You are on page 1of 426

DESIGN OF HIGHWAY OVERHEAD CANTILEVER-TYPE SIGN SUPPORT

STRUCTURES FOR FATIGUE LOADS

by

IAN E. HOSCH

FOUAD H. FOUAD, COMMITTEE CHAIR

TALAT SALAMA

VIRGINIA SISIOPIKU

HOUSSAM TOUTANJI

NASIM UDDIN

A DISSERTATION

Submitted to the graduate faculty of The University of Alabama at Birmingham, in partial fulfillment of the requirements for the degree of Doctor of Philosophy

BIRMINGHAM, ALABAMA

2009

Copyright by

Ian E. Hosch

2009

DESIGN OF HIGHWAY OVERHEAD CANTILEVER-TYPE SIGN SUPPORT STRUCTURES FOR FATIGUE LOADS

IAN E. HOSCH

DEPARTMENT OF CIVIL, CONSTRUCTION, AND ENVIRONMENTAL ENGINEERING

ABSTRACT

The 2001 edition of the American Association of State Highway and Transporta-

tion Officials (AASHTO) Standard Specifications for Structural Supports for Highway

Signs, Luminaires and Traffic Signals has been revised in its entirety through a major re-

search project conducted under the auspices of the National Cooperative Highway Re-

search Program (NCHRP 17-10). A major part of the revision included updated provi-

sions and criteria for extreme wind loads and new provisions and criteria on fatigue de-

sign. These provisions differ considerably from those in previous editions of the specifi-

cations, and have remained relatively unchanged in the 2009 edition.

The impact of the fatigue criteria on the design of overhead sign structures has not

been fully evaluated. The fatigue design loads do not adequately reflect the stresses gen-

erated on these structures from wind-induced fatigue loading. In addition, the provisions

do not account for the variety of support structures in design, each with different configu-

ration, sizes, shapes, and material properties that influence vibration behavior. As a re-

sult, the vulnerability of sign support structures to wind-induced fatigue loading is not

fully realized.

The main goal of the project was to conduct theoretical and experimental pro-

grams of study to evaluate the performance of cantilever-type highway overhead sign

support structures subjected to wind induced fatigue loads. A theoretical program was

iii

developed that took into account the variety of sign supports structures in design, as well

as addressed the vulnerability of these structures to wind-induced fatigue loading. The

experimental program was developed to evaluate the accuracy of the theoretical study. A

finite element analysis program was conducted to simulate the wind-induced loading en-

vironment and the response of sign support structures to this environment. Alterations

were made to the model that could not be done experimentally due to costs and time re-

straints. The results of the finite element analysis were compared to the theoretical and

experimental programs of study.

The developed information and criteria on fatigue design of sign support struc-

tures were used to develop fatigue design loads to provide an improved and more reliable

design method. Recommendations are made to update the current specifications to in-

clude more reliable fatigue loads.

Keywords: Cantilever-type sign support structures, fatigue design, fatigue load, natural

wind, truck-induced wind, vibration

iv

DEDICATION

The dissertation is dedicated to friends and family, most importantly to my

mother Sandra J. Haigh who gave me strong will and unconditional support during my

time as a student.

v

ACKNOWLEDGEMENTS

It was important to acknowledge the following people who helped with the re-

search preformed in this project:

Dr. Fouad H. Fouad who was my mentor and advisor,

Richard Hawkins and Dr. Jason Kirby who helped with the important technical

experimentation aspects and advice with presentations,

Daniel Jones and Ricky Love of the Alabama Department of Transportation who

helped with the instrumentation and field testing,

The Morrison Family, Brett and Dave, who helped specifically for making the ex-

tensions needed for the anemometers during field testing, and

Andrew Sullivan for setting up and reviewing the traffic recorders required with

the truck testing.

vi

TABLE OF CONTENTS

 

Page

ABSTRACT

 

iii

DEDICATION

v

ACKNOWLEDGMENTS

vi

LIST

OF

TABLES

xiii

 

LIST OF FIGURES

xvii

LIST

OF

ABBREVIATIONS

xxvi

CHAPTER

 
  • 1 INTRODUCTION

1

Problem Statement

1

Objective ..................................................................................................................

4

Study Initiatives

5

Work Plan

6

Summary

6

 

Project Tasks

7

11

Dissertation Outline

12

  • 2 LITERATURE REVIEW

16

Chapter Overview

16

16

Research

Review

19

DeSantis and Haig (1996)

19

 

Davenport (1961)

20

Kaczinski et al (1998)

22

 

Fouad et al (1997-2005)

23

Dexter and Ricker (2002)

24

 

Cook et al (1996)

25

vii

 

Page

Creamer et al (1979)

26

  • 3 FATIGUE PROVISIONS OF THE AASHTO SUPPORT SPECIFICATIONS

27

Chapter Overview

27

 

.........................................................................

27

Infinite-Life Approach

27

Predicting the Environment

29

Structural Excitation

35

Design Fatigue Equation for Natural Wind.....................................................36

Fatigue Load due to Truck-Induced Gust

37

  • 4 SIGN STRUCTURE

INSTRUMENTATION

39

Chapter Overview

39

39

Geometric Properties

 

40

Material Properties

42

 

43

Strain Gauges

47

Truss Chord Members......................................................................................48

Post Support

50

Anemometers

57

Accelerometers

........................................................................................

60

Data Acquisition System

63

  • 5 STRUCTURAL TESTING FOR NATURAL WIND

AND TRUCK-INDUCED WIND GUST

68

Chapter Overview

68

68

Pre-Determined Sample Size

68

Testing Day Ranking Schedule

69

Test Procedure

71

72

Test Procedure

74

  • 6 FATIGUE RESISTANCE

77

Chapter Overview

77

78

Failure Index

78

79

viii

 

Page

General Description of the Anchor Bolts

80

Fatigue Stress due to Natural Wind Gust

83

Fatigue Stress due to Truck-Induced Wind Gust

89

Evaluation of the Fatigue Stress in the Post-to-Base-Plate Connection ................93

General Description of the Post-to-Base-Plate Connection

93

Fatigue Stress due to Natural Wind Gust

95

Fatigue Stress due to Truck-Induced Wind Gust

103

Evaluation of the Fatigue Stress in the Chords

108

General Description of the Chords

108

Fatigue Stress due to Natural Wind Gust

112

Fatigue Stress due to Truck-Induced Wind Gust

119

Discussion of Results

123

  • 7 EXPERIMENTAL MODAL ANALYSIS

128

Chapter Overview

128

128

129

134

Spectral Analysis

135

Modal Shapes

140

Frequency Response Function

141

Modal Damping

149

  • 8 EXPERIMENTAL CALCULATION OF THE FATIGUE LOAD

DUE TO NATURAL WIND GUST

..........................................................................

156

Chapter Overview

156

156

Structural Excitation

158

Collected Sample Size

158

Usable Data Collected

159

Reduction of Structural Excitation Experimental Data

161

Structural Response

165

Data Offsetting

166

Strain Ranges

172

174

Theoretical Structural Analysis

175

Development of the Stress Element

184

Wind Pressure Calculation

189

191

Infinite-Life Approach

195

ix

 

Page

9

THEORETICAL CALCULATION OF THE FATIGUE LOAD

DUE TO NATURAL WIND GUST

..........................................................................

197

Chapter Overview

197

Significance of the Theoretical Program

197

Methodology

200

 

200

 

Supports Specifications ..................................................................................201

Experimentally Collected Data

 

207

Comparison between the Supports Specifications

and the Experimental Excitation

 

213

Structural Response

215

 

Response Power Spectral Density and the Root-Mean-Square

215

Vibration Response Spectrum

..............................................................................

220

 

Natural Frequency

222

Critical Damping Percentage

224

Peak-to-Peak Stress Range

226

Infinite-Life Approach

228

 

Vibration Response Spectrum for the Supports Specifications

234

  • 10 EXPERIMENTAL CALCULATION OF THE FATIGUE LOAD

DUE TO TRUCK-INDUCED WIND GUST

 

235

Chapter Overview

235

Fatigue Load Calculation Approach

235

Collected Sample Size

236

Truck Specimen

237

Structural Excitation

238

Structural Response

240

 

Stress

Range

241

 

244

 

Theoretical Structural Analysis

 

245

Development of the Stress Element

252

Wind Pressure Calculation

255

  • 11 THEORETICAL CALCULATION OF THE FATIGUE LOAD

DUE TO TRUCK INDUCED WIND GUST

 

258

Chapter Overview

258

Research Significance

258

Methodology

261

 

262

265

267

x

 

Page

  • 12 FINITE ELEMENT ANALYSIS

274

Chapter Overview

274

Model Development

.............................................................................................

275

Geometry

275

Element Type

279

Material Definition

279

280

Comparisons of FEA to Experimental Modal Analysis

282

Finite Element Structural Analysis

283

Natural Wind Gust

284

Truck-Induced Wind Gust

289

293

Natural Wind Gust

294

Truck-Induced Wind Gust

304

  • 13 PROPOSED FATIGUE PROVISIONS

311

Chapter Overview

311

Fatigue Load due to Natural Wind Gust

311

Discussion of Results

311

General Fatigue Design Equation for Natural Wind Gust

315

Detailed Fatigue Design Equation for Natural Wind Gust

316

Fatigue Load due to Truck-Induced Wind Gust

320

Discussion of Results

321

General Fatigue Design Equation for Truck-Induced Wind Gust

322

Detailed Fatigue Design Equation for Truck-Induced Wind Gust

324

Design Examples

327

Supports Specifications Fatigue Design Equations .......................................328

Natural Wind Gust

 

329

Truck-Induced Wind Gust

335

  • 14 SUMMARY AND CONCLUSIONS

338

Chapter Overview

338

Fatigue Resistance

340

Modal Analysis

343

344

General Fatigue Design Equation for Natural Wind Gust

345

Detailed Fatigue Design Equation for Natural Wind Gust

346

Fatigue Load due to Truck-Induced Wind Gust

351

General Fatigue Design Equation for Truck-Induced Wind Gust

351

Detailed Fatigue Design Equation for Truck-Induced Wind Gust

352

Finite Element Analysis Method

356

xi

 

Page

 

Contributions to the Practice

357

Recommendations for Future Research

358

LIST OF REFERENCES

359

APPENDIX

A INSTRUMENTATION IDENTIFICATION AND LAYOUTS

365

B

ANEMOMETER MOUNTING INSTRUCTIONS

374

C

ACCELEROMETER MOUNTING INSTRUCTIONS

376

D

RANKING SCHEDULE FOR NATURAL WIND TESTING

383

E

TRUCK-INDUCED WIND GUST TESTING PROCEDURE

396

xii

Table

LIST OF TABLES

Page

  • 2.1 Terrain Coefficients (13) -------------------------------------------------------------------- 21

  • 3.1 Terrain Coefficients (13) -------------------------------------------------------------------- 30

  • 4.1 Material Properties -------------------------------------------------------------------------- 43

  • 6.1 Supports Specifications Constant-Amplitude Fatigue Thresholds (1) --------------- 79

  • 6.2 Anchor Bolt Clearance Lengths ----------------------------------------------------------- 83

  • 6.3 Estimated Microstrain Range at Fatigue Wind for Anchor Bolts --------------------- 87

  • 6.4 Anchor Bolt Stress Range and Failure Index for Natural Wind ---------------------- 88

  • 6.5 Anchor Bolt Microstrain Ranges from the Truck Tests -------------------------------- 91

  • 6.6 Anchor Bolt Stress Range and Failure Index for Truck Tests------------------------- 92

  • 6.7 Estimated Microstrain Range at Fatigue Wind for Section AA --------------------- 100

  • 6.8 Section AA Stress Range and Failure Index for Natural Wind --------------------- 102

  • 6.9 Section AA (Normal) Microstrain Ranges for the Truck Tests --------------------- 104

    • 6.10 Section AA (Shear) Microstrain Ranges from the Truck Tests --------------------- 105

    • 6.11 Section AA Stress Range and Failure Index for Truck Tests ----------------------- 107

    • 6.12 Estimated Microstrain Range at Fatigue Wind for Chords ------------------------- 115

    • 6.13 Chord Stress Range and Category E Failure Index for Natural Wind ------------- 117

    • 6.14 Chord Stress Range and Category B Failure Index for Natural Wind ------------ 118

