xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001795798
007 cr mnu|||uuuuu
008 070712s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001584
Evaluation of AASHTO design specifications for cast-in-place continuous bridge deck using remote sensing technique
h [electronic resource] /
by Ebrahim Mehranipornejad.
[Tampa, Fla] :
b University of South Florida,
ABSTRACT: This research project concerns the construction, testing, and remote health monitoring of the first smart bridge structure in Florida, the East Bay bridge in Gibsonton, Hillsborough County. The East Bay Bridge is a four span, continuous, deck-type structure with a total length of 120' and width of 55'. The superstructure consists of an 18'' cast-in-place reinforced concrete slab, and is supported on pre-stressed pile bents, each consisting of 5 piles. The smart sensors used for remote health monitoring are the newly emerged Fabry --Perot (FP) Fiber Optic Sensors, and are both surface-mounted and embedded in the concrete deck.Static and Dynamic testing of the bridge were performed using loaded SU-4 trucks, and a finite element model for the bridge was developed for the test cases using commercial software packages. In addition, the smart sensors were connected to a data acquisition system permanently installed on-site. This system could be accessed through regular phone lin es, which permits the evaluation of the bridge behavior under live traffic loads.Currently, these live structural data under traffic loading are transmitted to Hillsborough County's bridge maintenance office to assist in the health evaluation and maintenance of the bridge.AASHTO LRFD Design Code has been investigated using analytical and laboratory test but no attempt has been made to verify its relative outlook with respect to Allowable Strength Design (ASD) and AASHTO Standard Specifications (LFD) in a real field test. The likely reason for could have been the lack of accurate and reliable sensing systems.The data collected as well as the analytical studies through out this research, suggest that current LRFD design specifications for deck-type bridges are conservative. The technology developed under this work will enable practical, cost-effective, and reliable systematic maintenance of bridge structures, and the study will provide a unique opportunity for future growth of this tech nology in the state of Florida and in other states and finally, long term collected data can be used to keep the design codes in check.
Dissertation (Ph.D.)--University of South Florida, 2006.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 246 pages.
Adviser: Ashraf Ayoub, Ph.D.
Remote structural monitoring.
Reinforced concrete-deck type bridges.
x Civil Engineering
t USF Electronic Theses and Dissertations.
Evaluation of AASHTO Design Spec ifications for Cast-In-Place Continuous Bridge Deck Using Remote Sensing Technique by Ebrahim Mehranipornejad A dissertation submitted in partial fulfillment of the requirement s for the degree of Doctor of Philosophy Department of Civil and En vironmental Engineering College of Engineering University of South Florida Major Professor: Ashraf Ayoub, Ph.D. Autar Kaw, Ph.D. Gray Mullins, Ph.D. Glen Besterfield, Ph.D. Kandethody Ramachandran, Ph.D. Date of Approval: March 31, 2006 Keywords: Fiber Optic, Strain Sens ors, Remote Structural Monitoring, Smart Structures, Fabry Perot, Reinforced Concrete-Deck Type Bridges Copyright 2006, Ebrahim Mehranipornejad
NOTE TO READER The original of this document contains color that is necessary for understanding the data. The original dissertation is on file with USF library in Tampa, Florida.
i TABLE OF CONTENTS LIST OF TABLES v LIST OF FIGURES vii ABSTRACT xix PREFACE xxi CHAPTER 1. INTRODUCTION 1.1 Introductory Background 1 1.2 Problem Statement 4 1.3 Objectives and Scope of Work 6 1.4 Overview of Following Sections 7 1.4.1 History of AASHTO Standard Specifications and AASHTO LRFD Code 7 1.4.2 Sensing Technology 8 18.104.22.168 Electrical Resistance Strain Gauge 9 22.214.171.124 Base-Line System 9 126.96.36.199 Global Position ing System, (GPS) 10 188.8.131.52 Hydrostatic Leve ling System (HLS) 10 184.108.40.206 Linear Variable Dis placement Transducers 11 220.127.116.11 Accelerometers 11 1.4.3 Fiber Optic Sensors 11 1.4.4 Fiber Optic Sensors Time Scale 13 18.104.22.168 Fabry-Pero t Interferometer 16 22.214.171.124 Fiber Bragg Grating Sensor 19 126.96.36.199 Long Gauge Fiber Optic Sensor 22 1.4.5 Sensing Systems 25 1.5 Literature Review on Applic ation of Fiber Optic Sensors 26 1.5.1 Low Coherence Fiber Optic Deformation Sensors 26 1.5.2 Long-Gauge Structural Moni toring of Civil Structures 26 1.5.3 Use of Fiber Reinforced Polymer Reinforcement Integrated with Fiber Optic Sensors for Concrete Bridge Deck Slab Construction 27 1.5.4 Test Model for the Fi rst Canadian Smart Highway Bridge 28 1.5.5 Using Fiber Bragg Gr ating Sensors to Monitor Pavement Structures 30
1.5.6 Using Sensors for Remote Field Test 30 1.6 Summary of Research Work and Implementation of Objectives 31 1.7 An Overview of Dissertation 32 CHAPTER 2. EXPERIMENTAL PROGRAM 2.1 Beams Fabrication for Laboratory Test 35 2.2 Laboratory Test Setup 36 2.3 Data Acquisition System Components (Hardware) 37 2.4 Concrete Surface Preparation 38 2.5 Installation of Beam on Reaction Frame 40 2.6 Laboratory Loading Condition 41 2.7 Conclusions 46 2.8 Proposed Remote Sensing System 47 2.9 Field Experiment 50 2.10 Determine Location of Sensors 51 2.10.1 Transverse Position of Sensors 51 2.10.2 Longitudinal Position of Axels on the Bridge Deck 52 2.10.3 Locations of Embedded a nd Surface Mount sensors 53 2.11 Field Readi ness and Planning 55 2.12 Methodology and Procedure 56 2.13 Surface Preparation of Steel Bars 57 2.13.1 Protecting the Sensors and Optical Fibers in the Slab 65 2.13.2 Protection of Fiber Optic Cables Out of Slab 66 2.13.3 Special Installation of FOS-B, P1 and P2 68 2.13.4 Protecting Fiber Optic Cables in PVC Conduits 74 2.14 Installation of Electric Powe r and Telephone on the Bridge 79 2.15 Truck Static Load-Test 81 2.16 Conclusions 84 CHAPTER 3. DESIGN AND ANALYSIS 3.1 Introduction 85 3.1.1 LRFD Code vs. AASHTO Standard Specifications 85 3.1.2 LFD Design Method 87 3.2 LRFD Design Method 90 3.3 Bridge Load Rating Usi ng Load Factor Method 92 3.3.1 Operating Rating 93 3.3.2 Inventory Rating 93 3.4 Steps in Designing of East Bay Road Bridge 96 3.4.1 General Specifications 96 3.4.2 Design Specifications 96 3.4.3 Design Method 97 3.4.4 Design Loading 97 3.4.5 Material Property 97 3.4.6 Code Distribution Width, 97 DE 3.5 Analysis 97 3.5.1 Service Moments 98 ii
