“STRUCTURAL : PSEA 1ANDBOOK THIRD EDITION EDITED BY i EDWIN H. GAYLORD, JR. l CHARLES N. GAYLORD 6 MEd a a aaa a , Vike er aStructural Engineering Handbook | |Other McGraw-! Handbooks of Interest Avallone and Baumeister » MARKS’ STANDARD HANDBOOK FOR MECHANICAL ENGINEERS Brady and Clauser » MATERIALS HANDBOOK Brater » HANDBOOK OF HYDRAULICS Callender « TIME-SAVER STANDARDS FOR ARCHITECTURAL DESIGN DATA Crocker and King + PIPING HANDBOOK Croft and Summers + AMERICAN ELECTRICIANS’ HANDBOOK Davis and Sorensen + HANDBOOK OF APPLIED HYDRAULICS DeChiara and Callender « TIME-SAVER STANDARDS FOR BUILDIN YPES Faherty and Williamson + WooD ENGINEERING AND CONSTRUCTION HANDBOOK Fink and Beaty « STANDARD HANDBOOK FOR ELECTRICAL ENGINEERS Gieck + ENGINEERING FORMULAS Harris » HANDBOOK OF NOISE CONTROL Harris and Crede » SHOCK AND VIBRATION HANDBOOK Hicks « STANDARD HANDBOOK OF ENGINEERING CALCULATIONS Higgins » HANDBOOK OF CONSTRUCTION ENGINEERING MAINTENANCE Higgins « MAINTENANCE ENGINEERING HANDBOOK Manas + NATIONAL PLUMBING CODE HANDBOOK McPartland + McGRAW-HILL’S NATIONAL ELECTRICAL CODE HANDBOOK Merritt + BUILDING DESIGN AND CONSTRUCTION HANDBOOK Merritt « STANDARD HANDBOOK FOR CIVIL ENGINEERS Merritt » STRUCTURAL STEEL DESIGNER'S HANDBOOK O'Brien +» CONTRACTOR'S MANAGEMENT HANDBOOK Parmley + FIELD ENGINEER'S MANUAL Parmley + STANDARD HANDBOOK OF FASTENING AND JOINING Rosaler and Rice » STANDARD HANDBOOK OF PLANT ENGINEERING. Rossnagel, MacDonald, and Higgins « HANDBOOK OF RIGGING Smeaton + SWITCHGEAR AND CONTROL HANDBOOK Tuma + ENGINEERING MATHEMATICS HANDBOOK Tuma + HANDBOOK OF NUMERICAL CALCULATIONS IN ENGINEERING For more information about other McGraw-Hill materials, call 1-800-2-MCGRAW in the United States. In other countries, call your nearest McGraw-Hill office.Structural Engineering Handbook Edited by Edwin H. Gaylord, Jr. Professor of Civil Engineering, Emeritus University of Illinois at Urbana-Champaign Charles N. Gaylord Late Professor of Civil Engineering, Emeritus University of Virginia Third Edition McGraw-Hill Publishing Company New York San Francisco Washington, 0.C. Auckland Bogota Caracas Lisbon London Madrid Mexico City Milan Montreal New Delhi San Juan SAo Paulo Singapore Sydney Tokyo TorontoLibrary of Congress Cataloging-in-Publication Data Structural engineering handbook | edited by Edwin H. Gaylord, Jr, Charles N. Gaylord. —3rd ed. Includes bibliographies and index ISBN 0.07-023188-5 1. Structural engineering—-Handbooks, manuals, ete. 1 Gaylord, Edwin Henry. II. Gaylord, Charles N. ‘TA635.77 1990 624.1—de20 89.2577 CIP Copyright © 1990 by McGraw-Hill, Ine. All rights reserved. Printed in the United States of America, Except as permitted under the United States Copyright Act of 1976, no part of this publication may be reproduced or distributed in any form or by any means, or stored in a data base or retrieval system, without the prior written permission of the publisher. 567890 DOCDOC 9876543 ISBN O-07-023188-5 ‘The editors for this book were Harold B. Crawford and Susan Thomas, the designer was Naomi Auerbach, and the production supervisor was Suzanne W, Babeuf, It was set in Caledonia by University Grap! Ine. Printed and bound by the R. R. Donnelley & Sons Company. Information contained in this work has been obtained by McGraw-Hill, Inc., from sources believed to be reliable. However, neither McGraw-Hill nor its authors guarantees the accuracy or completeness of any information published herein and neither McGraw-Hill nor its authors shall be responsible for any errors, “omissions, or damages arising out of use of this information. This work is published with the understanding that McGraw-Hill and its authors are supplying information but are not attempting to render engineering or other professional services. If such services are required, the assistance of an appropriate professional should be sought.Contents Preface xv Contributors — xvii Section 1 STRUCTURAL ANALYSIS Part 1, Fundamentals and Analysis of Elastic Trusses and Frames David A. Pecknold Introduction: Classification of Structures; Fundamentals of Analysis: Deserip- tion of Structural Configuration—Basic Principles—Static Determinacy and Sta- bility—Superposition—Symmetry; Energy Principles: Principle of Virtual Work—Principle of Minimum Potential Energy—Principle of Minimum Com- plementary Potential Energy—The Reciprocal Theorem; Analysis of Statically Determinate Structures: Forces: Plane and Space Trusses—Beams and Frames—Beam Dellections; Analysis of Statically Indeterminate Structures: Comparison of Force and Displacement Methods; Force Method: Structure Flexibility Coefficients—Dummy Unit-Load Method —Examples—Three- Moment Equation for Continuous Beams; Displacement Method: Structure Stiff- ness Coefficients—Member Stiffness Matrices—Member Fixed-End Forces— Portal Frame by Displacement Method—Slope-Deflection Equations-—-Moment Distribution—Matrix Formulation of the Displacement Method—Frame Ele- ment Stiffness Properties and Equivalent Loads Using Virtual Work Part 2. Continua: The Finite- ment Method William C. Schnobrich Finite-Element Method: Requirements for Finite-Element Analy sis—Discreti- zation of the Structure—Guidelines for Selection of Grid—Element Models— Plane-Stress and Plain-Strain Elements—Plane-Stress Analysis—Beam and Plate Bending—Shells and Combined Direct Stress and Bending—Three-Dimen- sional Problems—Element Stiffness Matrix and Nodal Force Vector—Equilib- rium Equations for the Assemblage—Solution for the Displacements—Solution for Element Strains and Stress. Section 2 COMPUTER APPLICATIONS IN STRUCTURAL ENGINEERING Steven J. Fenves The Structural Design Process—The Program-Development Process; Program- Development Tools: Structured Programming—Programming Languages— Flowcharting—Tools for Organizing Programs—Tools for Organizing Data— 11 11 1-46 24vi Contents Generic Software—Computer Hardware; Interaction with Computers: Modes of Interaction—Media of Interaction—Level of Interaction; Program Types and Sources: Program Types—Sources of Programs; Typical Applications: Analy- sis—Proportioning—Detailing—Preparation of Final Documents Section 3 EARTHQUAKE-RESISTANT DESIGN W. J. Hall Response of Simple Structures to Earthquake Motions—Karthquake Motions: Response Spectra: Elastic Systems——Design Response Spectra—Response Spec- tra for Inelastic Systems—Multi-Degree-of-Freedom Systems; Computation of Period of Vibration: Fundamental Mode—Higher Modes—Modal Participation Factors—Spring Constant for Equivalent Shear Beam; Design: General Consid- erations—Effects of Design on Behavior—Design Lateral Forces~—Seismic Forces for Overturning Moments and Shear Distribution—Damping—Gravity Loads—Vertical and Horizontal Excitation—-Unsymmetrical Structures in Tor- sion—Curtain-Wall Buildings—Core Walls—Parts of Buildings Effects of Soi Conditions—Detailing and Quality Control—Cost—Lifelines Section 4 FATIGUE, BRITTLE FRACTURE AND LAMELLAR TEARING W. H. Munse Fatigue of Structural Steel: Significance of Fatigue—Fatigue of Structural Steels—Factors Affecting Fatigue Strength—Structural Members—Riveted Connections—Bolted — Connections—Welded — Connections—Design Fatigue—Protecting Against Fatigue; Brittle Fracture of Structural Steel: Sig- nificance of Brittle Fractures—Initiation, Propagation, and Arrest—Factors Affecting Brittle Fracture—Fabrication—Inspection—Lamellar Tearing Section5 SOIL MECHANICS H. G. Larew Nature and Causes of Soil Deformation: Settlement—Frost Movements— Shrinkage—Subsidence—Soil_Deformation—Time-dependent Deformation; Strength Properties of Soils: Compressive Strength—Fffect of Confining Pres. sures—Transient and Repeated Loads; Factors Affecting Bearing Pressure: Allowable Bearing Pressure—Permissible Settlement—Elastic-plastic Defor- mation—Shear Failures—Consolidation—Time Rate of Settlement; Cuts and Embankments: Slope Stability; Compaction and Permeability: Compaction— Permeability; Appendix Section 6 SOILEXPLORATION Thomas H. Thornburn (deceased) and James W. Mahar General Foundation Conditions: Sources of Information—-Glacial Materials— Wind-laid Materials—Water-laid Materials—Organic Soils—Residual Soils— Frozen Soils; Exploratory Procedures: Test Pits—Borings—Sampling—Bore- hole Cameras—Boring Reports—Standard Penetration Tests—In Situ Strength Tests—Correlations with Surficial Geology—Geophysical Surveys—Load Tests—Special Observations—Preliminary Site Inspection—Fact-finding Sur- vey—Borings 31 5-1 6-1Contents Section 7 RETAINING STRUCTURES AND FOUNDATIONS Herbert O. Ireland Earth Pressures: Stresses in Earth Mass—Rankine’s Theory—Coulomb’s The- ory—Equivalent-Fluid Method—Trial-Wedge Method; Retaining Walls: Types, and Behavior—Determining Earth Pressure—Bases on Piles or Pi Bear- ing Capacity—Stability—Drainage—Other Considerations—Gravity Walls— Semi-gravity Walls—Cantilever | Walls—Counterfort Walls—Joints; Abut- ments; Bulkheads: Forces on Bulkheads—Penetration of Piles—Anchorage: Footing Foundations: Footings on Clay—Footings on Sand—Footings on Silt and Loess; Raft Foundations: Raft on Clay—Rafts on Sand; Pier Foundations: Open Excavations—Drilled Piers—Piers on Clay—Piers on Sand—Caisson Foundations—Foundation Requirements; Pile Found: ile-driving Equip- ment—Pile-driving Formulas—Pile Tests—Piles in Sand—Piles in Clay—Set- tlement of Pile Foundations—Laterally Loaded Piles—Batter Piles—Lateral Stability of Poles—Guy Anchors—Foundations Subjected to Uplift-—Improve- ment of Subsoil—Construction Problems Section 8 DESIGN OF STEEL STRUCTURAL MEMBERS: Lee C, Lim, William J. LeMessurier, and Hans William Hagen Types of Steel—Shapes; Tension Members: Concentrically Loaded Tension Members—Threaded Members—Member Types and Selection—Truss Mem- bers; Compression Members: Column Strength—Concentrically Loaded Col- umns—Effective Length—Amplification Factors and Frame Stability—Propor- tioning—Local Buckling—Lacing and Perforated Cover Plates—Tapered Columns—Slender Compression Elements; Beams: Allowable Stresses—Biaxial Bending—Shear—Deflection—Combined Bending and Compression; Plate Girders: Web—Flanges—Lengths of Flange Plates—Lateral Buckling— Requirements for Stiffeners—Combined Bending and Shear; Welded Connec- tions: Welding Processes—Weld Classification—Weldability—Flectrodes— Inspection—Fillet-welded — Joints—Groove-welded__ Joints—Concentrically Loaded Connections—Beam Seat Connections—Stiffened Beam-Seat Connec- tions—Framed Beam Connections—Moment-resistant Beam Connections; Riv- eted and Bolted Connections: Rivets—High-Strength Bolts—Installation of High-Strength Bolts—Inspection of High-Strength Bolts—Unfinished Bolts— Turned Bolts—Ribbed Bolts—Bearing Bolts—High-Strength Tension Control and Tension Set Bolts—Shear Connections—Eccentrically Loaded Connec- tions, Fasteners in Shear—Eccentrically Loaded Connections, Fasteners in Ten- sion——Flexible Beam-Seat Connections—Stiffened Beam-Seat Connections— Framed Beam Connections—Moment-resistant Beam Connections—Pinned Connections; Bearing Plates and Splices: Beam Bearing Plates—Column Bases—Compression-Member Splices Section9 PLASTIC DESIGN OF STEEL FRAMES Lynn S. Beedle and T. V. Galambos Inelastic Bending—Indeterminate Structures; Analysis: Theorems—Statical Method—Mechanisim Method—Moment Check—Instantaneous Center—Dis- tributed Loads; Deflection Analysis: Deflection at Ultimate Load—Deflection at Working Load; Design Requirements: Specifications—Loads and Forces—Gen- eral Design Procedure—Preliminary Design—Analysis; Secondary Design Con- siderations: Axial Force—Lateral Bracing—Local Buckling—Shear—Frame Instability; Connections: Corner Connections—Interior Beam-to-Column Connections vil 71 8-1vii Contents Section 10 DESIGN OF COLD-FORMED STEEL STRUCTURAL MEMBERS Roger A. LaBoube Materials—Shapes and Uses; Design: Section Properties—Thin Compression Elements—Uniformly Compressed Elements—Elements with Stress Gradi- ent—Elements with an Edge Stiffener—Elements with an Intermediate Stiffener—Members—Compression Members—Flexural _ Members—Beam Webs—Combined Bending and Axial Compression—Connections—Wall Studs—Ellects of Cold Forming on Steel Properties—Tests for Special Cases— Industrial Storage Racks—Shear Diaphragms—Folded-Plate and Shell Roofs Section 11 DESIGN OF ALUMINUM STRUCTURAL MEMBERS. John W. Clark Materials and Specifications: Shapes—Codes and Specifications—Applications and Alloys; Design of Tension Members: Yielding and Fracture—Welded Ten- sion Members; Design of Compression Members: Column Formulas—Lacing— Local Buckling of Plates, Legs, and Flanges in Edge Compression— Flat Plates in Edge Compression—Local Buckling of Round Tubes in m- pression—Welded Compression Members; Beams: Yielding and Ultimate Strength—Lateral-Torsional Buckling—Local Buckling of Beams—Shear Strength of Beam Webs; Plate Girders: Lateral Buckling—Design of Web— Vertical Stiffeners—Longitudinal Stiffeners—Bearing Stiffeners; Combined Loading: Combined Bending and Axial Load—Plate and Tube Buckling under Combined Stress; Connections: Riveted Connections—Bolted Connections— Welded Connections; Design for Repeated Loads Section 12 DESIGN OF REINFORCED CONCRETE STRUCTURAL MEMBERS William L. Gamble Concrete—Reinforcement—Specifications, Codes, and Standards—Strength Design and Working-Stress Design—ACI Load and Reduction Factors—Preci- sion—Rectangular Beams—Continuity—Doubly Reinforced Beams—Tee Beams—Special Beam Shapes—Shear and Diagonal Tension—Development and Anchorage of Reinforcement—Splices—Bar Cutoffs and Bend Points — Deflection—Column Design—Combined Compression and Bending—Column Splices—Columns