds of stress is intended only to show common and representative types of members and not to be all inclusive, 1.6 STEEL STRUCTURES Structures may be divided into three general categories: (a) framed structures, where elements may consist of tension members, columns, beams, and mem- bers under combined bending and axial load; (b) shell-type structures, where1.6 /STEEL STRUCTURES 19 axial stresses predominate; and (¢) suspension-type structures, where axial tension predominates the principal support system. Framed Structures Most typical building construction is in this category. The multistory building usually consists of beams and columns, either rigidly connected or having simple end connections along with diagonal bracing to provide stability. Even though a multistory building is three-dimensional, it usually is designed to be much stiffer in one direction than the other; thus it may reasonably be treated as a series of plane frames. However, if the framing is such that the behavior ‘of the members in one plane substantially influences the behavior in another plane, the frame must be treated as a three-dimensional space frame. Figure 1.6.1 Floor joists (plane trusses) and steel decking. (Photo by C. G. Salmon)20 1 /INTRODUCTION Figure 1.6.2 Space roof truss erected in sections; also shows plate girder at lower level containing vertical stiffener plates and special stiffening around rectangular holes through girder web. Upjohn Office Building, Kalamazoo, Michigan, (Photo courtesy of Whitehead and Kales Company) Industrial buildings and special one-story buildings such as churches, schools, and arenas, generally are either wholly or partly framed structures, Particularly the roof system may be a series of plane trusses (see Fig. 1.6.1), a space truss (see Fig. 1.6.2), a dome (see Fig. 1.6.3), or it may be part of a flat or gabled one-story rigid frame. Bridges are mostly framed structures, such as beams and plate girders (see Fig. 1.6.4), or trusses, usually continuous (see Fig. 1.6.5). Most of this text is devoted to behavior and design of elements in framed structures. Shell-Type Structures In this type of structure the shell serves a use function in addition to participation in carrying loads, One common type where the main stress is tension is the containment vessel used to store liquids (for both high and low temperatures), of which the elevated water tank is a notable example. StorageoS “S Figure 1.6.3 Dome roof, Brown University auditorium, Courtesy Bethlehem Steel Corporation Figure 1.6.4 Continuous orthotropic plate girder across Mississippi River at St. Louis, issouri, (Photo by C. G, Salmon) at22 1, /1NTRODUCTION Figure 1.65 Continuous truss bridge. Outerbridge Crossing, Staten Island, New York (Photo by C. G. Salmon) Figure 1.6.6 Cable-suspended roof for Madison Square Garden Sports and Entertain- ment Center, New York. Courtesy Bethlehem Steel Corporation.1.6/STEEL STRUCTURES 23 Figure 1.6.7 Suspension bridge: Golden Gate Bridge, San Francisco, California, (Photo by C. G. Salmon) bins, tanks, and the hulls of ships are other examples. On many shell-type structures, a framed structure may be used in conjunction with the shell. On walls and flat roofs the “skin” elements may be in compression while they act together with a framework. The aircraft body is another such example. Shell-type structures are usually designed by a specialist and are not within the scope of this text. Suspension-Type Structures In the suspension-type structure tension cables are major supporting elements. ‘A roof may be cable-supported, as shown in Fig. 1.6.6. Probably the most common structure of this type is the suspension bridge, as shown in Fig. 1.6.7. Usually a subsysiem of the structure consists of a framed structure, as in the stiffening truss for the suspension bridge. Since the tension element is the most efficient way of carrying load, structures utilizing this concept are increasingly being used. ‘Many unusual structures utilizing various combinations of framed, shell- type, and suspension-type structures have been built. However, the typical designer must principally understand the design and behavior of framed structures.24 © 1/ INTRODUCTION 1.7 SPECIFICATIONS AND BUILDING CODES Structural steel design of buildings in the United States is principally based on the specifications of the American Institute of Steel Construction (AISC) [1.5, 1.15]. AISC is comprised of steel fabricator and manufacturing companies, as well as individuals interested in steel design and research. The AISC Specifica- tions [1.5, 1.15] are the result of the combined judgment of researchers and practicing engineers. The research efforts have been synthesized into practical design procedures to provide a safe, economical structure. The advent of the digital computer in design practice has made feasible more elaborate design rules. The current specifications which are referred to throughout this book are the 1989 Specification for Structural Steel Buildings—Allowable Stress Design and Plastic Design [1.5] and the 1986 Load and Resistance Factor Design Specification for Structural Steel Buildings [1.15]. A specification containing a set of rules is intended to insure safety; however, the designer must understand the behavior for which the rule applies, otherwise an absurd, a grossly conservative, and sometimes unsafe design may result. The authors contend that it is virtually impossible to write rules that fully apply to every situation. Behavioral understanding must come first; application of rules then follows. No matter what set of rules is applicable, the designer has the ultimate responsibility for a safe structure. A specification when adopted by AISC is actually a set of recommenda- tions put forth by a highly respected group of experts in the field of steel research and design. Only when governmental bodies, such as city, state, and federal agencies, who have legal responsibility for public safety, adopt or incorporate a specification such as the 1986 Load and Resistance Factor Design Specification [1.15] into their building codes does it become legally official. The design of steel bridges is generally in accordance with specifications of the American Association of State Highway and Transportation Officials (AASHTO) [1.3]. This becomes a legal set of rules since it has been adopted by the states (usually the state highway departments have this responsibility). Railroad bridges are designed in accordance with the specifications adopted by the American Railway Engineering Association (AREA) [1.4]. In this case the railroads have the responsibility for safety and through their own. organization adopt the rules to insure safe designs. The term building code is sometimes used synonomously with specifica- tions. More correctly a building code is a broadly based document, either a legal document such as a state or local building code, or a document widely recognized even though not legal which covers the same wide range of topics as the state or local building code. Building codes generally treat all facets relating to safety, such as structural design, architectural details, fire protec- tion, heating and air conditioning, plumbing and sanitation, and lighting. On the other hand, specifications frequently refer to rules set forth by the architect or engineer that pertain to only one particular building while under construc-1.8,/ PHILOSOPHIES OF DESIGN: 25 tion. Building codes also ordinarily prescribe standard loads for which the structure is to be designed, as discussed in Sec, 1.4, The reader should not be disturbed by the interchangeable use of building code and specification, but should clearly understand that which is legally required for design and that which could be thought of as recommended practice, 1.8 PHILOSOPHIES OF DESIGN Two philosophies of design are in current use, the working stress design (referred to by AISC as Allowable Siress Design) and limit states design (referred to by AISC as Load and Resistance Factor Design). Working stress design has been the principal philosophy used during the past 100 years. During the past 20 years or so, structural design has been moving toward a more rational and probability-based design procedure referred to as “limit states” design. Haaijer [1.26] and Kennedy [1.27, 1.28] have presented the current status of the limit states concept and its use in design. Limit states design includes the methods commonly referred to as “ultimate strength design,” “strength design,” “plastic design,” “load factor design,” “limit design,” and the recent “load and resistance factor design” (LRFD). Structures and structural members must have adequate strength, as well adequate stiffness and toughness to permit proper functioning during the service life of the structure. The design must provide for some reserve strength above that which is needed to carry the service loads; that is, the structure must provide for the possibility of overload. Overloads can arise from chang- ing the use for which a particular structure was designed, from underestima- tion of the effects of loads by oversimplifications in structural analysis, and from variations in construction procedures. In addition, there must be a provision for the possibility of understrength. Deviations in the dimensions of members, even though within accepted tolerances, can result in a member having less than its computed strength. The materials (steel for members, bolts, and welds) may have less strength than used in the design calculations. ‘A steel section may occasionally have a yield stress below the minimum specified value, but still within the statistically acceptable limits. Structural design must provide for adequate safety no matter what philos- ophy of design is used. Provision must be made for both overload and understrength. The study of what constitutes the proper formulation of struc- tural safety has been continuing during the past 30 years [1.19, 1.20]. The main thrust has been to examine by various probabilistic methods the chances of “failure” occurring in a member, connection, or system. Rather than refer to “failure” the term “limit state” is preferred. Limir states means “those conditions of a structure at which it ccases to fulfill its intended function” [1.27]. Limit states are generally divided into two cate- gories, strength and serviceability. Strength (i.e, safety) limit states are such26 = 4/INTRODUCTION Frequency Resistance R Load @ Figure 1.8.1 Frequency distributions of load Q and resistance R. behavioral phenomena as achieving ductile maximum strength (i.e., plastic strength), buckling, fatigue, fracture, overturning, and sliding, Serviceability limit states are those concerned with occupancy of the building, such as deflection, vibration, permanent deformation, and cracking, Both the loads acting and the resistance (strength) of the structure to loads are variables that must be considered. In general, a thorough analysis of all uncertainties that might influence achieving a “limit state” is not practical, or perhaps even possible. The current approach to a simplified method using a Frequency Ongnpoy = Standard deviation of InnR/O) Failure 0 tin Ft/O)) in(R/Q) — | Figure 1.8.2 Reliability index B.1.8/ PHILOSOPHIES OF DESIGN = 27 probability-based assessment of structural safety uses first-order second-moment reliability methods [1.32]. Such methods assume the load Q and the resistance R are random variables. Typical frequency distributions of such random variables are shown in Fig. 1.8.1. When the resistance R exceeds the load Q there will be a margin of safety. Unless R exceeds Q by a large amount, there will be some probability that R may be less than @, shown by the shaded portion where the R and Q curves overlap. Structural “failure” (achievement of a limit state) may then be defined by comparing R with Q, or in logarithmic form comparing In(R/Q), as shown in Fig. 1.8.2. On both figures “failure” is the cross-hatched region. The distance between the failure line and the mean value of the function [In(R/Q)] is defined as a multiple 8 of the standard deviation o of the function, The multiplier # is called the reliability index. The larger is B the greater is the margin of safety. ‘As summarized by Pinkham [1.20], the reliability index PB is useful in several ways: 4. It can give an indication of the consistency of safety for various components and systems using traditional design methods. 2. It can be used to establish new methods which will have consistent margins of safety. 3. It can be used to vary in a rational manner the margins of safety for those components and systems having a greater or lesser need for safety than used in ordinary situations. In general, the expression for the structural safety requirement may be written as $B, 2 EQ, (1.8.1) where the left side of Eq. 1.8.1 represents the resistance, or strength, of the component or system, and the right side represents the oad expected to be carried. On the strength side, the nominal resistance R,, is multiplied by a resistance: (strength reduction) factor } to obtain the design sirength (also called usable strength or usable resistance). On the load side of the equation, the various load effects Q; (such as dead load, live load, and snow load) are multiplied by overload factors y, to obtain the sum Ly,Q, of factored loads. AISC — Load and ‘Resistance Factor Design (LRFD) During the 1980s the general “limit states design” approach has gained acceptance, culminating for steel design in the United States with the adoption of the 1986 AISC Load and Resistance Factor Design Specification [1.15], referred to throughout this book as AISC LRFD Specification. In Canada, limit states design for steel has been used since 1974 and since 1978 it has been the only method used; the latest edition is 1984 [1.18]_ The AISC LRFD Specification was developed under the leadership of T. V. Galambos at Washington University, St. Louis [1.21, 1.23]. The adapta-28 © 1,/INTRODUCTION tion of probabilistic methods to steel design and the development of the LRED Specification are explained by Galambos (1.22, 1.24] and by Galambos and Ravindra [1,23, 1.25]. ‘The general format of the LRFD Specification is given by Eq. 1.8.1. In general, Eq. 1.8.1 means that the strength (@R,) provided in design must at least equal the factored loads (27,0;) acting, The subscript i indicates that there will be terms for each type of load Q, acting, such as dead load (D), live load (), wind load (W), snow load (S), and earthquake load (£). The factors y, may be different for each type of load. * The AISC LRFD Specification is based on the following: 1. A probability-based model [1.23, 1.25. 