iii 3.5.2 Cracked Section Analysis 98 3.5.3 Uncracked Section Analysis 101 3.6 Application of Florida Legal Loads 101 3.7 Static and Dynamic Load Testing of Bridge 109 3.8 Finite Element Modeling 122 3.9 Beam Model Analysis S ubject to Static Load 124 CHAPTER 4. DATA COLLECTION 4.1 Remote Monitoring of the East Bay Road Bridge 134 4.2 Running the Softwar e, FISO Commander 137 4.2.1 Detect Comport for Laptop Communication Port 137 4.2.2 Detect Comport for Desktop Remote Connection 139 4.3 Using Different Versions of FISO Commander Software 142 4.4 Running FISO Commander v2 Software 144 4.5 Presentation of Remotely Collected Dynamic Data 151 4.6 Dynamic Response of the Bridge Subject to Live Traffic Load 176 4.7 Conclusions 181 CHAPTER 5. RESULTS AND COMPARISONS 5.1 Introduction 182 5.2 Evolution of the Topi c of this Dissertation 182 5.3 Selection of Sensing System for Bridge Instrumentation 185 5.4 Comparison Between the Mo st Commonly Used Sensors 186 5.4.1 Fiber Brag Gr ating and Long Gauge Fiber Optic Sensors 186 5.4.2 Fabry-Perot Fiber Optic Sensor 187 5.3 Site Specific Instrumentation 188 5.4 The Significance of Obje ctives of this Study 188 5.5 Discussions 190 5.6 Evaluation of Collected Data 192 5.7 Flexural Cracking in Bridge Concrete Deck 194 5.8 Evaluation of Design Specification 200 5.9 Damage Identification of the East Bay Road Bridge 206 CHAPTER 6. SUMMARY, CONCLU SIONS AND FUTURE WORK 6.1 Research Planning 209 6.1.1 Laboratory Test and Field Investigation 209 6.2 Laboratory Application of Sensors 210 6.3 Field Application of Sensors 210 6.4 Conclusions and Recommendations 210 6.4.1 Fine Tuning of LRFD Code 210 6.4.2 Cracked Section vs. Uncracked Section 211 6.4.3 Load Distribution Widt h (Code Tributary Width) 211 6.4.4 Discuss the Effect of Parapets and Traffic Barriers on Bridge Deck Stiffness 211 6.5 Future Studies 212
iv 6.5.1 Continuous Monitoring of the East Bay Road Bridge 212 6.5.2 Damage Identif ication of the East Bay Road Bridge 212 6.5.3 Weight-In-Mo tion (WIM) Systems 212 6.5.4 Wireless Sensors 213 6.5.5 Estimate of Bridge Life Expectancy 213 6.5.6 Development of New Bridge Management Systems Using Remote Health Monitoring Technique 214 REFERENCES 216 APPENDICES 219 Appendix A Flat Slab Design 220 ABOUT THE AUTHOR End Page
LIST OF TABLES Table 3.1 Yield Strength of Di fferent Grades of Steel 94 Table 3.2 Florida Legal Load and Design Live Load Trucks 95 Table 3.3 Moments due to Design Truck Loading, kip-ft 102 Table 3.4 Tensile Forces T s, due to Service Limit State Moment, kips 102 Table 3.5 Actual Stress, f s in Reinforcing Steel, ksi 103 Table 3.6 Steel Strain, of Cracked Section, 103 Table 3.7 Stress, of Uncracked Section, ksi 103 Table 3.8 Strains, of Uncracked Section, 104 Table 3.9 Dynamic Response of One Surface Mount Sensors 118 Table 3.10 Dynamic Response of One Embedded Sensor 118 Table 3.11 Dynamic Response of Two Sensor 119 Table 3.12 Dynamic Response of Four Sensors 119 Table 3.13 Dynamic Response of Eight Sensors, Two SU4 Trucks in Southbound Direction 120 Table 3.14 Dynamic Response of Eight Sensors, Two SU4 Trucks in Both Directions 121 Table 4.1 On Site Data Collection with a Laptop Computer from Static Load Test with SU4 Truck, GVW= 67,360 lbs 131 v
Table 4.2 Remotely Collected Data wit h a Desktop Computer from Static Load Test SU4 Truck, GVW= 67,360 lbs 132 Table 4.3 Verification of Remotely Collected Data with a Desktop Computer from Static Load Te st with SU4 Truck, GVW= 67,360 lbs 133 Table 4.4 Strain Values of Sens or CSM for Different Speeds 156 Table 4.5 Strain Values of Sensor F for Different Speeds 156 Table 4.6 Strain Values of Sensors CSM and F for Different Speeds 156 Table 4.7 Strain Values of S ensors H, CSM, P2, and F for Different Speeds 157 Table 4.8 Strain Values of Eight Sensors H, I, DSM, CSM, P1, P2, E and F for Different Speeds, Both Directions 158 Table 4.9 Strain Values of Eight Sensors H, I, DSM, CSM, P1, P2, E and F for Different Speeds, Southbound 159 Table 5.1 Strain Progression with Respect to Time, 207 vi
vii LIST OF FIGURES Figure 1.1 Structure of Optical Fiber 13 Figure 1.2 Components of Optical Fiber Cable Used with Sensors 14 Figure 1.3 Types of Fiber Optic Cables 15 Figure 1.4 Schematic Present ation of Fabry-Perot Sensors Components 16 Figure 1.5 Fabry-Perot Sensor Encaps ulated in Micro Capillary Tube 17 Figure 1.6 Principle of Fabry-Parot Strain Measuring System 19 Figure 1.7 Fabry-Parot Sensor 20 Figure 1.8 Fiber Bragg Grating Sensor 22 Figure 1.9 Fiber Bragg Grating Sensor System Components 23 Figure 2.1 Beam (1), 4 # 3 Rebars 36 Figure 2.2 Beam (2), 3 # 3 Rebar 36 Figure 2.3 Testing Framework 37 Figure 2.4 Conditioner Setup 37 Figure 2.5 Hydraulic Pump System 37 Figure 2.6 Hydraulic Jack System 37 Figure 2.7 RS 232 Communication Cable 38 Figure 2.8 Bus System 38 Figure 2.9 Computer Linked to Bus 38