With Biaxial Bending—Stairs—Wall Footings—Column Footings—Walls—Slabs—Structural Framing Systems Section 13 DESIGN OF PRESTRESSED-CONCRETE STRUCTURAL MEMBERS T. Y. Lin and Paul Zia Notation; Materials: Concrete—Steel—Grouting; Methods and Systems of Prestressing: Tensioning Methods—Pretensioning—Posttensioning Systems: Loss of Prestress: Elastic Shortening of Concrete—Creep—Shrinkage—Relax- ation in Steel—Slippage of Tendons during Anchoring—Friction—Effective Prestress—Elongation of Tendons; Analysis for Flexure: Basic Concepts— Stress in Steel—Cracking Moment—Ultimate Moment—Composite Sections; Design for Flexure: Preliminary Design—Elastic Design—Ultimate Design— Balanced-Load Design—Deflections; Shear, Bond, and Bearing: Principal Ten- sion—Web Reinforcement—Prestress Transfer Bond—Anchorage; Typical Sec- tions: Boam Sections—Span-Depth Ratios—Cable Layouts—Tendon Protection and Spacing—Partial Prestress—Combination of Prestressed and Reinforced Concrete; Continuous Beams: Continuous-Beam C Lines—Load-Balancing * Method—Ultimate Strength of Continuous Beams; Design Examples 10-1 11-14 aa 13-1Contents Section 14 DESIGN OF COMPOSITE MEMBERS Ira Hooper, Michael A. Grubb, and Ivan M. Viest Definitions—Elastic Properties of Cross Section—Plastic Strength of Beam Cross Section—Critical Column Stress—Combined Compression and Flex- ure—Shear Connectors—Unsymmetrical Steel Sections—Negative-Moment Sections—Deflections and Vibrations—Composite Columns; Building Design: ‘Assumptions—Design of Composite Beams—Selection of Steel Section Design of Shear Connectors—Design of Composite Columns; Bridge Design: Assumptions—Design of Composite Beams—Steel-Member Selection—Design of Shear Connectors Section 15 MASONRY CONSTRUCTION Walter L. Dickey Materials: Burned-Clay Units—Brick—Structural Clay Tile—Conerete Units— Dimensions—Mortar—Grout; Nomenclature; Unreinforeed Masonry; Rein- forced Masonry: Materials—Design—Allowable Stresses—Beams—Walls— Slender Walls—Columns—Connections—Diaphragms; Tests and Inspection; Compressive Strength of Masonry-—Field Tests; Detailing and Construction: Detailing—Concrete Foundations—Storing—Workmanship—Shoringand Scaf- folding—Prefabrication Section 16 TIMBER STRUCTURES Kenneth P. Milbradt Structural Properties of Wood: Anisotropic Nature of Wood—Elastie Con- stants—Directional Strength Properties—Factors Affecting Strength—Work- ing Stresses for Sawn Lumber—Glued-Laminated Lumber—Plywood; Fasten- ers: Bolts—Split Rings and Shear Plates—Truss Plates; Beams: Flexure— Shear—Bearing—Deflections—Lateral Stability —Continuous Spans—Pitched and Tapered Beams; Columns: Solid Columns—Box Columns—Spaced Col- umns—Beam Columns; Trusses: Proportions—Design of Members—Defl tions—Camber—Bracing—Trussed Joists; Arches: Three-Hinged Tudor ‘Arch—Two-Hinged Arches; Shell Structures: Domes—Barrel Vaults—Hyper- bolic Paraboloids Section 17 ARCHES AND RIGID FRAMES Thomas C. Kavanagh (deceased) and Robert C. Y. Young Nomenclature and Classification; Analysis: Assumptions—Kern Relationships— Finite-Element Analysis—Energy Methods; Design of Arches: General Proce- dure—Preliminary Selection of Shape—Approximations for Special Shapes— Intermediate Design—Approximations of Whitney Data—Final Design— Unsymmetrical Arches—Ultimate Design of Concrete Arches; Design of Frames: Steel Frames—Conerete Rigid-frame Bridges—Design—Arched Bents, Continuous Arches on Elastic Piers; Special Topies: Second-Order The- ory—Interaction of Arch and Deck—Buckling of Arches—Laterally Loaded Arches and Frames—Skewed Barrel Arches and Rigid-Frame Slabs; Construc- tion and Details; Concrete Arches and Frames—Steel Arches and Rigid Frames—Economics Section 18 STEEL AND CONCRETE BRIDGES Arthur L. Elliott General—Aestheties—Loads—Maximum Moments and Shear in Simple Spans—Positive Moments in Continuous Spans-—Negative Moments in Contin 14-1 15-1 16-1 17-1 18-1x Contents uous Spans—Shears in Continuow s—Impact—Wind—Other Loads— Design Methods; Steel Bridges: Floor Systems—Concrete Floors—Steel Floors—Floor Beams; Bearing and Expansion Details: End Bearings—Expan- sion Hangers—Deck Expansion Joints; Beam and Plate-Girder Bridges: Beam Bridges—Plate-Girder Bridges—Composite Beam _ Bridges—Contimious Spans—Spacing—Lateral _Systems—Deflection—Welded Plate Girders. Design of a Welded Plate Girder—Web Splices—Field Splices; Truss Bridges: Proportions—Loads and Stresses—Truss Members—Lateral Forces; Conerete Bridges: Camber, Plastic Flow, and Shrinkage; Slab Bridges: Simple Spans— Continuous Spans—Design of Bents—Typical Details, Continuous Slabs; T- beam Bridges: Economies—Design of a ‘T-beam Bridge—Design of Substruc- ture—Typical Details; Box-Girder Bridges: Economics—Proportions — Design—Substructure; Prestressed-Concrete Bridges: Standard Sections— Stresses—Path of Prestressing Force—Friction Losses—Ultimate Load—Web Reinforcerment—Uplift—Live-Load Deflection; Bridge Railings: Railing Design—Curbs and Sidewalks—Pedestrian Railings; Administration of a Bridge System: Maintenance and Inspection of Existing Bridges—Bridge Administration Section 19 STEEL-PLATE-DECK BRIDGES Roman Wolchuk Applications—Economic Considerations—Structural Behavior—Deck Plate— Design Stresses and Loads; Open-Rib Decks: Design of Ribs—Design of Floor Beams—Stresses in Ribs and Floor Beams; Closed-Rib Decks: Design of Closed Ribs; Box Girders: Analysis of Box Girders—Design in Accordance With Clas- sical Elastic Theory; Design with Consideration of Imperfections: Effect of Imperfections on Behavior of Steel Plating—Unstiffened Plate Panel under Axial Compression—Stiffened Plate Panel under Axial Compression—Plate Panel under Combined Shear and Axial Load—Load-Bearing Diaphragms; Construc- tion Details: Deck Details, Fabrication, and Erection—Erection of Box Girders; Wearing Surfaces: Seal and Tack Coating—Surface Courses; Railroad Bridges Section 20 CURVED STEEL I-GIRDER BRIDGES William N. Poellot, Jr. Design Considerations: Bridge Geometry—Thermal _Effects—Framing Arrangements—Number of Design Girders—Connection Details—Camber; Curved Girder Mechanics: Torsion, Warping Stresses—Radial Component of Flange Force—V-Load Approximation—Lateral Flange Bending—Centrifugal Force Effect; Diaphragm Forces Section 21 BUILDINGS—GENERAL DESIGN CONSIDERATIONS Stephen J. Y. Tang and Ian R. Chin Planning Building Structures: Selection of Structural Scheme—Spatial Require- ments—Wind Systems—Deflection—Structural Materials—Fire Resistance— Deterioration—Provision for Environmental-Control Systems—Limitations of Various Systems; Loads: Dead Load—Live Load—Snow Loads—Wind Loads— Seismic Loads; Floor and Roof Construction: Floor and Roof Systems—Floor Finish—Roofing; Wall Construction: Type of Walls—Nonbearing Walls—Bear- ing Walls—Windows; Stairs: Planning—Types—Framing—Steel Stairs—Con- crete Stairs—Escalators; Miscellaneous Considerations: Openings and Voids— Thermal and Seismie Movement 19-1 20-1 21-1Contents Section 22 INDUSTRIAL BUILDINGS James M. Fisher Design Philosophy —Identification of Client Requirements—Building Layout— Identifying Structural Requirements—Selection of Roofing and Wall Material— Selection of Bay Size; Framing Systems: Concrete-Framed Buildings—Wood- Framed Buildings—Steel-Framed Buildings—Lateral-Load Systems: Materials Handling; Crane Buildings: Fatigue—Crane Girder Design—Crane Columns Section 23 TALL BUILDINGS Morton H. Eligator, Sing L. Chu, and Louis A. Occhicone Framing; Bay Sizes—Columns—Elevator Shafts—Moving Stairs—Stairwells— Transfer Girders and Trusses; Bracing: Braced Bents, Rigid Frames, Shear Walls, and Interaction Systems—Tubular Frames, Tube Within a Tube, and Combinations—Fixed and Partially Fixed Joints in Steel Structures—Wind- Load Determination—Wind Deflection—Wind-Shear Dissipation—Approxi- mate Methods of Analysis—Computer Methods—Damping Systems—Thermal Effects Section 24 THIN-SHELL CONCRETE STRUCTURES David P. Billington Introduction to Design—Simplified Analyses—Basis for the Membrane ‘The- ‘ory—General Method of Analysis—Classification of Shells; Shell Walls: Behav- ior of Cylindrical Shell Walls—Cylindrical Tanks—Hyperboloids—Membrane Theory—Bending Theory—Examples—Behavior of Barrel-Shell Roofs—Long, Barrels—Short Barrels—Transverse Frames—Barrel-Shell Reinforcement; Folded Plates: Analysis of Folded Plates—Prestressed Folded Plates—Contin- uous Folded Plates; Translation Shells of Double Curvature: Membrane The- ory—Elliptic Paraboloids—Hyperbolie Paraboloids; Dimensioning; Stability and Safety; Construction; Appearance Section 25 SUSPENSION ROOFS Lev Zetlin and I. Paul Lew Examples of Suspension Structures; Design of Suspension Systems: Anchorage Forces—Dynamic Behavior—Single Cable under Uniformly Distributed Load—Configuration and Shapes of Suspension Structures; Double Layer of spension Systems—Structural Relations Notation—Preliminary Design of Double-Layer Cable System—Anah Double-Layer Cable System—Behavior of Pair-Set of Cables—Application to Preliminary Design of Cable Grids and Membranes—Load Combinations for Selection of Cables—Types of Cables—Fittings—Membranes Section 26 REINFORCED-CONCRETE BUNKERS AND SILOS German Gurfinkel Introduction—Bin Pressures—Emptying Pressures on Sheet 6—Emptying Pressures in Funnel-Flow Silos, ACI 313- sures on Horizontal Bottoms and Inclined Hoppers—Earthquake Forces; Wall Forces: Circular Silos—Rectangular and Polygonal Silos—Thermal Effects; Design of Walls: Minimum Thickness of Circular Walls—Maximum Crack Width—Walls in Tension—Walls in Tension and Flexure—Walls in Compres- n—Walls in Compression and Flexure—In-Plane Bending of Walls—Walls 22-1 23-1 24-1 25-1 26-1xii Contents Subjected to Thermal Stresses—Vertical Reinforcement—Details and Place- ment of Reinforcement; Design of Bottoms: Bottom Pressure—Plane Bottoms— Conical Hoppers—Pyramidal Hoppers—Hopper-Supporting Beams—Col- umns—Roofs—Failures—Dust Explosions in Grain Elevators and Flour Mills; Examples Section 27 STEEL TANKS Robert S. Wozniak Capacity—Shell Design—Bottom Plates—Concrete Ringwall— Roofs; Standpipes: Design—Anchorage—Foundations; Elevated “Tanks: Roofs—Bottoms—Baleony or Ring Girder—Columns—Single-Pedestal Tanks—Foundations; Accessories; Bins: Forces—Miscellaneous Details; Materials Section 28 TOWERS AND TRANSMISSION POLE STRUCTURES Alain H. Peyrot and Gene M. Wilhoite Types of Towers—Materials—Height Limitations—Loads—Candelabra; Free- standing Towers: Stresses—Foundations; Guyed Towers: Wind—Design of Guys—Ice Loading—Guy Tensioning—Guy Vibration—Design of Mast— Foundations—Erection; Transmission Towers: Types—Loads—Vibration— Stress Analysis—Steel Tension Members—Aluminum Tension Members—Steel Compression Members—Aluminum Compression Members—Limiting Slender- ness Ratios—Bolts—Tower Design—Foundations—Tower Tests; Pole Struc- tures: Design—Material—Pole Splices—Foundations Section 29 BURIED CONDUITS Raymond J. Krizek ‘Types of Conduits—Analysis and Design; Loads on Conduits: Loads on Ditch Conduits—Loads on Projecting Conduits—Loads on Conduits in Wide Ditches—Loads on Negative Projecting and Imperfect-Ditch Conduits—Sur- face Loads; Rigid Conduits: Supporting Strength—Bedding Classes for Trench Conduits—Bedding Classes for Embankment Installation— Monolithic Con- duits; Flexible Conduits: Ring Compression—Deflection—Pipe Arches— Arches on Rigid Foundations; Pressure Conduits: Flexible Pressure Conduits— Rigid Pressure Conduits; Modern Design Methodology: Elasticity Solution— Finite-Element Solution—CANDE (Culvert ANalysis and DEsign); Additional Design Considerations: Handling Criteria—Durability—Camber—-Wrappings and Coatings; Construction Considerations: Site Preparation—Bedding Fill Construction—Compaction Procedures—Strutting—Joints—Backpacking; Long-Span Corrugated-Metal Conduits Section 30 CHIMNEYS Shih-Lung Chu, Shu-Jin Fang, and Max Zar Materials—Diameter and Heights; Design Loads: Dead Loads—Wind Loads — Earthquake Forces—Pressure Differentials—Temperature Differentials—Nat- ural Frequency of Vibration; Steel Stacks: Allowable Stresses—Cone-to-Cylin- der Junction—Circumferential Stiffeners—Anchor Bolts—Base Ring for Anchor Bolts—Guyed Stacks—Braced Stacks—Resonant Vibrations; Rein- forced Concrete Chimneys: ACI Standard—Vibration due to Wind: Linings, Foundations 27-1 28-1 29-1 30-1Contents xiii APPENDIX Al Torsional Properties of Solid Cross Sections; Torsional Properties of Closed Thin-Walled Cross Sections; Torsional Properties of Open Cross Sections; Effec- tive-Length Coefficients for Columns; Buckling of Plates under Edge Stress; Stiffened Beam Webs; SI Conversion Factors Index follows the Appendix.Preface This Handbook provides engineers, architects, and students of civil engineering and architecture with an authoritative reference work on the planning and design of a variety of engineered structures. Among the structures covered are industrial buildings, tall buildings, bridges, thin-shell structures, arches, suspension roofs, tanks for liquid storage, bins and silos for granular materials, retaining walls, bulkheads, steel transmission towers and poles, chimneys, and buried conduits. Design in reinforced concrete, prestressed concrete, steel, composite con- struction, wood, aluminum, and masonry are covered. Sections on soil mechanics, soil exploration, and foundations, and a comprehensive treatment of structural analysis, give the designer the information likely to be needed for these phases of design. Earthquake-resistant design and design against fatigue, brittle fracture, and lamellar tearing are treated. In this third edition every section has been reviewed, and