1.32, 1.33]. 2. Calibration with the 1978 AISC ASD Specification. 3. Evaluation using judgment and past experience, along with design offices studies of representative structures. The development of probability-based load criteria by Galambos, Elling- wood, MacGregor, and Cornell (1.32, 1.33] led to the factored load combina- tions of the 1982 ANSI Standard [1.2]. The ANSI Standard [1.2] was devel- oped for use in design with al! structural materials. It is reasonable that the probability of various loads acting in combination, as well as the probability of overload with certain types of loads, should be unrelated to the material of which a structure is built. With this concept in mind, the 1986 AISC LRFD Specification adopted the ANSI factored load combinations. LRFD-A4.1 specifies that the following combinations shall be investigated: 14D (1.8.2) 12D + 1.6L + 0.5(L, or Sor R) (1.8.3) 1.2D + 1.6(L, or Sor R) + (0.5L or 08H) (1.8.4) 12D + 13W + 0.5L + 0.5(L, or S or R) (1.8.5) 1.2D + 15E + (0.5L or 0.25) (1.8.6) 0.9D - (1.3 or 1.5E) (1.8.7) The nominal service loads indicated by Eqs. 1.8.2 through 1.8.7 are D = dead load (gravity load from the weight of structural elements and permanent attachments) L = live load (gravity occupancy and movable equipment load) L, = roof live load W = wind load S = snow load E = earthquake load R = rainwater or ice load1.8 /PHILOSOPHIES OF DESIGN 29 Note that D, L, W, S, etc. are loads in a general sense, which includes bending moment, shear, axial force, and torsional moment. Sometimes these internal forces are called load effects. Thus, the symbol D means dead load, dead load moment, dead load shear, dead load axial foree, etc. The explana tion of the statistics relating to snow and wind load factors is given by Ravindra, Cornell, and Galambos [1.34]. AISC — Allowable Stress Design (ASD) The traditional method of the ATSC Specification has been allowable stress design (also called working stress design). In Allowable Stress Design (ASD) the focus is on service load conditions (i.c., unit stresses assuming an elastic structure) when satisfying the safety requirement (adequate strength) for the structure. The AISC 1989 Specification for Allowable Stress Design is referred to in this book as the ASD Specification. For Allowable Stress Design, Eq. 1.8.1 may be reformulated as follows: oR, ** 2509, (18.8) Y . In this philosophy all loads are assumed to have the same average variability. The entire variability of the loads and the strengths is placed on the strength side of the equation, To examine the equation in terms of Allowable Stress Design for beams, the left side would represent nominal beam strength M, divided by a factor of safety FS (equal to 7), and the right side would represent the service load bending moment M acting resulting from all types of loads. Thus, Eq. 1.8.8 would become Me M 1.8.9) a (1.8.9) ‘The term Allowable Stress Design implies an elastic stress calculation. Equa- tion 1.8.9 may be divided by I/c (Le., the moment of inertia J divided by the distance ¢ from the neutral axis to the extreme fiber) to obtain stress units. Thus, if one assumes the nominal strength M, is reached when the extreme fiber stress is the yield stress F, (ic., M, = F,J/c), Eq. 1.8.9 will become Fife M a 2 1.8.10 FS Ife = Ife (4:10) or £ Mc —. ot ee me [* ; | (1.8.11) In ASD the F,/FS would be the allowable stress F, and f, would be the computed elasiic stress under full service load. If the nominal strength M, had been based on achievement of a stress F,, less than F, because of, say, buckling, then the allowable stress F, would be £,/FS. Thus, the safety30 1 / INTRODUCTION criterion in ASD may be written 5 Fe he|b- Reams (1.8.12) The allowable stresses of the ASD Specification [1.5] are derived from the strength capable of being achieved if the structure is overloaded. When the section is ductile and buckling does not occur, strains greater than the “first yielding” strain «, = F,/E, can exist on the section (E, is the modulus of elasticity), Such ductile inelastic behavior may permit higher loads to be carried than possible if the structure had remained entirely elastic. In such cases the allowable stress is adjusted upward. When the strength is limited by buckling or some other behavior such that the stress does not reach yield stress, the allowable stress is adjusted downward. Serviceability requirements such as deflection limits are always investi- gated at service load conditions, whether the LRFD or the ASD design procedure has been used to satisfy safety requirements. AISC — Plastic Design ‘Traditionally, Part 2 of the AISC Specification was called Plastic Design. The 1989 Specification for Structural Steel Buildings [1.5] contains Plastic Design in Chapter N. Plastic Design is a special case of limit states design, wherein the limit state for strength is the achievement of plastic moment strength M,. Plastic moment strength is the moment strength when all fibers of the cross-section are at the yield stress F, (one side of the neutral axis in tension and the other side in compression). Plastic design does not permit using other limit states, such as instability, fatigue, or brittle fracture. The design philoso- phy as used by AISC applies to flexural members, including beam-columns, and for such members may be expressed by Eq. 1.8.1. Then letting R, = M, and y,/¢ = 1.7, Eq. 1.8.1 becomes M, = 1.739, (1.8.13) ‘The provisions for overload and for understrength are combined into a single factor 1.7 used for all gravity loads. The nominal strength must be the plastic moment strength M,. Since plastic design is a special case of limit states design and is covered more rationally in the AISC LRFD Specification, it is no longer treated as a special topic as in previous editions of this book. Plastic design becomes a component of LRFD. ‘The limit states design philosophy as codified in LRFD provides the designer with a more rational approach to design than has been available in ASD or Plastic Design, whose philosophies are outlined in the following paragraphs, Beedle [1.29] provides an excellent summary of, the advantages of using LRFD.1.9/ FACTORS FOR SAFETY —ASD AND LAFD GOMPARED 311 1.9 FACTORS FOR SAFETY — ASD AND LRFD COMPARED eX: ee eee Allowable Stress Design (ASD) The “safety factor” FS used in Eqs, 1.8,9 through 1.8.12 was not determined consciously by using probabilistic methods. The values used in the AISC ASD Specification [1.5] have been in use for many years and are the result of experience and judgment. It is clear that the safety required must be a combination of economics and statistics. Obviously, it is not economically feasible to design a structure so that the probability of failure is zero. Prior to the development of the AISC LRFD Specification [1.15], the AISC Specifica- tions from 1924 through 1978 did not give a rationale for the allowable stresses prescribed. One may state that the minimum resistance must exceed the maximum applied load by some prescribed amount. Suppose the actual load exceeds the service load by an amount AQ, and the actual resistance is less than the computed resistance by an amount AR. A structure that is just adequate would have R,-AR,=Q+40 R,(1- AR,/R,) = Q(1 + 40/2) (1.9.1) ‘The margin of safety, or “safety factor,” would be the ratio of the nominal strength R,, to nominal service load Q; or R 1 pea fs 2 Aeie (19.2) @ 1-AR/R, Equation 1.9.2 illustrates the effect of overload (AQ/Q) and understrength (AR,/R,,); however it does not identify the factors contributing to either. If ‘one assumes that occasional overload (AQ/Q) may be 40% greater than its nominal value, and that an occasional understrength (AR,/R,) may be 15% Jess than its nominal value, then 1404 14 1-015 0.85 — The above is an oversimplification but it shows a possible scenario for obtaining the traditional AISC value of FS — 1.67 used as the basic value in Allowable Stress Design. Dividing by 1.67 as indicated in Eq. 1.8.12 gives a multiplier of 0.60 on F, or Fe. ‘The basi¢ value of 1.67 is used for tension members and beams, and it is the lower bound for zero-length columns. A value of 1.92 is used for long columns, and values from 2.5 to 3 are used for connections. However, it must be noted that using these values for y/@ in Eq. 1.8.8 still leaves the “real” safety against “failure” unknown!32 1 /INTRODUCTION Load and Resistance Factor Design (LAFD) ‘As discussed in See. 1.8, the factors for overload are variable depending upon the type of load, and the factored load combinations that must be considered are those given by the ANSI Standard [1.2] and LRFD-Ad.1, and presented as Eqs. 1.8.2 through 1.8.7. The other part of the safety-related provisions is the factor, known as the resistance factor. The resistance factor » varies with the type of member and with the limit state being considered. Some representative resistance factors } are as follows: Tension Members (LRFD-D1) , = 0.90 for yielding limit state 6, = 0.75 for fracture limit state Compression Members (LRFD-E2) 4, = 0.88 Beams (LRFD-F1.2) , = 0.90 Welds (LRFD-Table J2.3) = same as for type of action; i.¢., tension, bending, etc. Fasteners (A325 bolts) (LRFD-Table J3.2) @=0.75 for tensile strength = 0.65 for shear strength In order to establish adequate safety using probabilistic methods the natural logarithm of the resistance R divided by the load Q, that is, In(R/Q) as shown in Fig. 1.8.2, may be treated as a random variable and is simpler than working with two groups ( and Q) of random variables as in Fig. 1.8.1. When in(R/Q) < 0, the limit state has been exceeded, and the shaded area in Fig, 1.8.2 is the probability of this event. The method used to develop LRFD uses the mean salues R,, and Q,, and the standard deviations op and 9g of the resistance and load, respectively. Frequently, the mean values and standard deviations can be estimated while the actual distributions cannot be obtained. ‘Thus, using the quantities that may be estimated the standard deviation o of the In(R/Q) may be approximated as Since oy = [v2 + V3 (1.9.3) where Va = o2/Ryy Vg = 00/ Om ‘The margin of safety is the distance from the origin to the mean, as shown. in Fig. 1.8.2, and is expressed as a multiple B of y¢x/o)- The distance1.9 / FACTORS FOR SAFETY — ASD AND LRFD COMPARED 33 representing the margin of safety may be approximated [1.16] as Bonnar = BY Va + Vg = I Rn/ Qn) (1.9.4) ‘Thus, the larger the distance the smaller the probability of reaching the limit state, The multiplier f is called the reliability index, The expression for 8 from Eq, 1.9.4 becomes _ W(R,,/Q,) More discussion of the development of Eq. 1.9.5 is given in the LRFD Commentary [1.16], by Ravindra and Galambos [1.25], and in NBS Special Publication 577 [1.32]. The treatment of the theory of probability is outside the scope of this book. Using the factored load combinations given by the ANSI Standard [1.2], the AISC Task Force and Specification Committee calibrated the LRFD Specification [1.15] to generally agree with past experience. Thus, the resis- tance factors @ were set in LRFD with the objective of obtaining the following values of f: (1.9.5) Objective Load Combinations Reliability Index 8 Dead load + live load (or snow load) 3.0 for members 4.5 for connections Dead load + live load + wind load 2.5 for members Dead load + live load + earthquake load 1.75 for members Because of a lower probability of wind or earthquake occurring with full gravity load, the reliability index 2 was made lower for those cases. The 8 values for connections were made higher than for members in order to make connections stronger than members as has been traditionally done, ‘The LRFD Specification development combined the following: 1. Probability model 2. Calibration to the 1978 AISC ASD Specification 3. Judgment and past experience 4. Design office studies of representative structures The LRFD Specification (LRFD-A4.) uses six factored load combina- tions, given as Eqs. 1.8.2 through 1.8.7. This was necessary to account for each of the separate loads (dead, live, roof, wind additive to gravity, wind opposite to gravity, and earthquake) acting at its maximum lifetime value. The loads other than dead load will act at the “arbitrary point-in-time” value. The “arbitrary point-in-time” value is that value which can be expected to be on the structure at any time. The arbitrary point-in-time value of live load (L)34 1/INTRODUCTION might be as low as one-quarter of its mean maximum lifetime load but its distribution widely varies. The arbitrary point-in-time wind (1) is the maxi- mum daily wind, The “lifetime maximum” is taken as the 50-year recurrence value. Thus, each factored load combination and its corresponding load occur- ring at its 50-year maximum are as follows: Load at its Lifetime LRFD Eq. Load Combination (50 year) Maximum (A4-1) 14D Dead load D during construction; other loads not present (A4-2) 12D + 16L + 05S Live load L (A4-3) 12D + 1.68 + (0.8 or 0.5L) Roof load; ie., snow load S or rain R other ‘than ponding effect (A4-4) 12D +13W + 05L +055 Wind load W additive to dead load (A4-5) 12D + 15E + (05L or 0.28) Earthquake load E additive