    • 6.15 Chord 1 Microstrain Ranges from the Truck Tests ----------------------------------- 119

xiii

Table

Page

  • 6.16 Chord 2 Microstrain Ranges from the Truck Tests ----------------------------------- 120

  • 6.17 Chord 3 Microstrain Ranges from the Truck Tests ----------------------------------- 120

  • 6.18 Chord 4 Microstrain Ranges from the Truck Tests ----------------------------------- 121

  • 6.19 Chord Stress Range and Category E Failure Index for Truck Tests ---------------- 122

  • 6.20 Chord Stress Range and Category B Failure Index for Truck Tests --------------- 123

  • 6.21 Experimental Stress Ranges for Natural Wind Tests --------------------------------- 124

  • 6.22 Experimental Stress Ranges for Truck-Induced Wind Tests ------------------------ 124

    • 7.1 Experimental Natural Frequencies ------------------------------------------------------ 139

    • 7.2 Modal Damping Results ------------------------------------------------------------------ 155

    • 8.1 Collection Dates and Times for Natural Wind ---------------------------------------- 158

    • 8.2 Development of the Wind Directionality Unit Vector ------------------------------- 165

    • 8.3 Exposure Area Breakdown --------------------------------------------------------------- 178

    • 8.4 Drag Coefficients -------------------------------------------------------------------------- 179

    • 8.5 Height Coefficients ------------------------------------------------------------------------

180

  • 8.6 Effective Area ------------------------------------------------------------------------------ 182

  • 8.7 Material Properties at Section AA and Section BB ----------------------------------- 185

  • 8.8 Relevant Stress and Strain Equations --------------------------------------------------- 187

  • 8.9 Example of Wind Pressure Magnitude Calculation at Location A ----------------- 189

  • 9.1 Terrain Coefficients (13) ----------------------------------------------------------------- 201

  • 9.2 Natural Frequency Comparison --------------------------------------------------------- 224

    • 10.1 Order of Truck Runs and Truck Speed ------------------------------------------------- 236

    • 10.2 Peak-to-Peak Strain Ranges ------------------------------------------------------------- 244

xiv

Table

Page

  • 10.3 Exposure Truss Area Breakdown -------------------------------------------------------

248

  • 10.4 Drag Coefficients ------------------------------------------------------------------------- 249

  • 10.5 Effective Area ----------------------------------------------------------------------------- 250

  • 10.6 Material Properties at Section AA and Section BB ---------------------------------- 253

  • 10.7 Relevant Stress and Strain Equation --------------------------------------------------- 254

  • 10.8 Truck-Induced Wind Gust Pressure ---------------------------------------------------- 256

  • 11.1 Natural Frequency Comparison ---------------------------------------------------------

272

  • 12.1 Material Properties ------------------------------------------------------------------------

280

  • 12.2 Natural Frequencies of the FEA Model------------------------------------------------ 281

  • 12.3 FEA and Experimental Modal Analysis Comparison ------------------------------- 283

  • 12.4 Normal Strain from FEA at Section AA ----------------------------------------------- 288

  • 12.5 Comparison of FEA with Experimental Strain Range ------------------------------- 288

  • 12.6 Normal Strain from FEA at Section AA ----------------------------------------------- 292

  • 12.7 Comparison of FEA with Experimental Strain Range ------------------------------- 293

  • 12.8 Modal Analysis of the Aluminum Structure ------------------------------------------ 299

  • 12.9 FEA Comparison for Natural Wind Gust --------------------------------------------- 301

    • 12.10 FEA Comparison for Truck-Induced Wind Gust ------------------------------------- 310

      • 13.1 Experimental and Theoretical Results for Natural Wind Gust --------------------- 312

      • 13.2 Comparison of Results from Experimental and Theoretical Programs ----------- 321

      • 13.3 Design Case Description for Natural Wind Gust ------------------------------------- 330

      • 13.4 Design Case Results for Natural Wind Gust ------------------------------------------ 331

      • 13.5 Design Case Description for Truck-Induced Wind Gust ---------------------------- 335

xv

Table

Page

  • 13.6 Design Case Results for Truck-Induced Wind Gust --------------------------------- 337

  • 14.1 Experimental Stress Ranges for Natural Wind Tests -------------------------------- 341

  • 14.2 Experimental Stress Ranges for Truck-Induced Wind Tests ----------------------- 342

  • 14.3 Results of Back-Calculation Verification --------------------------------------------- 357

xvi

Figure

LIST OF FIGURES

Page

  • 1.1 I-65/I-565 interchange, Exit 340B. --------------------------------------------------------- 3

  • 1.2 Failure of cantilever sign support. ---------------------------------------------------------- 3

  • 1.3 Close-up of fractured anchor bolt.---------------------------------------------------------- 4

  • 1.4 Cantilever-type highway overhead sign structure. --------------------------------------- 7

  • 2.1 Research paper breakdown. ---------------------------------------------------------------- 17

  • 2.2 Referenced papers in the research paper breakdown spreadsheet. ------------------- 18

  • 3.1 Wind velocity PSD for annual mean wind velocity 11 mph. -------------------------- 31

  • 3.2 Force PSD for annual mean wind velocity 11 mph. ------------------------------------ 33

  • 3.3 Force PSD using limit-state wind velocity 37 mph. ------------------------------------ 35

  • 4.1 Cantilever-type highway overhead sign support structure. ---------------------------- 40

  • 4.2 Elevation view of the structure configuration. ------------------------------------------ 41

  • 4.3 Plan view of structure configuration. ----------------------------------------------------- 41

  • 4.4 Side view of structure configuration. ----------------------------------------------------- 42

  • 4.5 Strain gauged post support in laboratory. ------------------------------------------------ 44

  • 4.6 Strain gauged truss overhang in laboratory. --------------------------------------------- 45

  • 4.7 Set anchor bolts in structure foundation. ------------------------------------------------- 45

  • 4.8 Truss overhand attachment. ---------------------------------------------------------------- 46

  • 4.9 Sign attachment. ----------------------------------------------------------------------------- 46

xvii

Figure

Page

  • 4.10 Finished instrumented structure. ---------------------------------------------------------- 47

  • 4.11 Truss-to-pole support strain gauges. ----------------------------------------------------- 48

  • 4.12 Chord strain gauges. ------------------------------------------------------------------------ 49

  • 4.13 Chord strain gauges on the assembled structure. --------------------------------------- 50

  • 4.14 Pole support strain gauges. ---------------------------------------------------------------- 51

  • 4.15 Rosette and uni-axial strain gauges on shaft. ------------------------------------------- 52

  • 4.16 Instrumented post support.----------------------------------------------------------------- 52

4.17

Anchor bolts

before installation. ---------------------------------------------------------- 54

  • 4.18 Anchor bolts and foundation. ------------------------------------------------------------- 54

  • 4.19 Anchor bolt strain

gauges. ----------------------------------------------------------------- 55

  • 4.20 Anchor bolt strain

gauging preparation. ------------------------------------------------- 55

  • 4.21 Strain gauged anchor bolt. ----------------------------------------------------------------- 56

  • 4.22 Instrumented anchor bolts on assembled structure. ------------------------------------ 56

  • 4.23 WindSonic ultra-sonic wind and direction sensor. ------------------------------------- 57

  • 4.24 Overview of anemometer layout. --------------------------------------------------------- 58

  • 4.25 AN-1 and AN-2 above traffic lane.------------------------------------------------------- 58

  • 4.26 Anemometers AN-3 and AN-4 above post. --------------------------------------------- 59

  • 4.27 Accelerometer locations for cantilever structure. -------------------------------------- 61

  • 4.28 Accelerometer location AC-2.

------------------------------------------------------------ 64

  • 4.29 Accelerometer location AC-3. ------------------------------------------------------------ 64

  • 4.30 AC-2 and AC-3 view. ----------------------------------------------------------------------- 65

  • 4.31 Accelerometer mounting block. ----------------------------------------------------------- 65

xviii

Figure

Page

  • 4.32 Mounting block at AC-1. ------------------------------------------------------------------- 66

  • 4.33 Typical test setup with van data acquisition system. ----------------------------------- 66

  • 4.34 Van data acquisition system and computer setup. -------------------------------------- 67

    • 5.1 Truck and driver for truck-induced wind test. ------------------------------------------- 73

    • 5.2 JAMAR Trax Flex HS recorders. --------------------------------------------------------- 75

    • 5.3 Typical truck-induced wind gust runs. --------------------------------------------------- 76

    • 6.1 Anchor bolt connection. -------------------------------------------------------------------- 80

    • 6.2 Anchor bolt layout. -------------------------------------------------------------------------- 81

    • 6.3 Anchor bolt identification. ----------------------------------------------------------------- 82

    • 6.4 All anchor bolt microstrain range vs. wind velocity. ----------------------------------- 84

    • 6.5 Transformed independent variable for AB-8. ------------------------------------------- 86

    • 6.6 Estimated strain range at fatigue wind for AB-8. --------------------------------------- 86

    • 6.7 Estimated Microstrain range at fatigue wind for anchor bolts. ------------------------ 87

    • 6.8 Failure index of anchor bolts for natural wind. ------------------------------------------ 89

    • 6.9 Depth chart of anchor bolt strain range in truck tests. ---------------------------------- 90

      • 6.10 Failure index of anchor bolts for truck tests. -------------------------------------------- 92

      • 6.11 Fillet-welded tube-to-transverse plate connection. ------------------------------------- 93

      • 6.12 Post support strain gauges. ---------------------------------------------------------------- 94

      • 6.13 Strain gauge locations at Section AA. --------------------------------------------------- 94

      • 6.14 45° rosette arrangement with coordinate axis. ------------------------------------------ 96

      • 6.15 Normal microstrain range vs. wind velocity. ------------------------------------------- 97

      • 6.16 Rosette microstrain range vs. wind velocity. ------------------------------------------- 98

xix

Figure

Page

  • 6.17 Transformed independent variable for SGR-AA-11. ---------------------------------- 99

  • 6.18 Estimated strain range at fatigue wind for SGR-AA-11----------------------------- 100

  • 6.19 Estimated microstrain range at fatigue wind for Section AA. ---------------------- 101

  • 6.20 Failure indexes for natural wind of Section AA. ------------------------------------- 103

  • 6.21 Depth chart of Section AA normal strain in truck tests. ----------------------------- 105

  • 6.22 Depth chart of Section AA shear strain in truck tests. ------------------------------- 106

  • 6.23 Failure index for Section AA truck tests. ---------------------------------------------- 107

  • 6.24 Chord-to-plate weld. ---------------------------------------------------------------------- 108

  • 6.25 Chord-to-column bolted connection. --------------------------------------------------- 109

  • 6.26 Chord instrumentation layout. ---------------------------------------------------------- 110

  • 6.27 Chord labeling. ---------------------------------------------------------------------------- 110

  • 6.28 Chord strain gauge labeling. ------------------------------------------------------------- 111

  • 6.29 Chord microstrain ranges vs. wind velocity. ------------------------------------------ 113

  • 6.30 Transformed independent variable for SG-C2-5. ------------------------------------ 114

  • 6.31 Estimated strain range at fatigue wind for SG-C2-5. -------------------------------- 115

  • 6.32 Estimated microstrain range for fatigue wind of chords. ---------------------------- 116

  • 6.33 Failure indexes of chord gauges for natural wind. ----------------------------------- 118

  • 6.34 Depth chart of chord strain in truck tests. --------------------------------------------- 121

  • 6.35 Comparison of results between natural wind and truck gusts. --------------------- 125

    • 7.1 Accelerometer locations for cantilever structure. ------------------------------------ 130

    • 7.2 Accelerometer location AC-2. ---------------------------------------------------------- 131

    • 7.3 Accelerometer location AC-3. ---------------------------------------------------------- 131

xx

Figure

Page

  • 7.4 Overview of accelerometers AC-2 and AC-3. ---------------------------------------- 132

  • 7.5 Mounting block at AC-1. ---------------------------------------------------------------- 133

  • 7.6 Forced event used for modal analysis. ------------------------------------------------- 137

  • 7.7 Structural response to the forced event. ----------------------------------------------- 138

  • 7.8 Modes 1 and 3 from AC-2-Y. ----------------------------------------------------------- 139

  • 7.9 Modes 2 and 4 from AC-2-Z. ----------------------------------------------------------- 140