many have been revised extensively. The section on structural analysis has been completely rewritten so as to emphasize computer-based techniques, and the treatment of finite-element idealization of various types of structures, including recommendations as to suitable types of elements, has been expanded. The section on steel structural members has been expanded to include coverage of the new Load and Resistance Factor Design spec- ification (LRFD). Design of composite beams and girders according to the LRFD specification is also covered, and a detailed description and an example of the autostress design procedure for highway bridges is given. There is also a new section on curved steel I-girder bridges. A discussion of lamellar tearing and suggestions on how to avoid it has been added to the section on fatigue and brittle fracture. Coverage of fatigue is updated, with comprehensive tables for determining the fatigue life of various types of joints. The section on cold-formed steel members has been rewritten so as to conform to changes in the 1986 edition of the American Iron and xvxvi Preface Steel Institute specification, which is an extensive revision of earlier editions. The sections on structural members of reinforced concrete, prestressed concrete, masonry, wood, and aluminum have also been revised to conform to the latest design specifications. The 30 sections have been written by 45 contributors, They have pre- sented their material in ready-to-use form wherever possible. To this end, derivations of formulas are omitted in all but a few instances and many worked-out examples are given. Background information, descriptive matter, and explanatory material have been condensed or omitted. Because each section treats a subject which is broad enough to fill a book in itself, the contributors have had to select that material which in their judgment is likely to be most useful to the greatest num- ber of users. However, sources of additional material are noted for most of the topics which could not be treated in sufficient detail Each section was edited to minimize duplication, to arrange the con- tents of the book in a logical order, and to see that important topics were not overlooked. The contributors’ painstaking efforts, their coop- eration in the review and editing of their work, and their patience dur- ing the time it has taken to complete the third edition are gratefully acknowledged. Edwin H. Gaylord, Jr.Contributors Lynn$. Beedle Director, Fritz Engineering Laboratory, Lehigh University (Plastic Design of Steel Frames) David P. Biltington Professor of Civil Engineering, Princeton University (Thin-Shell Con- crete Structures) lanR.Chin Senior Consultant, Wiss, Janney, Elstner Associates, Inc., Chicago, IL (Build- ings—General Design Considerations) Sing L. Chu. Senior Associate Partner, Weiskopf and Pickworth, Consulting Engineers, New York, NY (Tall Buildings) JohnW. Clark Formerly Technical Adviser, Alcoa Laboratories, Alcoa Center, PA (Design of Aluminum Structural Members) WalterL. Dickey Consulting Engineer, Higgins Brick Company, Los Angeles, CA (Masonry Construction) Morton H.Eligator Partner, Weiskopf and Pickworth, Consulting Engineers, New York, NY (Tall Buildings) Arthur L, Eliott Bridge Engineer, Sacramento, CA (Bridges) Steven J. Fenves Sun Company University Professor of Civil Engineering, Carnegie-Mel- lon University (Computer Applications in Structural Engineering) James M. Fisher Vice President, Computerized Structural Design, Inc., Milwaukee, WI (Industrial Buildings) T.V. Galambos Professor of Civil and Mineral Engineering, University of Minnesota- Twin Cities (Plastic Design) William L. Gamble Professor of Civil Engineering, University of Mlinois at Urbana-Cham- paign (Design of Reinforced-Concrete Structural Members) Michael A. Grubb AISC Marketing, Inc., Pittsburgh, PA (Design of Composite Members) German Gurfinkel Professor of Civil Engineering, University of Illinois at Urbana-Cham- paign (Reinforced-Concrete Bunkers and Silos) Hans William Hagen Partner, LeMessurier Consultants Inc., Consulting Engineers, Cam- bridge, MA (Design of Steel Structural Members) William J. Hall Professor and Head of Civil Engineering, University of Illinois at Urbana- ‘Champaign (Earthquake-Resistant Design) Ira Hooper Seelye Stevenson Value & Knecht, New York, NY (Design of Composite Members) Herbert ©. Ireland Professor Emeritus of Civil Engineering, University of Illinois at Urbana-Champaign (Retaining Structures and Foundations) xvilxviii Contributors Thomas C. Kavanagh (deceased) Former Vice-President, Iffland Kavanagh Waterbury, P.C., New York, NY (Arches and Rigid Frames) Raymond J. Krizek Professor of Civil Engineering, Northwestern University (Buried Conduits) Roger A. LaBoube Associate Professor, University of Missouri-Rollla, Rolla, MO (Design of Steel Cold-Formed Members) H.G.Larew Professor of Ci William J. LeMessurier Senior Partner, LeMessurier Consultants, Inc., Consulting Engi- neers, Cambridge, MA (Design of Steel Structural Members) |. Paul Lew Vice President, Lev Zetlin Associates, Inc., Engineers and Designers, New York, NY (Suspension Roofs) Lee ¢. Lim President, Lim Consultants, Inc., Consulting Engineers, Cambridge, MA (Design of Steel Structural Members) T.Y. Lin Professor Emeritus of Civil Engineering, University of California, Berkeley (Design of Prestressed Concrete Structural Members) James W. Mahar Resident Engineer, Department of Civil Engineering, University of Illi- nois at Urbana-Champaign (Soil Exploration) Engineering, University of Virginia (Soil Mechanics) Kenneth P. Milbradt Associate Professor Emeritus of Civil Engineering, linois Institute of Technology (Timber Structures) W.H.Munse Professor Emeritus of Civil Engineering, University of Hlinois at Urbana- Champaign (Fatigue, Brittle Fracture, and Lamellar Tei Louis A. Occhicone Senior Associate Partner, Weiskopf and Pickworth, Consulting Engi- neers, New York, NY (Tall Buildings) David A.Pecknold Professor of Civil Engineering, University of Hlinois at Urbana-Cham- paign (Structural Analysis—Fundamentals and Analysis of Elastic Trusses and Frames) Alain H. Peyrot_ Professor of Civil Engineering, University of Wisconsin-Madison (Towers and Transmission Poles) William N. Poellot, Jr. Project Manager, HDR-Richardson Gordon, Inc., Pittsburgh, PA (Curved Steel I-Girder Bridges) William C. Schnobrich Professor of Civil Engineering, University of Illinois at Urbana- Champaign (Structural Analysis—Continua: The Finite-Element Method) Shih-Lung Chu Associate and Head, Structural Analytical Division, Sargent and Lundy, Engineers, Chicago, IL (Chimneys) Shu-din Fang Associate and Assistant Head, Structural Analytical Division, Sargent and Lundy, Engineers, Chicago, IL (Chimneys) Stephen J. ¥.Tang Professor Emeritus of Architecture, University of Oregon (Buildings— General Design Considerations) Thomas H. Thombum (deceased) Former Consulting Soils Engineer, Las Vegas, NV (Soil Exploration) Wan M. Viest Consultant, Bethlehem, PA (Design of Composite Members) Gene M. Wilhoite Transmission Line Consultant, Chattanooga, TN (Towers and Transmis- sion Pole Structures) Roman Woichuk Consulting Engineer, Jersey City, NJ (Steel-Plate-Deck Bridges) Robert S. Wozniak Senior Engineer, Chicago Bridge and Iron Company, Oak Brook, IL (Steel Tanks) Robert C. Y. Young Vice-President, Iffland Kavanagh Waterbury, P.C., New York, NY (Arches and Rigid Frames)Contributors xix Max Zar Former Partner and Manager of Structural Department, Sargent and Lundy, Engineers, Chicago, IL (Chimneys) Lev Zettin President, Zetlin/Argo, New York, NY (Suspension Roofs) Paul Zia Professor of Civil Engineering, North Carolina State University (Design of Pre- stressed Concrete Structural Members)ABOUT THE EDITORS Epwin H. Gaytorp, Professor Emeritus of Civil Engincering with the University of Illinois at Urbana-Champaign, is a registered structural engineer in Illinois. The late CHarLes N. GayLorD was Professor Emeritus of Civil Engineering at the University of Virginia.Section 1 Structural Analysis Part 1. Fundamentals and Analysis of Elastic Trusses and Frames DAVID A. PECKNOLD Professor of Civil Engineering, University of Illinois at Urbana- ‘Champaign INTRODUCTION In order to perform a structural analysis, the structural engineer invariably must idealize, or model, the actual structure, to a greater or lesser degree, depending on the complexity of the structure itself, the purpose for which the analysis is being made, and the compu- tational tools at hand. Computational tools which are easily available make it feasible to model structures for analysis so that structural response can be calculated well into the nonlinear range, taking account of inelastic material behavior and geometry changes under load. Dynamic response to, for example, earthquake, wind, and wave loadings can be routinely deter- mined. Such analyses can provide useful information on postelastic stress redistribution, can identify potential failure mechanisms, and can provide better estimates of safety mar- gins. In many cases, however, analyses are made under the assumption that the structure behaves in a linear elastic manner, that the deflections are small, and that the loads are applied slowly. Even if these assumptions are not all reasonable approximations in a given case, a linear elastic static analysis is often made as a first step, since such an analysis often provides useful information in its own right and also provides a context within which the results of more sophisticated analyses can be evaluated and understood. Structural modeling for analysis must be done with a clear understanding of the struc- tural load-carrying mechanisms which are involved and an appreciation for essential behavior versus unimportant detail; in some cases modeling is relatively routine, while in others it may require a great deal of skill and experience. 4. Classification of Structures Structures can be classified for analytical purposes in a number of ways. They can be regarded naturally in many cases as consisting of a number 141-2 Structural Analysis of structural components each of which may carry a relatively simple system of forces. This decomposition into simpler components can be done at several levels; thus a struc- ture may be regarded as consisting of a number of subassemblages or substructures, each of which may, in turn, consist of other, simpler substructures; until finally at the lowest level the basic components are simple structural members or elements. This hierarchical decomposition is useful, both for computational purposes and for conceptual, descriptive Purposes. Structural members or elements can be classified in terms of the essential force systems which they carry and in terms of their shapes and dimensions. One possible classification scheme is as follows: 1, Line elements: bars, rods, columns, or cables in which the cross-sectional dimen- sions are small compared to the length of the member, and the force system is axial, either tension or compression 2. Surface elements (two-dimensional equivalent of line elements): panels, deep beams, and tension and compression membranes in which the thickness is small compared to the surface dimensions, and the force system is two-dimensional and lies in the plane of the midsurface 3. Line elements which carry load by bending: beams or arches 4. Surface flexural elements: plates or shells 5. Hybrid elements: for example, beam columns, which are subject to both axial load and bending Structural components which are of general shape and are subjected to three-dimen- sional stress systems are solids. FUNDAMENTALS OF ANALYSIS 2. Description of Structural Configuration ‘The state or configuration of the structure may be defined in terms of the degrees of freedom of the structure. These are defined as a set of independent coordinates, or pieces of geometric information, minimum in number, required to uniquely define the shape and orientation of the structure, A rigid body in space has six degrees of freedom, which can be taken as three cartesian components of displacement and three rotations. A planar truss has 2 degrees of freedom: the two carte- sian components of displacement of each joint, where j is the number of joints. A planar frame has 3j degrees of freedom: two displacements and a rotation for each joint, In each of these examples, the degree of freedom is a natural, physical choice used to describe the deformed configuration. However, the degrees of freedom can be chosen in a some- what more abstract way, which is very useful in some cases. For example, a degree of freedom may simply be an amplitude or multiplier of a special deflected shape or config- uration of the structure. A good example in structural dynamics is the set of natural modes of vibration of the structure. A completely general deflected shape of the structure can be specified as a linear combination of these mode shapes. The amplitude of each of these mode shapes in that linear combination is a degree of freedom. A simply supported beam provides another example. The rotations at the supports at either end of the bearn can be defined as the (two) degrees of freedom for the beam. Alternatively, the deflected shape can be described in terms of the superposition of two special shapes: a symmetric deflec- tion shape in which the rotations at the ends of the beam are of equal magnitude but in ‘opposite rotational senses, and an antisymmetric deflection shape in which the rotations are of equal magnitude and the same rotational sense. The amount of each of these shapes which is present in the general shape can be selected as a degree of freedom. In these latter cases, the independent shapes (the mode shapes, or the symmetric/antisymmetric beam deflection shapes) play the same role as the cartesian unit vectors do in the case of the truss joint displacements. Other Structural Variables. In the case of, for example, a planar truss under load, it is obvious that if the two cartesian displacement components of every joint are known, then the position of each joint in the loaded configuration is known. Thus, member lengths in the deformed configuration can be determined, leading directly to member elongations, or strains. From these strains and knowledge of the material behavior, the stresses, or alternatively member axial forces, can be determined. Thus, specification of the values of the degrees of freedom leads directly to determination of the other structural variables of interest. In the procedure just described, the degrees of freedom are viewed as theBasic Principles 1-3 fundamental variables, and the strains and stresses as derived variables. However, since unique relationships between these variables exist for elastic structures, it follows that one may consider, for example, the member forces to be the fundamental variables, and the joint displacements and strains to be derived variables. The choice of different fun- damental variables leads to different general methods of structural analysis. The displace- ment method and the force method are the two best-known such general methods. Various hybrid methods have also been developed, in particular, for the finite-element analysis of plates and shells. 