viii Figure 2.10 Two Components Epoxy 39 Figure 2.11 FOS-N Installed 39 Figure 2.12 Load Assembly 40 Figure 2.13 Suspended Brackets 40 Figure 2.14 Jacks Bottom Plates 41 Figure 2.15 Jacks Top Plates 41 Figure 2.16 Digital Caliper Assembly 41 Figure 2.17 Digital Caliper 41 Figure 2.18 Experimental Beam Cross Section 43 Figure 2.19 Load and Deformation Graph 45 Figure 2.20 Tension Cracks 46 Figure 2.21 Propagatio n of Cracks 46 Figure 2.22 Cracks Directly U nder the Load and on the Side 46 Figure 2.23 Cracks Directly Under the Load 46 Figure 2.24 Proposed Remote Sensing System 48 Figure 2.25 Profile of the East Bay Road Bridge 49 Figure 2.26 FP Sensors Bonded to Reinforcing Steel 49 Figure 2.27 FP Sensors in Conduit 49 Figure 2.28 Elevation View of Old East Bay Road Bridge 50 Figure 2.29 Transverse Positions of Wheels on the Bridge Deck 51 Figure 2.30 Longitudinal Spacing of Axels in SU4 Truck 52 Figure 2.31 Locations of 16 S ensors on Two Northbound Lanes 54 Figure 2.32 Surface Bonded Sensor s ASM, BSM, CSM, DSM and P1 and P2 Sensors, Slig htly Below the Surface 55
ix Figure 2.33 Sensors C, D, E, F, T1 (Bottom) and G, H, I, J & T2 (Top) 55 Figure 2.34 Position of Sensors on Bottom Mat Rebar 58 Figure 2.35 M-Bond 5 Minutes Adhesive 59 Figure 2.36 M-Coating and Neutralizer 59 Figure 2.37 Area of Rebar to Place the Sensor on 59 Figure 2.38 Sensor Secured on Rebar 59 Figure 2.39 A Very Small Drop of 5-Minutes Epoxy Placed on In coming Optical Fiber 60 Figure 2.40 Correct and In correct Procedure for Sensors with Epoxy 60 Figure 2.41 Mylar Tape was Applied to Sensor to Keep it in Good Contact with Rebar 61 Figure 2.42 Final Procedural Step s of Sensor Installation 61 Figure 2.43 Placing Mylar Tape on Sensor Optic Fiber 62 Figure 2.44 Sensor Wrapped in Nitrite Rubber and Placed in Conduit 62 Figure 2.45 Single Channel Data Logger Reads the Strain of Sensor in nm 62 Figure 2.46 PVC Conduit Protection 66 Figure 2.47 Seal PVC Conduit 66 Figure 2.48 Secure PVC Conduit To Rebar 66 Figure 2.49 Placing Fiber Optic in Conduit 66 Figure 2.50 G, H, I and J Sensors in Conduit Exiting the Forms 67 Figure 2.51 C, D, E and F Sensors in Conduit Exiting the Forms 67 Figure 2.52 FO Cables in the Box 67 Figure 2.53 Forms are Removed 67
x Figure 2.54 Cables are Safely Out of Bridge Slab 68 Figure 2.55 Box Hous ing the Cables 68 Figure 2.56 Bedding for P1 Sensor 69 Figure 2.57 Bedding for P2 Sensor 69 Figure 2.58 Edge Bedding for P1 & P2 70 Figure 2.59 Edge Bedding for P1 & P2 with PVC 70 Figure 2.60 Beddings are Pr epared and Ready to Install P1 and P2 Sensors and Optical Cable 70 Figure 2.61 Sensor P1 is Installed 71 Figure 2.62 Sensor P2 is Installed 71 Figure 2.63 P1 and P2 Sensors 71 Figure 2.64 P1 and P2 Sens ors out of the Slab at the Edge of the Bridge 71 Figure 2.65 Material were Used to Install Sensors P1 and P2 on the Deck Over Bent 2 72 Figure 2.66 Prot ective Box 72 Figure 2.67 Final Step, P1 and P2 Sensors in the Protective Box 72 Figure 2.68 Sensors Housed in the Protective Boxes 73 Figure 2.69 Accessibility Probl em at the Bridge Edge 75 Figure 2.70 Accessibility Probl em to Install Conduits 75 Figure 2.71 Scaffold Installation from the Top 75 Figure 2.72 Scaffold Insta llation from the Bottom 75 Figure 2.73 Scaffold Installation in Progress 76 Figure 2.74 Cover the Sca ffold with Wood Planks 76 Figure 2.75 Scaffold Installation is Completed 76
xi Figure 2.76 Scaffold Installation is Approved for Use 76 Figure 2.77 Potential for Damage to Fiber Optic Cables and Sensors 77 Figure 2.78 Unprotected FO Cables were Damaged 77 Figure 2.79 Attaching Condu it to the Bridge 78 Figure 2.80 Fishing Cabl es Through Conduit 78 Figure 2.81 Conduits Entering DMI 78 Figure 2.82 Conduits are A ttached to the Bridge and Connected to the DMI 78 Figure 2.83 FO Cables are Guided Th rough Conduit into DMI System 78 Figure 2.84 FO Cables and Sensors are in DMI Box 78 Figure 2.85 Telephone Line is Secured on the Bridge 79 Figure 2.86 Electric Line is Secured on the Bridge 79 Figure 2.87 Six Cases of Static Truck Load-Test 80 Figure 2.88 Marking Locations of the Sensors on the Deck Topside 82 Figure 2.89 Case 1, Span 1, Northbound 82 Figure 2.90 Case 2, Span 2, Northbound 82 Figure 2.91 Case 3, Span 1 and 2, Trucks in Tandem, Northbound 83 Figure 2.92 Case 4, Span 2, Two Trucks Side-By-Side, Northbound 83 Figure 2.93 Case 5, Span2, Two Trucks Side-By-Side, Northbound and Southbound 83 Figure 2.94 Case 6, Span1, Two Trucks Side-By-Side, Northbound and Southbound 83 Figure 2.95 On Site Direct Da ta Collection via RS-232 84 Figure 3.1 Notional HS20-44 Truck, Axles, Wheel Spacing and Weights of Each Axle 89
Figure 3.2 Alternate Military Loading 89 Figure 3.3 Lane Load on Continuous S pan for Positive Moment 89 Figure 3.4 Lane Load on Continuous Span for Negative Moment 90 Figure 3.5 LRFD Design Load Combinatio ns (HL-93), Positive Moment 90 Figure 3.6 LRFD Code, Design Load to Produce Maximum Negative Moment 91 Figure 3.7 Distributions Width this figure Shows Actual DE Cross Section of East Bay Road Bridge Superstructure 91 Figure 3.8 Graph of (+) 0. 009 and (-) 0.023 inches of Deflections due to HS20 Design Truck Loading 104 Figure 3.9 Graph of (+) 631.8 and (-) 882.3 Moments due to 0.64-kip/ft Uniform Lane Load 105 Figure 3.10 Graph of (+) 1568.3 a nd (-) 989.7 Moments due to SU2 Truck Loading (Moment Values are Shown in Table 3.1) 105 Figure 3.11 Graph of (+) 2774.9 a nd (-) 1893.3 Moments due to SU3 Truck Loading (Moment Values are Shown in Table 3.1) 106 Figure 3.12 Graph of (+) 3019.4 a nd (-) 2050.1 Moments due to SU4 Truck Loading (Moment Values are Shown in Table 3.1) 106 Figure 3.13 Graph of (+) 1685.7and (-) 1752.8 Moments due to C3 Truck Loading (Moment Values are Shown in Table 3.1) 107 Figure 3.14 Graph of (+) 2467 and (-) 2350 Moments due to C4 Truck Loading (Moment Values are Shown in Table 3.1) 107 Figure 3.15 Graph of (+) 1900.8 and (-) 2268.8 Moments due to C5 Truck Loading (Moment Values are Shown in Table 3.1) 108 Figure 3.16 Graph of (+) 2830.4 a nd (-) 2262.5 Moments due to HS20 Truck Loading (Moment Values are Shown in Table 3.1) 108 Figure 3.17 Bridge Load Te st with SU4 Trucks 110 Figure 3.18 Experimenta l Strain Contour Line s, Load Case 1 Truck Positioned at Mid Span 1. Units in 111 xii