    • 7.10 Quadrature picking of mode 1. ---------------------------------------------------------- 143

    • 7.11 Mode 1: torsion about support shaft. --------------------------------------------------- 144

    • 7.12 Quadrature picking of mode 2. ---------------------------------------------------------- 145

    • 7.13 Mode2: vertical rocking. ----------------------------------------------------------------- 145

    • 7.14 Quadrature picking of mode 3. ---------------------------------------------------------- 146

    • 7.15 Mode 3: horizontal truss twist. ---------------------------------------------------------- 147

    • 7.16 Quadrature picking of mode 4. ---------------------------------------------------------- 148

    • 7.17 Mode 4: outward and inward clamping. ----------------------------------------------- 148

    • 7.18 Typical truck transient event from accelerometers. ---------------------------------- 150

    • 7.19 Exponential decay of the transient truck event. -------------------------------------- 152

    • 7.20 Trendline

of extracted peak amplitudes. ----------------------------------------------- 153

  • 8.1 Wind rose diagram of all data. ----------------------------------------------------------- 160

  • 8.2 Wind rose diagram of usable data. ------------------------------------------------------ 161

  • 8.3 Coordinate system. ------------------------------------------------------------------------ 164

  • 8.4 Strain values per 0.5 ph wind velocity intervals. -------------------------------------- 167

  • 8.5 Transformed regressor. ------------------------------------------------------------------- 168

xxi

Figure

Page

  • 8.6 Parabolic curve fit.------------------------------------------------------------------------- 169

8.7

Parabolic offset. ---------------------------------------------------------------------------- 170

  • 8.8 90 mph projection using offsetting procedure. ---------------------------------------- 170

  • 8.9 Peak-to-peak range. ----------------------------------------------------------------------- 172

    • 8.10 45° rosette arrangement with coordinate axis. ---------------------------------------- 174

    • 8.11 Area breakdown of the front face. ------------------------------------------------------ 176

    • 8.12 Area breakdown of East side face. ----------------------------------------------------- 177

    • 8.13 Area breakdown of West side face. ---------------------------------------------------- 177

    • 8.14 Height coefficient stepped profile. ----------------------------------------------------- 181

    • 8.15 Free body diagram. ------------------------------------------------------------------------ 184

    • 8.16 Cross section of post at Section AA and Section BB. ------------------------------- 186

    • 8.17 Strain gauge locations at Section AA and Section BB.------------------------------ 188

    • 8.18 Typical stress element. ------------------------------------------------------------------- 188

    • 8.19 Wind velocity vs. wind pressure for Section AA. ------------------------------------ 191

    • 8.20 Wind velocity vs. wind pressure for Section BB. ------------------------------------ 192

    • 8.21 Wind velocity vs. wind pressure for Rosettes. ---------------------------------------- 192

    • 8.22 Wind velocity vs. wind pressure for all sections. ------------------------------------ 193

    • 8.23 Transformed wind velocity vs. wind pressure. --------------------------------------- 194

    • 8.24 Wind velocity vs. wind pressure trendline. ------------------------------------------- 195

    • 8.25 Fatigue load due to natural wind. ------------------------------------------------------- 196

      • 9.1 Highway overhead sign support structures. ------------------------------------------- 198

      • 9.2 Wind velocity PDS for annual mean wind velocity. --------------------------------- 202

xxii

Figure

Page

  • 9.3 Force PDS for annual mean wind velocity. ------------------------------------------- 204

  • 9.4 Force PDS using limit-state wind velocity 37 mph. --------------------------------- 206

  • 9.5 Experimental wind velocity PDS. ------------------------------------------------------ 209

  • 9.6 Logarithmic transformation of average wind velocity PDS. ----------------------- 210

  • 9.7 Best fit line approximating the average wind velocity PDS. ----------------------- 211

  • 9.8 Theoretical plot of the experimental average wind velocity PDS. ----------------- 212

  • 9.9 Approximation of the experimental average wind pressure PDS. ----------------- 213

    • 9.10 Comparison of experimental to Supports Specifications PDS. --------------------- 214

    • 9.11 Structural response to wind pressure excitation. ------------------------------------- 218

    • 9.12 Response of n SDOF systems to common excitation input. ------------------------ 220

    • 9.13 RMS wind pressure VRS for 1.82% damping. --------------------------------------- 221

    • 9.14 First modal shape equal to 1.61 Hz. ---------------------------------------------------- 222

    • 9.15 RMS wind pressure VRS with damping equal to 0.5%. ---------------------------- 225

    • 9.16 Response PDS with damping equal to 0.5%. ----------------------------------------- 225

    • 9.17 VRS of peak-to-peak amplitude for 1082% damping.------------------------------- 228

    • 9.18 Peak-to-peak VRS plots versus average wind velocity. ----------------------------- 229

    • 9.19 Linear transformation of the peak-to-peak VRS. ------------------------------------- 231

    • 9.20 Fatigue load comparison between experimental and VMS. ------------------------ 231

    • 9.21 Fatigue wind pressure VRS for 1.82% damping. ------------------------------------ 232

    • 9.22 Fatigue wind pressure VRS for damping range. -------------------------------------- 233

    • 9.23 Fatigue wind VRS of Supports Specifications (2% damping). --------------------- 234

    • 10.1 Truck and driver for truck-induced wind test. ---------------------------------------- 237

xxiii

Figure

Page

  • 10.2 Anemometer layout and orientation. --------------------------------------------------- 241

  • 10.3 Strain gauge locations at Section AA and Section BB.------------------------------ 242

  • 10.4 Maximum peak-to-peak strain range. -------------------------------------------------- 243

  • 10.5 Underneath exposed horizontal area. -------------------------------------------------- 247

  • 10.6 Exposed underneath truss area breakdown. ------------------------------------------- 248

  • 10.7 Free body diagram. ----------------------------------------------------------------------- 252

  • 10.8 Cross section of strain gauge location Section AA and Section BB. -------------- 254

  • 10.9 Typical stress element. ------------------------------------------------------------------- 255

  • 11.1 Highway overhead sign support structures. ------------------------------------------- 259

  • 11.2 Vertical truck-induced wind gust impulses. ------------------------------------------- 264

  • 11.3 Structural response time history due to the Control impulse. ---------------------- 267

  • 11.4 Response of n SDOF systems to common excitation input. ------------------------ 268

  • 11.5 SRS for vertical truck-induced wind gust. -------------------------------------------- 269

  • 11.6 Second modal shape of cantilever structure. ------------------------------------------ 271

  • 12.1 FEA model of cantilever-type sign support structure. ------------------------------- 276

  • 12.2 Sign-to-truss connection. ---------------------------------------------------------------- 277

  • 12.3 Truss-to-post connection. ---------------------------------------------------------------- 277

  • 12.4 Foundation connection. ------------------------------------------------------------------ 278

  • 12.5 Modal shapes of the first five modes from FEA. ------------------------------------- 281

  • 12.6 Average wind pressure PDS for FEA input. ------------------------------------------ 296

  • 12.7 VRS used for the FEA analysis. -------------------------------------------------------- 298

  • 12.8 VRS for the aluminum FEA model with 0.5% damping. --------------------------- 299

xxiv

Figure

Page

  • 12.9 VRS PDS response to wind pressure excitation. ------------------------------------- 302

    • 12.10 SAP2000 response PDS. -----------------------------------------------------------------

303

  • 12.11 Smaller frequencies of the SAP2000 response PDS. -------------------------------- 303

  • 12.12 Transient excitation function for FEA input. ----------------------------------------- 306

12.13

SRS used for the FEA analysis. --------------------------------------------------------- 308

  • 13.1 VRS for 1.5% damping ratio. ----------------------------------------------------------- 319

  • 13.2 Fatigue wind pressure VRS for damping range. -------------------------------------- 319

  • 13.3 SRS for vertical truck-induced wind gust. -------------------------------------------- 326

  • 13.4 Direction of wind loading for natural wind gust. ------------------------------------- 330

  • 13.5 VRS for Case 1 and Case 2. ------------------------------------------------------------- 331

  • 13.6 VRS for Case 3. --------------------------------------------------------------------------- 332

  • 13.7 VRS for Case 4. --------------------------------------------------------------------------- 332

  • 13.8 VRS for Case 5 and Case 6. ------------------------------------------------------------- 333

  • 13.9 Direction of wind loading for truck-induced wind gust. ---------------------------- 336

  • 14.1 Comparison of results between natural wind and truck gusts. --------------------- 344

  • 14.2 VRS for 1.5% damping ratio. ----------------------------------------------------------- 350

  • 14.3 Fatigue wind pressure VRS for damping range. -------------------------------------- 350

  • 14.5 SRS for vertical truck-induced wind gust. -------------------------------------------- 355

xxv

LIST OF ABBREVIATIONS

AASHTO

American Association of State Highway Transportation Officials

AC

accelerometer

ALDOT

Alabama Department of Transportation

CAFL

constant amplitude fatigue limit

DFT

Discrete Fourier transform

DIA

diameter

DOF

degree-of-freedom

DOT

Department of Transportation

FEA

finite element analysis

FFT

fast Fourier transform

FT

Fourier transform

FRF

frequency response function

NA

not available

NCHRP

National Cooperative Highway Research Program

PDS

power density spectrum

RMS

root-mean-square

SDOF

single degree-of-freedom

SG

strain gauge

SGR

strain gauge rosette

xxvi

S-N

stress vs. number of cycles

SRS

shock response spectrum

UAB

University of Alabama at Birmingham

VMS

variable message sign

VRS

vibration response spectrum

1

CHAPTER 1

INTRODUCTION

Problem Statement

Presently, there is a lack of research involving the development of fatigue loads

specifically for sign support structures. Likewise, there is little knowledge and under-

standing of this subject. In addition, there is a lack of accountability for the variety of

sign support structures in design each with different configuration, sizes, shapes, and ma-

terial properties that influence vibration behavior. Sign support structures are highly

flexible with very low damping properties which make them highly susceptible to wind-

induced fatigue loading, and as such the vulnerability of these structures is not com-

pletely realized. As a result, currently published fatigue design equations do not ade-

quately reflect the stresses generated on these structures from wind-induced fatigue load-

ing.

The 2009 AASHTO Standard Specifications for Structural Supports for Highway

Signs, Luminaires and Traffic Signals (1-6) (hereafter referred to as Supports Specifica-

tions) include fatigue design criteria for cantilevered overhead sign structures that could

significantly impact the design of these structures. Fatigue loadings due to four wind

phenomena are generally prescribed in the specifications:

Natural wind gust,

2

Truck-induced wind gust,

Galloping, and

Vortex shedding.

Fatigue loads due to natural wind and truck-induced wind gusts are applicable to

overhead sign structures and are the focus of this study. Galloping may occur in overhead

sign structures, but 4-chord trusses (which are analyzed in this study) are not susceptible

to galloping. Vortex shedding is not applicable to the overhead sign structures of this

study.

In April 2006, a cantilevered overhead sign structure located at the I-565 and I-65

Interchange, Exit 340B, (Figure 1.1) failed due to fatigue of the anchor bolts. Fatigue

fracture of the anchor bolts due to combined axial stresses and bending in the bolts was

noted. The photographs in Figure 1.2 show the cantilever overhead sign that failed. A

close-up of the bolt fracture is shown in Figure 1.3. The bolt layout was designed using

an earlier version of the Supports Specifications that did not consider fatigue design.

The main goal of this project was to evaluate the performance of cantilever-type

highway overhead sign structures subjected to wind-induced fatigue loads resulting from

natural wind and truck-induced wind gusts, and to develop more reliable fatigue loads for

design. The project involved detailed theoretical as well as experimental programs to ad-

dress this issue.

3

Nashville I-565 I-65 Huntsville Decatur Exit 340B Birmingham
Nashville
I-565
I-65
Huntsville
Decatur
Exit
340B
Birmingham

FIGURE 1.1 I-65/I-565 interchange, Exit 340B.

3 Nashville I-565 I-65 Huntsville Decatur Exit 340B Birmingham FIGURE 1.1 I-65/I-565 interchange, Exit 340B. FIGURE

FIGURE 1.2 Failure of cantilever sign support.

4

4 FIGURE 1.3 Close-up of fractured anchor bolt. Objective The main objectives of this work can

FIGURE 1.3 Close-up of fractured anchor bolt.