3. Basle Principles _In the solution to any structural analysis problem, the structural variables must satisfy three basic physical laws, which are described here in the context of a truss: Equilibrium: The member forces (stresses) and the externally applied joint loads must satisfy the equations of static (or dynamic) equilibrium. Compatibility: The member elongations (strains) and the joint displacements must be geometrically compatible; that is, structural continuity must be preserved. Constitutive laws: The axial force (stress) and elongation (strain) of each member must satisfy the behavior laws for the particular material of which the member or strue- ture is constructed. For the simple planar truss shown in Fig. 1, these relationships are as follows: Ai + feos. a = Py fesin a + fe = Py a ey = they ey = thy COS a + Uy Sina @) ey = Uys (3) where f; = axial tension ‘a4 = elongation of member k tu, ty = cartesian components of displacement of joint é Pur Py = external loads acting on joint i Yy4.Pya Member 4 cape ued Ay = Cross-sectional area of member k les Vieeve E, = Modulus of elasticity of member k ig=V Fig.1 Planar truss.Structural Analysis In the general case of a plane or space truss, Eqs. (1) to (3) can be written in the form HE = p (4) e= Cu 6) f=Ee (6) in which the degrees of freedom are contained in the displacement vector u, the member forces (stresses) in f, and the member elongations (strains) in e. If the plane truss has j joints, b members, and r > 3 restrained degrees of freedom, the equilibrium matrix H is (2j — r X b), the compatibility matrix Cis (b X 2j — r), and the matrix E, which contains member stiffness information, is (b X b). Equations (4) to (6) display the general structure of the equations which appear throughout structural analysis, from simple trusses to finite-element models of very complex structures. In the latter case, (usually numerical) integrations over element volumes are necessary. The force vector Hf is often called the internal resisting force vector i. It consists simply of the member internal forces resolved into the directions of the global degrees of freedom. 4. Static Determinacy and Stability Several important special cases can be identified from Eqs. (4) to (6). If the equilibrium matrix H of a plane truss is square (2) — r = b) and nonsingular, the truss is statically determinate and statically stable; the member forces f can be determined from the equations of joint equilibrium, Eq. (4), and the member elon- gations and joint displacements can be subsequently determined. If 2j — r > b, the truss is unstable; there are not enough bar forces to satisfy statics for all possible loading con- ditions. If 2j — r < b, the truss is statically indeterminate. It is possible that a truss is statically unstable even if 2j — r = b. This condition corresponds to a critical geometric arrangement of members and/or supports which produces instability." Instability, as cussed here, refers to the possibility of (possibly small) rigid-body motions of the entire structure, or of substructures within the structure, taking place, without requiring mem- ber deformations. Such rigid-body modes are called kinematic modes. This type of insta- bility must be distinguished from elastic instability or buckling. If the truss is statically indeterminate, the compatibility and stress-strain relations must be utilized in the solution. In this case, the joint displacements can be expressed as u=K"p (7) in which the matrix K = HEC is the structure stiffness matrix. The relation H™ = C can be verified for the example shown in Fig. 1 and Eqs. (1) to (3). It is true in general, as will be shown subsequently. The matrix E is square, symmetric, and positive definite. As a result, the structure stiffness matrix K possesses the same properties. Consequences of Static Indeterminacy. In statically indeterminate structures, particu- lar systems of internal forces can exist in the absence of external loads. These systems of internal forces are sometimes called prestress, residual stress, autostress, or self-stress. (It is possible to stabilize structures which possess kinematic modes by means of prestress.) The classification of a structure as statically determinate or indeterminate is not, as in the case of some classification systems in structural analysis, merely for the purposes of anal- ysis. The behavior of statically indeterminate structures differs fundamentally from that of determinate structures. Some of these differences in behavior, and their consequences, can be inferred from Eqs. (4) to (7). For example, in a statically determinate structure undergoing small displacements, the internal forces are determined directly and uniquely from the external loads. If the loads are zero, so are the internal forces. No states of -esid- ual stress are possible. Thus, in a statically determinate structure no internal stresses are generated by, for example, differential temperature changes, differential support move- ments, or a lack of fit of the structural members (.c., differences between ideal and actual member dimensions). Conversely, in an indeterminate structure, all these effects can result in locked-in stresses, which can be of considerable magnitude. On the other hand, it is evident from Eqs. (4) to (7) that the relative member stiffnesses influence the internal force distributions in indeterminate structures. This has consequences in member pro- portioning for design, which must usually be carried out in an iterative design/analysis cycle. A second major consequence of this property of indeterminate structures is that, when inelastic behavior occurs, resulting in localized stiffness changes in the structare, a redistribution of internal stresses takes place if the structure can tolerate the associated deformations. Alternative load paths can therefore be developed in indeterminate struc- tures once significant inelastic action occurs. Determinate structures, in contrast, have no capacity for significant stress redistribution, and thus do not have any extra safety margin.Symmetry 1-5 5. Superposition A fundamental property of linear systems is that solutions can be superposed. That is, if loads p,, ps produce displacements us, us, respectively, then a load Pi: + pz produces a displacement u, + u,. This can, in fact, be regarded as the definition of a linear system. Therefore, the principle of superposition applies for small displace- ments of elastic systems. This principle is used in developing many of the methods of linear structural analysis and also as a routine calculation technique. 6. Symmetry Symmetry properties can be used as an organizing concept for under- standing the behavior of particular classes of structures. Additionally, these properties of structures are routinely used in analysis, in order to reduce the amount of computation which must be done, especially in the case of large structures. Symmetry is often recog- nized and utilized by the analyst when the external load also possesses some degree of symmetry. However, although the structure must possess some particular symmetry properties in order for these methods to be utilized, the external load need satisfy no particular symmetry conditions. ‘The simple portal frame shown in Fig. 2a possesses reflection symmetry about a vertical plane x-x’. The applied loading can be decomposed into two loading conditions, one sym- metric with respect to the symmetry plane, and one antisymmetric with respect to this plane. x P/2 jx PIA Pe yx PZ ' i ul + x x x (a) p/2 1X pre i* ! Only vertical Ix, Rotation Gnd displacement horizontal {} stones t=" displacement | + | allowed i 1 Il 1 1 i x x (b) Fig. 2 Use of symmetry: (a) Decomposition of load into symmetric and antisymmetric com- ponents. (b) Symmetric and antisymmetric structures of reduced size. One-half of the original structure, the reduced structure, can then be analyzed under two different sets of boundary conditions at the plane of symmetry, as shown in Fig. 2b, and the results superposed. Although reflection symmetry is the most commonly occur- ring and most commonly utilized type of structural symmetry, there are many other kinds which occur in structures. Rotational, or cyclic, symmetry, for example, describes the synthesis of a structure by rotation of a substructure or repeating unit about an axis. Even when advantage is not taken of symmetry in performing the structural analysis, the nec- essary symmetry in the computed results provides an additional check on the correctness of the results. ENERGY PRINCIPLES Energy principles provide alternative ways of viewing some of the basic results and meth- ods of structural analysis.? They also lead to extremely powerful techniques for obtaining approximate solutions, for example, in the finite-element analysis of plates and shells and general two- and three-dimensional solids.1-6 Structural Analysis, 7, Principle of Virtual Work The principle of virtual work states that a body (or mechan- ical system) is in equilibrium under the action of a system of forces if, and only if, the virtual work done by the system of forces during an arbitrary virtual displacement is equal to zero. This principle can be regarded as an alternative, parallel statement of the con- ditions of equilibrium. It is not restricted to any particular material behavior assumption, and so it is applicable to inelastic as well as elastic systems, (a) (b) Fig. 3 Principle of virtual work for a particle. (a) Rigid particle in two dimensions under a system of forces. (b) Resultant force and virtual displacement. The concepts involved in this principle can be illustrated by considering the equilib: m of a rigid particle in a plane (Fig. 3). The particle is considered to have zero dimen sions, i.e., it is a point in two-dimensional space. A system of forces F,, F;, . . . F, acts on the particle, The resultant of this system of forc = Dt F,. The virtual displacement 6a is defined in general as an arbitrary reversible displacement consistent with the con straints. The virual displacement is arbitrary in both magnitude and direction. Thus, since the rigid particle in the plane has two degrees of freedom, there are two indepeadent pieces of information necessary to specify the virtual displacement. Rather than magni- tude and direction, these could alternatively be taken as the components éu, and Su, of 6ii with respect to cartesian axes (x, y). The work done by one of the forces F, were it to move through the displacement 8% would be F, - 6i, Therefore, the virtual work 8W done by the system of forces is 6=W = Ly, (F, - 6%) = R - 52. The statement of the principle of virtual work is that the particle is in equilibrium if, and only if, 6W = 0 for an arbitrary 8ii. The latter requirement is essential, as it precludes the special choice in which 64 is orthogonal to R. Therefore 5W = 0 requires that R = 0 for equilibrium. The virtual work can also be written in the form 6W = (Ef, Fy) 6, = (Ef Fy) 6u,. Thus 8W = 0 is completely equivalent to Dj., F, = 0 and Si, Fy = 0, the two equations of static equi- librium in the plane. One equation of equilibrium is generated corresponding to each degree of freedom o? the system. The coefficient of each virtual displacement in the expression for 6W is called a generalized force. A rigid body with finite dimensions, shown in Fig. 4, has three degrees of freedom i two dimensions, two translations éu,, 5u, in the plane, and a rotation 50, about an axis normal to the plane through point O. Thus the components of the virtual displacement Fig. 4 Rigid-body virtual displacementsPrinciple of Virtual Work 1-7 éi, at the point of application (x,, y;) of one of the forces F can be expressed in terms of the three independent quantities 41, ty, 60, as duty buy ou, — Ys buy + 5, oO, (8) 6, The virtual work is now given by the expression 6W = Dv, (F, + 6%), or 3W = (Dp, By) + bu, + (Efe, By) + dy + [ee (-Fat xt) | + 60, which leads to the three equations of static equilibrium in the plane: D!