Figure 3.19 Experimenta l Strain Contour Lines, Load Case 2 Truck Positioned at Mid Span 2. Units in 112 Figure 3.20 Experimental Strain Contour Lines, Load Case 3, Trucks Positioned at Mid Span 1 and Mid Span 2. Units in 113 Figure 3.21 Experimental Strain Contour Lines, Load Case 4, Trucks Positioned at Mid Span 2, Both Trucks are in North Direction. Units in 114 Figure 3.22 Experimental Strain Contour Lines, Load Case 5, Trucks Positioned at Mid Span 2, Northbound and Southbound. Units in 115 Figure 3.23 Experimental Strain Contour Lines, Load Case 6, Trucks Positioned at Mid Span 1, Northbound and Southbound. Units in 116 Figure 3.24 Dynamic Strain 117 Figure 3.25 Analytical St rain Contour Lines for SU4 Truck on Span 2. Units in 123 Figure 3.26 Experimen tal Strain Contour Lines, Load Case 2, 123 Figure 3.27 Moment Diagram fo r Beam Model for Case 1 Static Load Test 125 Figure 3.28 Moment Diagram fo r Beam Model for Case 2 Static Load Test 126 Figure 3.29 Moment Diagram fo r Beam Model for Case 3 Static Load Test 127 Figure 4.1 Locating the Sensors on Deck Topside 134 Figure 4.2 Marking the Locations of the Sensors on Deck Topside 134 Figure 4.3 SU4 Truck was Placed with the Tires of the Middle Rear Axle Placed on the Mark Over the Sensors Over Span 2 134 Figure 4.4 Field Truck Load Test and on Site Real Time Data Collection 135 Figure 4.5 Graph of Heavy Trucks on the Bridge 136 xiii
xiv Figure 4.6 Graph of Cars and Light Trucks on the Bridge 136 Figure 4.7 Initializing DMI 138 Figure 4.8 Modem Communication Initialisation Dialogue Box 138 Figure 4.9 Conditioner Initia lization Error (FTI-10) 139 Figure 4.10 Configur e Conditioner 140 Figure 4.11 Configure Transducer/Sensors Assignment 141 Figure 4.12 Direct Acquisition with Graph 141 Figure 4.13 Delay Acquisition 142 Figure 4.14 Dynamic Graph 4-Channels 143 Figure 4.15 Dynamic Graph 1-Channel 143 Figure 4.16 Dynamic Graph 2-Channels 143 Figure 4.17 Dynamic Graph 3-Channels 143 Figure 4.18 Connection Initialization 145 Figure 4.19 Modem Se tup Dialogue Box 145 Figure 4.20 Application Selection Dialogue Box 145 Figure 4.21 System Conf iguration Information 146 Figure 4.22 Gauge List and Channel Setting 147 Figure 4.23 Graphic Acquisition for Selected Sensors 148 Figure 4.24 Graphic Confi guration Dialogue Box 149 Figure 4.25 Basic and Custom Color Chart 149 Figure 4.26 Warning to Protection Unsaved Data 149 Figure 4.27 Memory/Delay Acquisition 150 Figure 4.28 File Acquisition Ap plication of v2 Software 151
Figure 4.29 Channels H (Green), I (Blue), T1(R ed), and T2 (Yellow); Strain and Temperature Gr aphs. (Speed is Unknown) 153 Figure 4.30 Graph of Strain Values for Channels E (Yellow) and I (Blue) Where T1 (Red) and T2 (Pink) are Graphs of Temperature for Top and Bottom of Slab 153 Figure 4.31 Graph of Response from P1 and P2 Sensors 154 Figure 4.32 Response of E and P2 Sensors 154 Figure 4.33 Response of D, E and F Sensors to a Tandem Condition 155 Figure 4.34 Dynamic Load Test, Two SU 4 Trucks in Tandem at 10 mph, Over a Single Sensor, CSM = 16.5 and 18.5 160 Figure 4.35 Dynamic Load Test, Two SU 4 Trucks in Tandem at 10 mph, F = 15 and 16 160 Figure 4.36 Dynamic Load Test, Si ngle SU4 Truck at 10 mph, CSM = 17.5 and F = 11 161 Figure 4.37 Dynamic Load Test, Two SU 4 Trucks in Tandem at 10 mph, CSM = 14 and 9.4 and F = 15 and 13.5 161 Figure 4.38 Dynamic Load Test, Two SU 4 Trucks in Tandem at 10 mph, H = 18.5 and 14.5, CSM = 11. 5 and 6.5, P2 = 13 and 15, F = 12 and 14 162 Figure 4.39 Dynamic Load Test, Si ngle SU4 Truck at 10 mph, H= 14 and 10, I = 14 and 6.5, DSM = 18 and 9, CSM = 19 and 5, P1 = 8.5, P2 = 10.5, E = 14.5 and 6.5, F= 15 and 8.5 162 Figure 4.40 Dynamic Load Test, Two SU 4 Trucks in Tandem at 10 mph, G = 2, H = 13, I = 9, J = 9, P1 = 5, P2 = 7.5, ASM = 2.5, BSM = 5, CSM = 6, DSM = 6, C = 6, D = 4, E = 4.5, F = 0.5 163 Figure 4.41 Dynamic Load Test, Si ngle SU4 Truck at 20 mph, CSM = 13.5 163 Figure 4.42 Dynamic Load Test, Two SU 4 Trucks in Tandem at 20 mph, CSM = 12 and 13.5 164 xv
Figure 4.43 Dynamic Load Test, Two SU 4 Trucks in Tandem at 20 mph, F = 12.5 and 14 165 Figure 4.44 Dynamic Load Test, Si ngle SU4 Truck at 20 mph, CSM = 12, F = 11.5 165 Figure 4.45 Dynamic Load Test, Two SU 4 Trucks in Tandem at 20 mph, CSM = 11.5 and 12, F = 9 and 13 166 Figure 4.46 Dynamic Load Test, Si ngle SU4 Truck at 20 mph, H = 11.5, I = 8.5, DSM = 13, CSM = 6.5, P1 = 5.5, P2 = 5.5, E = 5.5, F = 3 166 Figure 4.47 Dynamic Load Test, Si ngle SU4 Truck at 20 mph, H = 10, CSM = 6.5, P2 = 6.5, F = 5 167 Figure 4.48 Dynamic Load Test, Two SU 4 Trucks in Tandem at 20 mph, H = 3.5 and 11, I = 5.5 and 10.5, DSM = 13.5 and 1.0, CSM = 12.5 and 3 167 Figure 4.49 Dynamic Load Test, Si ngle SU4 Truck at 20 mph, H = 11, I = 10.5, DSM = 13.5, CSM = 12.5, P1 = 8, P2 = 8.5, E = 17.5, F = 12 168 Figure 4.50 Dynamic Load Test, Si ngle SU4 Truck at 20 mph, DSM = 19.5,and H = 10 168 Figure 4.51 Dynamic Load Test, Two SU 4 Trucks in Tandem at 30 mph, CSM = 13.5 and 15.5 CSM is a Surface Mount Sensor 169 Figure 4.52 Dynamic Load Test, Two SU 4 Trucks in Tandem at 30 mph, F = 17, and 15 170 Figure 4.53 Dynamic Load Test, Two SU 4 Trucks in Tandem at 30 mph, CSM = 7 and 11, F = 8 and 14 170 Figure 4.54 Dynamic Load Test, Two SU 4 Trucks in Tandem at 30 mph, CSM = 6 and 7.5, F = 7 and 10, 171 Figure 4.55 Dynamic Load Test, Si ngle SU4 Truck at 30 mph, CSM = 13 171 xvi