to dead load (A4-6) 0.9D — (1.3 or LSE) Wind load W or earthquake load E opposite to dead load fote; Where snow S is used in the above equations, the meaning is snow S OR oof live load £, OR rain R other than ponding. Comparison of LAFD with ASD for Tension Members ‘The comparison of safety obtained for tension members in design by the two AISC methods is indicative of the general design result expected. Direct comparisons will be more difficult in design of other types of members because the nominal strengths R, are not necessarily the same in the two methods, For tension members acted upon by gravity dead and live loads, the resistance factor ¢ = 0,90, and using Eq, 1.8.3 gives for LRFD 12D + 16L = 0.90R, (1.8.3) 1.33D + 1,78L=R, LRFD In Allowable Stress Design (ASD) the factor of safety FS = 1.67 for axial tension, which gives from Eq, 1.8.8 where ¥/¢ is the factor of safety R,/167=50=D+L [1.8.8]1.9/ FACTORS FOR SAFETY —ASD AND LAFD COMPARED = 35 Figure 1.9.1 Comparison of Load and Resistance Factor Design with Allowable Stress Design for tension members, or 1.67) + 167L=R, ASD Next, dividing Eq. 1.8.3 by Eq. 1.8.8 gives LRFD 1.33) +1.78L _ 0.8 + 1.07(L/D) ASD — 1.67) + 1.67L 14+ (L/D) Since this is a gravity load comparison, LRFD Eq. (A4-1) [Eq 1.8.2] must also be used as L/D approaches zero, Thus, Eq. 1.8.2 gives (19.6) 14D = 0.90R, (1.8.2] 1.56D = R, LRED Dividing LRFD by ASD gives LRFD 156D 0.93 ASD 16D+167L 1+ (L/D) (19.7) Equations 1.9.6 and 1.9.7 are shown plotted in Fig, 1.9.1. The design of tension members will be about the same in both LRFD and ASD when the live load to dead load ratio (L/D) is about 3, As the L/D ratio becomes lower (that is, dead load becomes more predominant) there will be economy in using LRFD. With the L/D ratio larger than 3, ASD will be slightly more economical, but rarely by more than about 3%.36 © 1 /INTRODUCTION 1.10 WHY SHOULD LRFD BE USED? The many advantages of LRFD are well-expressed by Beedle [1.29], whose listing is the basis of the following: 1. LRED is another “too!” for structural engineers to use in stecl design. ‘Why not haye the same tools (variable overload factors and resistance factors) available for steel design as are available for concrete design? 2, Adoption of LRFD is not mandatory but provides a flexibility of options to the designer. The marketplace will dictate whether or not LRFD will become the sole method. 3. ASD is an approximate way to account for what LRFD does in a more rational way. The use of plastic design concepts in ASD has made ASD such that it no longer may be called an “elastic design” method. 4. The rationality of LRFD has always been attractive, and becomes an incentive permitting the better and more economical use of material for some load combinations and structural configurations. It will also likely lead to having safer structures in view of the arbitrary practice under ASD of combining dead and live loads and treating them the same. 5. Using multiple load factor combinations should lead to economy. 6. LRED will facilitate the input of new information on loads and load variations as such information becomes available. Considerable knowledge of the resistance of steel structures is available, On the other hand our knowledge of loads and their variation is much less. Separating the loading from the resistance allows one to be changed without the other if that should be desired. 71. Changes in overload factors and resistance factors ¢ are much easier to make than to change the allowable stress in ASD. 8. LRFD makes design in all materials more compatible. The variability of loads is actually unrelated to the material used in the design. Future specifications not in the limit states format for any material will put that material at a disadvantage in design. 9. LRFD provides the framework to handle unusual loads that may not be covered by the Specification. The design may have uncertainty relating to the resistance of the structure, in which case the resistance factors may be modified. On the other hand, the uncertainty may relate to the loads and different overload factors may be used. 10. Future adjustments in the calibration of the method can be made without much complication. Calibration for LRFD was done for an average situation but might be adjusted in the future. 11. Economy is likely to result for low live load to dead load ratios. For high live load to dead load ratios there will be diseconomy but a low amount. 12. Safer structures may result under LRFD because the method should lead to a better awareness of structural behavior.1.11 / ANALYSIS OF THE STRUCTURE 37, 13, Design practice is still at the beginning with regard to serviceability limit states; however, at least LRFD provides the approach, The rationality of LRFD and its many advantages as outlined by Beedle [1.29] over ASD are indicative that the design philosophy will relegate ASD to the background in the next few years. It is important, however, that the designer understand both philosophies of design because many structures will continue to be designed by ASD and the designer may frequently need to evaluate structures of the past. Lindsey [1.31] provides an interesting article on the recent advances in stecl design, and Heger [1.30] provides some interesting thoughts on the difficulties of bridging the gap between the theory of statistics and probability and the real world of actual structures. 1.11 ANALYSIS OF THE STRUCTURE In general, the structural analysis to obtain the service loads (or load ¢/fects bending moment, shear, axial force, and torsional moment) on the members is performed the same for LRFD as for ASD. Elastic methods of structural analysis are used except when the limit state is the plastic collapse mechanism as described in Chapter 10. A first-order analysis (1.14, p. 21] is sufficient in usual framed structures that are braced against sway. In a first-order analysis equilibrium equations are based on the original geometry of the structure, This means that the designer is assuming the internal forces (moments, shears, ete.) are not sufficiently affected by the change in shape of the structure to justify more complicated analysis. When the elastic displacements are small com- pared to the dimensions this approximation will be satisfactory. ‘The most common situation where a second-order effect must be consid- ered is in the multistory structure that must rely on the stiffnesses of the interacting beams and columns to resist sway under lateral loading (wind and/or earthquake). This is the so-called unbraced frame. In this case the lateral displacement A (also called sway or drift) causes additional bending moments due to the gravity loads (ZP) acting at the displaced position A. The analysis must include this PA secondary effect. There are varying degrees of sophistication that are used in analysis to include the PA effects. In both ASD and LRFD the second-order effects may be computed as a part of the analysis or they may be accounted for using formulas in the Specifications (1.5, 1.15] or ‘Commentaries [1.6, 1.16]. ‘The emphasis in this book is on designing the members to have adequate strength and proper serviceability, rather than on structural analysis. The reader is referred to Wang and Salmon [1.14] for basic structural analysis topics: Other references specifically related to analysis for unbraced frames are given in Chapter 14. Most examples in this book use given service loads, or service load. effects, acting on the structural member to be designed. These values are to be assumed the result of structural analysis. The service loads are to be factored using Egs. 1.8.2 through 1.8.7 for design in LRFD but used as given in ASD.38 1, / INTRODUCTION SELECTED REFERENCES 1. 12 13, 14. 1.5. 16. 17. 18. 19. 1.10, Ll 1.12. 113. 1.14. Hans Straub, A History of Civil Engineering. Cambridge, MA: M.LT, Press, 1964 (pp. 173-180). ANSL American National Standard Minimum Design Loads for Buildings and Other Structures, AS$.1-1982. New York: American National Standards Insti- tute, Ine. (1430 Broadway, New York, NY 10018), 1982, AASHTO. Standard Specifications for Highway Bridges, 14th ed. Washington, DC: The American Association of State Highway and Transportation Officials (444 North Capitol Street, N.W., Suite 225, Washington, DC 20001), 1989. AREA. Manual for Railway Engineering, Chapter 15 (Specifications for Steel Railway Bridges). Chicago, IL: American Railway Engineering Association, 1980. AISC. Specification for Structural Stee! Buildings—ANowable Stress Design and Plastic Design (June I, 1989). Chicago: American Institute of Steel Construction (400 North Michigan Avenue, Chicago, IL 60611-4185), 1989. AISC. Commentary on the Specification for Structural Steel Buildings—Allowable Stress: Design and Plastic Design (June 1, 1989). Chicago, IL: American Institute of Stecl Construction, 1989. AISC. Manual of Stee! Construction, Allowable Stress Design, 9th ed. Chicago, IL: American Institute of Steel Construction, 1989. ASTM. Standard Specification for General Requirements for Rolled Steel Plates, Shapes, Sheet Piling, and Bars for Structural Use, A6-88c. Philadelphia, PA: ‘American Society for Testing and Materials, 1988. National Research Council of Canada. Canadian Structural Design Manual, Supplement No. 4 to the National Building Code of Canada, Ottawa, Canada, 1983. HL S. Lew, Emil Simiu, and Bruce Ellingwood. “Loads,” Chapter 2, Building Structural Design Handbook, Richard N. White and Charles G. Salmon, Ed. New York: John Wiley & Sons, 1987 (pp. 9-43). AISI. Specification for the Design of Cold-Formed Steel Structural Members. ‘Washington, DC: American Iron and Steel Institute, September, 1980. AISI. Cold-Formed Steel Design Manual, Part I, Specification; Part I, Commen- tary; Part II, Supplementary Information; Part IV, Iltustrative Examples; and Part V, Charis and Tables. New York: American Iron and Steel Institute, 1982. SUL Standard Specifications Load Tables & Weight Tables for Steel Joists & Joist Girders. Myrtle Beach, SC: Steel Joist Institute (Suite A, 1205 48th Avenue North, Myrtle Beach, SC 29577), 1988, Chu-Kia Wang and Charles G. Salmon. Introductory Structural Analysis. Engle- wood Cliffs, NJ: Prentice-Hall, Inc,, 1984, Load and Resistance Factor Design 1.15. 1.16. AISC. Load and Resistance Factor Design Specification for Structural Stee! Build- ings. Chicago, IL: American Institute of Steel Construction, September 1, 1986. AISC. Commentary on the Load and Resistance Factor Design Specification for Structural Steel Buildings (September 1, 1986). Chicago, UL: American Institute of Steel Construction, 1986.117. 118. 1.19. 1.20, 121, 1,22, 1.23. 124. 1.25. 1.26. 127. 1.28. 1.29, 1.30. 1.31. 1.32. 1.33. 134, SELECTED REFERENCES = 39 AISC. Manual of Stee! Construction, Load & Resistance Factor Design. Chicago, IL: American Institute of Steel Construction, 1986. CSA. Steel Structures for Buildings—Limit States Design, CAN3-S16.1-M84. Rexdale, Ontario, Canada: Canadian Standards Association, 1984. ASCE Task Committee on Structural Safcty of the Administrative Committee on. Analysis and Design of the Structural Division. “Structural Safety—A Literature Review,” Journal of the Structural Division, ASCE, 98, ST4 (April 1972), 845-884. Clarkson W, Pinkham. “Design Philosophies,” Chapter 3, Building Structural Design Handbook, Richard N, White and Charles G. Salmon, Ed. New York: John Wiley & Sons, 1987 (pp. 44-34). C. W, Pinkham and W. C. Hansell. “An Introduction to Load and Resistance Factor Design for Stecl Buildings,” Engineering Journal, AISC, 15, 1 (First ‘Quarter 1978), 2-7. T, V, Galambos. “Load Factor Design of Steel Buildings,” Engineering Journal, AISC, 9, 3 (July, 1972), 108-113. Theodore V. Galambos and M. K. Ravindra. “Proposed Criteria for Load and Resistance Factor Design,” Engineering Journal, AISC, 15, 1 (First Quarter 1978), 8-17. Theodore V. Galambos, “Load and Resistance Factor Design,” Engineering Journal, AISC, 18, 3 (Third Quarter 1981), 74-82 Mayasandra K. Ravindra and Theodore V. Galambos. “Load and Resistance Factor Design for Steel,” Journal of the Structural Division, ASCE, 104, ST9 (September 1978), 1337-1353. Geerhard Haaijer. “Limit States Design—A Tool for Reducing the Complexity of Steel Structures,” paper presented at AISC National Engineering Conference, March 4, 1983. D. I. Laurie Kennedy. “Limit States Design of Steel Structures in Canada,” Journal of Structural Engineering, ASCE, 110, 2 (February 1984), 275-290. D. J, Laurie Kennedy. “North American Limit States Design,” Proceedings, The 1985 International Engineering Symposium on Structural Steel. Chicago, IL: American Institute of Steel Construction, May 22-24, 1985. Lynn §. Beedle: “Why LRFD?” Modern Steel Construction, AISC 26, 4 (4th ‘Quarter, 1986), 30-31. Frank J. Heger. “Proposed AISC LRFD Design Criteria,” Journal of the Struc- tural Division, ASCE, 106, ST3 (March 1980}, 729-734, Disc, 106, ST12 (Decem- ber 1980), 2576-2577. Stanley D. Lindsey, “Advances in Steel Design and Construction,” Proceedings, Solutions in Steel, The National Engineering Conference. Chicago, IL: American Institute of Steel Construction, June 12-14, 1986. Bruce Ellingwood, Theodore V, Galambos, James G. MacGregor, and C. Allin Cornell. Development of a Probability Based Load Criterion for American National Standard AS8, NBS Special Publication 577, Washington, DC: US Department of Commerce, National Bureau of Standards, June 1980. Theodore V. Galambos, Bruce Ellingwood, James G. MacGregor, and C. Allin Cornell. “Probability Based Load Criteria: Assessment of Current Design Prac- tice,” Journal of the Structural Division, ASCE, 108, STS (May 1982), 959-977. Mayasandra K. Ravindra, C. Allin Comell, and Theodore V. Galambos. “Wind and Snow Load Factors for Use in LRFD,” Journal of the Structural Division, ASCE, 104, ST9 (September 1978), 1443-1457.40 1/ INTRODUCTION Wind Loads 1.35. ASCE, Task Committee on Wind Forces, Committee on Loads and Stresses, Structural Division. “Wind Forces on Structures,” Journal of the Structural Division, ASCE, 84, ST4 (Tuly 1958) (Preliminary Report); and Final Report, Transactions, ASCE, 126, pt, I (1961), 1124-1198. 