-,Fy = 0, Efa:Fy = 0, and EM, = 0. The body under consideration need not be rigid for the foregoing relations to apply. The virtual displacement was simply chosen as if the body were rigid, but this virtual displacement can just as well be imposed on a deformable body. Thus, imposition of rigid body virtual displacements on a deformable body generates the overall equations of static equilibrium. A deformable body may also be subjected to virtual-displacement patterns which, if they were actual displacements, would distort the body. Such virtual-displacement pat- terns provide information about the internal forces, For example, the simple plane truss shown in Fig, 1 has two degrees of freedom w,,, tuys. Let virtual displacements 5ti,4, diy, he imposed on the structure. These joint, or nodal, virtual displacements define ‘a deflected shape of the entire structure which is a perturbation from the equilibrium posi- tion, as illustrated in Fig. 5. Such a virtual-displacement pattern would, if it were actually imposed on the structure, produce length changes in some or all of the structural mem- bers. Therefore, the internal forces, which would tend to resist such changes, must con- tribute to the virtual-work expression in this case. It is important to note that the forces do not change during the virtual deformation of the structure; they are the forces existing in the equilibrium position. Equilibrium position Fig. 5 Virtual displacement from equilibrium position. Since the joints of the truss can be regarded as rigid particles, the foregoing relations can be directly applied to them. Equations (1) are the equilibrium equations for joint 4. Thus the virtual work can be expressed as BW = Ip — Fi + foc08 a)] + btey4 + [Pye — (fe sin a + fh] + diy, (9) This expression can be separated into two components, the first involving the external loads, called the external virtual work W,q, and the second involving the internal forces, called the internal virtual work 6W. bWou = Pas + thes + Pye * Stiys (10a) 8Wan = Lf - bttaa + fe * (Bttey COS a + Stig Sin a) + fy + Bey) = li: ber + fa bee + fi + bes] (0b) In the general case, Eqs. (4) to (6), the virtual work can be written as 6W = bu" - (p — Hf) (1d) and the external and internal virtual work can be written as bWoa = bu p (12a)1-8 Structural Analysis Comparison of the latter two expressions in Eq. (12a) shows that ¢ = Hu. Since ¢ Cu from Eq, (5), it follows that C = H. The principle of virtual work can therefore be written in the form. bWiy = —duT - HE = —8(H7u)" - F = —beT Compatibility 4 out p = be" fF t Equilibrium In Eq, (13), the internal force f and the external load p satisfy the equations of equilib- rium, Eq. (4). The (vistual) joint displacements 5u and the corresponding (virtual) mem- ber elongations ée satisfy the compatibility equations, Eq. (5). Material behavior descrip- tions are not involved in this relation, and it therefore applies to inelastic materials as well. Equation (13) may be regarded as a general reciprocal relation which relates any compatible set of kinematic variables (5u, 5e) and any equilibrated set of static variables (p, £). For some purposes, it is convenient to take these compatible kinematic variables (du, ée) as the actual displacements and elongations, and the equilibrated static variables (p, f) as a conveniently chosen auxiliary load system. This is the basis for the unit dummy load method (Art. 12). 8. Principle of Minimum Potential Energy The principle of minimum potential energy treats the displacements or degrees of freedom of the structure as primary variables. It is closely related to the displacement or stiffness methods of structural analysis. The (total) potential energy I is equal to the sum of the (internal) strain energy U and the load poten- tial, or external potential energy, 0. These quantities are defined as follows. Strain Energy. The strain energy is the stored energy of deformation. For a one- dimensional structural member with an internal force vs. elongation relation as shown in Fig. 6, the strain energy is the work done by the force in producing the current elonga- tion, Thus, U(e) = i ‘f (2) dz, the area under the curve up to the current elongation e. The gradual change of force with elongation is taken into account in this calculation ‘This quantity involves not only the material response but also the cross-sectional area and length of the member. It is more convenient instead to define a strain-energy density, .e., strain energy per unit volume, Uy which involves only the material response and not the dimensions of the member. Thus, in terms of stress ¢ and strain ¢, the strain-energy density Up(@) = ic o() dé. For a linear clastic material in a state of uniaxial stress U,(€) = lo Yale = ¥Ee where Kis the modulus of elasticity. The first-order change in strain energy due to a change in elongation or strain is denoted 8U and is given by the expression dU = foe as shown in Fig. 6; thus the relation U = —3W,, between the first-order change in strain energy and the internal virtual work. 3 Su- $8480 Su=F8e Internal force F Elongation e Stroin energy U Fig.6 Strain energy.Principle of Minimum Potential Energy 1-9 ‘These relations can be generalized to multiaxial states of stress, resulting in 1 0-3 f toi + oy + at + -Jav (14a) la =4fe-eav BU = f foe bee + 0486, + tap dry + © ]dV (14b) = | 6? odV F The expression for U given in Eq, (14a) is valid for three-dimen: systems. The expression for 6U in Eq. (14b) is general. Expressions for U and 6U for various types of structural members are obtained by insert- ing in Eq, (14) the appropriate assumptions regarding strain distributions in each partic- ular case, For example, in conventional beam theory, it is assumed that plane sections remain plane. The strains ¢, are therefore constrained to vary linearly through the beam depth, s0 & = G9 — 2 +k — Y + Kin which the x-axis is the longitudinal axis of the beam and the y- and 2-axes aré the principal axes of the cross section, éo is the axial strain at the beam centroidal axis, and x, = 1/R, and x, = 1/R, are the beam curvatures for bending in the x-z and the z-y planes, respectively. This strain distribution describes a beam under axial force and biaxial bending about its principal axes. In addition, transverse shear deformations are neglected. Therefore, inserting these constraints in Eq. (14) and decom- posing the volume integration into an integration over the cross section followed by inte- gration along the length of the beam, the following expressions result. nal linear elastic a Useam = ; i [Neo + May + Muc,| dx (15a) <0 i eS i IN, deo + My xy ++ M, bx,] dx (15b) a0 in which Nat feds My= finda Mm [yoda (16) For the truss-type structural system described by Eqs. (4) to (7), the strain energy is given by the expression U = Het f = Mu"(CTEC)u = Yu"Ku (17) This expression for strain energy in terms of the stiffness matrix K is general. External Potential Energy. For conservative external loadings, a potential energy of the loads can be defined such that @ = —u"- p, Potential Energy. ‘The total potential energy Il of the structural system is I = U+ Itis regarded as a function of the displacements, and therefore of the degrees of freedom, of the structure. The principle of minimum potential energy provides the criterion for selecting the actual displacements of the structure (i.e., those which also satisfy equ rium) from the set of all possible displacements which satisfy the constraints (boundary conditions). It may be stated in the following way: Of all displacements which satisfy the constraints (boundary conditions on displacements) those which also satisfy equilibrium make the potential energy an absolute minimum. The potential energy is T= kulku — utp (1s) Ifa small displacement increment du is taken from the equilibrium configuration u, the potential energy changes from II to II + ATI. The change in potential energy is equal to AM = 6u"(Ku ~ p) + 46u"Ku (19) = Ol + 46° According to the principle, AII must be positive since any displacement away from the equilibrium position results in an increase in potential energy. Therefore, for an equilib-1-10 Structural Analysis: rium position, SIL = 0, which leads to the equilibrium equations of the stiffness method Ku = p, The additional requirement for stable equilibrium is that 5°11 > 0, which provides analytical conditions for the study of structural stability. Application of the minimum potential energy approach generates the same equations as does the virtual work approach. Both approaches lead to powerful approxmation methods. 9, Principle of Minimum Complementary Potential Energy The principal of minimum com- plementary energy treats the stresses (internal forces) as primary variables. It is closely related to the force or flexibility methods of structural analysis. It provides the criterion for selecting the actual stresses (i.e., those which correspond to compatible strains) from the set of all possible stresses which satisfy equilibrium and stress boundary conditions, It may be stated in the following way: Of all stresses (internal forces) which satisfy equi- librium and stress boundary conditions, those which also satisfy the associated strain com- patibility conditions make the complementary potential energy an absolute mininwum. 410. The Reciprocal Theorem The reciprocal theorem relates two separate loaded config- urations of a linearly elastic structure.' Figure 7 shows two separate loading conditions acting on the same linearly elastic structure. The displacements are shown in the figure with subscripts which have the following meaning: dy is the displacement at point i due to the action of load P,, The displacements are the components of the resultant structural displacements in the directions of the loads at the same points; that is, the respective force and displacement variables are work-conjugate. Py P2 on Sas Nagai day Lood case 1 Load case 2 Fig. 7 The reciprocal theorem. The reciprocal theorem states that P,dy, = Pad). It is sometimes stated in terms of unit loads, ie., for the special case P, = Pz = 1, where it is known as Maxwell’s law. Moments and rotations can be considered in this relationship as well as forces and displacements. The terms which appear must be work-conjugate pairs. The two configurations need not be restricted to single loads. ‘The generalization to multiple loads is usually called Betti’s law. It can be stated in the following form. The work done by the first loading system acting through the displacements produced by the second loading is equal to the work done by the second loading system acting through the displacements produced by the first loading. The main use of these relationships is in demonstrating symmetry of stiffness and flexibility coefficients and in developing a method (the Miller-Breslau principle?) for constructing influence lines. ANALYSIS OF STATICALLY DETERMINATE STRUCTURES, Statically determinate models of structures are often useful. For example, a truss model in which each member carries only axial force can capture the important aspects of struc- tural behavior in many cases, even though the assumption of frictionless pin joints con- necting the members is far from accurate. Further, the detailed analysis of members or of substructures is often carried out as a final (postprocessing) step after the analysis of the full (statically indeterminate) structure. In this case, the indeterminate structural analysis provides sufficient information so that the detailed member substructure analysis can proceed as if the member or substructure were statically determinate. Forces 11, Plane and Space Trusses Ideal trusses have frictionless pin joints at which the cen- troidal axes of connecting members meet (no eccentricities), and are subjected to loads only at these joints. As a result, only axial force, either tension or compression, is carriedBeams and Frames 1-11 by each member. Of these assumptions, the critical one is that loads are applied only at the joints, Calculations with both a rigid-jointed frame model and an ideal truss model of the same structure with loads applied only at the joints generally yield axial forces which agree closely with each other. Additional internal forces (shears and bending moments) exist in the frame, but these are due only to compatibility effects and are not required for equilibrium. The additional stresses due to these bending effects are called secondary stresses and are often in the range of 15 to 20 percent, or less, of the stresses due to the axial forces. Plane and space trusses are described in general by Eqs. (4) to (6). For plane trusses, the criterion for static determinacy is that b + r = 2j, in which b = number of members, r = number of independent reaction components, and j = number of joints. For space trusses, the criterion for static determinacy is that b + r = 33. The degree of static inde- terminacy is (b + r) ~ 2 and (b + r) — 3j, respectively, when these indices are positive. These criteria determine whether or not the complete set of reactions and internal forces can be determined from the equations of statics, i.e., they relate to overall static deter minacy. It is possible that external reactions can be determined, or that some member forces can be determined, even if the structure is statically indeterminate overall, as determined by these criteria. Certain special truss configurations, called