Figure 4.56 Dynamic Load Test, Two SU 4 Trucks in Tandem at 30 mph, H = 11.5 and 11.5, CSM = 4.5 and 5.5, F = 5 and 10.5 172 Figure 4.57 Dynamic Load Test, Single SU4 Truck at 30 mph, H = 7, CSM = 9, P2 = 6.5, F = 3 172 Figure 4.58 Dynamic Load Test, Single SU4 Truck at 30 mph, G = 3, H = 6.5, I = 9, J = 8.5, DSM = 15, CSM = 1.5, ASM = 1.5, BSM = -1.5, P1 = 4.5, P2 = 4, C = 2.5, D = 4, E = 10.5, F = 16 T1 = 12.55, T2 = 18.15 0 C 173 Figure 4.59 Dynamic Load Test, Single SU4 Truck at 35 mph, H = 11, I = 8.5, DSM = 4, CSM = 9.5, 174 Figure 4.60 Dynamic Load Test, Single SU4 Truck at 35 mph, P1 = 7, P2 = 6, E = 6, F= 12 174 Figure 4.61 Dynamic Load Test, Single SU4 Truck at 35 mph, H = 11, I = 8.5, DSM = 4, CSM = 9.5, P1 = 7, P2 = 6, E = 6, F = 12 175 Figure 4.62 Dynamic Load Test, Si ngle SU4 Truck at 40 mph, G = 0, H = 1, I = 1.5, J = 5, DSM = -2, CSM = 1, ASM = -1, DSM = 1, P1 = 1, P2 = 0.5, C = 1.5, D = 1, E = 1, F = 1 T1= 12.4, T2 = 18.25 0 C 175 Figure 4.63 Dynamic Strain Readings in Real Time Recorded on February 4, 2005 at 4:30 pm, H = 5, I = 7, E = 7, F = 5.5 176 Figure 4.64 Dynamic Strain Readings in Real Time, Recorded on June 14,2005 at 11:39 am, I = 13, F = 9.5 177 Figure 4.65 Dynamic Strain Readings in Real Time, Recorded on November 28, 2005 at 2:50 am H = 13, F = 13 178 Figure 4.66 Dynamic Strain Readings in Real Time, Recorded on January 15,2006 at, 9:28 am D = 11, E = 19, F = 28 179 Figure 4.67 Dynamic Strain Readings in Real Time, Recorded on November 30, 2005 at 7:41 am, E = 21.5, P2 = 11 180 Figure 4.68 Dynamic Strain Readings in Real time, Recorded on January 2, 2006 at 7:45 am, F=28, P1=13 180 Figure 5.1 Moment-Curvature Rela tionship for Bridge Section 194 xvii
xviii Figure 5.2 Stress at Cracking 196 Figure 5.3 Axial Force vs. Average Strain for and Axially Loaded Reinforced Specimen 198 Figure 5.4 Strain Contours with In clusion of Barrier Walls 203 Figure 5.5 Strain Contours without Inclusion of Barrier Walls 203 Figure 5.6 Damage Index of East Bay Road Bridge 208
xix EVALUATION OF AASHTO DESIGN SPECIFICATIONS FOR CAST-IN-PLACE CONTINUOUS BRIDGE DECK USIN G REMOTE SENSING TECHNIQUE Ebrahim Mehranipornejad ABSTRACT This research project concerns the construction, testing, and remote health monitoring of the firs t smart bridge structure in Florida, the East Bay bridge in Gibsonton, Hillsbor ough County. The East Bay Bridge is a four span, continuous, deck-type structure with a to tal length of 120and width of 55. The superstructure consists of an 18 cast-i n-place reinforced concrete slab, and is supported on pre-stressed pile bents, each consisting of 5 piles. The smart sensors used for remote health monitori ng are the newly emerged Fabry Perot (FP) Fiber Optic Sensors, and are both surface-mounted and embedded in the concrete deck. Static and Dynamic testing of the br idge were performed using loaded SU4 trucks, and a finite element model for the bridge was developed for the test cases using commercial software packages. In addition, the smart sensors were connected to a data acquisition system permanently installed on-site. This system could be accessed through regul ar phone lines, which permits the evaluation of the bridge behavio r under live traffic loads.
xx Currently, these live structural data under traffic loading are transmitted to Hillsborough Countys bridge maintenance office to assist in the health evaluation and maintenance of the bridge. AASHTO LRFD Design Code has been investigated using analytical and laboratory test but no attempt has been made to verify its relative outlook with respect to Allowable Strengt h Design (ASD) and AASHTO Standard Specifications (LFD) in a real field test The likely reason for could have been the lack of accurate and reliable sensing systems. The data collected as well as the analytical studies through out this research, suggest that current LRFD desi gn specifications for deck-type bridges are conservative. The technology developed under this work will enable practical, cost-effective, and reliable systematic ma intenance of bridge structures, and the study will provide a unique opportunity for fu ture growth of this technology in the state of Florida and in ot her states and finally, long term collected data can be used to keep the design codes in check.
xxi PREFACE The author would like to express his deepest gratitude to Dr. Ashraf Ayoub for his invaluable direction, profound patience, guidance, constant support and encouragement during the course of this re search. I am gratef ul for the helpful suggestions and comments of th e members of my committee, with special thanks to Dr. Autar Kaw for his ongoing support, encouragement and assistance. I am also very thankful to Dr. Amir Ayoub for his technical support and encouragement. The author gr eatly appreciates the support and help provided by the technical team of Hillsborough County, Florida. My sincerest thanks go to Mr. Bernardo Garcia, P.E. Assistant Hills borough County Administrator. Without Mr. Garcias wisdom and help, this highl y innovative and unique research would have not been possible. I am also gratef ul to Mr. Robert Go rdon, P.E. Director of the Public Works Department, Mr. Sco tt Catrell, Director of the Engineering Division; James Harper, Structural E ngineer, Bridge Section, and Mr. Maize Monroe, Project Manager. I am indebted to Mr. Larry Watts, Director of the Maintenance Department and Mr. Wayne Pool e, the East Bay Bridge truck load test support team Supervisor for his field support during the load test. The author is also grateful of the hel p provided by the technical team of the RocTest-Fiso, Technologies, the manufacturer of fi ber optic sensors and readout equipment with associated software. I am particularly thankful to Mr. Jean-Marie Br h, Mr.
xxii Frank Cossette, Mr. Stephane-Eric Thivierg e, Dr. Marco Quirion and Mr. Rene Malo,. The informational data provided by technical team TBE Group, particularly Mr. Robert Heck is greatly appreciated. Last but not the least, I would lik e to extend my deepest thanks for the boundless support, encouragement, inexhaustib le love and understanding of my wonderful wife Carol, our three great ch ildren Ebrahim II, Kavon, and Sarah and our four exceptional grandchi ldren, Kayla, Kavon II, Rosario II, and Cyrus, my dearly departed mother Roghi eh, my brother Abol, and my sister in law Peggy, throughout my study. To them I dedicate this dissertation. The financial support of this research study was provided by a grant from Hillsborough County, Florida, with matching funds from The University of Missouri. The findings and opinions in th is manuscript are t hose of the author and not necessarily those of the sponsoring organizations.