1.36. Kishor C. Mehta. “Wind Load Provisions ANSI #A58.1-1982,” Journal of ‘Structural Engineering, 110, 4 (April 1984), 769-784, 1.37. Theodore Stathopoulos, David Surry, and Alan G. Davenport. “Effective Wind Loads on Flat Roofs,” Journal of the Structural Division, ASCE, 107, ST2 (February 1981), 281-298. Snow Loads 1.38. Michael J. O'Rourke and Ulrich Stiefel. “Roof Snow Loads for Structural Design,” Journal of Structural Engineering, ASCE, 109, 7 (July 1983), 1527-1537. 1.39, J.T. Templin and W. R. Schriever, “Loads due to Drifted Snow,” Journal of the Structural Division, ASCE, 108, ST8 (August 1982), 1916-1925, 1.40. Michael O'Rourke, Wayne Tobiason, and Evelyn Wood, “Proposed Code Provi- sions for Drifted Snow Loads,” Journal of Structural Engineering, 112, 9 (Septem- ber 1986), 2080-2092. 1.41. Michael J. O'Rourke, Robert Redfield, and Peter von Bradsky. “Uniform Snow Loads on Structures,” Journal of the Structural Division, ASCE, 108, ST12 (December 1982), 2781-2798, 1.42. Michael J. O'Rourke, Robert S. Speck, Jr., and Ulich Stiefel. “Drift Snow Loads on Multilevel Roofs,” Jounal of Structural Engineering, ASCE, 11, 2 (February 1985), 290-306. Earthquake Load 1.43, SEAOC, Recommended Lateral Force Requirements and Commentary. San Fran- ciseo, CA: Seismology Committee, Structural Engineers Association of Califor nia, 1974, 1,44, Anil K. Chopra and Emesto F. Cruz. “Evaluation of Building Code Formulas for Earthquake Forces,” Journal of Structural Engineering, ASCE, 112, 8 (August 1986), 1881-1899, 1.45, Anil K. Chopra. Dynamics of Structures, A Primer. Berkeley, CA: Earthquake Engineering Research Institute, 1980. 1.46. R. W. Clough and Joseph Penzien. Dynamics of Structures. New York: McGraw- Hill, 1975. 147. Applied Technology Council (ATC), Associated with the Structural Engineers ‘Association of California. Tentative Provisions for the Development of Seismic Regulations for Buildings (ATC 3-06) (NBS Publication 510). Washington, DC: Center for Building Technology, National Bureau of Standards, June 1978. 1,48, UBC. Uniform Building Code, Whittier, CA; International Conference of Build- ing Officials, 1985,Chapter 2 Steels and Properties 2.1 STRUCTURAL STEELS During most of the period from the introduction of structural steel as a major building material until about 1960, the steel used was classified as a carbon. steel with the ASTM (American Society for Testing and Materials) designation A7, and had a minimum specified yield stress of 33 ksi. Most designers merely referred to “‘steel” without further identification, and the AISC specification prescribed allowable stresses and procedures only for the A7 type of steel. Other structural steels, such as a special corrosion resistant low alloy steel (A242) and a more readily weldable steel (A373), were available but they were rarely used in buildings. Bridge design made occasional use of these other steels. Today (1989) the many steels available to the designer permit use of increased strength material in highly stressed regions rather than greatly increase the size of members. The designer can decide whether maximum rigidity or least weight is the more desirable attribute. Corrosion resistance, hence elimination of frequent painting, may be a highly important factor. Some steels oxidize to form a dense protective coating that prevents further oxidation (corrosion), acquiring a pleasing even-textured dark red-brown appearance, Since painting is not required, it may be economical to use these “weathering steels” even though the initial cost is somewhat higher than traditional carbon steels. Certain steels provide better weldability than others; some are more suitable than others for pressure vessels, either at temperatures well above or well below room temperatures, * 4a42 2/STEELS AND PROPERTIES John Hancock Center, Chicago, showing exterior diagonal bracing. (Photo by C. G. Salmon) Structural steels are referred to by ASTM designations, and also by many proprietary names. For design purposes the yield stress in tension is the reference quantity that specifications, such as AISC, use as the material property variable to establish strength or allowable stress. The term yield ‘stress ig used to include either “yield point,” the well-defined deviation from perfect elasticity exhibited by most of the common structural steels; or “yield strength,” the unit stress at a certain offset strain for steels having no well-defined yield point. Today (1990) steels are readily available having yield stresses from 24 to 100 ksi (170 to 690 MPa). Steels for structural use in hot-rolled applications may be classified as carbon sieels, high-strength low-alloy steels, and alloy steels. The general requirements for such steels are covered under ANSI/ASTM A6 SpecificationTABLE 2.1.1 PROPERTIES OF STEELS USED FOR BUILDINGS AND BRIGGES F, ‘Minimum & Maximum yield Tensile thickness ASTM AG. | stress strength for plates groups* asTMt ksi ksi in. for designation =} (MPa)* (MPa)* (mm) shapes A36 2 $8-80 Over 8 = (220) (400-550) (200) 36 58-80 Tos all 250) (400-550) (200) A33 Grade B 35 (240) 60 (415) A2a2 a 3 Over 1} tod 45 290) (435) (40 to 200) 46 67 Over} tot 3 1s) (460) (20 to.40) 30 70 To} 12 (345) (480) 20) ‘AdAL Discontinued 1989; replaced by A372 ‘A500 Grade A 33-228) 45 (310) Round Grade B 42 (250) 58 (400) Round Grade C 46 (317) 62. (427) Round Grade A 39 (269) 45 (310) ‘Shaped Grade B 46 17) 58 400) ‘Shaped Grade C 50 (345) 62 (427) Shaped ASOL 36 (250) 38 (400) ASIA 0 100-130 Over 2} to 6 (620) (690-895) {65 to 150) 100 110-130 To2t (690) (760-895) 3) A829 a2 0-85. To} 1 290) (a14—s86) a3) ‘A510 Grade 40 40 275) 55 (380) Grade 45 45 (310) 60 (415) Grade 50 50 G45) 65 (450) A572 Grade 42 a2 60 To6 All 2%) (413) (iso) Grade 50 0 65 Tos all (B45) (450) qao0) Grade 60 60 7s Tok 12 (415) (20) (32) Grade 65 65 80 Toit 1 (450) (550) Gd AS88 42 63 Over 5 to 8 - (290) (435) (125 to 200) a6 ‘r Over 4 to 5 - Gis) (460) (200 to 125) 30 70 Tod All (345) (485) «i00) A606 45 (310) 65 (430) 50 (345) 70 (480)44° 2/STEELS AND PROPERTIES TABLE 2.1.1 Continued &, Maximum: Tensile thickness ASTM AB strength for plates groups" asTut isi in. tor designation (MPa)* (mm) shapes: A607 Grade 45 45 G10) 60 (410) Grade 50 $0 (340) 65 (450) Grade 55 55 (380) 70 (480) Grade 60 60 410) 18 (820) Grade 65 65 (450) 80(350) Grade 70 70 (480) 85 (590) AGLI Grade C BQH) 48 330) Grade D 40 275) 52 (360) Grade E 80 (550) 82 (565) ‘A618 Grades I & I 50 (348) 70 (485) To} in, walls Grade 11 50 (348) 65 (450) ‘A709 Grade 36 36 ‘58-80 To4 All (230) (400-550) (100) Grade 50 30 65 Tos All G45) (asa) (100) Grade SOW 50 0 Tod all Ga) (483) 100) Grade 100 0. 100-130 Over 24 tod. & 100W (620) (690-898) (64-102) Grade 100 100 110-130 To2s &100W (690) (760-895), (64) —— ‘YAM steels listed are approved under the AISC Specifications (1.5, 1.15] except A611 and A709. “ASTM AS [1.8] places structural rolled shapes (W, M, S, HP, C, MC, and L) in Groups 1 through 5 according to size for tensile property classification. All rolled flanged sections having at least fone cross-section dimension 3 in. (7$ mm) or greater are included. The size basis for groups is approximately the web thickness corresponding to the maximum thickness for plates, with the thinnest web sections in Group 1, The specific sections included in each group are given in ASTM A6 [1.8] and in the AISC Manuals [1.7, 1.17]. All SI values are those given in the particular ASTM Specification. [1.8]. Table 2.1.1 lists the common steels, their minimum yield stresses, and tensile strengths, Their common uses are given in Table 2.1.2. ‘Carbon Steels Carbon steels are divided into four categories based on the percentage of carbon: low carbon (less than 0.15%); mild carbon (0.15-0.29%); medium carbon (0.30-0.59%); and high carbon (0.60-1.70%), Structural ear- bon steels are in the mild carbon category; a steel such as A36 has maximum2.1 /STRUCTURAL STEELS 45 TABLE 2.1.2 USES OF VARIOUS STEELS ASTM" designation ‘Common usage ‘A36 Carbon steel General structural purposes; botted and welded, mainly for buildings ‘A33 Carbon Steel ‘Welded and seamless pipe (A242 High-strength ‘Welded and bolted bridge construction where corrosion low-alloy steel resistance is desired; essentially superseded by A709, Grade SOW ‘ASO ‘Cold-formed welded and seamless round, square, rectangular, or special shape ‘Carbon steel structural tubing for bolted and welded general structural purposes ASOL ‘Hot-formed welded and seamless square, rectangular, round, or special shape ‘Carbon steel structural tubing for bolted and welded general structural purposes 'ASLA Alloy steel, quenched Plates in thicknesses ef 6 in, (180 mm) and under, primarily for welded bridges: and tempered largely superseded by A109 for bridges A529) Plates and bars} in, (13-ram) and less in thickness Carbon steet or diameter and Group 1 shapes [1.8] for use in bolted and welded metal building system frames and trusses A570 ‘Hot-rolled sheet and strip cut in lengths or coils; for Carbon steel cold-formed sections [maximum thickness 0.229 in. (6 mm] ‘A372 High-strenath, ‘Structural shapes, plates, sheet pling, and bars for Towsalloy, eolumbiuan or vanadium steel bolted and welded buildings; welded bridges in Grades 42 and SO-only; essentially superseded by A709, Grade $0 A588 Structural shapes, plates, and bars for welded buildings and bridges where weight High-strength savings or added durability are nceded; atmospheric corrosion resistance is low-alloy steel about four times that of A36 stecl: essentially superseded by A709, Grade 50 for bridges ‘A606 Hot and cold-rolled sheet and strip in lengths or coils; for cold-formed sections, Hiigh-strength where enhanced durability is desired; atmospheric corrosion resistance for low-alloy steel ‘Type 2 at least twice that of earbon steel; and for Type at least four times that of carbon steel “A607 ot and cold-rolled sheet and stip in lengths or cols; High-strength low-alley for eold-formed sections, where greater strength and weight savings are important, columbium or ‘atmospheric corrosion resistance (without copper) is the same 2s carbon steel; vanadium steel ‘with coppes, corrosion resistance is twice that of carbon stect AGLL ‘Cold-rolled sheet in cut lengths or coils for making Carbon steel ‘eolé-formed sections ‘A618 Hotformed welded and seamiess square, rectangular, round, or special shape Highestrength structural tubing for bolted and welded general structural purposes: Jow-alloy steel Grade Tl has corrosion resistance about twice that of carbon steel: Grade 1 has corrosion resistance about four times that of carbon steel: ‘ Grade III for enhanced corrosion resistance may have copper specified ‘A709 Carbon: high: ‘Structural shapes, plates, and bars in Grades 36, $0, and SOW for use in bridges: ‘strength, low-alloy: plates in Grades 100 and 100W for use:in bridges: when supplementary and quenched and requirements are used, requirements of A36, AST2, AS#B, and AS14 are tempered alloy ‘exceeded: Grades SOW and 100W are weathering steels es “All steels listed are approved under the AISC Specifications [1.5, 1.15] except A611 and A709.46 2 STEELS AND PROPERTIES ———Tensile strength, F, Heat-treated constructional alloy steels; A514 quenched ‘and tempered allay ste! Minimurn yiald strength Fy = 100ksi High strength, low ‘alloy carbon steels; Ag72 Carbon steels; ASS Stress, kips per square inch 005 = 10s o20 «402 8=©=— 0.30 0.35 Strain, inches per inch Figure 2.1.1 Typical stress-strain curves. carbon varying from 0.25 to 0.29% depending on thickness. Structural carbon steels exhibit definite yield points as shown on curve (a) of Fig. 2.1.1. Increased carbon percent raises the yield stress but reduces ductility, making welding more difficult. The carbon steels given in Table 2.1.1 are A36 [2.1], A53 [2.2], A500 [2.9], ASO [2.10], A529 [2.12], A570 [2.13], A611 [2.18] and A709 [2.20], Grade 36 High-Strength Low-Alloy Steels ‘This category includes steels having yield stresses from 40 to 70 ksi (275 to 480 MPa), exhibiting the well-defined yield point shown in curve (b) of Fig. 2.1.1, the same as shown by carbon steels. The addition to carbon steels of small amounts of alloy elements such as chromium, columbium, copper, manganese, molybdenum, nickel, phosphorus, vanadium, or zirconium, improves some of the mechanical properties. Whereas carbon steels gain their strength by increasing carbon content, the alloy elements create increased strength from a fine rather than coarse microstructure obtained during cooling of the steel. High-strength low-alloy steels are used in the as-rolled or normalized condi- tion; ie, no heat treatment is used. The high-strength low-alloy steels of Table 2.1.1 are A242 [2.3], A441 (2.6, AST2 [2.14], AS88 [2.15], A606 [2.16], AGOT [2.17], A618 [2.19], and A709 (2.20), Grades 50 and SOW.22/FASTENERSTEELS 47 Alloy Steels Low-alloy steels may be quenched and tempered to obtain yield strengths of 80 to 110 ksi (550 to 760 MPa). Yield strength is usually defined as the stress at 0.2% offset strain, since these steels do not exhibit a well-defined yield point. A typical stress-strain curve is shown in Fig. 2.1.1, curve (c). These steels are weldable with proper procedures, and ordinarily require no addi- tional heat treatment after they have been welded. For special uses, stress relieving may occasionally be required. Some carbon steels, such as certain pressure vessel steels, may be quenched and tempered to give yield strengths in the 80 ksi (S50 MPa) range, but most steels of this strength are low-alloy steels. These low-alloy steels generally have a maximum carbon content of about 0.20% in order to limit the hardness of any coarse-grain