simple trusses and compound trusses, which are statically determinate overall, do not require a simultaneous solution of the equations of joint equilibrium, Eq, (4). A simple truss can be defined as one which can be assembled by starting with a simple triangular-shaped panel and then connecting new joints one by one, using two new members for each joint. A compound truss consists of simple-truss substructures. For these two special configurations, the reactions and member forces can be determined by use of the method of joints and the method of sections, These two meth- ods are also useful as methods for checking results of analyses of indeterminate structures. Although statics does not provide enough information to uniquely determine member forces in this case, spot checks should be made to ensure that equilibrium is satisfied by the computed reactions and member forces 12. Beams and Frames Rigid frames, or moment frames, consist of members connected together by joints which are assumed to be capable of transferring moments between connecting members. In some cases, more complicated structural models may be used, in which some of the joints are assumed to be frictionless pins or in which there are other types of force releases. Each such release limits or constrains the internal force system at a given location in the structure and therefore provides additional information which aug- ments the equations of statics. These additional equations are sometimes called equations of condition. oe plane frames, the criterion for static determinacy is that 3b + r = 3j + ¢, in which ¢ = number of releases and the remaining symbols are as defined in Art. 11. For space frames, the criterion for static determinacy is that 6b + r = 6j + c. The degree of static indeterminacy is (3b + r) — (3j + ¢) and (6b + r) — (6j + c), respectively, when these indices are positive. These criteria determine whether or not the complete set of reac- tions and internal forces can be determined from the equations of statics; i.e., they relate to overall static determinacy. As with trusses, itis possible that external reactions can be determined (if r = 3 + ¢ for plane frames or r = 6 + ¢ for space frames) or that some member forces can be determined, even if the structure is statically indeterminate over- all, as determined by these criteria. Reaction Calculations Using Virtual Work. Throughout structural analysis there are two alternative approaches: (1) direct application of the basic principles (equilibrium, compatibility, constitutive laws), and (2) a work-energy approach, using virtual work as the fundamental tool. In the case of statically determinate structures, reactions can be determined by directly using the equations of static equilibrium. When the structure is complex, perhaps with a number of releases, freebodies of substructures as well as free- bodies of the entire structure will have to be considered using the equilibrium approach. Alternatively, virtual work can be used to determine external reactions by applying suit- ably chosen rigid-body virtual displacements which violate the constraint corresponding to the reaction under consideration, i.e., so that the appropriate reaction does some vir- tual work. For complex, but statically determinate, structures with force releases, the virtual-work approach is likely to be more convenient than the direct equilibrium approach.1-12 Left Running Head (8) 4 Hs, mI Virtual displacement pattern ta I Fig. 8 Frame reaction using virtual R work, L Figure 8 shows « simple portal frame which is pinned at the base and which has a moment release at one of the beam-to-column connections. In order to calculate the ver- tical reaction R, a virtual-displaceent pattern is chosen in which the right-hand support is moved vertically. The resulting pattern is such that the angle between column and girder at the moment release can change, but the angle between column and girder at the rigid joint must remain a right angle. The virtual work of the reaction is R » éu, while the virtual work of the applied distributed load is —wH - KH/L - éu. These are both contributions to the external virtual work 6W,,. The internal virtual work is zero in this case, since the virtual-displacement pattern is that of a rigid hody. Thus, by setting 5W = 5Wen = 0, the reaction is found to be R = wi?/2L. One of the characteristics of the virtual-work approach is that it is often possible to avoid calculating auxiliary information which is not of direct interest. The analysis of a frame can be considered to consist of the determination of the external reactions and the internal forces in each member at the joints of the frame. The subse- quent determination of internal forces (and stresses) in a given member at a given mem- ber cross section (or at all cross sections) can be considered to be beam analysis. The relations between load, shear, and bending moment in the beam are determined by exam. ining a freebody of a differential element of the beam (Fig. 9), which gives dV ‘de 7 PO) (20) dM Se 7 VG) + mie) where p(x) = distributed lateral load m(x) = distributed applied moment ‘The sign convention is shown in Fig. 9. plxidx : TT eC TIC CO |) mean mix) dx x Fig. 9 Load, shear, and bending moment sign conventionBeam Defiections 1-13, Equations (20) provide expressions for the slopes of the shear and moment diagrams at any axial coordinate in the beam. They can be integrated between two arbitrary cross sections of the beam x = a and x = b to provide the following relations, ‘ ve) — vie) = [ple de + Pa nb b M(b) — Mia) = f Vea) dx + (a) dr + 2 Max which show that the change in shear between two sections is equal to the resultant exter- nal load acting on the beam between those two sections; and the change in bending moment between two sections is equal to the area under the shear diagram plus the resul- tant external moment acting on the beam between those two sections. 43, Beam Deflection Methods for the calculation of beam deflections are of two gen- eral types: they are based either on (1) the integration of the differential equation for beam deflections (using double integration or one of the many analogs for it, such as the moment-area method or conjugate beam); or, (2) work-energy methods, for example the dummy unit-load method, Differential Equation for Beam Deflections. Equations (20) relate load to shear and hear to bending moment. Two additional differential relations which incorporate strain- lacement compatibility and the (elastic) stress-strain relation complete the deserip- tion of beam response. The two equations are (21) dem ds” EI ae ae tan where EI = flexural stiffness of beam @ = angle between x-axis and tangent to beam axis If rotations of the beam axis are small, # < 1, the curvature dé/ds can be approximated as dé/dr, and the slope dy/dx = tan 0 as 0 so that these equations can be written aa Mie) dx (22) dy _ ae Equations (22) are often combined into the single relation dy _ M(x) EL Fora statically determinate beam, the bending moment can be determined by statics, and then the deflected shape from this equation can be determined by double integration or anumber of alternative equivalent methods. Table 1 gives some standard beam formulas, showing in most cases maximum values of shear, moment, slope, and deflection. Example. For the first case in Table 1, the bending moment M(x) = M(1 — x/L). Two successive integrations of d°y/dx? = M(a)/BI gives vy) Constants ©; and C2 are found from the boundary conditions y(0) = 0 and y(Z) = 0, Then 2 2 r Mol? x 2) _ (x woe [-2(F) +9(2) -(@)] ‘When the boundary conditions are such that the beam is not statically determinate, Eqs. (20) and (22) must be solved together rather than sequentially Classical Analogs for Double Integration, Equations (20) and (22) have a symmetrical structure which suggests some of the analogs which have been developed for calculatingyeis'o +40 ZB esg000 #44 Tusg0=¥ 40 31% 9070 = ag raves ® z ey ae ae oe zn . alse ae wees ag Zurs a me Teev'o= x49 1 13 x00 : es Fin 2900 F wouo0e0 ES is 209ys aiming, ‘sweeg 20) sejnuuoy | T1eVL 144aetna 2g = % acres emgaae es LO * 0 %8="@ na w= Ww =a ¥S z oe oe 7929°0=% 40 eon ser 72°S 8 13 os oz iy, eck = YOU 2 Loy = Fin 1£100°0 ee sar S raeee a z ae 289-Y, 2 Oye =: Sy 0-8 9-9 yaw 7784 WW = o's ean eros sa age cam 729M (o+ae) Sy *"S = = + oh 0: 892% 2 20g za ese = og) #L. ence (arog) AA's T3261 _ 24 2 8p = cee one 0: 89 =%9 Z.¥s woag pan 1302 13261 ’ 2% assy, esee ne "ae oer ee oc (Penuquog) sweog 10; sejnuod 1 TTEVL 1451302 yy sees a eG fe Be. Bg ao Z-.85 oo am ie 7" ‘ © 7 : : TE BOE v4 13h2 vg Baty a Es 5 ° 7 ° ’ ney, 9, a 5 ft ae em"? ance inte eho i" Hele | dt. ao Eames aa ae Tw = 8H m Jt i nz 3 232 _ oy Boy 6 zw son? " 2 a =" + A wwoeg s0na3v09 4 vojjsaag] Eades Q) wewow dy -0us aunsanaig, (Wenuquog) sueeg 10) senuios | 3TEVL 1-16Oey \ | Teov'o= x40 1308 _y mL OF Lug a mae eze'0=* 10 an * Jesoroo0-"o S ec ov zmEzvor0=*W i : Soe —_ er Oaa ee 33024 a Ob oy Ss 1a : oan Alpyo=%40 sa e'S : =, 6€200°0 =" ae 8 -e a uses Sach Tezv'o=% 40 = 1388 2 _.ey 2 as 13 0 = xOUK .* ame 4% vso00- > Uae (o+7€) ea’? 2 05 292X410 2 Loy eee as TRE Lug ams one : as 1a ro = OM, 2 eeeeeram FE zeso00=14 3h -=8W £ exio 13 Ly, cena ney 132 soug aw"? ow TON nee zow senaiiyuoo paddosg, (Ponuquog) euwed 205 seul’ b TEVL 117‘Structural Analy: beam deflections. In particular, it should be noted that Eqs. (22) are very similar to the equations for shear and moment, Eqs. (20), if @ is thought of as shear, y as a bending moment, and M/EI as a distributed load in some auxiliary, fictitious beam. This is the basis for the conjugate beam method.’ Other well-known methods, such as the moment-area method! also use this analogy. The moment-area method is cast in a geometrical form which emphasizes visualization of the deflected shape. It is usually stated in the form of two moment-area principles, which specify procedures for calculating (1) changes A9 in the beam tangent angle between two beam cross sections, and (2) deflection of the beam at a given section with respect to the tangent drawn from another section. The methods used for calculating these quantities in the moment-area method are equivalent to Eqs. (21) for calculating changes in shear and bending moment in the beam, using the curva- ture M/El as the load. These methods can be used to calculate deflections due to general inelastic strains by substituting the actual curvature, however it is caused, for the term M/EL The Dummy Unit-Load Method. Deflection calculations can also be carried out very conveniently using work-energy methods. The dummy unit-load method is probably the most versatile of this class of methods. Alternatives such as Castigliano’s second theorem! are equivalent to the dummy unit-load method. The dummy unit-load method employs the principle of virtual work, displayed in Eq. (13) for trusses. The actual strains and displacements of the structure are selected as the compatible set of kinematic variables. The equilibrated set of static variables is chosen as an external load of unit magnitude, the dummy unit-load (placed on the structure so that it is work-conjugate to the desired displacement), and its associated internal forces. Thus the expression for the extemal vir- tual work reduces to du" - p = 1 - A, where A is the desired displacement. The dummy unit-load method can then be expressed in the general form os in which the dummy unit load and its associated stresses satisfy equilibrium, and the displacement A results from the actual strains « The displacement which is calculated is actually a generalized displacement—any type of deformation response (rotations, relative deflections, etc.) can be calculated by apply- ing the appropriate work-conjugate dummy unit load. A rotation can be calculated using the dummy unit load as a concentrated moment. Relative deflections can be calculated using pairs of dummy unit loads. ‘A characteristic of the dummy unit-load method is that unwanted intermediate results need not usually be calculated. On the other hand, a single calculation gives the gener- alized displacement at only one location in the structure, and if a more complete descrip- tion of the deflected shape is desired, other approaches may be preferable. In order to carry out calculations using the dummy unit-load method, the internal vir- tual-work expression given in general form in Eq. (23) must be made more specifc. Internal-Virtual-Work Expressions. The internal-virtual-work expressions for specific types of structural members are developed from Eq. (23) by imposing the appropriate yematic assumptions. For beams subjected to axial force and biaxial bending, the result is Eq. (15). If torsional rotations and transverse shear deformations are also included, the internal virtual work in a single member consists of the contributions odv (23) L BWin = — J (Ne Bean + M, Bk, + M, be. + V, by, + Vedy. + Tha) de (24) ‘The internal virtual work for the structure is obtained by summing member contributions. For members of a truss, the axial force N, is constant along the member length, and Stao deo dx = be. For a truss, therefore, the internal virtual work for the whole structure is — 3, N, de, as indicated in Eq. (13). The dummy unit-load formula givea by Eq. (23) can therefore be written as in 4=D [me + ii (me + vy + 19) de (25) . eno for the special case of planar bending, in which the (constant) axial force n, bending moment m, shear force v, and torsional moment ¢ are the internal force system associatedBeam Deflections 1-19. with the dummy