1 CHAPTER 1. INTRODUCTION 1.1 Introductory Background In 1993, American Association of State Highway and Transportation Officials (AASHTO 1989) Subcommittee on Bridges and Structures responded to interest in developing new-updated AASHTO bridge specifications with accompanying commentary. The goal was to develop more comprehensive specifications that woul d eliminate any gaps and inconsistencies in the Load Factor Design-based format (AASHTO 197 3) of standard specifications by incorporating the latest in bridge re search and technology. The decision was made to develop these specifications in a Load and Resistance Factor Designbased format (AASHTO 1993) which takes the variability of the structural elements into account thr ough the application of statistical methods. The LRFD specifications were approved by AASHTO for use as alternative specifications to the AASHTO Standard Specifications for highway Bridges (LFD). The AASHTO LRFD was evolved based on percepti on of gaps and inconsistencies, nonuniform margin of safety and less reliabilit y in LFD design specifications across a wide variety of structures. To validate these downside issues raised about LFD design standard specifications and verify acclaimed out look of AASHTO LRFD design method,
2 research of literature in related t opic, laboratory experiment and actual field load test of bridges designed by LRFD-based format are necessary. Further continuous monitoring of bridges designed by LR FD-based format deserve closer attention. The merit of this process is to generate sufficient data for analysis of structural behavior of the br idge subject to long term various truck loading conditions, stresses induced by large temperature change and extreme natural events such as hurricanes in Florida and earthquakes elsewhere. The existing East Bay Road Bridge in Gibsonton, Hillsborough County, Florida was candidate for replacement with a new four continuous span concrete bridge. Hillsborough County provided funds to install 16 fiber optic sensors (FOS), ten of which were embedded in the concrete during construction and four were surface mounted on the underside of bridge deck after completion of construction. It was decided to conti nuously monitor, observe and record behavior of the bridge under the effect of the traffic and environment for two years. Periodic monitoring at the time of two years inspection cycle will generate a history on structural behavior of the bridge. At the completion of construction, six surface mount strain sensors were installed on the bridge. Prior to opening the bridge to daily tra ffic, the bridge was subject to a series of static load tests. The static l oad test resulted in the strain values that were used to investigate and evaluate the design of the bridge under AASHTO LRFD (AASHTO 1994) design specific ations and AASHTO LFD (AASHO 1931) standard specifications. This is the primar y objective of this dissertation. The results of field static test were compar ed with an analytical model of the structure
3 to define the degree of reliability and conser vative state of the bridge design by AASHTO LRFD design specifications and AASHTO LFD standard specifications. In addition to static and dynamic tru ck load test, continuous monitoring of the bridge will be performed to obtain real live data (strain values) to compare with the design strain values described in chapter 3. This comparison will help the bridge engineers and facilities management to understand the actual condition of the bridge and its level of performance. Continuous health monitoring of bridge structures is a new area that has been driven by the necessity of efficient structural condition assessment. Presently, repair and replacement decisions for the bridges are based on highly subjective visual observations (Van Daveer 1975; Chase and Washer 1997). According to Aktan et al. (1996), subjecti ve or inaccurate condition assessment has been identified as the most critical technical barrier to the effective management of bridges, which results in annual $3 billion maintenance cost in the US (Chase and Washer 1997). Neverthe less, an earlier study (Catbas et al. 1998) has confirmed that more than 40 per cent of the bridges in the U.S. are functionally obsolete or structurally defici ent due to corrosion, scour, or subjective and inaccurate observations and data collection. In addition, several bridges have experienced major damage or collapse recently due to extreme events (e.g. earthquakes, hurricanes). Often, inaccurate structural condition assessment has lead to unfounded decisions to replace numerous reinforced concrete bridges possessing significantly large number of re maining safe operating service life.
4 With the advent of todays new technol ogies, existing and new structures can now be instrumented for evolution and veri fication of the code that they have been designed with. The measuring and monitoring systems can be conveniently operated and controlled from a remote central monitoring station that is located several miles away from the field. Sensors are placed at several critical locations along the structure, and send structural information (e.g. strains, stresses, accelerations) to the central station. The stru cture is thus thought of as a smart system that is c apable of sending information that can be used in evaluation and verification of design code an d specifications wh ile at the same time would be providing warni ngs before any major failure. Several types of advanced sensors ar e used for remote monitoring and damage detection. Fiber Optic strain Sensors (FOS) are the most commonly used, especially in Canada by the ISIS center (2001). The so-called WiMMS accelerometers have been developed at the Blume Earthquake Engineering Center at Stanford University (Stras er and Kiremidjian 1998). In addition, miniature micro-electro-mechanica l systems (MEMS) or smart dust accelerometers have been also used. 1.2 Problem Statement AASHTO LRFD Design Code (AASHTO 1944) has been investigated (Shahawy 1996) using analytical and labor atory test but no attempt has been made to verify its relative outlook with respect to Allowable Strength Design, ASD (AASHO 1931) and AASHTO Standard Specific ations, (AASHO 1931) in a real
5 field test. The likely reason for that is the lack of accurate and reliable measuring systems. Literature has noted mixed opinions regarding vague interpretation, difficult, time consuming calculations, which lead to excessively conservative results for LRFD Design code in compar ison with analytical models, laboratory test and prediction by AASHTO Standard S pecifications, LFD (Shahawy 1996). The effective repair and rehabilitation of a bridge depends on understanding of its structur al condition. This understanding begins with bridge inspection (Haque 1997). Scheduled periodic bridge inspections are tailored to detect and assess structural damages for the purpose of maintenance and replacement. Inspection and damage a ssessment based on visual observations are highly objective (Van Daveer 1975; Chase and Washer 1997). Collected data on the condition of bridges are used to determine needs for repair or replacement, and to form models of future needs. Numerical Condition Ratings (NCR) assigned to structural elements dur ing visual inspection are qualitative condition ratings and determine the level of need for repair and rehabilitation. Condition ratings are impr ecise, and the ratings ar e only a subset of the information collected during a bridge inspection (Hearn and Shim 1997). To have a better understanding of bridge condi tion and verify visual observations, various methods of nondestructive evaluations (NDE) have been developed to detect the extent of deter ioration and damage to the br idge elements. One such NDE is a static load test to determi ne structural strength and load carrying capacity of the bridge. However, neither periodic visual inspection nor random nondestructive evaluation can detec t the initiation and propagation of