microstructure (martensite) that may form during heat treating or welding, thus reducing the danger of cracking. ‘The heat treatment consisis of quenching (rapid cooling with water or oil from at least 1650°F (900°C) to about 300-400°F); then tempering by reheating to at least 1150°F (620°C) and allowing to cool. Tempering, even though reducing the strength and hardness somewhat from the quenched material, greatly improves the toughness and ductility. Reduction in strength and hardness with increasing temperature is somewhat counteracted by the occurrence of a secondary hardening, resulting from precipitation of fine columbium, titanium, or vanadium carbides. This precipitation begins at about 950°F (510°C) and accelerates up to about 1250°F (680°C). Tempering at or near 1250°F to get maximum benefit from precipitating carbides may result in entering the transformation zone, thus producing the weaker microstructure that would have been obtained without quenching and tempering. In summary, the quenching produces martensite, a very hard, strong, and brittle microstructure; reheating reduces the strength and hardness somewhat while increasing the toughness and ductility. For more detailed information concerning the metallurgy of the quenching and tempering process, the reader is referred to the Welding Handbook [2.22]. The quenched and tempered alloy steels of Table 2.1.1 are A514 [2.11] and A709 [2.20], Grades 100 and 100W. 2.2 FASTENER STEELS The detailed treatment of the design of threaded fasteners appears in Chapter 4. A brief description of the materials used for bolts appears in the following paragraphs. A307 [2.4], Carbon Steel Bolts and Studs, 60,000 psi Tensile Strength This material is used for what are Commonly referred to as “machine bolts.” ‘These are usually used only for temporary installations. Included are Grade A bolts for general applications, which have a minimum tensile strength of 60 ksi48 2/STEELS AND PROPERTIES (415 MPa); and Grade B bolts for flanged joints in piping systems where one or both flanges are cast iron. The Grade B bolts have a maximwn tensile strength limitation of 100 ksi (700 MPa). No well-defined yield point is exhibited by these bolts, and no minimum yield strength (for instance, 0.2% offset strength) is specified. : A325 [2.5], High-Strength Bolts for Structural Steel Joints This quenched and tempered medium carbon steel is used for bolts commonly known as “high-strength structural bolts,” or high-strength bolts. This mate- rial has maximum carbon of 0.30%. It is heat-treated by quenching and then by reheating (tempering) to a temperature of at least 800°F. This steel behaves in a tension test more similarly to the heat-treated low-alloy steels than to carbon steel, It has an ultimate tensile strength of 105 ksi (733 MPa) (1} to 14 in.-diam bolts) to 120 ksi (838 MPa) (} to 1-in.-diam bolts). Its yield strength, measured at 0.2% offset, is prescribed at 81 ksi (566 MPa) minimum for 1} to 1+-in.-diam bolts, and 92 ksi (643 MPa) for bolts 3 to 1 in. diam (see Table 4.1.1). ‘A449 [2.7], Quenched and Tempered Steel Bolts and Studs ‘These bolts have tensile strengths and yield stresses (strength at 0.2% offset) the same as A325 for bolts 14 in. diam and smaller; however, they have the regular hexagon head and longer thread length of A307 bolts. They are also available in diameters up to 3 in, The AISC Specifications (1.5, 1.15] permit ‘use of A449 bolts only for certain structural joints requiring diameters exceed- ing 14 in. and for high-strength anchor bolts and threaded rods. A490 [2.8], Heat Treated, Steel Structural Bolts, 150 ks! (1035 MPa) Tensile Strength This material has carbon content that may range up to 0.53% for 11 in.-diam bolts, and has alloying clements in amounts similar to the AS14 [2.11] steels. After quenching in oil the material is tempered by reheating to at least 90°F. The minimum yield strength, obtained by 0.2% offset, ranges from 115 ksi (803 MPa) (over 2} in. to 4 in. diam) to 130 ksi (908 MPa) (for 24 in. diam and under). Galvanized High-Strength Bolts In order to provide corrosion protection, A325 bolts may be galvanized. Hot-dip galvanizing requires the molten zinc temperature to be in the range of the heat treatment temperature; thus, the mechanical properties obtained by heat treatment may be diminished. In general, only Type 1 A325 bolts should be hot-dip galvanized. If Type 2 are galvanized the mechanical properties must2.9 / WELD ELECTRODE AND FILLER MATERIAL 49 Tater be checked. Whenever galvanized bolts are used, the nuts must be “oversized,” If the nuts are also galvanized, they must be “double oversized.” Steels having tensile strength in the range of 200 ksi or higher are subject to hydrogen embrittlement when hydrogen is permitted to remain in the steel and high tensile stress is applied. The introduction of hydrogen occurs during the pickling operation of the galvanizing process and the subsequent “sealing- in” of the hydrogen and zinc coating [2.23]. The minimum tensile strength of A325 bolts is well below the critical 200 ksi range. On the other hand, A490 bolts have a minimum tensile strength of 170 ksi, a value considered too close to the critical range. Thus, galvanizing of A490 bolts is not permitted. 2.3 WELD ELECTRODE AND FILLER MATERIAL ‘The detailed treatment of welding and welded connections appears in Chapter 5. The electrodes used in shielded metal arc welding (SMAW) (see Sec. 5.2) also serve as the filler material and are covered by AWS A5.1 and A5.5 Specifications [2.24]. Such consumable electrodes are classified E60XX, EV0XX, E80XX, E90XX, E100XX, and E110XX. The “E” denotes electrode. ELECTRODES USED FOR WELDING® Process Shielded Submerged Gas Metal Flux Cored Matal arc Are Are Arc Welding Welding «Welding = Welding (SMAW) (SAW) (GMAW) = (ECAW) Minimum Minimum AWSAS.1 AWSA5.17 AWSAS.i8 AWS AB.20 yield tensile OrA5.S Or A523 Or AS.2B Or AG. ‘stress strength (ksi) (MPa) (ksi) (MPa) E60XX E6XT-X 50 MS Glmin 425 F6X-EXXX 30 M5 «62-80 425-550 ET0XX ERTSX = ETNT-X 6 415 T2min 495 FIX-EXXX 6 415 70-90 485-655 ES0XX 67460 80min 550 FEN-EXXX ESXT 68 470 80-100 $50-690 ERS0S 65 450 80min 550 E100XX 87 600 100min 690 FLOX-EXXX, $8 60S 100-130 690-895 ‘ERLOOS: 9 620-100 min 690 EIOXT $3 G05 100-120 690-830 Euox 97 67 «10min 760 FLIX-EXXX 98 675 110-130 760-895 ERIS, 98 675 0 min 760 EUXT 98 675-110-125 760-860 “Filler metal requirements given by AWS D1 [2,24], Table 4.1.1 to mateh the various structural stecls,50 2,/ STEELS AND PROPERTIES The first two digits indicate the tensile strength in ksi; thus the tensile strength ranges from 60 to 110 ksi (414 to 760 MPa). The “X's” represent numbers indicating the usage of the electrode. For submerged are welding (SAW) (see Sec. 5.2), the electrodes which also serve as filler material are covered by AWS A5.17 and A5.23, and are designated F6X-EXXX, F7X-EXXX, F8X-EXXX, F9X-EXXX, FIOX-EXXX, and F11X-EXXX. The “F” designates a granular flux material that shields the weld as it is made. The first of the two digits following the “F” indicate the tensile strength (6 means 60 ksi), while the second digit gives the Charpy V-notch impact strength. The “E” and the other X’s represent numbers relating to the use. For gas metal are welding (GMAW) | (see Sec. 5.2) and flux cored are welding (FCAW) (see Sec. 5.2) the electrodes are designated ER70S-X and E7XT-X, respectively. The number 70 (or 7) is the tensile strength in ksi. The yield stresses and tensile strengths of the commonly used electrodes are given in Table 2.3.1. 2.4 STRESS-STRAIN BEHAVIOR (TENSION TEST) AT ATMOSPHE! TEMPERATURES Typical stress-strain curves for tension are shown in Fig. 2.1.1 for the three categories of steel already discussed: carbon, high-strength low-alloy, and heat-treated high-strength low-alloy. The same behavior occurs in compression when support is provided so as to preclude buckling, The portion of each of the stress-strain curves of Fig. 2.1.1 that is utilized in ordinary design is shown enlarged in Fig, 2.4.1. The stress-strain curves of Fig. 2.1.1 are determined using a unit stress obtained by dividing the load by the original cross-sectional area of the specimen, and the strain (inches per inch) is obtained as the elongation divided by the original length. Such curves are known as engineering stress-strain curves and rise to a maximum stress level (known as the tensile strength) and then fall off with increasing strain until they terminate as the specimen breaks. Insofar as the material itself is concerned the unit stress continues to rise until failure occurs. The so-called true-stress/true-strain curve is obtained by using the actual cross-section even after necking down begins and using the instanta- neous incremental strain. The engineering stress-strain curve permits the practical use of the original (before necking down) cross-section. Stress-strain curves (as per Fig. 2.4.1) show a straight line relationship up to a point known as the proportional limit, which essentially coincides with the yield point for most structural steels with yield points not exceeding 65 ksi (450 MPa). For the quenched and tempered low-alloy steels the deviation from a straight line occurs gradually, as in curve (c), Fig. 2.4.1. Since the term yield point is not appropriate to curve (c), yield strengih is used for the stress at an offset strain of 0.2%; or alternatively, a 0.5% extension under load, as shown in2.4 /STRESS-STRAIN BEHAVIOR (TENSION TEST) 51 0.5% Extension under load yield strength, F, = 100 ksi 0.2% Offat yield strength, F, = 100 ksi For F, = 100 ksi; typical for steels F, > 65 (0.2% Offset (0,002 in/in.) For F, = 50 ksi; typical for most structural steels with F, < 65 ksi ‘Stress, kal Strain-hardening range 10 max. tensile strength 0.020 0.028 Figure 2.4.1 Enlarged typical stress-strain curves for different yield stresses. Fig. 2.4.1. Yield stress is the general term to include the unit stress at a yield point, when such exists, or the yield strength. The ratio of stress to strain in the initial straight line region is known as the modulus of elasticity, or Young’s modulus, E, which for structural steels may be taken approximately as 29,000 ksi (200,000 MPa). In the straight-line region loading and unloading results in no permanent deformation; hence it is the elastic range. The service load unit stress in steel design is always intended to be safely within the proportional limit, even though in order to ascertain safety factors against failure or excessive deformation, knowledge is required of the stress-strain behavior up to a strain about 15 to 20 times the maximum elastic strain. For steels exhibiting yield points, as curves (a) and (b) of Fig. 2.4.1, the large strain for which essentially constant stress exists is known as the plastic range. The load and resistance factor design method consciously uses this range. The higher strength steels typified by curve (c) of Fig. 2.4.1 also have a52 ° 2//STEELS AND PROPERTIES region that might be called the plastic range; however, in this zone the stress is continuously increasing as strain increases instead of remaining constant. Thus, this region is not permitted by AISC to be used for steels whose yield stress exceeds 65 ksi. For strains greater than 15 to 20 times the maximum elastic strain the stress again increases but with a much flatter slope than the original elastic slope. This increase in strength is called strain hardening, which continues up to tensile strength. The slope of the stress-strain curve is known as the strain-hardening modulus, E,,. Average values for this modulus and the strain €,, at which it begins have been determined [2.25] for two steels as:, A36 steel, E,, = 900 ksi (6200 MPa) at <,, = 0.014 in. per in.; and for A441, E,, = 700 ksi (4800 MPa) at ,, = 0.021 in. per in. The strain-hardening range is not consciously used in design, but certain of the buckling limitations are conser- vatively derived to preclude buckling even at strains well beyond onset of strain hardening. ‘The stress-strain curve also indicates the ductility. Ductility is defined as the amount of permanent strain (Le., strain exceeding proportional limit) up to the point of fracture. Measurement of ductility is obtained from the tension test by determining the percent elongation (comparing final and original cross-sectional areas) of the specimen. Ductility is important because it permits yielding locally due to high stresses and thus allows the stress distribution to change. Design procedures based on inelastic behavior require large inherent ductility, particularly for treatment of stresses near holes or abrupt change in member shape, as well as for design of connections. 