unit load; and e, «, 7, @ are the actual deformations causing the gener- alized deflection A, The summation is over all members in the structure. The deflection formula, Eq. (25), separates the total deflection into distinct compo- nents; axial length changes, and flexural, shear, and torsional deformations. Furthermore, the actual deformations need not be caused by loads—they may be due to, for example, temperature changes in the structure. In the case when the deformations are produced by loads, the actual deformations are related to the actual internal forces N, M, V, T by NL M (eel AE i ac q The effective shear area A, which appears in Eq. (26) accounts for the nonuniform dis- tribution of shear stress over the beam cross section. For rectangular cross sections, it is often taken as A, = A/1.2. For WF sections, a good approximation is A, = Aye. The treatment of torsional deformations does not include warping torsion effects. e (26) Example. To calculate the vertical deflection at midspan of the simple, statically determinate truss shown in Fig. 10, a unit load is placed at point C as shown, For trusses, Eq. (25) reduces to mee 7) If the member elongations are due only to member internal forces, then Eq. (27) specializes further to ny Nil - asda (28) The calculations are summarized in Table 2. The member length changes due to the actual load are shown in the column headed “‘e.” The member forces ny due to the dummy unit load are shown in the column headed ‘'n.” The quantity ng can be interpreted as the vertical deflection of joint C due toa unit length change of member k with all other members remaining unchanged in length. Thus if member AF were to be elongated by, say, 0.1 in (due to any cause), then joint C would deflect verti- cally by an amount (—0.707)(0.1) ='—0. (the negative sign indicates that the deflection is in 20K 20K 20K ef VG ti El sti Nx by jongations e = SACK 7 MS BE Member forces nq c v 1 (b) Fig. 10 Truss deflection by dummy unit-load method. (a) Actual load. (b) Durnmy unit load.1-20 Structural Analysis TABLE 2 Deffection of Truss of Fig. 10° Member N, ips —_L,in n ne, in AB,DE — +20.0 120 Ls +05 +0.040 BC,CD — +40.0 120 20 +10 +0.080 FG,GH 30.0 120 2.0 -0.5 +0.030 AR EH ~42.4 1697 9 25 =0.707 +0.068 BG.DG ~141 1697 10 -0.707 +0.057 BE,DH — +10.0 120 1.0 +05 +0.020 +00 120 10 +10 ++0.000 ne +0.59 E = 30,006 ksi for all members. the opposite direction to the unit load, ic., upward in this case). The vertical component of the deflec- tion of joint Cis calculated as 0.59 in. Example, The calculation of deflections due to flexure is illustrated by the simple portal frame shown in Fig. 114. The horizontal deflection of joint C is sought. Therefore a duromy unit load is placed at C in the horizontal direction (Fig. 11). If the contributions of axial deformations and shear deformations to the dellection are disregarded, Eq. (25) becomes nM a-¥( finda) (29) in which M/EI are the member curvatures due to internal bending moments, and m are the bending moments corresponding to the dummy unit load. The flexural stiffness EI is taken as constant and is the same for all members. The curvatures are shown in Fig, 11, and the bending moments due to the dummy unit load in Fig. Le. Curvatures Moments m (b) (ec) Fig. 11 Frame deflection by dummy unit-load method.4-21 [Gwen tue + 2 | [owe way tue ‘ . - (rug Ou) TW | (hu -Ou) TN Cuz sou t= | (tw w2 TN | (bu Ow) WA (w—ews)ou] ew+eweyew] & wow) te ates aww | ow+twartw? awit eww wud w= engi ew 2 | a HyowZt Tew = | w+enz) row? Thou & ewe aww & a 1 Cn-ewewe nw = aww wenn Jwue eww aww w ae a a a q ow w A u)ou0 ow as ‘swoubo1p Ww 21)090u04 ‘suosbo1p |W 109U!7 yosenen f aTavL1-22 Structural Analysis: Values of typical integrals which appear in Eq. (29) are given in Table 3. For the frame of Fig. 11, they have the values aoHl"/EI for AB; wLE®/6EI for BC; and 0 for CD. Thus the horizontal component of deflection of joint C arising solely from bending of the frame is Ac = (&wH'VEN(1 + 0.8L/H). ANALYSIS OF STATICALLY INDETERMINATE STRUCTURES. The two fundamen:al methods of structural analysis are the force method and the dis- placement method. The force method is also sometimes called the compatibility method or the flexibility method. The displacement method is also sometimes called the equilib- rium method or the stiffness method. 14, Comparison of Force and Displacement Methods The force method and the displace- ment method are dual methods. Each step or procedure in one has its mirror-image equiv- alent in the other. In the force method, equilibrium is satisfied at all stages, and the gov- ering equations of the method result from the imposition of the conditions of compatibility. In the displacement method, compatibility is satisfied at all stages, and the governing equations of the method result from the imposition of the equations of equilib- rium. The force method is closely related to the principle of minimum complementary potential energy, ard the displacement method is closely related to the principle of min- imum potential energy. The displacement method is the basis for the overwhelming majority of production computer codes for structural analysis, because of a number of practical computational advantages. Nevertheless, the concepts embodied in the force method are fundamental to a full understanding of structural theory and analysis. The three-span continuous beam shown in Fig. 12 illustrates the fundamental concepts of these methods. Force Method. Consider a structure which is statically indeterminate to the nth degree. For the beam shown in Fig. 12a, n = 2. A statically determinate stable structure, Pou Ue XRy kay xu 270 _ k : XRe . wa Z Laue he R,=0 Usz0 (al (b) Fig. 12 Comparison of (a) force and (b) displacement methods.Comparison of Force and Displacement Methods 1-23 the primary structure, is produced by introducing sufficient releases. The generalized forces corresponding (in a work-conjugate sense) to these releases are the redundants. ‘The primary structure is now subjected to n + 1 loading cases: the first of these is the external loading condition; each of the remaining n loading cases corresponds to a unit value of one redundant with all other redundants set equal to zero. The generalized dis- placements at the releases in each of these loading cases are defined as the flexibility coefficients. The flexibility coefficent f, is the generalized displacement work-conjugate to redundant R, due to a unit value of redundant R, with all other redundants equal to zero. ‘A linear combination of the n + 1 loading cases is constructed to give expressions for the total generalized displacement at each release, as indicated in Fig. 12a. For this exam- ple these two generalized displacements are the vertical displacements at supports C and D, which must be zero in the indeterminate beam. The equations for this example are Auk, + fuk + Dy fuRy + faaR, + Ds Thus the equations of the force method express compatibility conditions which are sat- fied by choosing the appropriate combination (j.e., the values of the redundants R,) of the n loading cases to superpose on the external load case. In general, the equations of the force method have the form F-R+D=D" (31) or F-R=D where F = symmetric, positive definite structure flexibility matrix D=D'-D-D = generalized displacement of primary structure due to external load D' = specified generalized displacement at releases (often equal to zero) Displacement Method. Consider a structure which is kinematically indeterminate to the nth degree. This means that there are n unknown degrees of freedom U,. Kinematic indeterminacy in the displacement method is the counterpart to static indeterminacy in the force method, In the beam shown in Fig. 12b, the degree of kinematic indeterminacy (the number of degrees of freedom) isn = 2. There is, in general, no particular correla- tion between the degree of indeterminacy of a structure and the number of degrees of freedom which it has. The first step in the displacement method is to produce a kinematically determinate structure by fixing all degrees of freedom. This fixed-end structure is the entity in the displacement method which is the dual of the primary structure in the force method. It is usually highly statically indeterminate. For convenience, external loads are separated into two types of loads: element loads, which are applied directly to the members, and joint loads which act directly on the joints, The fixed-end structure is now subjected to n + 1 load cases. ‘The first consists of the external member loads acting on the fully restrained structure. Each of the remaining n loading cases corresponds to a unit value of one of the degrees of freedom with all other degrees of freedom equal to zero. The generalized forces at the degrees of freedom in each of these loading cases are defined as the stiffness coefficients. The stiffness coefficient k, is the generalized force work-conjugate to the degree of freedom U, which is due to a unit value of degree of freedom U, with all other degrees of freedom equal to zero. A linear combination of the n + 1 cases is constructed to give expressions for the total generalized force at each degree of freedom. The element loads give rise to a set of gen- eralized forces P, often called the fixed-end forces or fixed-end moments. The general- ized force at each degree of freedom must be equal to the specified (externally applied) generalized force—the joint load. In the example shown in Fig. 12b, the two degrees of freedom are joint rotations at B and C. The equations are (30) kU, + kyla + By = Pt 0 (32) 0 kyiU, + kyaUs + Py = Pt Thus the equations of the displacement method express equilibrium conditions which are satisfied by choosing the appropriate combination (i.e., the values of the degrees of free- dom U) of the n loading. cases to superpose on the external load case.1-24 Structural Analysis In general, the equations of the displacement method have the form K-U+P=P° (33) K-U=P where K = symmetric, positive definite structure stiffness matrix P = P’—P - P = generalized force in fixed-end structure due to element loads P° = specified (externally applied) generalized force at degrees of free- dom, i.e., joint loads Force Method The steps involved in the force method are (1) selection of redundants; (2) calculation of generalized displacements in the primary structure for the external loading Dy; (3) cal- culation of the generalized displacements (the flexibility coefficients f,) for the n unit redundant loading cases; (4) imposition of compatibility conditions work-conjugate to the redundants; (5) solution of the simultaneous equations for the values of the n redundants Rg (6) back substitution for member forces and deformations. Difficulties arise in setting up a procedure for making a standardized, systematic choice of redundants for general structures. This has led to a strong preference for displacement methods in production computer codes for structural analysis. Nevertheless, for specific classes of structures, the force method can be very effective and efficient. 15. Structure Flexibility Coefficients The generalized displacements D, and f; of the pri- mary structure can be calculated using any method for calculating displacements in stat- ically determinate structures. Both direct methods and work-energy methods are avail- able. The dummy unit-load method is illustrated in the following article. 16. Dummy Unit-Load Method Exampi Statically Indeterminate Truss. The planar truss shown in Fig. 13 is two degrees stat- ically indeterminate (r = 3, j = 8, b = 15; b + r — 2 = 2). The redundants R, and R; are chosen as the internal bar forces (tension positive) in members CF and CH. Therefore, releases are inserted in members CF and CH, consisting of infinitesimal cuts. The gener- alized displacements which are work-conjugate to the redundants are therefore the rel- ative displacements (i.e., opening or closing) of the cuts. Whenever internal generalized forces (as opposed to external reactions) are chosen as redundants, the corresponding generalized displacements are relative quantities of this type. The relative displacements due to the external ‘oads are D, and D>. = In order to calculate the relative displacement Dy the required dummy unit-load system is a pair of unit forces which induce bar forces n}. Thus D, = Enle, where ¢, are the member elongations in the primary structure due to the external actions. If these external actions consist only of mechanical loads, then e, = N,L/A,E,, where N, are the bar forces due to the external loads. The same dummy unit-load system is used to calculate the rel- ative displacements f,; and f;2, due to the loadings R, = 1, Ry = 0 and R, = 0, R = 1, respectively. This dummy unit-load system happens to be the same as the load system for R, = 1, Ry = 0. The relative displacements D,, fy, and fyy are calculated in an analogous ‘way using a dummy unit-load system consisting of a pair of unit forces which induce bar forces n?. This second dummy unit-load system happens to be the same as the load system for R, = 0, Rz = 1, The formulas for the flexibility coefficients shown in Fig. 13 verify that fi > 0, for > 6, and fis = fx. Table 4 summarizes the calculations. The redundants R, and R, are found from 0.01732 0.00200] | R,| _ | 0.1224 0.00200 0.01732] | Ry 0.1224 [i] -[s2 | ‘The total member forces in the indeterminate truss are found as a final step from N, = + njRy + niRy, Statically Indeterminate Frame. ‘The portal frame shown in Fig, 14 is statically inde- terminate to the first degree. The horizontal thrust at the base of the columns is chosen as the redundant R; (the sign convention chosen is such that a negative sign for Ry in cates a positive thrust). The dummy unit-load system which is required in order to cal-Dummy Unit-Load Method Examples 1-25 20K 20K 20K et te ba PS ao. U8 Dummy unit loadings ee : a Elongations “ 5 R,=0 Ry 0 TABLE 4 Statically Indeterminate Truss (Fig. 13) by Force Method" ee ee eee ae a Ar fee 4B 120-15 4300 0.080) +300 120 15 +300 0.080 $30.0 120 20 +400 —0.707 089 —0.0566 0.001 $355 120 20 +400 0.707 0.080 0.0566 0.001 +355 120 20 ~300 -0.707 0.060 0.0424 0.001 rots 129_ 20 —300 0.107 —0.060 0.0424 o.oo =345 169.7 25 424 =0.098 = 424 169.7 25 424 =0.096 =a 169.7 10-141 +10 0.080 ~0,0800 0.00566 18 1697 V0 =14. +10 0.080 —0,0800 0.00366 78 120° 1.0 4100 ~0.707 0.040 0.0283 0.002 $55 1201.0 +100 0.707 0.040 0.0283, 0.002 +55 120. 