6 deterioration in structural elements unt il the damages are serious and often not repairable. Continuous moni toring of a bridge is known to instantly detect the onset of damage in a bridge; associ ated with over stress by heavy load, corrosion and structural elements section losses. 1.3 Objectives and Scope of Work There are two objectives of this resear ch. First objective is the short-term application of newly emerged sensor as a tool for evaluation of AASHTO Design guidelines, and that is, to investigate the new bridge design method of LRFD and old design method of LFD, evaluate and verify the assumptions and parameters considered in design of the East Bay Road Bridge. We then compare the design of the bridge with the data obtained from the sensors installed in the bridge during construction. This data is also used to investigate and verify the results of LFD method of bridge load capacity rating. Second objective is to develop a new methodology for damage detection and cost life cycle evaluation of bridges. Wh ile the primary purpose of fitting the East Bay Road Bridge with sensors was to investigate AASHTO LRFD design specifications, the strain measuring syst em was permanently left in the structure to provide an opportunity for a long term monitori ng of the bridge condition and to develop a new methodology for damage dete ction and life cycle evaluation of the bridges. To achieve these goals, the l ong-term application of sensors to build strain history of the bridge by continuous or periodic monitoring and evaluation of collected data is necessary. Needless to say that the analysis of collected data
7 over a long period of time would be a valuable tool for diagnostic measures such as safety assessment, damage detection and rehabilitation of existing bridge. Verification of both objectives, the fiel d load testing of East Bay Bridge in addition to damage detection and evaluation of bridge condition are presented in the following chapters. 1.4 Overview of Following Sections Section 1.4.1 describes the need for a new more advanced bridge design code AASHTO LRFD, the history and development of AASHTO LRFD and technical papers written on the topic. Section 1.4.2 describes sensing technology encompassing development of di fferent sensors in a chronological order, e.g., from early basic sensors to more advanced fiber optic sensors leading to the three most commonly used sensors; Fabry Perot Interferometer, Fiber Brag Grating and Long gauge sensors. Section 1.4.3 describes the application of sensors th rough literature review on related topics. 1.4.1 History of AASHTO Sta ndard Specifications and AASHTO LRFD Code AASHO, American Association of Stat e Highway Officials, the standard specifications was formed in December 12, 1914. In 1921, AASHO organized the bridge and structures committ ee to develop and compile design specifications until the first edition of standard specifications, published in 1931 and followed by 1935, 1941, 1944, 1949, 1953, 1957, 1961, 1965, 1969, 1973,
8 1977, 1983, 1989, 1992, and 1996 revised editions. In 1973-revised edition, the letter T was added to AASHO to form, the American Association of State Highway and Transportation Officials, AASHTO. In 1993, AASHTO adopted the Load and Resistance Factor design (LRFD) specifications for bridge design and published the first edition of design specifications in 1994. AASHTO approved t he LRFD specification to be used as an alternative specification to the AASH TO Standard Specifications (LFD) for Highway Bridges. Additional versions ( editions) were developed and latest appeared in 2005. The methodology and philosophy of AASHTO LRFD Design specifications and AASHTO Standard Specifications, LFD are presented in Chapter 3. 1.4.2 Sensing Technology With the emergence of measuri ng technology, the use of traditional measuring and monitoring devices and systems have been gradually phasing out. These systems were not fully capable of continuous measuring stresses and monitoring of structures. Some of these systems consisted of several parts and components and were time consum ing and difficult to handle during installation. Some systems requir ed more than one specialized person for equipment installation and se tup. Frequent monitoring of structure with these types of measuring systems was not economically feasible. Amongst these systems include triangulation, water level, vibrating string, dial gages, invar wires, and mechanical extensometers, Base -line system, global positioning system,
9 strain gage-base system, linear variabl e displacement transducer (LVDT), accelerometer, etc. This review would briefly describe a few of the systems that are still occasionally implemented in moni toring of some specific situation. 188.8.131.52 Electrical Resistance Strain Gauge Electric resistance sensor is a device whose electrical resistance varies in proportion to the amount of strain in the device. The most widely used gauge is the bonded metallic strain gauge. The metallic strain gauge consists of a very fine wire or, more commonly, metallic fo il arranged in a grid pattern. The grid pattern maximizes the amount of metallic wire or foil subject to strain in the parallel direction. The cross sectional ar ea of the grid is minimized to reduce the effect of shear strain a nd Poisson Strain. The grid is bonded to a thin backing, called the carrier, which is attached directly to the test specim en. Therefore, the strain experienced by the test specimen is transferred directly to the strain gauge, which responds with a linear change in electrical resistance. Strain gauges are available commercially with nomi nal resistance values from 30 to 3000 with 120, 350, and 1000 being the most common values. 184.108.40.206 Base-Line System Base-line system consists of high st rength piano wire, pulley and weight. This system is only capable of measuring deflection due to static load. This system is not suited for dynamic monitori ng since the vibration of piano wire and constant tension weight w ould prevent accurate deflect ion measurements. Digital
10 Calipers and Linear Variable Displacem ent Transducer (LVDT) are used to measure the vertical deflection relative to Base Line. This system was used as deflection monitoring system in H-3 North Halawa Valley (Lee, 1995). 220.127.116.11 Global Positi oning System, (GPS) Deflection (deformation) of Br idge structural elements have been monitored using strain gages. Strain gage capability is limited to measuring deflection due to static load. In this system, in addition to the sensors installed on structure, one sensor must be locat ed on a fixed and stable reference point near the structure. All sensors including the reference sensor must have antenna and communicate with at least four GPS satellites. This system can only process one reading in every ten seconds therefore, it is not recommended for dynamic and seismic applications, (Celebi 2002) 18.104.22.168 Hydrostatic L eveling System (HLS) The hydrostatic leveling system is bas ed on the classical physical law of connected vessels. The vessels are made of calibrated glass beakers connected with transparent plastic tubes. Since the water level within the tubes always remains on a horizontal plane, ve rtical displacements can be deduced from the difference of the water levels between the defo rmed and the initial position of the structure. Vibration generated by the traffic does not influence measurements because of the great iner tia of the HLS. The error made on deflections for the overall system is about .5 mm.
11 22.214.171.124 Linear Variable Displacement Transducers Linear Variable Displacement Trans ducers, LVDTs are used to measure high frequency of relative displacement between two points on a bridge. They are capable of measuring def lection of bridge elements but require a fixed and stable reference point. They are not recommended for seismic application since a fixed object on the ground would not rema in stable during a seismic activity. 126.96.36.199 Accelerometers Accelerometers are used to measure deflection in structural members subject to dynamic loading. Deflection values are obtained by double numerical integration of accelerati on. Literatures have noted unreliability in deflection results due to integration process and undetected anomalies in the sensors recorded values, (Celibi and Sanli 2002). 1.4.3 Fiber Optic Sensors The new generation of high tech sensors render the aforementioned sensing systems obsolete. These new families of sensing sensors are technologically highly comp lex and expensive to manuf acture. But, their high cost is quickly offset by their physical simplicity to handling, versatility and easy installation. These sensors are known as fiber optic sensors. Several different fiber optic sensors have been developed in rec ent years, from a simplest form of measuring an on-off state to highly co mplex sensors capable of measuring a wide range of wavelengths.