Objective

The main objectives of this work can be enumerated in the following three initia-

tives:

  • 1. Perform theoretical and experimental studies to evaluate the performance of canti- lever-type highway overhead sign support structures subjected to wind induced fatigue loads.

  • 2. Use this information to develop fatigue design loads to provide an improved and more reliable design method.

  • 3. Compare the developed criteria with the Supports Specifications (1), and recom- mend new equations for the specifications.

5

Study Initiatives

Study initiatives of this study to accomplish the main objectives are enumerated

as follows:

  • 1. Complete a detailed theoretical fatigue loading analysis program and perform calculations to develop an accurate generalized fatigue design model to ac- count for the variety of sign support structures in design.

  • 2. Instrument one cantilever-type highway overhead sign structure with the nec- essary instrumentation needed to evaluate fatigue loading on the structure, and take field measurements under different natural wind and truck induced wind gust conditions.

  • 3. Evaluate the fatigue resistance of the structure to wind loading using the ex- perimentally obtained measurements. Determine areas of the structure that are most susceptible to fatigue and evaluate their performance under fatigue con- ditions, and establish sections of the structure best suitable for calculating fa- tigue loads from the experimental data.

  • 4. Develop design fatigue load criteria for cantilever-type highway overhead sign support structures from the experimental measurements.

  • 5. Compare the experimental fatigue load criteria to the results and conclusions of the theoretical program.

  • 6. Create finite element models and check the developed fatigue load criteria (experimental and theoretical) for accuracy by evaluating stress at critical lo-

6

cations of the structure. Make alterations to the FEA model and evaluate the

precision of the methodology developed in the theoretical program.

  • 7. Perform fatigue load calculations in accordance with the Supports Specifica- tions and assess the “accuracy” of Supports Specifications fatigue provisions with the developed fatigue loading criteria.

  • 8. Propose design recommendations as to fatigue load considerations for over- head sign structures based on the developed loading criteria.

Work Plan

Summary

A detailed theoretical program to evaluate fatigue loading on sign support struc-

tures was developed. The theoretical program addressed the variety of sign support struc-

ture used in design each with different configurations, sizes, shapes, and material proper-

ties. Finite element models were developed to perform loading simulations of the devel-

oped fatigue criteria and stresses at critical locations were evaluated.

An experimental program was developed to check the accuracy of the loading

provisions developed theoretically. A cantilever-type highway overhead sign structure

(shown in Figure 1.4) was selected for field measurement. The structure was instru-

mented with electric strain gages to determine stresses at critical locations. Accelerome-

ters and anemometers were mounted on the structure to determine structural dynamic

properties and wind velocities behavior. Two types of applied loads were considered in

the testing program:

7

  • a) Natural wind gust, and

  • b) Truck induced wind gust.

The experimental program placed specific emphasis on the instrumentation and

measuring of forces in base plate vicinity including the anchor bolts. Recorded data was

analyzed and compared to developed analytical methods, used to verify the procedures of

Supports Specifications provisions, and to propose practical wind-induced fatigue design

recommendations for overhead sign structures.

7 a) Natural wind gust, and b) Truck induced wind gust. The experimental program placed specific

FIGURE 1.4 Cantilever-type highway sign structure.

Project Tasks

The project involved an extensive analytical program to evaluate the fatigue loads

on sign support structures. An experimental program was performed to check the accu-

8

racy of the theoretical program on these structures. Fatigue loading criteria were devel-

oped and recommendations were proposed to provide a more reliable fatigue design

method. The objectives for this project were accomplished through the following tasks:

Task 1: Analytical Studies

A theoretical model was developed to address vibration behavior of sign support

structures to wind-induced fatigue loading. The model was created in such a way as to

account for the variety of sign support structures, each with different sizes, shapes, con-

figurations and material properties. Emphasis was placed on the vibration of the structure

due to wind-induced fatigue loading. The design fatigue load was determined based on

the dynamic characteristics of the structures. The infinite-life approach to fatigue design

was performed.

The loading criteria were used as input for computer software to check critical

stresses in the main members, and for evaluating the accuracy of the loading criteria. The

SAP 2000 v. 10 finite element analysis (FEA) computer software was used (11). A three-

dimensional full-scale cantilever structure was modeled in the FEA program. The model

was based on the shop drawings provided by ALDOT, and was same structure used in the

experimental program for field measurements and testing. A modal analysis was con-

ducted as well as static and dynamic loading simulations including material alterations.

Task 2: Site Selection

The site selection of the experimentally tested cantilever structure was located at

9

Task 3: Sign Structure Instrumentation

The support structure was instrumented with strain gauges, accelerometers, and

anemometers. Electric strain gages were placed at maximum stress locations on the verti-

cal and horizontal structural members of the overhanging section, and in the vicinity of

the base plate to determine strains under the different loading conditions. Accelerometers

were placed on the post and the cantilever overhang to determine the major dynamic

properties. Wind speed readings were recorded using anemometers placed 4 ft (1.22 m)

above the top of the post (to prevent shielding effects) for ambient readings, and along

the cantilever section for truck induced gusts measurements. Each transducer provided

simultaneously measured time history streamlines.

Task 4: Structural Testing

The overhead sign structure was tested under loading conditions: a) Natural wind

gust, and b) Truck induced wind gust. The natural wind data was taken over an extended

time period in an effort to capture the predominant natural wind gusts. Wind data from

the National Weather Service near Huntsville, AL, in the form of the annual mean wind

velocity for the area, was determined to help schedule testing days, and to compare to the

wind measurements taken to distinguish if the measured results were representative of the

wind environment. A standard semi–trailer vehicle was used to apply the truck induced

wind gusts. Varying truck speeds were used and the corresponding structural response

data was recorded.

10

Task 5: Experimental Data Reduction

The data was analyzed using the critical stresses determined in the structural

members and anchor bolts corresponding to wind velocity excitation. An evaluation of

the fatigue resistance was performed first to understand the distribution of stresses in the

structure, and to determine areas of the structure that were most vulnerable to fatigue

damage. A modal analysis was also performed using the experimental measurements.

Structural dynamic characteristics of the structure were determined such as the modal

frequencies, modal shapes, and damping characteristics. The analysis was used in the de-

velopment of the theoretical program as well as to verify finite element models.

Fatigue loading was developed from the data analysis after accomplishing an un-

derstanding of the structural behavior to wind loading and the distribution of stresses. The

infinite life approach to fatigue design was utilized and applied. The data collected of

natural wind gust were analyzed as a randomly occurring continuous load, whereas the

truck induced data were evaluated based on principles related to transient loading envi-

ronments.

Task 6: Design Recommendations

Design fatigue load recommendations for natural wind and truck-induced gusts

were developed after the completion of Task 5. The comparison of the analytical results

to the experimental data was utilized and fatigue design criteria were developed that en-

compassed the results of these efforts. The recommendations addressed the “accuracy” of

Supports Specifications fatigue provisions for cantilever-type sign structures.

11

Task 7: Design Examples

The effect of the proposed provisions was assessed and explained by performing

fatigue load calculations for design. The examples compared fatigue loads using the fa-

tigue provisions of Supports Specifications and the fatigue loads according to the pro-

posed guidelines of this study of both experimental and theoretical means.

Task 8: Project Report

A report summarizing Tasks 1 through 7 was prepared.

Significance and Benefits

The fatigue provisions of Supports Specifications have not been fully used or

evaluated by most state DOTs, and as such, insufficient information is available on the

design of structural supports using the current fatigue provisions of the specifications.

The provisions were developed primarily through analytical methods, and therefore an

extensive comparison of the provisions to fatigue loading through experimental meas-

urements has not been performed. In addition, the provisions are only applicable to spe-

cific types of support structures, and changes in size, shape, and material properties

would result in significant different vibration behavior and subsequent fatigue load. A

study is needed to develop fatigue loads that apply to all type of support structures, and

reflect reliable loading criteria to be used in design. A methodology is needed that reveals

the vulnerability of sign supports structures to wind-induced fatigue loading, from which

the engineer can reliably make alterations to the structural design.

12

Dissertation Outline

An outline of the dissertation is provided. The outline identifies the enclosed

chapters with a brief description of the work involved. The dissertation is presented in the

following outline:

Chapter 1 Introduction: An introduction to the study is presented. The objectives

of the project are stated with a breakdown of the tasks needed to accomplish the

projects objectives.

Chapter 2 Literature Review: A review of information in the literature related to

the project is provided. The most relevant material is presented in a form as to de-

scribe the brief historical development of the fatigue guidelines for support struc-

tures.

Chapter 3 Fatigue Provisions of the AASHTO Supports Specification: An intro-

duction to the fatigue provisions for natural wind and truck-induced gusts in the

Supports Specifications is provided. A historical description on the development

of the specification is provided.

Chapter 4 Sign Structure Instrumentation: A description of the support structure

instrumentation for experimental testing is described and illustrated.

13

Chapter 5 Structural Testing for Natural Wind and Truck-Induced Gusts: A de-

scription on the development of the testing program is provided. The testing pro-

cedure is explained for natural wind and truck-induced gusts.

Chapter 6 Fatigue Resistance: A detail of the fatigue resistance analysis of the

support structure using the experimentally collected data is provided. The analysis

helped to understand the distribution of stress in the structure, and to determine

the proper approach to calculate the fatigue load from the experimental data.

Chapter 7 Experimental Modal Analysis: A modal analysis on the structure was

performed using the experimentally collected data. The analysis helped to under-

stand the dynamic behavior of the structure needed primarily for the analytical

program.

Chapter 8 Experimental Calculation of the Fatigue Load due to Natural Wind

Gust: The procedure used in the reduction of data collected experimentally for

natural wind gust. The methodology and calculation of the fatigue load due to

natural wind is presented and described.

Chapter 9 Theoretical Calculation of the Fatigue Load due to Natural Wind Gust:

An analytical program was developed to calculate the fatigue load due to natural

wind guts. The method was developed to account for the variety of sign support

14

structures in design, each with different configurations, sizes, shapes, and material

properties.

Chapter 10 Experimental Calculation of the Fatigue Load due to Truck-Induced

Gust: The procedure used in the reduction of data collected experimentally for

truck-induced wind gust. The methodology and calculation of the fatigue load due

to truck-induced gusts is presented and described.

Chapter 11 Theoretical Calculation of the Fatigue Load due to Truck-Induced

Gust: An analytical program was developed to calculate the fatigue load due to

truck-induced wind guts. The method was developed to account for the variety of

sign support structures in design, each with different configurations, sizes, shapes,

and material properties.

Chapter 12 Finite Element Analysis: The finite element analysis program is pre-

sented. The analysis was conducted to verify the structural analysis used to calcu-

late the fatigue load. The analytical programs were also verified by changing the

material properties to address the accuracy of the proposed methodology.

Chapter 13 Proposed Fatigue Provisions: The proposed fatigue design equations

for natural wind and truck-induced gusts are presented. Design examples are pro-

vided to compare with the current provisions in the Supports Specifications and to

clarify the proper use of the proposed design equations.

15

Chapter 14 Conclusions: The developed conclusions are presented based on the

completed work of this project.

List of References: All references used in the dissertation are provided.

Appendix: Other relevant information is attached in the appendices. This included

handouts, instrumentation layouts, and other useful information in the completion

of the project. Ranking schedules are also included that helped in scheduling test-

ing days between ALDOT and the UAB research team.

16

CHAPTER 2

LITERATURE REVIEW

Chapter Overview

An extensive literature review was performed on fatigue of overhead highway

sign support structures. A spreadsheet was developed to categorize the reviewed studies

by breaking down the research into important areas of interests. Interests included rec-

ommendations for future research, vibration analysis, and instrumentation of support

structures. The cataloging helped to develop and systemize the research program. Also

provided in the literature review chapter are descriptions of the most relevant studies that

played a crucial role in the development and progress of the fatigue provisions of the

Supports Specifications, as well as other projects that were considered important to the

research of this project.

Research Paper Breakdown

Various research papers and documentations on sign support structures were re-

viewed before developing the research program. Important aspects that had relevancy to

this project were identified during the review process. Each of the documents reviewed

were tagged and categorized regarding the predetermined interests of this study. A

spreadsheet was created that helped to label the reviewed studies with respect to the

tagged aspects. Figure 2.1 displays the spreadsheet showing the properties of interests for

17

this project and the color coded categorizing method. The numbers at the top of the

spreadsheet reference the papers (Figure 2.2) from which the color coded categories were

indentified (7-10, 12, 15-17, 19, 20, 33, 36, 37, 50, 52, 53, 55, 59, 60, 64). Only the most

important references related to supports structures are listed in the figure. When reference

to a particular subject was needed during the project, the spreadsheet was used by utiliz-

ing the legend with respect to the property of interest, and identifying the documentation

that contained information on the subject. This process helped to allow the research to

perform smoothly and efficiently.