140 +00 —0.707 =0.707 0.000 0.002 0.002 0.002 90 169.7 10 + 00 +10 0.000 0.00566 +63 169.7 1.0 + 00 +10 9.000 0.00566 4.5.3 1224 “0.1204 OOITS® 6.008 0.01732 aoe De “Force units are kips length units are faches; E = 30,000 kat1-26 Structural Analysis culate the horizontal displacements D, and f;, at the support D in the primary structure consists of a unit horizontal force at D. If only flexural deformations of the frame are taken into account, the horizontal displacement at the support D in the primary structure can be calculated as indicated in Fig. 14. Table 3 gives formulas for integrals [MM dr. Using these results, D, = wL°H/12E,,, and f,, = 2H°/3EJ, + H*L/E,J,. Thus the recundant force R, is oe oe eer in which a = (Eyly/L)/(E.1,/H) is a measure of beam-to-column relative stiffness. The max- imum positive and maximum negative bending moments in the frame are shown in Fig. 15 as a function of the beam-to-column relative stiffness. Shear and axial deformations can be taken into account by adding the appropriate terms shown in Eq, (25). 47. Three-Moment Equation for Continuous Beams The three-moment equation is a special force-method solution for continuous beams in which the bending moments over the inte- rior supports are chosen as redundants. The general form of the equation relates bending moments in the continuous beam at three adjacent interior supports. This results in a standard approach for continuous beams which can be easily applied and which gives well-conditioned simultaneous equations with a narrow bandwidth (the bandwidth is 3). H Curvatures Let wi = Sept Moments, Km Ry=1 Fig. 14 Statically indeterminate portal frame by force method.‘Three-Moment Equation for Continuous Beams 1-27 ive moment Negative moment 0.0 02 0.4 06 08 1.0 a Ta Beom—to-column stiffness ratio Fig. 15 Effects of proportions on maximum bending moments in hinged-base portal frame. ‘The primary structure formed by releasing internal bending moments over all interior supports consists of a number of adjacent simply supported spans (Fig. 16). Therefore, the basic building block in this approach is a simply supported beam, and the generalized displacements D, and f, of the force method are easily constructed from calculated rota- tions at the end supports, A small core of standard results suffices for many problems of practical interest (Fig, 17). In this regard the three-moment equation possesses the advantageous characteristics usually associated with displacement methods The example shown in Fig, 16 is'a four-span continuous beam with equal span lengths Land constant stiffness EI in all spans, The external load P is applied at the center of the end span AB. The flexibility coefficients f,,, fu, fare the relative rotations at the releases for the unit-load system R, = 1, Ry = 0, Ry = 0 (Fig. 16c). These are found from Fig. IT as fir = L/SEI + L/3EL, fy = L/GEI, fy, = 0. The two contributions to fi, are the end rotations at B of spans AB and BC, respectively, under the pair of unit moments Ry = 1 Ifthe spans AB and BC have different lengths and elastic properties, then fi, = (L/3ED ay + (L/3ENac far = (L/GEDgc, and fy, = 0. The generalized displacement Dy of the pri- mary structure is the rotation at B of span AB, due to the external load P. This is given in Fig. 17 as PL?/16EI. Additional loads are treated in the same way. Once the redundants R, are calculated, the bending moment diagram for the continuous beam is constructed using superposition, as shown in Fig. 16. Displacement Method The steps involved in the displacement method are (1) identification of the degrees of freedom; (2) calculation of generalized forces P, in the fully restrained structure due to element loads (the fixed-end forces); (3) calculation of generalized forces (the stiffness cocfiicients k,) due to the n unit degree of freedom displacement cases; (4) imposition of equilibrium conditions work-conjugate to the n degrees of freedom (assembly of stiffness equations); (5) solution of the simultaneous equations for the values of the n degrees of freedom U, (6) back substitution for the member forces and deformations.1-28 Structural Analysis The displacement method is generally preferred over the force method as the basis for general-purpose structural analysis computer codes. This is primarily due to the fact that the operations associated with the displacement method can be systematically compart- mentalized so that all structural elements, whether they represent trusses, frames, plates, shells, or solids, can be handled in the same way and the same procedures and algorithms can be utilized. Details which are specific to a particular type of structural member can be confined to a small separate part of the computer code. 18. Structure Stitiress Coefficients The stiffness coefficients ky of a structure are deter- mined by direct summation of stiffness coefficients of individual members. This is illus- trated by the planar truss shown in Fig. 18. The two degrees of freedom U;, U; are defined as the horizontal and vertical displacements, respectively, of joint 4 of the truss. The stiffness coefficients are determined from the basic relations for the truss shown in. Eqs. (1) to (3). Compatibility relations shown in Eq. (2) determine member elongations given the imposed joint displacements; constitutive relations (clastic behavior) shown in Eq. (3) determine member forces from elongations; and equilibrium relations shown in Eq, (1) determine the stiffness coefficients from the member forces. ‘The stiffness coefficients are displayed in Fig. 18 in matrix form, in which the separate 2U3EL 6 0 Ri) [prec veel 2u38 Yoel] |R, 0 0 WOE! Ry 0 1 = 48) [10004 Ra|= PE! 39].) casser 3} |-.0067e1] (d) + 0.0268PL, - 0.0067PL }004 PL 9.1998PL a) 0.0268PL 0.1004 Samoa © Fig. 16 ‘Three-moment equation for continuous beam: (a) 4-span continuous beam, (5) mary structure, (c) unit-load systems, (d) compatibility equations of force method, (c) constn tion of bending-moment diagram.Member Stiffness Matrices 1-29 contributions of the three members appear as member stiffness matrices. Member 2 is oriented at an angle a with respect to the cartesian coordinate system (x, y). Member stiffness matrices are usually expressed in a local or element coordinate system, which in this case would have a local x-axis pointing along the member axis. A general coordinate transformation is then used to describe member stiffness information in the global coor- dinate system. 49. Member Stiffness Matrices The member stiffness matrix is the basic building block of the displacement method. Member stiffness information’ is required for every type of structural member (model) from which a given structural model is to be synthesized. Planar Truss. The stiffness matrix shown in Fig. 19 for a general planar truss member is referenced to four degrees of freedom in order to cover the most general situation. The four degrees of freedom u{, uj, uj, u; are defined with respect to the member axes (the element coordinate system). The truss member has no stiffness associated with displace- ments normal to its centerline as long as the displacements are small. The transformation matrix T expresses the relationship between the displacements in the two coordinate sys- tems asu’ = T- u, Since u’™ - p’ = u? - p, the relation between forces in the two coordinate systems fol- lows. The transformation w=T-u p=T-p (34) is known as a contragradient transformation. The stiffness matrix of the member in the global coordinate system is then expressed as keT-k-T (35) For a planar frame the degrees of freedom at a joint include a rotation as well as two displacements. The transformation matrix T is then a 6 X 6 rather than a4 X 4 matrix, which is obtained from the matrix shown in Fig. 19 by inserting rows and columns cor- responding to degrees of freedom 3 and 6, with 1 on the diagonal and 0 elsewhere. Planar Frame. Meraber response in the displacement method is separated into two component parts: (I) the response of the fixed-end member (i.e., with member degrees of freedom constrained) subjected to member loads, and (2) the response of the unloaded member to the degrees of freedom. The stiffness coefficients are related to this latter part of the member deformation. Either the direct approach or the work-energy approach can be used to develop the stiffness information for one-dimensional members (line ele- ments). For plates, shells, and solids, however, the work-energy approach is indispensable. 1 1 L/eer L/BEL L/3EL L/6EI Aaa) L, EL w Pp wey ve “w32aer = PL2v16E1 CPLR NBEL a wih/eael | er LL Pab(a+2b) L,EI = Pab (b+2a) 6EIL 6EIL Fig. 17 Support rotations in a simply supported beam,1-30 Structural Analysis u Joint (a) (b) [ ,]_ Ag i o], Aaée [eee cota soa Ades [2 J ee | eC LO Ol ace tran geet |earea|D 1 (c) Fig. 18 Structure stiffness coefficients: (a) Unit degrees of freedom and definition of struc- ture stiffness coefcients.(b Joint equilibrium, (c) Structure stfivess assembled from mem- ber contributions, The stiffness coefficients for a plane beam can be developed using the approach illus- trated in Fig. 20. The degrees of freedom corresponding to axial displacements are not shown since the axial and bending responses uncouple from each other. The determina- tion of the stiffness coefficients involves the solution of a statically indeterminate beam, which can be carried out by the force method. In order to determine the stiffness coef. ficients ks and k,s, the force method gives [ee] le] - [3] and for the determination of the stiffness coefficients ky, and ky, le ellel-[2]000 Element coordinate system (x's ¥) ile cee 4) Global coordinate system (x, y) ae eae up ce cosa, $= sna Wye G 0 ee | jus 00 (@) (b) Fig. 20. Coefficients for plane beam: (a) stiffness, (b) flexibility. 1311-32 Structural Analysis In these two sets of simultaneous equations, the flexibility coefficients are assumed to be known and the stiffness coefficients are sought. The coefficient matrix (the flexibili matrix) is the same in both cases, and so they can be combined into Ui Allis l-[ot] © ome from which it is found that ky, = fj!. The flexibility coefficients can be found by using the dummy unit-loed method in which, for example, shear deformations can be included as well as flexural deformations. The full 4 X 4 stiffness matrix is found by using equilibrium and the fact that the sti ness matrix must be symmetric, Equilibrium requires that ky = Tas ~ kuy and kye = T, sa, where Te = |# 4) and symmetry requires that kq, = kf,. Therefore the 4 X 4 stiffness matrix for the planar beam is a | TakTh, Toke 5 k ae re (36) Figure 21 shows the explicit form of this stiffness matrix expanded to a 6 X 6 matrix to include the axial degrees of freedom. Direct Method for Determining Member Response. The response of one-dimensional (line) elements can also be determined directly from the differential relations based on equilibrium, compstibility, and the constitutive law. For a planar beam in which only flexural deformations are considered, Eqs. (20) and (22) are the starting point. If shear deformation effects are to be considered, the second differential relation in Eq, (22) is modified to include the transverse shear strain y, which is assumed to be constant through the depth of the beam. In addition, the constitutive relation between shear force and shear strain is needed. The result is (37) In this theory, sometimes called Timoshenko beam theory, the rotation 0 of the line which is originally normal to the centerline differs from the rotation dy/dx of the centesline by the shear angle y. The solutions of Eqs. (20) and (37) give the complete response of the beam. The por- tion of the response involving the applied load p(x) determines the fixed-end forces. The portion of the response for p(x) = 0 determines the stiffness coefficients. Figure 22 shows the solutions of Eqs. (20) and (37) for p(x) = 0 and with nodal degrees of freedom w, = (0), uz = 6(0), us = y(L), u = O(L). The internal generalized beam forces are determined from Eq. (37) as V(x) = A,Gy(x), M(x) = EI dé(x)/dx, and the nodal generalized forces are determined from p; = V(0), ps = —M(0), ps = — VIL), pa = M(L). The coefficients of the generalized coordinates ty, us, tty in these expressions for ps, Po, Ps, Pa are the stiffness coefficients shown in Fig. 21. Axial displacements are included by expanding the stiffness matrix and inserting the axial stiffness coefficients as in Fig. 21 The analytical expressions for the beam deformations y(x), @(z), (x) in terms of the degrees of freedom involve shape fictions, shown in Eq, (38) for the beam deflection u(x). 4 u(x) = IM, NSN) [:] (38) uy