2.5 MATERIAL TOUGHNESS ‘The use of steels having higher strength than A36 without heat treatment has resulted in problems relating to lack of ductility and material fracture [2.27]; at least the use of such steels requires the structural engineer to be more conscious of material behavior. In structural steel design, toughness is a measure of the ability of steel to resist fracture; i.e, to absorb energy. According to Rolfe [2.26], material toughness is defined as “the resistance to unstable crack propagation in the presence of a notch.” Unstable crack propagation produces brittle fracture, as opposed to stable crack growth of a subcritical crack from fatigue. For uniaxial tension, toughness can be expressed as the total area under the stress-strain curve out to the fracture point where the diagram terminates. Since uniaxial tension rarely exists in real structures, a more useful index of toughness is based on the more complex stress condition at the root of a notch, Notch toughness is the measure of the resistance of a metal to the start and propagation of a crack at the base of a standard notch, commonly using the Charpy V-notch test. This test uses a small rectangular simply supported beam having a V-notch at midlength. The bar is fractured by a blow from a swinging25/MATERIAL TOUGHNESS 53 Temperature, °C -30 20-10 0 +10 Ductility transition temperature at 16 ft-lb, ay 17°F Energy, fttbs. esesees -2 0 «#20 +80 #80 Temperature, °F Figure 2.5.1 Transition temperature curve for carbon steel obtained from Charpy ‘V-noich impact tests. (Adapted from Ref. 2.25) pendulum, The amount of energy absorbed is calculated from the height the pendulum raises after breaking the specimen. The amount of energy absorbed will increase with increasing temperature at which the test is conducted. Though the Charpy V-notch test has been a common means of determin- ing notch toughness, other fracture criteria and more recently fracture mechan- ics have been used [2.26]. Barsom and Rolfe [2.28] and Barsom [2.29, 2.30] have excellently presented the important factors relating to fracture of steel. Figure 2.5.1 shows the typical relationship between temperature and the transition from ductile to brittle behavior, such as one may obtain from the ‘Charpy V-notch test. The temperature at the point where the slope is steepest (point A of Fig. 2.5.1) is the transition temperature. Since brittleness and ductility are qualitative terms, the various structural steels have different requirements for ductility at various temperatures depending on their service ‘environment (loading, temperature, stress and strain levels, loading rate, and cyclic loading). For example, a moderate amount of ductility may be required for ordi- nary structures where very low temperatures are not expected; in such cases, 15 ft-lb has commonly been the energy absorption required. The temperature obtained from the test results curve of Fig. 2.5.1 at 15 ft-lb would be about 17°F, known as the ductility transition temperature. This would indicate that the material may be expected to be brittle when service temperatures are below 17°F. 2.6 YIELD STRENGTH FOR MULTIAXIAL STATES OF STRESS Only when the load-carrying member is subject to uniaxial tensile stress can the properties from the tension test be expected io be identical with those of the structural member. It is easy to forget that yielding in a real structure is54 — 2,/STEELS AND PROPERTIES usually mot the well-defined behavior observed in the tension test. Yielding is commonly assumed to be achieved when any one component of stress reaches the uniaxial value F,. For all states of stress other than uniaxial, a definition of yielding is needed, These definitions, and there are frequently several for a given state of stress, are called yield conditions (or theories of failure) and are equations of interaction between the stresses acting. Energy-of-Distortion (Huber - von Mises -Hencky) Yield Criterion This most commonly accepted theory gives the uniaxial yield stress in terms of the three principal stresses. The yield crilerion* may be stated o = 3[(0, — 03)? + (0) — 08)? + (05 — 04)'] (26.1) where 6,, 6,, 6, are the tensile or compressive stresses that act in the three principal directions; i.e., the stresses that act in the three mutually perpendicu- lar planes of zero shear, and o, is the “yield stress” that may be compared with the uniaxial value F,. For most structural design situations, one of the principal stresses is cither zero or small enough to be neglected; hence Eq.2.6.1 reduces to the following for the case of plane stress (all stresses considered are acting in a plane) o2 = 0} + of — 0; (2.6.2) When stresses on thin plates are involved, the principal stress acting transverse to the plane of the plate is usually zero (at least to first-order approximation). Flexural stresses on beams assume zero principal stress perpendicular to the plane of bending, Furthermore, structural shapes (Fig. 1.5.1) are comprised of thin plate elements, so that each is subject to Eq, 2.6.2. The plane stress yield criterion, Eg. 2.6.2, is the one used throughout the remaining chapters where needed, and is illustrated in Fig. 2.6.1. Shear Yield Stress The yield point for pure shear can be determined from a stress-strain curve with shear loading, or if the multiaxial yield criterion is known, that relation- ship can be used. Pure shear occurs on 45° planes to the principal planes when "See Fred B. Seely and James ©. Smith, Advanced Mechanics of Materials, Ind ed. (New York: John Wiley & Sons, Inc., 1952), pp. 76-91.2.8 / YIELD STRENGTH FOR MULTIAXIAL STATES OF STRESS = 55 on reoy 3 = 0 t A t 4—+O+-y = % oO mate, wi ue are shee stress condition a—- F< “8 Hydrostrie-sest condition Figure 2.6.1 Energy-of-distortion yield criterion for plane stress, 0, = —o,, and the shear stress + = 9, Substitution of o, = —o, into Eq. 2.6.2 gives 0? = of + of — o,(—a,) = 307 (2.6.3) a, = 7 = ,/¥3 = shear yield (2.6.4) which indicates that the yield condition for shear stress acting alone is 1, = 4,/V3 = 0.580, (2.6.5) Poisson's Ratio, » When stress is applied in one direction, strains are induced not only in the direction of applied stress but also in the other two mutually perpendicular directions. The usual value of y: used is that obtained from the uniaxial stress condition, where it is the ratio of the transverse strain to longitudinal strain56 2,/ STEELS AND PROPERTIES under load. For structural steels, Poisson’s ratio is approximately 0.3 in the elastic range where the material is compressible and approaches 0.5 when in the plastic range where the material is essentially incompressible (i.¢., constant resistance no matter what the strain). Shear Modulus of Elasticity Loading in pure shear produces a stress-strain curve with a straight line portion whose slope represents the shear modulus of clasticity. If Poisson's ratio w and the tension-compression modulus of elasticity E are known, the shear modulus G is defined by the theory of elasticity as E o= ta (2.6.6) which for structural steel is just over 11,000 ksi (75,800 MPa). 2.7 HIGH TEMPERATURE BEHAVIOR The design of structures to serve under atmospheric temperature rarely in- volves concern about high temperature behavior. Knowledge of such behavior is desirable when specifying welding procedures, and is necessary when concérned with the effects of fire. When temperatures exceed about 200°F (93°C) the stress-strain curve begins to become nonlinear, gradually eliminating the well-defined yield point. The modulus of elasticity, yield strength, and tensile strength all reduce as temperature increases, The range from 800 to 1000°F (430 to 540°C) is where the rate of decrease is maximum. While each steel, because of its different chemistry and microstructure, behaves somewhat differently, the general rela- tionships are shown in Fig. 2.7.1. Steels having relatively high percentages of carbon, such as A36, exhibit “strain aging” in the range 300 to 700°F (150 to 370°C). This is evidenced by a relative rise in yield strength and tensile strength in that range. Tensile strength may rise to about 10% above that at room temperature and yield strength may recover to about its room tempera- ture value when the temperature reaches 500 to 600°F (260 to 320°C), Strain aging results in decreased ductility. ‘The modulus of elasticity decrease is moderate up to 1000°F (540°C); thereafter it decreases rapidly. More importantly, at temperatures above about 500 to 600°F (260 to 320°C), steels exhibit deformation which increases with increasing time under load, a phenomenon known as creep. Creep is well known in concrete structures; and its effect in steel, which does not occur at atmospheric temperatures, increases with increasing temperature, ‘Other high temperature effects are (a) improved notch impact resistance up to about 150 to 200°F (65 to 95°C), as discussed in Sec. 2.5; (b) increased2.7, /HIGH TEMPERATURE BEHAVIOR = 57 et tT ia 1.0 F 2 § a ost aaa Eee os rer a ast BSE ae L gee ast Ln oO 400 800 1200 1600, 2000 Temperature, °F (a) Average Effect of Temperature on Yield Strength Z = 1.0 bee fee os fot cz 5 are °° 5 i 04 aoe ele o2 400, 800 1200 1600 2000 Temperature, °F Mb) Average Effect of Temperature on Tensile Strength 06 ot a2 Ratio of High-Temperature Medulus of Elasticity to That at Reom Temperature 200 400 600 800 1000 Temperature, °F (c) Typical Effect of Temperature on Madulus of Elasticity Figure 2.7.1 Typical effects of high temperature on stress-strain curve properties of structural steels, (Adapted from Ref. 2.25)58 —2/ STEELS AND PROPERTIES brittleness due to metallurgical changes, such as carbide precipitation dis- cussed in Sec. 2.1, begins to occur at about 950°F (510°C); and (c) corrosion resistance of structural steels increases for temperatures up to about 1000°F (540°C). Most steels are used in applications below 1000°F, and some heat treated steels should be kept below about 800°F (430°C). 2.8 COLD WORK AND STRAIN HARDENING After the strain ¢, = F,/E, at first yield has been exceeded appreciably and the specimen is ‘unloaded, reloading may give a stress-strain relationship differing from that observed during the initial loading. Elastic loading and unloading results in no residual strain; however, initial loading beyond the yield point such as to point A of Fig. 2.8.1 results in unloading to a strain at point B, A permanent set OB has occurred. The ductility capacity has been reduced from a strain OF to the strain BF. Reloading exhibits behavior as if the stress-strain origin were at point B; the plastic zone prior to strain hardening is also reduced. ‘When loading has occurred until point C is reached, unloading follows the dashed line to point D; ie., the origin for a new loading is now point D. The length of the line CD is greater, indicating that the yield point has Tensile strength Ideal olastie-plastic stresy-strain relationship Fracture Stress Strain hardening: Figure 2.8.1 Effects of straining beyond the elastic range.2.8/ COLD WORK AND STRAIN HARDENING — 59 ‘Yield paint i ‘Yiold point increase from strain hardening strain Duetility after strain harder and strain aping Figure 2.8.2 Effect of strain aging after straining into strain-hardening range and unloading. increased, The increased yield point is referred to as a strain hardening effect; the ductility remaining when loading from point D is severely reduced from its original value prior to the initial loading. The process of loading beyond the elastic range to cause a change in available ductility, when done at atmo- spheric temperature, is known as cold work. Since real structures are not loaded in uniaxial tension-compression the cold work effect is much more complex and any theoretical study of it is outside the scope of the text. ‘When structural shapes are made by cold-forming from plates at atmo- spheric temperature, inelastic deformations occur at the bends. Cold working into the strain hardening range at the bend locations increases the yield strength, which design specifications may permit taking into account. The Specification for the Design of Cold Formed Steel Structural Members* [1.11] has such provisions. Upon unloading and after a period of time, the steel will have acquired different properties from those represented by points D, C, and F of Fig. 2.8.1 by a phenomenon known as sirain aging. Strain aging, as shown in Fig. 2.8.2, produces an additional increase in yield point, restores a plastic zone of constant stress, and gives a new strain hardening zone at an elevated stress. The original shape of the stress-strain diagram is restored but the ductility is reduced, The new stress-strain diagram may be used as if it were the original for analyzing cold-formed sections, as long as the ductility that remains is sufficient. The corner regions of cold-formed shapes generally would not require high ductility for rotational strain about the axis of the bend. “Referred to henceforth as the AISI Specification,60 © _2/STEELS AND PROPERTIES Stress relieving by annealing will eliminate the effects of cold work should that be desired. Annealing involves heating to a temperature above transfor- mation range and allowing slow cooling; a recrystallization occurs to restore the original properties. Bittence [2.31] provides an excellent summary of the basics of heat treating. 