Referenced Papers Property of Interest 1 2 3 4 5 6 7 8 9 10 11
Referenced Papers
Property of Interest
1
2 3
4 5
6 7
8
9 10
11
12
13 14
15
16 17
18 19
20
Cantilever-Type Highway Sign Support Structure
Cantilever-Type VMS Support Structure
Bridge-Type Highway Sign Support Structure
Bridge-Type VMS Support Structure
High Mast Structure
Traffic Signal Structure
Luminare Structure
Strain Guage Transducers
Anemometer Transducers
Accelerometer Transducers
Pressure Transducers
Galloping
Vortex Shedding
Anchor Bolt Study
Natural Wind Gusts: Fatigue Loading
Natural Wind Gusts: Fatigue Resistance
Truck-Induced Wind Gusts: Fatigue Loading
Truck-Induced Wind Gusts: Fatigue Resistance
Horizontal Loading for Truck Gust
Suction Pressure for Truck Gust
Variation in Truck Type
Spectral Analysis
Natural Frequencies of Support Structures
Damping of Support Structures
Modal Shapes of Support Structures
Steel Structure
Aluminum Structure
Design Example
Legend
Involved Experimental Program
Invovled Analytical Program
Involved Experimental & Analytical Programs
Includes Equation

FIGURE 2.1 Research paper breakdown.

18

Referenced Papers

 

1

Cook, Ronald A., Bloomquist, D., Agosta, A.M., Taylor, K.F., Wind Load Data for Variable Message Signs . Report Number 0728-9488. Florida Department of Transportation, Research Management Center, April, 1996.

2

Azzam, D., Fatigue Behavior of Highway Welded Aluminum Light Pole Support Structures. Dissertation, University of Adron, May 2006.

 

Creamer, B. M., Frank, K. H., Klingner, R. E., Fatigue Loading on Cantilever Sign Support Structures from Truck Wind

3

Gusts. Research Report Number 209-1F. Texas State Department of Highways and Public Transportation, Transportation Planning Division, April, 1979.

4

Dexter, R. J., Ricker, M. J., Fatigue-Resistant Design of Cantilevered Signal, Signs, and Light Supports . NCHRP Report 469, The Transportation Research Board, Washington D.C. 2002

5

Zalewski, B., Huckelbridge, A., Dynamic Load Environment of Bridge-Mounted Sign Support Structures. Report No. ST/SS/05-002. Ohio Department of Transportation, Office of Research and Development, September 2005.

 

Foutch, D.A., Kim, T.W., LaFave, J.M., Rice, J.A. Evaluation of Aluminum Highway Sign Truss Designs and Standards

6

for Wind and Truck Gust Loadings. Research Report N0. 153. Illinois Department of Transportation, Bureau of materials and Physical Research, December 2006.

7

Ginal, S. Fatigue Performance of Full-Span Support Structures Considering Truck-Induced Gust and Natural Wind Pressures. Thesis, Marquette University. December 2003.

8

Kaczinski, M.R., Dexter,R.J., and VanDien, J.P. Fatigue Resistant Design of Cantilevered Sign, Signal and Light Supports. NCHRP Report 412. Transportation Research Board. Washington D.C. 1998.

9

South, S.M. Fatigue Analysis of Overhead Sign and Signal Structures. Report No. 115. Illinois Department of Transportation, Bureau of Materials and Physical Research. May 1994.

10

Fouad, F.H., Calvert, E.A., and Nunez, E. Structural Supports for Highway Signs, Luminaires, and Traffic Signals. NCHRP Report 411, Transportation Research Board. Washington D.C. 1998.

11

Edwards, J.A., and Bingham, W.L. Deflection Criteria for Wind Induced Vibrations in Cantilever Highway Sign Structures. Report No. FHWA/NC/84-001, Center for Transportation Engineering Studies, North Carolina State University,

12

DeSantis, P.V., and Haig, P.E., Unanticipated Loading Causes Highway Sign Failure. Proceedings of ANSYS Convention,

1996.

 

Albert, M.N., Manuel, L., Frank, K.H., and Wood, S.L., Field Testing of Cantilevered Traffic Signal Structures under

13

Truck-Induced Gust Loads . Report No. FHWA/TX-08/0-4586-2, Center for Transportation Research, University of Texas at Austin, 2007.

14

Cali, P., and Covert, E.E., On the Loads on Overhead Sign Structures in Still Air by Truck Induced Gusts. Wright Brothers Facility Report 8-97, Massachusetts Institute of Technology.

15

Ramy, A.S., Fatigue Resistant Design of Non-Cantilevered Sign Support Structures. Thesis, University of Alabama at Birmingham, 2000.

16

Fisher, J.W., Nussbaumer, A., Keating, P.B., and Yen, B.T., Resistance of Welded Details Under Variable Amplitude Long- Life Fatigue Loading. NCHRP Report 354, The Transportation Research Board, Washington D.C., 1993.

17

Irwin, H.P., and Peeters, M. An Investigation of the Aerodynamic Stability of Slender Sign Bridges, Calgary. LTR-LA- 246, national Research Council Canada-Aeronautical Establishment, 1980.

 

McDonald, J.R., Mehta, K.C., Oler, W., and Pulipaka, N., Wind Load Effects on Signs, Luminaires and Traffic Signal

18

Structures.

Texas Department of Transportation Report No. 1303-1F, Wind Engineering Research Center-Texas Tech

University, Lubbaock, TX, 1995.

 

Gilani, A.S., Chavez, J.W., and Whittaker, A.S., Fatigue-Life Evaluation of Chaneable Message Sign Structures, Volume

19

 

1 - As Built Structures.

Report No. UCB/EERC-97/10, Earthquake Engineering Research Center, University of California,

Berkeley, CA, 1997.

 

Kashar, L., Nester, M.R., Johns, J.W., Hariri, M., and Freizner, S., Analysis of the Catastrophic Failure of the Support

20

Structure of a Changeable Message Sign.

Structural Engineering in the 21st Century, Proceedings of the 1999 Structures

Congress, New Orleans, LA, 1115-118, 1999.

FIGURE 2.2 Referenced papers in the research paper breakdown spreadsheet.

Many aspects that were considered important by the researchers were not found in

the literature and were therefore excluded from the spreadsheet. Some aspects were sim-

ply dropped from the spreadsheet as the research progressed because of relevancy,

whereas other aspects were added which resulted in the breakdown shown in the Figure

19

The research paper breakdown served as a useful tool during the development of

the research program. Relevant information on gathered literature documents was easily

and quickly identified to extract information. The breakdown also helped to identify areas

where research is needed or is lacking. Only the most prominent papers as they related to

the research performed with this project was included in the breakdown. A complete list

of reviewed papers is provided in the List of References at the end of the dissertation.

Research Review

A description of the reviewed studies that played a crucial role in the development

of the fatigue provisions in the Supports Specifications are listed as follows. Other pro-

jects included in the literature review were considered important to this research and are

described.

DeSantis and Haig (1996)

The fatigue provisions for truck-induced wind gust in the Supports Specifications

were initially based on this study. The research focused on a cantilever-type overhead

VMS support structure. The structure failed due to fatigue loading for which prompted

the study. After the structure was replaced, large deflections were observed due to wind

gust created from passing trucks. In the analysis, the researchers assumed that the veloc-

ity of the wind gusts onto the structure was equal in magnitude to the speed of the truck,

in addition to a gust factor equal to 1.3 to account for head winds. The wind pressure was

calculated using the fundamental fluid mechanics relationship between wind force and

the square of the wind velocity. The resulting wind pressure was doubled to account for

20

the upward deflection of the sign plus the downward deflection due to the pull of gravity.

The concluded value represented a pressure range to be used for fatigue design, and is

shown in Eq. 2.1.

P

TG

= 36.6

C d
C
d

(psf)

where

P

TG

= design fatigue pressure load due to truck -induced wind gust

C

  • d = drag coefficient

[Eq. 2.1]

The researchers performed a FEA model of the structure using the software pack- age ANSYS. The cantilevered end of the structure was observed in the field to deflect vertically about 1 ft (0.305 m) in length after exposed to wind gust from passing trucks. Experimentation was not performed to validate the tip deflection other than visual obser- vation. To help verify the observations, the researchers back-calculated the wind pressure that would theoretically produce a 1 ft (0.305 m) deflection, and inputted the pressure into the ANSYS program of the modeled structure. The results matched the wind pres- sure calculated from Eq. 2.1 and was concluded as the appropriate design load based on this comparison (16).

Davenport (1961)

Davenport’s research was not focused on overhead sign support structures spe- cifically, but rather focused on the characteristics of wind velocity. His research has pro- vided an accurate model for simulating wind velocity behavior. The model was used in

21

the development of the fatigue provisions for natural wind gust in the Supports Specifica- tions. It simulated the randomness along with the gustiness and turbulence commonly associated with wind velocity. The model was developed in the form of a power density spectrum generally used for predicting randomly occurring events. The power density spectrum was created using experimentally measured wind velocity time histories gath- ered from sites located around the world. An empirical formulation was developed as a function of the mean wind velocity, and is shown in Eq. 2.2. Terrain coefficients were also indentified the formulation process, and are shown in Table 2.1 (13).

S

v

(

f

) =

4

V

κ

10

2

2

x

f

(1

+

4

2

x

)

3

[Eq. 2.2]

where

S

v

(

f

) =

wind velocity power spectral density at any height

f = frequency

  • V 10

= mean wind velocity at a stadard height of 10 meters above ground leve

l

κ

=

surface drag coefficient (Table 2.1)

x

=

1200 f

2

V

10

with

f

2

V

10

in cycles per meter.

TABLE 2.1 Terrain Coefficients (13)

Type of Surface

κ

α

Open unobstructed country (e.g., prairie-type grass- land, arctic tundra, desert)

0.005

0.15

Country broken by low clustered obstructions such as trees and houses (below 10 m high)

0.015 - 0.020

0.27 – 0.31

Heavily built-up urban centers with tall buildings

0.050

0.43

22

Kaczinski et al (1998)

The research performed by Kaczinski reported in NCHRP Report 412 formed the

framework of the fatigue provisions for galloping, vortex shedding, natural wind gust,

and truck-induced wind gusts in the Supports Specifications. The majority of his work

related to natural wind and truck-induced wind gusts was entirely theoretical. By using

the research performed by DeSantis and Haig, the fatigue provisions for the truck-

induced gust were created. Equation 2.1 was recommended as the appropriate pressure

load to use for truck-induced wind gusts.

Davenport’s wind velocity power density spectrum was used to create the fatigue

provisions for natural wind gust. The velocity spectrum was transformed into a wind

force spectrum. The spectrum was applied as input into an FEA program to generate a

stress response spectrum for four particular overhead sign support structures. The stress

spectrum was formed at different locations in the modeled structure and compared. The

fatigue load was then developed using the infinite-life approach. Equation 2.3 was devel-

oped from the results and was recommended as the appropriate pressure load to use for

natural wind gust.

P

NW

= 5.2

C I

d

F

(psf)

where

P

NW

= design fatigue pressure load due to natural wind gusts

C

  • d = drag coefficient

  • I F

= importance factor

[Eq. 2.3]

23

Report 412 also focused on developing S-N curves (stress vs. number of cycles)

for anchor bolt connection details. The results were used to create the constant-amplitude

fatigue thresholds currently available in the Supports Specifications. Recommendations

for anchor bolt design and structural analysis were provided. In addition, Report 412 in-

troduced “Importance Factors” to be used with the Supports Specifications. Fatigue de-

sign examples are provided for different types of supports structures including overhead

sign, traffic signals, and luminaires. The examples provided a thorough design procedure

using the proposed fatigue design equations. Stresses at critical details were calculated

and compared to the fatigue thresholds (52).