2.9 BRITTLE FRACTURE As has been discussed in several sections, steel that is ordinarily ductile can become brittle under various conditions. Barsom (2.29, 2.30], Barsom and Rolfe [2.28], and Rolfe [2.26] have provided an excellent summary of fracture and fatigue control for structural engineers. Rolfe [2.26] defines brittle fracture as “a type of catastrophic failure that occurs without prior plastic deformation and at extremely high speeds.” Fracture behavior is affected by temperature, loading rate, stress level, flaw size, plate thickness or constraint, joint geometry, and workmanship. Effect of Temperature Notch toughness, as determined by the Charpy impact transition temperature curves (see Sec. 2.5), is an indication of the susceptibility to brittle fracture. Temperature is a vital factor in several ways: (a) the value below which notch toughness is inadequate; (b) in the 600 to 800°F (320 to 430°C) range causes formation of brittle microstructure; and (c) over 1000°F (340°C) causes precipitation of carbides of alloying elements to give more brittle microstruc- ture. The other temperature factors have already been discussed in earlier sections. Effect of Multiaxial Stress The complex stress condition found in usual structures, particularly at joints, is another major factor affecting brittleness. The Primer on Brittle Fracture [2.32] has provided an excellent rational presentation of this and forms the basis for what follows. The engineering stress-strain curve is for uniaxial stress; prior to fracture a necking down occurs, as shown in Fig. 2.9.1a. If biaxial lateral loading as shown in Fig. 2.9.1b could be applied, plastic necking down could be suppressed to the point where the bar would break in a brittle manner without elongation and without reduction in area, The fracture stress based on the unreduced cross-sectional area would be the same high value as that based on the necked-down cross section in the uniaxial tension case. The unit stress would be far above the nominal maximum tensile strength of the engineering stress-strain curve, which is always computed on the basis of original cross section. Also the effects of notches have been alluded to in the discussion of notch toughness in Sec. 2.5. The notch serves somewhat the same purpose as the2.9,/ BRITTLE FRACTURE 61 Biaxial transverse hoadting prevants pecking down ' (a) Threaded test bar (b) Threaded test bar moar fracture pear fracture {uniaxial loading) {triaxial loading) Figure 2.91 Uniaxial and triaxial loading. theoretical triaxial loading of Fig. 2.9.1b, in that it restrains plastic flow which otherwise would occur and thus at some higher stress may likely fail in a brittle manner. Figure 2.9.2 shows the effect of a notch in a tensile test specimen. The cross-sectional area at the base of the notch corresponds to the area of the original specimen of Fig. 2.9.1b. The reduced section tries to become narrower as the axial tension increases, but is resisted by the diagonal pull that develops in the corners, as shown in Fig. 2.9.2, The test bar will fail at high stress by brittle fracture. Notches can occur in real structures by use of unfilleted corners in design or from improperly made welds that may crack. Such occurrences can lead to brittleness. Notches and cracked welds can, however, be minimized by good design and welding procedures. Lt Tees seetional area as bar of Fig, 2.9.1 Lateral restraint Figure 2.9.2 Effect of notch on uniaxial tension test.62 2. / STEELS AND PROPERTIES ‘Unusual configurations and changes in section should be made gradually so the stress flow lines are not required to make abrupt changes. Whenever the complexity is such as to give rise to three-dimensional stresses, the tendency for brittleness increases, Castings, for instance, have the reputation for brittle- ness. Primarily this is because of the built-in three-dimensional continuity. Multiaxial Stress Induced by Welding In general, welding creates a built-in restraint that gives rise to biaxial and triaxial stress and strain conditions, which result in brittle behavior. To illustrate, consider the loaded simply supported beam of Fig. 2.9.3, which in turn supports a plate in tension. Due to flexure, the bottom flange of the beam is in tension; therefore, the stress at point A is uniaxial tension (neglecting the small effects of beam width and attachment of flange to web). Application of the tension plate using angles and bolts puts the flange bolts and the angles essentially in uniaxial tension and the bolt which passes through the suspender plate in shear, so that there is no appreciable effect on the stress at point A. In other words, the stress conditions in the connection of Fig. 2.9.3a are approxi- mately uniaxial in nature. Next, consider the tensile suspender plate welded to the tension flange of the beam, as in Fig, 2.9.3b. The stress at point A is now biaxial because of the direct attachment to the flange at that point. The weld region, therefore, is subject to triaxial stress; biaxial from the directly applied loads, plus the resistance to deformation along the axis of the welds resulting from continu- ous attachment (Poisson's ratio effect). The design of welded joints should (a) Bolted joint (b) Welded joint Figure 2.9.3 Comparison of stress conditions in bolted and welded joints,2.9/BRITTLEFRACTURE 63 consider the possibilities of brittleness due to three-dimensional stressing. The subject of damellar tearing is treated in Sec. 2.10. Effect of Thickness As discussed in Sec. 2.6, if plane stress exists, such as with thin plates where stress in the direction transverse to the plane of the plate may be disregarded, the third dimensional effect is climinated. For thick plates, because of the three-dimensional effects, the Sendeary for brittleness increases, Fro the waterial. apes (ASTM A6 [1.8], Groups 4 and so-called “jumbo shapes,” exhibit low fracture toughness at the core of the thick i flange to web junction and the center of the web adjacent to it, according to Fisher and Pense [2.33]. This low fracture toughness may cause brittle failures when these heavy W shapes are used as tension members. For this reason their use is intended only for compression members [2.34]. Effect of Dynamic Loading The stress-strain properties referred to so far have been for static loading slowly applied. More rapid loading, such as that of forge drop hammers, earthquake, or nuclear blast changes the stress-strain properties, Ordinarily, TABLE 2.9.1 THE ELEMENT OF RISK: FACTORS TO ANALYZE IN ESTIMATING SERIOUSNESS OF BRITTLE FRACTURE (FROM REF. 2.32) 1. What is the minimum anticipated service temperature? The lower the temperature, the greater the susceptibility to brittle fracture. 2. Are tension stresses involved? Brittle fracture occur only under condition of tensile stress, 3, How thick is the material? The thicker the steel, the greater the susceptibility to brittle fracture. 4, Is there three-dimensional continuity? Three-dimensional continuity tends to re- strain the steel from yielding and increases susceptibility to brittle fracture. . Are notches present? The presence of sharp notches increases susceptibility to brittle fracture. |. Are multiaxial stress conditions likely to occur? Multiaxial stresses will tend to restrain yiclding and increase susceptibility to brittle fracture. Is loading applied at a high rate? The higher the rate of loading, the greater susceptibility to brittle fracture, . Is there a changing rate of stress? Britile fracture occurs only under conditions of increasing rate of stress, 9. Is welding involved? Weld cracks can act as severe notches.64 = 2,/ STEELS AND PROPERTIES the increased strain rate from dynamic loading increases the yield point, tensile strength, and ductility. At temperatures about 600°F (320°C) there is a moderate decrease in strength. Some increased brittleness has been noted with high strain rate, but it seems principally associated with other factors already discussed, such as notches where stress concentrations exist and the tempera- ture effect on toughness. The more important factor relating to dynamic load application is not that a rapid increasing strain rate occurs, but that it is combined with a rapid decreasing strain rate. The effect of stress variation is discussed in the section on fatigue. Table 2.9.1, from Ref. 2.32, provides a list of factors “to help determine whether or not the risk of brittle fracture is serious and requires special design considerations.” 2.10 LAMELLAR TEARING Lamellar tearing is a form of brittle fracture occurring “in planes essentially parallel to the rolled surface of a plate under high through thickness loading.” [2.35]. In a highly restrained welded joint “ thru-thickness” strains are induced by weld metal shrinkage. The localized strains due to weld metal shrinkage can be several times larger than yield point strains. Since the stresses due to service loads are well below the yield stress, the strains due to such loads are not believed to initiate or propagate lamellar tears, The subject of lamellar tearing has received considerable attention since the early 1970s, resulting in a tendency for structural engineers to blame lamellar tearing for many brittle fractures. The AISC has provided an excel- lent summary of the phenomenon [2.36]. Thornton [2.37] has provided design and supervision procedures to minimize lamellar tearing. For more detailed treatment, the reader is referred to Kaufman, Pense, and Stout [2.35] and Holby and Smith [2.38]. As a result of the hot rolling operation in manufacture, steel sections have different properties in the direction parallel to rolling (see Fig, 2.10.1), in the transverse direction, and in the “thru-thickness’ direction. In the elastic range, both the rolling and transverse directions exhibit similar behavior, with the elastic limit for the transverse direction being only slightly below that for the rolling direction. The ductility (strain capability), however, in the “thru-thick- ness” direction may be well below that for the rolling direction. Generally, I-shaped steel sections are adequately ductile when loaded either parallel or transverse to the rolling direction. They will deform locally to strains greater than the yield strain (F,/E,), carrying load with some of the material acting at the yield stress and bringing adjacent material into partici- pation if added strength is needed. When, however, the strain is localized for instance in the “thru-thickness” direction at one thick flange of a section, a restrained situation exists because the strain cannot redistribute from the flange through the web to the opposite flange. The large localized “ thru-thick-2.10/LAMELLAATEARING — 65 Transverse ireetion Figure 2.10.1 Definition of direction Z 2 Thru-thigkness direction terminology. (From Ref. 2.36) ness" strain may exceed the yield point strain, causing decohesion and leading to a lamellar tear. Figure 2.10.2 illustrates the relationship of a lamellar tear to a welded joint. The condition of connection restraint is not related to continuity as referred to by structural engineers in the analysis of a statically indeterminate rigid frame. The restraint potentially giving rise to lamellar tearing is internal joint restraint that inhibits the large unit strains resulting from weld shrinkage. Referring to Fig. 2.10.3, when the weld shrinkage occurs in the “thru-thick- © Figure 2.10.2 Joints showing typical lamellar tears resulting from shrinkage of large welds in thick material under high restraint. (From Ref. 2.36)66 © 2,/STEELS AND PROPERTIES Susceptible detat Improved detait Susceptible detail Improved detail © Figure 2.10.3 Susceptibility to a tearing can be reduced by careful detailing of welded connections, (From Ref, 2.36) ness” direction, the material being connected becomes susceptible to lamellar tearing. The weld detail should be made so that weld shrinkage occurs in the rolling direction. References 2.36 and 2.39 suggest ways of avoiding the problem, 2.11 FATIGUE STRENGTH ‘Repeated loading and unloading, primarily in tension, may eventually result in failure even if the yield stress is never exceeded. The term fatigue means failure under cycli¢ loading, It is a progressive failure, the final stage of which is unstable crack propagation. The fatigue strength is governed by three variables; (1) the number of cycles of loading, (2) the range of service load stress (the difference between the maximum and minimum stress), and (3) the initial size of a flaw, A flaw is a discontinuity, such as an extremely small crack. In welded assemblies, a flaw could be the “notch” intersection of two elements or a “discontinuity” such as a bolt hole. Flaws may be the result of2.12,/ CORROSION RESISTANCE AND WEATHERING STEELS 67 poorly made welds, rough edges resulting from shearing, punching, or flame cutting, or small holes. Such flaws may be of no concern; however, under many cycles of loading the flaw (notch effect) may give rise to a crack that increases in length with each cycle of load and reduces the section carrying the load, consequently increasing the stress intensity on the uncracked part. The fatigue strength is more dependent on the localized state of stress than is the static strength. Fatigue is always a service load consideration; the actual service load state of stress is what determines crack propagation. ‘The minimum stress and the grade of steel have no apparent affect on the number of cycles to failure. On the other hand, the specimen geometry, including the surface condition and internal soundness of the weld, have a significant effect. These factors are reflected in the Structural Welding Code [2.24] rules for welded structure design. Recent work by Zuraski and Johnson [2.40] evaluating the remaining life in steel bridges has shown that under certain conditions repeated stressing in steel sections can actually increase their fatigue life. This phenomenon, known as coaxing, was first studied by Sinclair [2.41] and results from repeatedly stressing near but below the fatigue limit and gradually increasing the stress. The AISC Specifications [1.5, 1.15] in Appendix K4 prescribe no fatigue effect for fewer than 20,000 cycles, which is approximately two applications a day for 25 years. Since most loadings in buildings are in that category, fatigue is generally not considered. The exceptions are crane runway girders and structures supporting machinery. Fatigue is always considered in the design of highway bridges, which are expected to have in excess of 100,000 cycles of loading. Volume 1 of the Welding Handbook (2.22, p. 402] shows several good examples of the fatigue relationships for welded plate girders and cover-plated ‘beams. Extended discussion of fatigue is given by Barsom and Rolfe [2.28]. 2.12 CORROSION RESISTANCE AND WEATHERING STEELS Since the earliest uses of steel, one of the important drawbacks was that painting was required to prevent the deterioration of the metal by corrosion (rusting). The lower-strength carbon steels were inexpensive but very vulnera- ble to corrosion, Corrosion resistance may be improved by the addition of copper as an alloy clement. However, copper-bearing carbon steel is too expensive for general use, High-strength low-alloy steels have several times [2.43] the corrosion resistance of structural carbon steel, with or without the addition of copper, as shown in Fig. 2.12.1. The high-strength low-alloy steels do not pit as severely as carbon steels and the rust that forms becomes a protective coating to prevent further deterioration. With certain alloy elements the high-strength low-alloy steel will develop an oxide protective coating that is pleasing in68 2 STEELS AND PROPERTIES Structural carbon stee! § 2 ti Suuctrl eran tl a ‘ with copper as ae Cr-Si-Cu-P steel ge (high-strength, lowslloy) 12345678910 Time, years Figure 2.12.1 Comparative corrosion of steels in an industrial atmosphere. Shaded areas indicate range for individual specimens. (Adapted from Ref. 2.43) appearance and is described as follows*: “It is a very dense corrosion—actu- ally a deeply colored brown, red, purple... . It has a texture and color which cannot be reproduced artificially—a character only nature can give, as with stone, marble, and granite.” When steels are to be unpainted and left exposed they are called weathering steels, ‘As might be expected, the corrosion properties of any steel, including the weathering steels, are dependent on the chemical composition, the degree of pollution in the atmosphere, and the frequency of wetting and drying of the steel. Since its first major use in 1958, for the Administrative Center for Deere & Company jin Moline, Illinois, the use of weathering steel has received considerable attention. At first such steels were specified under ASTM A242 which as previously discussed is very general, allowing a wide variation in chemistry, With the adoption of A388 steel in 1969, and A709 in 1975, A242 is now essentially obsolete. A588_is_gencrally used eathering steel in-buildings and_ A709 Grades S0W and 100W_ for weath teel in bridges (see Table 2.1.2). _ Fabrication and erection of weathering steel requires care. Unsightly gouges, scratches, and dents should be avoided. Painting, even for identifica- tion, should be minimized, since all marks must be removed after the erection is completed. Scale and discoloration from welding also must be removed. The extra expense resulting from fabrication and erection is offset by the elimina- tion of painting at intervals during the life of the structure. * Architectural Record, August 1962.SELECTED REFERENCES 69° ‘The practice of using weathering steels, including the results of 30 years experience, has been summarized by Coburn [2.42], who presents the following, “rules”: 1. For optimum performance in the unpainted condition the structure should be boldly exposed to the elements. 2. The development of the protective oxide film is best achieved under normal exposure wherein. the surfaces are wet at night by dew forma- tion and dry during daylight hours. 3. Because this wet-dry cycle cannot occur when the steel is buried in the soil or immersed in water, regardless of its type, the protective oxide will not form and the performance will-resemble that_of mild, carbon, steel exposed to the same conditions. SELECTED REFERENCES eee 21, 22, 23. 24. 25. 26. 27. 2.8, 29, 2n. ASTM. Specification for Structural Steel (A36-8%c). Philadelphia, PA: American Society for Testing and Materials, 1988. ASTM. Specification for Pipe, Steel, Black and Hot-Dipped, Zinc~Coated Welded and Seamless (A53-88a). Philadelphia, PA: American Society for Testing and Materials, 1988. ASTM. Specification for High-Sirength Low-Alloy Structural Steel (A242-88). Philadelphia, PA: American Society for Testing and Materials, 1988. ASTM. Specification for Carbon Steel Bolts and Studs, 60000 psi Tensile Strength (A307-88a). Philadelphia, PA: American Society for Testing and Materials, 1988. ASTM, Specification for High-Strength Bolts for Structural Steel Joints (A325-86). Philadelphia, PA: American Society for Testing and Materials, 1986. ASTM. Specification for High-Strength Low-Alloy Structural Manganese Vana- dium Steel (A441-85), Philadelphia, PA: American Society for Testing and Materials, 1985, ASTM. Specification for Quenched and Tempered Steel Bolts and Studs (A449-86). Philadelphia, PA: American Society for Testing and Materials, 1986, ASTM. Specification for Heat-Treated Steel Structural Bolts, 150 ksi Minimum Tensile Strength (A490-85); also Specification for High-Strength Steel Bolts, Classes 10.9 and 10.9.3, for Structural Steel Joints. [Metric] (A490M-85). Philadelphia, PA: American Society for Testing and Materials, 1985. ASTM. Specification for Cold-Formed Welded and Seamless Carbon Steel Struc- tural Tubing in Rownds and Shapes (AS00-84). Philadelphia, PA: American Society for Testing and Materials, 1984. |. ASTM. Specification for Hot-Formed Welded and Seamless Carbon Steel Struc- tural Tubing (AS01-88). Philadelphia, PA: American Society for Testing and Materials, 1988. ASTM.-Specification for High-Yield-Sirengih, Quenched and Tempered Alloy Steel Plate, Suitable for Welding (AS14-88). Philadelphia, PA: American Society for Testing and Materials, 1988. 2.12. ASTM, Specification for Structural Steel With 42 ksi [290 MPa] Minimum Yield Point (I /2 in. [13 mm] Maximum Thickness) (A5S29-88), Philadelphia, PA: ‘American Society for Testing and Materials, 1988.70 2 STEELS AND PROPERTIES 213. ASTM. Specification for Steel, Sheet and Strip, Carbon, Hot-Rolled, Structural Quality (A570-88). Philadelphia, PA: American Society for Testing and Materi- als, 1988, 2.14, ASTM, Specification for High-Strength Low-Alloy Columbiven-Vanadium Steels of Structural Quality (A572-88c). Philadelphia, PA: American Society for Testing and Materials, 1988, 2.15. ASTM. Specification for High-Strength Low-Allay Structural Steel with 50 ksi [345 MPa} Minimum Yield Point to 4 in. [100 mm] Thick (AS88-88a). Philadelphia, PA: American Society for Testing and Materials, 1988. 2.16. ASTM. Specification for Steel, Sheet and Strip, High-Strength, Low-Alloy, Hot- Rolled and Cold-Rolled, with Improved Atmospheric Corrasion Resistance (A606- $85). Philadelphia, PA: American Society for Testing and Materials, 1985. 2.17. ASTM. Specification for Steel, Sheet and Strip, High-Strength, Low-Alloy, Columbium or Vanadium, or Both, Hot-Rolled and Cold-Rolled (607-85). Philadelphia, PA: American Society for Testing and Materials, 1985. 2.18, ASTM. Specification for Steel, Sheet, Carbon, Cold-Rolled, Structural Quality (AG11-85). Philadelphia, PA: American Society for Testing and Materials, 1985. 2.19, ASTM. Specification for Hot-Formed Welded and Seamless High-Strength Low- Alloy Sirctural Tubing (A618-88), Philadelphia, PA: American Society for Test- ‘ing and Materials, 1988. 2.20, ASTM. Specification for Structural Steel for Bridges (A709-88a). Philadelphia, PA; American Society for Testing and Materials, 1988, 2.21, ASTM. Specification for Steel Sheet and Strip, High-Strength, Low-Alloy, Hot- Rolled, and Steel Sheet, Cold-Rolled, High-Strength, Low-Allay, with Improved Formability (ATLS-88). Philadelphia, PA: American Society for Testing and ‘Materials, 1983, 2.22, AWS. Welding Handbook, 8th ed., Vol. 1, Welding Technology, Miami, FL: ‘American Welding Society (550 N.W. LeJeune Road, P.O. Box 351040, Miami, FL 33135), 1987, 2.23, Research Council on Structural Connections. Commentary on Specifications for Structural Joints Using ASTM A325 or A490 Bolts. Chicago, IL: American Institute of Steel Construction, November 13, 1985. 2.24. AWS. Structural Welding Code—Steel, 11th ed., Effective January 1, 1988 (ANSI/AWS DI1.1-88). Miami, FL: American Welding Society, 1988. 2.25, R. L. Brockenbrough and B. G. Johnston. Stee! Design Manual, Pittsburgh, PA: United States Steel Corporation, 1968, Chap. 1. 2.26, S. T. Rolfe. “Fracture and Fatigue Control in Steel Structures,” Engineering Journal, AISC, 14, 1 (Ist Quarter 1977), 2-15, 2.27. K. A. Godfrey, Ir. “High Strength Steel: Crisis or No?”, Cini Engineering, May 1985, 50-53. 2.28. John M, Barsom and Stanley T. Rolfe, Fracture and Fatigue Control in Structures —Applications of Fracture Mechanics, 2nd ed. Englewood Cliffs, NJ: Prentice- Hall, Inc., 1987. 2.29, John M, Barsom, “Material Considerations in Structural Steel Design,” Proceed- ings, National Engineering Conference & Conference of Operating Personnel. Chicago, IL: American Institute of Steel Construction, April 29-May 2, 1987, 1-1 through 1-15. 2.30, J. M. Barsom, “Material Considerations in Structural Steel Design,” Engineering Journal, AISC, 24, 3 (3rd Quarter 1987), 127-139.231. 2.32, 2.33, 2.34, 2.35. 2.37. 2.38, 2.39. 2.40. 241. 2.42, SELECTED REFERENCES = 71 John ©. Bittence. “The Basics of Heat Treating —What It Does—How It Works —Where to Specify It,” Machine Design, January 24, 1974, 106-111; February 7, 1974, 117-121. A Primer on Brittle Fracture, Booklet 1960-A, Steel Design File, Bethlehem Steel Corporation, Bethlehem, PA, John W. Fisher and Alan W. Pense, “Experience with Use of Heavy W Shapes in Tension,” Engineering Journal, AISC, 24, 2 (2nd Quarter 1987), 63-77. “The Use of Jumbo Shapes in Non-Column Applications,” Engineering Journal, AISC, 23, 3 Grd Quarter 1986), 96. E. J. Kaufman, A. W. Pense, and R. D. Stout. “An Evaluation of Factors Significant to Lamellar Tearing,” Welding Journal, 60, March 1981, Research. Supplement, 43s~49s. . “Commentary on Highly Restrained Welded Connections,” Engineering Journal, AISC, 10, 3 Grd Quarter 1973), 61-73. Charles H. Thornton. “Quality Control in Design and Supervision Can Eliminate Lamellar Tearing,” Engineering Journal, AISC, 10, 4 (4th Quarter 1973), 112-116. E. Holby and J. F.-Smith, “Lamellar Tearing—The Problem Nobody Secms to Want to Talk About,” Welding Journal, 59, February 1980, 37-44. “Causes and Prevention of Lamellar Tearing,” Civil Engineering, April, 1982, 4-75. P, ID, Zuraski and J, E, Johnson, “Research on the Remaining Life in Steel Bridges," Proceedings, ASCE Speciality Conference on Probabilistic Mechanics and Structural Reliability, Berkeley, CA, January 11-13, 1984, New York: Ameri- can Society of Civil Engineers, 1985, 414-418, G. M. Sinclair. “An Investigation of the Coaxing Effect in Fatigue of Metals,” Proceedings, ASTM, 52 (1952), 743-758. ‘Seymour Cobum, “Theory and Practice in Use of Weathering Stecls,” Proceed- ings, National Engineering Conference & Conference of Operating Personnel. ‘Chicago, IL: American Institute of Steel Construction, April 29-May 2, 1987, 14-1 through 14-25, .C. P, Larrabee, “Corrosion Resistance of High-Strength Low-Alloy Stecls as Influenced by Composition and Environment,” Corrosion, 9, August 1953, 259-271,Chapter 3 Tension Members 3.1 INTRODUCTION Tension members are encountered in most steel structures. They occur as principal structural members in bridge and roof trusses, in truss structures such as transmission towers and wind bracing systems in multistoried build- ings. They frequently appear as secondary members, being used as tie rods to stiffen a trussed floor system or to provide intermediate support for a wall girt system. Tension members may consist of a single structural shape or they may be built up from a number of structural shapes. The cross-sections of some typical tension members are shown in Fig. 3.1.1. In general, the use of single structural shapes is more economical than the built-up sections. However, built-up members may.be required when (a) the tensile capacity of a single rolled section is not sufficient, (b) the slenderness ratio (the ratio of the unbraced length Z to the minimum radius of gyration r) does not provide sufficient rigidity, (c) the effect of bending combined with the tensile behavior requires a larger lateral stiffness, (d) unusual connection details require a particular cross-section, or (e) esthetics control, 3.2 NOMINAL STRENGTH The strength of a tension member may be described in terms of the “limit states” that govern. The controlling strength limit state for a tension member will be either (a) yielding of the gross cross-section of the member away from the connection, or (b) fracture of the effective net area (i.c., through the holes) at the connection [3.1]. 723.2/NOMINAL STRENGTH = 73 Structural steel framework at intermediate floor level. This level suspended from roof space truss system with tension sods, Photo by C. G. Salmon) . — fe. JIL al pound Pat bar angie cr fo Double ae Starred noe Section Channa Double Waseetion (Amerenn chen! channels (widefange) —Scancare al | reo Led Built-up box sections Figure 3.1.1 Cross-section of typical tension members,74 3/ TENSION MEMBERS (a) Elastic stresses {b) Ultimate condition Figure 3.2.1° Stress distribution with holes present. When the limit state is general yielding of the gross section over the member length, as for a tension member without holes (ie, with welded connections), the nominal strength 7, may be expressed T= BA, (32.1) where, = yield stress A, = gross cross-sectional area For tension members having holes, such as for rivets or bolts, the reduced cross-section is referred to as the met area. Holes in a member cause stress concentrations (nonuniform stresses); for example, a hole in a plate will give rise to a stress distribution at service load as shown in Fig. 3.2.la. Theory of elasticity shows that tensile stress adjacent to a hole will be about three times the average stress on the net area: However, as each fiber reaches yield strain ¢, = F,/E,, its stress then becomes a constant F, with deformation continu- ing with increasing load until finally all fibers have achieved or exceeded the strain ¢, (Fig. 3:2.1b). When the limit state is a localized yielding resulting in a fracture through the effective net area of a tension member having holes, the nominal strength T,, may be expressed T,= EA, (3.2.2) where F, = tensile strength (see Fig. 2.1.1) A, = effective net area = UA,, (see Secs. 3.4 and 3.5) net area U = efficiency factor (Table 3.5.1). Because of the rise in resistance when the tensile strain becomes large, known as strain hardening (see Sec. 2.4), the actual strength of a ductile tension member may exceed that indicated by Eq, 3.2.1 [3.1], However, at such large elongation due to general yielding along the entire Jength of the member, the ends of the member may move apart too far and cause distress to the structure; thus, the member no longer serves its intended purpose. Either unrestrained yielding or fracture through the reduced section at holes may limit the structural usefulness of the member, Traditionally, a higher margin of