Fouad et al (1997-2005)

The research by Fouad et al in NCHRP project 17-10 and 17-10(2) (32, 33)

evaluated fatigue design of sign support structures. The research was compiled into a

specification based on the Allowable Stress Design (ASD) methodology, the basis of

which the current Supports Specifications were formed. The impact of the new specifica-

tions was analyzed with significant conclusions. The research brought together important

fatigue concerns such as galloping, vortex shedding, natural wind, and truck-induced

gusts. Many design examples were made which provided the framework for engineers to

use for fatigue evaluations. The work performed by Fouad et al compiled all relevant in-

formation on fatigue of sign support structures into a single stand-alone document. It now

resides as a well established document for fatigue design that is specific to sign support

structures, from which helped to propel future enhancements and innovations to fatigue

design of these structures. The researchers also provided valuable information on trans-

24

forming the specification into a Load and Resistance Factor Design (LRFD) methodology

in the future. The focus was to first develop a much needed stand-alone specification us-

ing the ASD method, and then upgrade to LRFD in the future (21-35).

Dexter and Ricker (2002)

The research performed by Dexter and Ricker reported in NCHRP Report 469

formed the framework of the current Supports Specifications. Their research involved

experimental and analytical programs. They concluded that the fatigue provisions for

natural wind gusts in the Supports Specifications were accurate. However, the truck-

induced gusts were highly conservative as compared to their experimentally determined

values. A reduction in the truck provisions were recommended by Eq. 2.4. The reduction

did not reflect the experimental values gathered in the report because they were signifi-

cantly less the fatigue provisions, and it was believed the committee would not accept

them. The authors were also concerned with the type of structures used in the analysis.

The fatigue load was determined from analysis on variable message signs which are sig-

nificantly heavier than conventional signs. The reduction of the fatigue design equation

for truck-induced gusts shown as Eq. 2.4 was recommended.

P =

TG

18.8

C

d

I

f

(psf)

where

P

TG

= design fatigue pressure load due to truck -induced wind gust

C

  • d = drag coefficient

  • I F

= importance factor

[Eq. 2.4]

25

Report 469 also provided a reduction to the fatigue load to accommodate for

height of the structure above the ground. They recommended a linear reduction in wind

pressure starting from 19.7 ft (6.00 m) above ground level to zero at approximately 32.8

ft (10.0 m) above ground level. A recommended area on the support structure to apply the

truck load was also provided. It was proposed to apply a uniformly distributed load to the

12 ft (3.66 m) area of the structure that produces the maximum stress range. This was

recommended as opposed to applying the load to portions of the structure that is not di-

rectly above the traffic lane (16).

Cook et al (1996)

The research performed by Cook et al is indentified in this review due to its rele-

vancy to fatigue design due to truck-induced gusts. Their research determined the truck-

induced wind gusts through direct measurement of wind pressure. Pressure transducers

were placed on a bridge spanning a highway to measure the wind pressure from passing

trucks. Importantly, the bridge was extremely rigid so that the measured values were not

influenced by the type of structure. Twenty-three loading events were measured and re-

corded.

The results provided a clear view of the wind behavior created from passing

trucks. It revealed a biaxial loading event with a vertically applied pressure and a hori-

zontally applied pressure. A suction event was also recorded as the truck passed under-

neath the structure. The results indicated a ramped loading impulse. Variation of the

truck-induced wind gust with respect to height above ground level was also measured. A

10% pressure reduction per foot above 17 ft (5.18 m) above ground level was found (12).

26

Creamer et al (1979)

The work performed by Creamer et al was very influential for future research.

Their research focused on fatigue induced loading due truck wind gusts. Where Cook’s

determined the truck wind pressure directly using pressure transducers, Creamer’s work

determined the truck wind pressure indirectly. They measured strain variations at critical

locations of a cantilevered structure due to truck wind gusts, and back calculated for wind

pressure. A ramped impulse function was developed and inputted into a FEA model of

the tested structure. The loading simulation produced similar strains on the structure as

measured experimentally (10).

27

CHAPTER 3

FATIGUE PROVISIONS OF THE AASHTO SUPPORTS SPECIFICATIONS

Chapter Overview

An overview of the development of the natural wind and truck-induced wind

gusts fatigue provisions within the AASHTO Standards Specifications is provided. This

was believed to be an important and required task so that a complete understanding of the

developed fatigue provisions was obtained before addressing their accuracy. The provi-

sions for natural wind were formed indirectly through theoretical calculations. They were

developed using the infinite-life approach to fatigue design. Spectral analysis formed the

framework of the development procedure. The truck-induced wind gust fatigue provi-

sions were developed through theoretical and experimental observations.

Fatigue Load due to Natural Wind

Infinite-Life Approach

The method used to develop the natural wind provisions was based on spectral

analysis in collaboration with the infinite-life approach to fatigue design. The provisions

are adequate within certain limitations, as they were formed using four particular catego-

rized structural types, each of which were chosen to represent the population of the cate-

gory selected. The structural response of two overhead signal support structures, one can-

tilever-type overhead sign support structure, and one luminaire support structure to natu-

28

ral wind excitation were analyzed, the transmitted stresses of each were averaged, and the

code was developed from the averaged results (52).

The AASHTO fatigue code was based on the constant amplitude fatigue limit

(CAFL). Tests to developed plots of the stress vs. the number of cycles (S-N curves) for

common connection details were used to determine this value. Fatigue stress ranges that

occurred below the CAFL were estimated to have an infinite life. Stresses ranges at criti-

cal locations such as welded or bolted connections must be under the CAFL to be in

compliance with the code. Changes in the design must be made if they are not to lower

the fatigue stresses to fall below the CAFL.

The CAFL tests were based on constant amplitude, whereas amplitudes that occur

due to natural wind are random with variable amplitudes. To account for variable ampli-

tude loading, the infinite-life approach adopted the findings of NCHRP Report 354 where

variable amplitude fatigue tests were performed on a variety of full scale connection de-

tails. From the results, it was determined that failure would still occur if the variable am-

plitudes exceeded the CAFL of the connection detail 0.05% of the time. However, if the

CAFL was exceeded 0.01% of the time, the specimen was said to have an infinite life

(19). In view of the results, the infinite-life approach involved estimating the stress

ranges that would occur onto the structure with a 0.01% exceedence probability, includ-

ing dynamic amplification. These ranges were referred to as the fatigue limit-state load

ranges. The structure was designed such that the limit-state stress ranges did not exceed

the CAFL of the detail in question.

29

Predicting the Environment

The first step in determining the fatigue limit-state load ranges to develop the

AASHTO fatigue provisions was to estimate the structural excitation due to natural wind.

This involved predicting the natural wind environment for which the structure is to be

exposed. This was done using a spectral analysis. Simulating the natural wind force can

be difficult due to the randomness, turbulence and gustiness of natural wind velocity. A

static load that represents the force due to wind can be easily calculated, as well as a dy-

namic load in terms of a periodic wave such as a sine wave. This type of examination

does not account for the gustiness and turbulent nature of the wind itself. In most cases, a

gust factor is used to account for the random nature of wind excitation, specifically for

structural capacity predictions. A. G. Davenport developed a wind velocity power density

spectrum (PDS) curve (13) shown in Eq. 3.1 that simulated the turbulent nature of wind

velocity.

S

v

(

f

) =

4

V

κ

10

2

x

2

f

(1

+

4

x

2

)

3

[Eq. 3.1]

where

S

v

(

f

) =

wind velocity power density spectrum at any height

f = frequency

  • 10 = mean wind velocity at a stadard height of 10 meters above ground level = surface drag coefficient (Table 3.1)

V

κ

x

=

1200 f

2

V

10

with

f

2

V

10

in cycles per meter.

30

TABLE 3.1 Terrain Coefficients (13)

Type of Surface

κ

α

Open unobstructed country (e.g., prairie-type grass- land, arctic tundra, desert)

0.005

0.15

Country broken by low clustered obstructions such as trees and houses (below 10 m high)

0.015 - 0.020

0.27 – 0.31

Heavily built-up urban centers with tall buildings

0.050

0.43

Random excitation is best analyzed over a frequency spectrum. PDS curves are

commonly used for dynamic loading analysis, but more specifically suited to random vi-

bration. In random excitations, there are no periodic systems within the excitation that

can be analyzed at specific frequencies because the excitation is applied randomly over a

spectrum of frequencies. To account for this, the natural wind fatigue analysis involved

an examination of the total energy, or power, in the random excitation over a frequency

band. A typical PDS curve is plotted with the ordinate in units of the parameter (force,

acceleration, velocity, displacement, etc.) squared divided by the bandwidth (e.g., N 2 /Hz),

and the abscissa in units of frequency (Hz). The peaks on a PDS curve identify the fre-

quency range(s) at which the majority of the energy lies within the particular excitation.

The area under a PDS curve is equal to the mean square value, and the square root of this

area is equal to the root-mean-square (RMS) value: a value of which plays an important

role in determining the magnitude of loads that are transmitted onto a structure from vi-

bration (39, 40).

Davenport developed his wind velocity PDS curve from at least 70 experimental

wind velocity data collections from various locations around the world. His intention was

to develop a model which simulated the turbulence and gustiness of wind velocity. He

developed Eq. 3.1 from the 70 experimental data collections. The equation is a function

31

of wind velocity frequency with respect to a mean wind velocity at a specified height. His

formulation is plotted in Figure 3.1 for frequencies ranging from 0 to 10 Hz, an open ter-

rain (see Table 3.1), and an annual mean wind velocity of 11 mph (5 m/s).

31 of wind velocity frequency with respect to a mean wind velocity at a specified height.

FIGURE 3.1 Wind velocity PDS for annual mean wind velocity 11 mph.

Once the behavior of the wind velocity environment was estimated, the PDS was

transformed into a wind force PDS by using principles related to fluid mechanics. An in-

duced force onto a structure due to natural wind, referred to as drag, is proportional to

wind velocity squared, as shown by the Eq. 3.2

F

D

=

1

2

C AV

ρ

d

  • 2 [Eq. 3.2]

32

where

F

D

ρ

=

= drag force density of air

=

1.22

kg

m

3

C

  • d = drag coefficient

A = area of exposed surface

  • V = wind velocity at any height.

By utilizing the proportionality between drag pressure and wind velocity, and accounting

for the turbulent wind velocity and its variance about the mean wind velocity, a wind

pressure PDS was developed from Davenport’s wind velocity PDS shown in Eq. 3.3. The

plotted equation is shown in Figure 3.2 for an annual mean wind velocity of 11 mph (5

m/s) and normalized for exposed area and the drag coefficient.

S

F

(

f

)

= ρ

2

2

C

d

2

A V

2

S

v

(

f

)

[Eq. 3.3]

The PDS curve accounts for the gustiness and turbulence of wind velocity over a

spectrum of frequencies, based on an averaged wind velocity taken at a specified height

above ground level. Since most support structure are at or around 32.8 ft (10.0 m) in

height, the PDS curve was well suited for these types of structures. Yet, the PDS can be

used at any particular height by using the power law profile for approximating variation

in wind velocity with height, as shown in Eq. 3.4:

V

=

V

10

αz

α

[Eq. 3.4]

33

where

V

= wind velocity at height z

= surface coefficient in Table 4 z = height above ground.

α

33 where V = wind velocity at height z = surface coefficient in Table 4 z

FIGURE 3.2 Force PDS for annual mean wind velocity 11 mph.

In this case, where the objective was concentrated on formulating a design code

for fatigue wind, the wind velocity variable in the pressure PDS equation was taken at the

standard height of 32.8 ft (10.0 m) above ground level, and kept uniform across the ex-

posed façade. The purpose of which was to provide a simplified design equation for

commercial use. Some conservative formulation exists as the wind velocity typically in-

creases from the ground level upwards (13).

34

Once the natural wind environment was estimated, the next step was to apply the

PDS to the infinite-life approach in determining the limit-state stress ranges. Since the

force spectrum was based primarily on the annual mean wind velocity, the wind velocity

that was exceeded 0.01% of the time was found and referred to as the limit-state wind

velocity. The force spectrum was calculated using the limit-state wind velocity and used

to calculate the structural response to determine the limit-state stress range.

Wind velocity is random in nature, but it can be predicted though statistical rela-

tionships. It has been found through many experiments that the magnitude of the wind

velocity vector will form a Rayleigh distribution (14, 52, 54). By using the Rayleigh dis-

tribution, the 0.01% exceedence probability can be found through the relationship in Eq.

3.5 based on the annual mean wind velocity.

P E
P
E

(

v

) =

π

v

2

e

4

V

2

where

P

E

(

v

)

=

probability

  • v =

wind velocity corresponding to the probability

  • V =

mean wind velocity.

[Eq. 3.5]

An analysis was conducted to determine which annual mean wind velocity to use in

Eq. 3.5 to determine the limit-state wind velocity (wind velocity with a 0.01% ex-

ceedence probability). The annual mean wind velocities of major U.S. cities were ana-

lyzed. It was found that an annual mean wind velocity of 11 mph (5 m/s) was exceeded in

only 19% of the U.S. cities analyzed and was therefore chosen. By plugging in 11 mph in

Eq. 3.5, and solving for the wind velocity corresponding to the 0.01% probability, a limit-

35

state wind velocity was found to be equal to 37 mph (17 m/s). The force spectrum was

then formed using the limit-state wind velocity (see Figure 3.3) and was used as the natu-

ral wind velocity prediction model for structural excitation.

35 state wind velocity was found to be equal to 37 mph (17 m/s). The force

FIGURE 3.3 Force PDS using limit-state wind velocity 37 mph.

Structural Excitation

Once the excitation was determined, the spectral density of the response was cal-

culated. This was done through finite element analysis for the signal, sign, and luminaire

support structures. Each structure had different dynamic characteristics such as natural

frequency and critical damping percentages, all of which are based on the size, shape, and

material of the structure. The response PDS was calculated in units of stress squared di-

36

vided by Hz. The area underneath the response PDS curve is equal to the variance of the

response about the mean. The square root of the area is equal to the RMS. If the mean is

equal to zero, then the RMS is equal to the standard deviation (58, 62). Since the re-

sponse curves of support structures are predominately controlled by a single frequency of

vibration, the developed response PDS curves were narrow-banded at the natural fre-

quency of the structures. Therefore, the stress range was calculated as a constant ampli-

tude in the form of a sinusoid. For any sinusoid, the peak stress amplitude is equal to the

square root of two times the RMS (assuming a zero mean for analysis purposes). The

stress amplitude was then multiplied by two to account for a peak-to-peak stress range.

The RMS was determined from the response PDS for each structure, and the effective

stress range was calculated using Eq. 3.6

S

eff

r

= 2.8σ

rms

where

S

eff

r

σ

rms

= effective stress range or limit -state stress range = RMS of the response

[Eq. 3.6]

Design Fatigue Equation for Natural Wind

Equation 3.7 was calculated for the four type of structures analyzed. The analysis was normalized for the drag coefficient and exposed area. The result ranged from 3.6 to 6.3 psf (170 to 300 Pa) using 1% and 2% damping percentages. An average was taken and used for the code. The final equation (Eq. 3.7) was multiplied by the drag coefficient, importance factor, and a wind factor to account for other annual mean wind velocities.

37

P

NW

= 5.2

C

d

I

F

  v  

11

2

where

P

NW

= design fatigue load due to natural wind, psf (Pa)

C

  • d = drag coefficient

  • I F

= importance factor

  • v = annual mean wind velocity other than 11 mph (5 m/s)

[Eq. 3.7]

Fatigue Load due to Truck-Induced Gust

The truck-induced gust fatigue provision was developed as an impulse loading

occurring over a finite length of time. The loading was created when semi-trailer trucks

passed underneath the structure causing the structure to vibrate. The vibration generated

stresses in the structure that can potentially accumulate fatigue damage over time.

The Supports Specifications design equations for truck-induced gusts was origi-

nally developed from the work performed by Desantis and Haig (16), and later revised

from the work presented in NCHRP Report 469 (14). It assumes that the wind loading

onto the structure was equivalent to the speed of the passing truck, plus a gust factor of

1.3 to account for head wind. For instance, a truck traveling at 65 mph (29.1 m/s) would

produce a 65 mph (29.1 m/s) wind onto the structure. The result was an 18.3 psf (876 Pa)

magnitude static pressure. The code only accounted for a vertical force applied upward

(ground up) onto the structure. Assuming that the upward motion of the structure result-

ing from the truck gust was equivalent to a proceeding downward motion, the 18.3 psf

(876 Pa) was doubled to account for a peak-to-peak range equaling 36.6 psf (1760 Pa).

NCHRP Report 469 investigated the fatigue load due to truck gusts performed by

researchers on VMS cantilever structures. The research indicated that the current code

38

was too conservative based on their conclusions. A reduction of the loading was recom-

mended by the following equation:

P

TG

= 18.8

C

d

I

F

  V  

65

2

where

P

TG

= design fatigue load due to truck gust, psf (Pa)

C

  • d = drag coefficient

  • I F

= importance factor

  • V = truck speed, mph (m/s)

[Eq. 3.8]

The reduction was determined based on experimental evidence from VMS struc-

tures. Strain measurements on a cantilevered VMS support structure were recorded dur-

ing random truck events. An equivalent static load was calculated that would produce the

same strain range measured experimentally. The resulting load was to be applied verti-

cally to all horizontal areas (underside of the structure) along a 12 ft length (3.7 m), or

equal to the width of a travel lane. A reduction in the load with respect to height of the

structure above ground level was also permitted. It was discovered the pressure load de-

creased linearly with height, with the maximum occurring at 19.7 ft (6 m) above ground

level to zero at 32.8 ft (10 m).

39

CHAPTER 4

SIGN STRUCTURE INSTRUMENTATION

Chapter Overview

A cantilever-type highway overhead sign structure was selected for field meas-

urement. The structure was instrumented with electric strain gages to determine stresses

at critical locations. Accelerometers were mounted on the structure to determine struc-

tural dynamic behavior, and anemometers were attached to measure wind velocity vec-

tors (magnitude and direction). Each transducer provided time history streamlines that

were collected simultaneously using a data acquisition system. A detailed description of

the instrumentation program is provided in this chapter.

Cantilever-Type Highway Overhead Sign Support Structure

A four chord truss cantilever-type highway overhead sign support structure was

chosen for analysis. The structure was located at northbound Exit 340B on highway I-65

near Decatur, Alabama USA. A picture of the structure evaluated for this project is

shown in Figure 4.1.

40

40 FIGURE 4.1 Cantilever-type highway overhead sign support structure. Geometric Properties The truss overhang was made

FIGURE 4.1 Cantilever-type highway overhead sign support structure.

Geometric Properties

The truss overhang was made of 3.5 in (88.9 mm) diameter steel pipe supported

by a 24 in (610 mm) steel pipe shaft support. The shaft was support by a 1.25 in (31.8

mm) thick by 35 in (889 mm) diameter base plate with eight 1.5 in (38.1 mm) anchor

bolts connected to a concrete foundation. The structural configurations are shown in Fig-

ure 4.2 through Figure 4.4.

41

14 ft 25 ft 10.5 ft Four Chord 2 ft Cantilever Truss 11.5 ft 4 ft
14 ft
25 ft
10.5 ft
Four Chord
2 ft
Cantilever
Truss
11.5 ft
4 ft
4 ft
Post
Sign

26.75 ft

FIGURE 4.2 Elevation view of structure configuration.

3 ft

4 ft

10.75 in
10.75 in

37.75 ft

FIGURE 4.3 Plan view of structure configuration.

42

3 ft

FIGURE 4.4 Side view of structure configuration.

Material Properties

The post and truss section was made of API-5L-X52 steel pipe. All plates were

structural steel ASTM A572 Gr. 50. The anchor bolts were AASHTO M314-90 Gr. 55

(essentially the same as ASTM F1554 Gr. 55). The W-shape and T-shape sections used

for the sign-to-truss connection were made of A572 Gr. 50 steel. The sign was aluminum

wrought alloy designated as 6061-T6. The concrete pile foundation was conventional

concrete with #9 size rebar Gr. 60. All of the major required material properties of the

designations given are listed in Table 4.1.

43

TABLE 4.1 Material Properties

     

Modulus of

Yield

Tensile

Material

Members

Material

Designation

Elasticity

(psi)

Stress

(psi)

Stress

(psi)

Steel Pipe

Truss, Post

API-5L-X52

29,000,000

52,000

66,000

 

Truss and

ASTM A572

     

Steel Plate

Base plates

Gr. 50

29,000,000

50,000

65,000

   

AASHTO

     

Steel Rod

Anchor Bolts

M314-90 Gr.

29,000,000

55,000

75,000

55

Aluminum

Sign

6061-T6

10,000,000

37,000

42,000

 

Pile

4, 000 psi

     

Concrete

Foundation

Compressive Strength

3,600,000

NA

NA

 

Pile

ASTM A706

     

Rebar

Reinforcement

Gr. 60

29,000,000

60,000

80,000

Instrumentation Process

There was a unique advantage associated with this project. The researchers were

able to instrument the structure before and during assembly, which allowed for a thor-

oughly detailed instrumentation program. Limitations associated with an already assem-

bled structure in the field were avoided as a result. All strain gauges, except on the anchor

bolts, were placed on the structure at the ALDOT laboratory in Montgomery, Alabama

USA before transportation to the assembly site (Figure 4.5 – 4.6). The anchor bolts were

instrumented on site the day after the foundation was poured and the anchor bolts set

(Figure 4.7). The post support was placed on the foundation and bolted to the anchor

bolts after the anchor bolts were gauged. A period of 14 days (28 days after pouring)

44

transpired after the post support was erected for concrete curing of the foundation. After

the curing stage was completed, the truss overhang and sign was attached to the post sup-

port (Figure 4.8 – 4.9). A lift bucket was used to place the accelerometers and anemome-

ters in their appropriate preconceived locations. The finished instrumented structure is

shown in Figure 4.10.

44 transpired after the post support was erected for concrete curing of the foundation. After the

FIGURE 4.5 Strain gauged post support in laboratory.

45

45 FIGURE 4.6 Strain gauged truss overhang in laboratory. FIGURE 4.7 Set anchor bolts in structure

FIGURE 4.6 Strain gauged truss overhang in laboratory.

45 FIGURE 4.6 Strain gauged truss overhang in laboratory. FIGURE 4.7 Set anchor bolts in structure

FIGURE 4.7 Set anchor bolts in structure foundation.

46

46 FIGURE 4.8 Truss overhang attachment. FIGURE 4.9 Sign attachment.

FIGURE 4.8 Truss overhang attachment.

46 FIGURE 4.8 Truss overhang attachment. FIGURE 4.9 Sign attachment.

FIGURE 4.9 Sign attachment.

47

47 FIGURE 4.10 Finished instrumented structure. Strain Gauges Uni-axial strain gauges and single-plane rectangular strain gauge

FIGURE 4.10 Finished instrumented structure.

Strain Gauges

Uni-axial strain gauges and single-plane rectangular strain gauge rosettes (meas-

uring in three primary directions: 0°, 45°, and 90°) were used to measure the strain re-

sponse of the structure. They were placed on the structure at locations where strains were

most critical, and especially applicable to determining the causative load. These locations

included: truss chord members near the truss-to-support connection, bottom of the pole

support member near the base plate, and the anchor bolts. A list of all strain gauges and

their locations along with a schematic detailing these locations are provided in Appendix

A.

48

Truss Chord Members

A total of 16 uni-axial strain gauges were used on the truss chord members lo-

cated near the truss-to-support connection: four gauges per chord placed along the cir-

cumference of the member. These gauges measured strain in a single direction, oriented

along the longitudinal length of the chord member, in an attempt to capture tensile and

compressive strain readings. A schematic of the gauge locations are shown in Figure

4.11, with the actual gauged structure at this location shown in Figure 4.12. A picture of

the gauges on the assembled structure is shown in Figure 4.13.

Placed on each Chord at 90º around Chord Circumference Centerline of 1 st Truss Panel 3’-
Placed on each
Chord at 90º
around Chord
Circumference
Centerline
of 1 st Truss
Panel
3’- 4.5”

FIGURE 4.11 Truss-to-pole support strain gauges.

49

49 FIGURE 4.12 Chord strain gauges.

FIGURE 4.12 Chord strain gauges.

50

50 FIGURE 4.13 Chord strain gauges on the assembled structure. Post Support Uni-axial and rosette strain

FIGURE 4.13 Chord strain gauges on the assembled structure.

Post Support

Uni-axial and rosette strain gauges were used for the post support. The gauges

were attached to the structure in the laboratory before shipment to the assembly site. Two

levels, located nearest to the base plate, were gauged along the circumference of the

member in accordance to the schematic in Figure 4.14. This was done in an attempt to

capture tensile and compressive strains near the base plate. Forty-five degree rectangular

rosettes were used to measure transverse shear and torsion strains. A close up of Section

AA viewing a rosette and Section BB located 4 in (102 mm) above Section AA viewing a

uni-axial strain gauge is shown in Figure 4.15. Two levels were chosen for redundancy

issues in the case of strain gauge measurement failure since the structure was tested out-

51

side, and the testing duration lasted over multiple months. A picture of the assembled in-

strumented post support is shown in Figure 4.16

Section AA Section BB Uni-axial Strain Gauge Rosette Strain Gauge 4 in Section BB Section AA
Section AA
Section BB
Uni-axial Strain Gauge
Rosette Strain Gauge
4 in
Section BB
Section AA
12 in
Base Plate Weld

FIGURE 4.14 Pole support strain gauges.

52

52 FIGURE 4.15 Rosette and uni-axial strain gauges on shaft. FIGURE 4.16 Instrumented post support.

FIGURE 4.15 Rosette and uni-axial strain gauges on shaft.

52 FIGURE 4.15 Rosette and uni-axial strain gauges on shaft. FIGURE 4.16 Instrumented post support.

FIGURE 4.16 Instrumented post support.

53

Anchor Bolts

Strain gauges were attached to the anchor bolts in the field, after the concrete

foundation concrete was placed. A picture of the anchor bolts before placement into the

concrete foundation is shown in Figure 4.17, and after placement in the foundation before

gauging in Figure 4.18. The material specifications of the anchor bolts are provided in

Table 4.1. The concrete foundation reached 11 ft (3.35 m) into the ground, with the rein-

forcement cage reaching 10.5 ft (3.20 m) deep.

A total of eight anchor bolts were used for the structural foundation connection.

Uni-axial strain gauges were attached to each anchor bolt in the orientation shown in Fig-

ure 4.19, totaling eight strain gauges. The gauges were attached as close as possible to the

foundation due to assembly procedures and leveling processes. The gauges were located

tangent to the outside rim of the anchor bolt group formation, with the sensitive end ori-

ented along the longitudinal length of the bolt. The threads at the gauge location were

shaved (Figure 4.20) and cleaned for gauge attachment. This was done for each bolt. A

picture of a finished instrumented anchor bolt is shown in Figure 4.21, and of the assem-

bled structure in Figure 4.22.

54

54 FIGURE 4.17 Anchor bolts before installation. FIGURE 4.18 Anchor bolts and foundation.

FIGURE 4.17 Anchor bolts before installation.

54 FIGURE 4.17 Anchor bolts before installation. FIGURE 4.18 Anchor bolts and foundation.

FIGURE 4.18 Anchor bolts and foundation.

55

Support Base Plate Strain Gauge Anchor Bolt Foundation Orientation
Support
Base Plate
Strain Gauge
Anchor Bolt
Foundation
Orientation

FIGURE 4.19 Anchor bolt strain gauges.

55 Support Base Plate Strain Gauge Anchor Bolt Foundation Orientation FIGURE 4.19 Anchor bolt strain gauges.

FIGURE 4.20 Anchor bolt strain gauging preparation.

56

56 FIGURE 4.21 Strain gauged anchor bolt. FIGURE 4.22 Instrumented anchor bolts on assembled structure.

FIGURE 4.21 Strain gauged anchor bolt.

56 FIGURE 4.21 Strain gauged anchor bolt. FIGURE 4.22 Instrumented anchor bolts on assembled structure.

FIGURE 4.22 Instrumented anchor bolts on assembled structure.

57

Anemometers

Anemometers were used to measure wind velocity and direction. The WindSonic

ultra-sonic wind and direction sensor was used for this application (see Figure 4.23). The

anemometer calculated speed and direction by measuring the time needed for generated

sound pulses to travel from one transducer to the other within the air gaseous medium. A

combined total of six anemometers were required for wind measurements of natural wind

and truck induced wind gusts. Two were placed above the post support structure, and four

on the truss section of the support structure. Steel extensions were fabricated to fit the

anemometers to the structure. A list of the anemometers identification and locations are

provided in Appendix A, along with a schematic detailing their locations. An overview of

their locations can also be seen in Figures 4.24 – 4.26. The mounting instructions pre-

pared for ALDOT are given in Appendix B.

57 Anemometers Anemometers were used to measure wind velocity and direction. The WindSonic ultra-sonic wind and

FIGURE 4.23 WindSonic ultra-sonic wind and direction sensor.

58

58 FIGURE 4.24 Overview of anemometer layout. FIGURE 4.25 AN-1 and AN-2 above traffic lane.

FIGURE 4.24 Overview of anemometer layout.

58 FIGURE 4.24 Overview of anemometer layout. FIGURE 4.25 AN-1 and AN-2 above traffic lane.

FIGURE 4.25 AN-1 and AN-2 above traffic lane.

59

59 FIGURE 4.26 Anemometers AN-3 and AN-4 above post. Anemometers AN-1 and AN-2 were used to

FIGURE 4.26 Anemometers AN-3 and AN-4 above post.

Anemometers AN-1 and AN-2 were used to measure the wind created from trav-

eling trucks. They were placed along the centerline of the traffic lane above where the

trucks traveled. The distance was 30 ft (9.14 m) from the centerline of the post support.

The anemometers were attached onto the truss structure using steel extensions as to not

60

be blocked by the sign. Two anemometers were used for redundancy issues, as well as for

providing differing planes of measurement: AN-1 measured the horizontal plane whereas

AN-2 measured the vertical plane. This enabled a three dimensional wind velocity vector

to be measured and used in the data analysis.

Anemometers AN-3 and AN-4 were used to measure the ambient wind environ-

ment. The anemometers were attached to a steel extension that reached 4 ft (1.22 m)

above the structure to avoid any effects for wind dynamics (turbulence, vortices, etc.)

caused by the structure. Two anemometers were used for redundancy as well as for pro-

viding differing measurement planes: AN-3 measured the vertical plane whereas AN-4

measured the horizontal plane. This enabled a three dimensional wind velocity vector to

be measured and used in the data analysis. Each anemometer provided wind velocity and

direction angle measured from an identified North direction marked on the instrument

with measurement recordings in compass bearings. For all anemometers, the North com-

pass on the instrument was directed from the front façade of the sign, opposing the direc-

tion of traffic.

Accelerometers

Accelerometers were used to obtain the acceleration response of the structure.

They were of the piezoelectric type that uses a piezoelectric crystal mounted to a small

mass from which the voltage output is converted to acceleration. Each accelerometer had

a maximum capacity of 96.5 ft/sec 2 (3 G). The data was used to determine major dynamic

characteristics of the structure such as:

61

Natural frequencies and periods

Modal shapes

Critical damping percentages

The locations of the accelerometers were strategically chosen, as the accelerome-

ter only gave the natural frequency of the member upon which it rests. A combined total

of six unidirectional accelerometers were required for the structure. They were placed at

particular locations to measure each possible degree of freedom in vibration direction of

the structure. Accelerometer locations are shown in Figure 4.27.

AC-1 AC-2 Centerline Post AC-3 Centerline Truss
AC-1
AC-2
Centerline
Post
AC-3
Centerline
Truss

FIGURE 4.27 Accelerometer locations for cantilever structure.

62

Three accelerometers were placed at location AC-1 (see Figure 4.27) to measure

the vertical (perpendicular to the direction of traffic), horizontal (parallel to the direction

of traffic), and longitudinal (transverse to the direction of traffic) accelerations of the post

member. Two accelerometers were placed at location AC-2 to measure the vertical and

horizontal directions (Figure 4.28). One accelerometer was placed at location AC-3 to

measure the horizontal direction for indentifying torsion behavior of the overhang truss

(Figure 4.29). An overview of locations AC-2 and AC-3 is shown in Figure 4.30.

The accelerometers were fixed to the mounting surface by means of a two

threaded screws. Electrical insulation between the accelerometers and the test surface was

built into the transducer device. All mounting surfaces were flat to avoid distortion that

may produce strains which could affect the accelerometer’s response. Carefulness was

taken when screwing the accelerometers to the attachment surface as to not overreach the

torque recommended by the manufacturer.

Mounting blocks were used at each location for accelerometer attachment (see

Figure 4.31). For instance at location AC-1 where three accelerometers were needed, all

three were attached to one single attachment block. A picture of this location with the at-

tached mounting block is shown in Figure 4.32. Once the truss section of the structure

was erected, the AC-1 accelerometers were screwed onto the mounting block with the

sensitive ends (measurement direction) oriented in the proper directions. The same type

of block was used at locations AC-2 and AC-3. Due to the round surface at these loca-

tions (steel pipe truss web member), a small flat steel plate was welded to the side of the

members for which the mounting blocks and accelerometers were attached. The size of

the plate was manufactured as small as possible, with enough space to accurately and se-

63

curely mount the accelerometers and the plate to the web member, but without the possi-

bility of creating significant additional wind drag. The initial mounting instructions pre-

pared for ALDOT are given in Appendix C.

Data Acquisition System

All instrumentation was wired into a data acquisition system. The data was con-

verted into engineering units by the data acquisition system, filtered, and stored onto the

hard drive of a portable computer. Data was then saved onto Maxwell CD-R data storage

disks to be distributed after testing. A white van was used to hold the data acquisition

system and computer during testing. It was driven to and parked underneath the sign

structure on the side of the highway for testing. Figure 4.33 shows a typical test setup

with the van and all instrumentation wiring fed through a side hole in the van and hooked

to the acquisition system. A close up of the data acquisition system and computer is

shown in Figure 4.34.

The data acquisition was capable for collecting all data from instrumentation si-

multaneously, which was required for the fatigue tests. The maximum number of chan-

nels used was as follows:

Strain Gauges: 48 channels

Anemometers: 8 channels

Accelerometers: 6 channels

Total Channels = 62

64

64 FIGURE 4.28 Accelerometer location AC-2. FIGURE 4.29 Accelerometer location AC-3.

FIGURE 4.28 Accelerometer location AC-2.

64 FIGURE 4.28 Accelerometer location AC-2. FIGURE 4.29 Accelerometer location AC-3.

FIGURE 4.29

Accelerometer location AC-3.

65

65 FIGURE 4.30 AC-2 and AC-3 view. FIGURE 4.31 Accelerometer mounting block.

FIGURE 4.30 AC-2 and AC-3 view.

65 FIGURE 4.30 AC-2 and AC-3 view. FIGURE 4.31 Accelerometer mounting block.

FIGURE 4.31 Accelerometer mounting block.

66

66 FIGURE 4.32 Mounting block at AC-1. FIGURE 4.33 Typical test setup with van data acquisition

FIGURE 4.32 Mounting block at AC-1.

66 FIGURE 4.32 Mounting block at AC-1. FIGURE 4.33 Typical test setup with van data acquisition

FIGURE 4.33 Typical test setup with van data acquisition system.

67

67 FIGURE 4.34 Van data acquisition system and computer setup. It was important to be specific

FIGURE 4.34 Van data acquisition system and computer setup.

It was important to be specific about the sampling rate of the data acquisition sys-

tem. The vibration behavior of sign support structures are typically around 1 to 10 Hz,

including all modes of vibration. In view of this, data were collected at 60 samples per

second for all instrumentation, which was at least six times greater than the highest an-

ticipated modal frequency of 10 Hz. The Nyquist frequency was therefore 30 Hz, allow-

ing all frequencies recorded below 30 Hz to be accurate, avoiding aliasing and other

prominent sampling errors. A digital low-pass Butterworth filter was used and set at 20

Hz to filter out unwanted high frequencies from the collected data.

68

CHAPTER 5

STRUCTURAL TESTING FOR NATURAL WIND AND TRUCK-INDUCED WIND GUST

Chapter Overview

The sign support structure was tested under loading conditions: a) Natural wind

gust, and b) Truck induced wind gust. The wind data was taken over an extended time

period in an effort to capture the predominant natural wind gusts. Wind data from the Na-

tional Weather Service near Huntsville, Alabama USA, in the form of the annual mean

wind velocity for the area, was determined to help schedule appropriate testing days be-

tween ALDOT and the UAB research team, and to compared to the wind measurements

taken for accuracy so that a proper representation of the natural wind environment was