12 Fiber Optic strain Sensors are in gener al better suited for structural health monitoring of the bridges than accelerome ters, LVDTs, HLS, GPS, etc., since they can be easily bonded to reinforcing bars and embedded in the structure, and they can provide a complete strain histor y including strains from concrete curing, construction loads and in-situ service loads, and creep and thermal changes. FOS sensors have proven to be accurate, inexpensive, and easy to use. Fiber optic sensors have numerous advantages: small size, lightweight, long-term stability, large selection of gauge length, corrosion-resistance, wide variety of packaging for surface mount ing and embedment in the structure, distributed capability, immunity to electromagnetic and radio frequency interference, and multiple xing capabilities among others. Their main advantage though lies in their remote sensing capabilities. Fiber optic sensors are manufactured eit her as discrete or distributed type. Discrete sensors come as short and long gauges. Fiber Bragg-grating and SOFO are examples of discrete shortgauge and distributed long-gauge sensors, respectively. Discrete sensors detect changes at locations where they are installed while distributed sensors detec t changes at several locations in the structure. Short-gauge s ensors are highly influenced by presence of cracks related to their locations in structure (local stress) and thus do not represent global behavior of the stru cture (e.g., deflection).
1.4.4 Fiber Optic Se nsors Time Scale Fiber optic, in a very basic form bu t yet revolutionary, was developed in 1950. The system was basically light confin ement within two layers of glass. In 1960, the laser light source was introduced in to the system. The refinement of optical fiber manufacturing methods and use of LED (light emitting diode) as a light source became practical in 1970. In 1980, optical fiber was widely used in telecommunication systems. In 1990, optical fiber wa s used in instrumentation and commercially available sensors. In 1995, the application of optical fiber on site in highway bridges bec ame possible. The following Figure 1.1 through Figure 1.9 have been reproduced and recreated with permission from RocTest, Canada. 13 50 m Coating / Buffer Cladding Core An optical fiber consists of three pr incipal elements, arranged concentrically: Figure 1.1 Structure of Optical Fiber Coating / Buffer: This is the first nonoptical layer around the cladding, typically consists of one or more layers of a pol ymer that protects the silica structure against physical or environmental damages.
Cladding: This is the first optical la yer around the core. T he cladding creates an optical wave-guide that confines the light. Cladding is usually made of silica. Core: This is the central section made of s ilica. It is the hi gh transmitting region of the fiber. A Simplex cable is a tight-buffered Optical Fiber Glass reinforced with Kevlar fiber strands and then covered with a PVC outer jacket 14 50 m PVC cable jacket Kevlar strengthening fibers Mechani cal buffer Cladding Core Figure 1.2 Components of Optical Fiber Cable Used with Sensors Fiber optics are manufactured in singlemod e and multimode fiber, Figure 1.3(a) and Figure1.3(b). Each one has different light signal transmission and properties. In single-mode fiber, only the fundamental mode is propagated, it travels straight through the fiber without reflection at the core-cladding boundary. It has higher bandwidth, 5 to 10 microns co re diameter and 125 microns cladding diameter.
(a) Singlemode Fiber In multimode fiber, higher-order m odes are propagated in addition to the fundamental. The different modes travel in curved, wavelike paths. It has lower bandwidth, 50 to 100 microns core diamet er and 125 microns cladding diameter. (b) Multimode Fiber Figure 1.3 Types of Fiber Optic Cables In the following sectio n 188.8.131.52, 184.108.40.206 and 220.127.116.11, the general configuration, working principle and application of these three commonly used sensors are presented. The most comm only used (FOS) for health monitoring of bridge structures are: (1) the Fabry Perot Interferomet er (FPI), (2) the Fiber Bragg Grating (FBG), and (3) the long-gauge sensors. 15
18.104.22.168 Fabry-Perot Interferometer Figure 1.4 is the schematic of Fabr y-Perot fiber optic sensor depicting components of sensor, direction of signal an d the light source. The light hits the mirror and reflects back to the readout unit (e.g., DMI-16). The principle of interferometer is a unique feature to Fabr y-Perot sensor. Inte rferometer is an optical instrument that allows two beams of light deriv ed from a single source (and thus of the same frequency and in pha se at identical distances from the source) to traverse paths whose difference in length determines the nature of the interference pattern obtained when the beams are allowed to interfere. The wavelength of light can be measured if t he path length difference is known, and vice versa. Optical Fiber Mirror Fabry-Perot Cavity Optical Fiber Mirror Fabry-Perot Cavity Figure 1.4 Schematic Presentation of Fabry-Perot Sensors Components Figure 1.5 is the schematic of an encapsulated Fabry-Perot fiber optic sensor. In this figure, the actual co mponents of the Fabry-Pe rot are shown in a 10mm micro capillary tube. The reflect ed light is traveling toward the readout 16
unit. The magnitude of strain is functi on of ratio of change in cavity length and gage length. Lg d Strain Measurement is achieved by measuring the Fabry-Perot cavity length using white light in terferometer. Optical fiber FP cavity Mirror Micro capillary To Readout 10 mm d Lg To Readout 10 mm d d Lg Figure 1.5 Fabry-Perot Sensor Encap sulated in Micro Capillary Tube Fabry-Prot interferometer (FPI) manufactu red by Roctest is basically consisted of two multimode optical fibers, 50 to 12 5 microns thick facing each other. The two fibers are placed inside a 200 microns diameter gl ass micro-capillary. The tips of fiber ends facing each other are coated with Semi-reflective coating acting as mirrored reflectors. The space separat ing the two mirrors is called the cavity length. Light from a broadband source is aimed at one arm of a 2 x 2 coupler and directed toward the Fabry-Prot gauge along an incoming multi-mode optical fiber. Light reflected in the FPI is wa velength-modulated in accordance with the cavity length. The reflected light signal travel through the fiber into a read-out unit. At this point, the light tr avels through a white-light cross17
18 correlator (Fizeau Interferometer), and detected by a linear Charged-Coupled Device (CCD) array with a pixel arrangement that allows for 1:10,000 resolution. Finally, the incoming fiber that transports light to the gauge is mechanically decoupled or isolated from the strain sensing fiber (Figure 1.5) The Fabry-Prot interferometer (FPI) gauge converts strain into cavity length variations measurements (Figure1.6). Fabry-Pro t has been used to monitor the behavior of several structures such as Morristo wn bridge in Vermont (Benmokrane, et at. 2003), and the Joffre bridge in Sherbrooke, Canada (Choquet et al. 2000) among others. The principle of this measuri ng system is shown in Figure 1.7. FabryPerot (FP) sensor has the follo wing unique characteristic: While a calibration process is required for each sensor, in-line FP sensors provide low thermal sensitivity because t he cavity is in air, combined with a welldefined gauge length and relati vely high strength. Since the FP sensor is decoupled from the surrounding micr o-capillary, it avoids creep that might arise from t he use of adhesives. Sensed information is the Fabry-Perot cavity length, which is also an absolute parameter. The output does not depend directly on t he total light intensity levels, losses in the connecting fibers and couplers, or recalibration or re-initialization of the system. Fabry-Perot strain gage uses a multi-mode fiber instead of a single mode fiber. Fabry-Perot sensors are easier to splice, repair and connect. Their transducers would lose less light when subjected to bending.
Figure 1.6 Principle of Fabry-Pe rot Strain Measuring System 22.214.171.124 Fiber Bragg Grating Sensor A fiber Bragg Grating (FBG) can be fabricated from a continuous germanium doped fiber core, surrounded by germanium-doped silica. The grating portion consists of a modulation in the index of refraction along a length of continuous fiber core. A change in length of the grating is due to mechanical or thermal strain in the host material. The change in the length of grating is detected as a shift in the wavelength of the reflected lig ht. Bragg grating measures strain based on wavelength shift. 19
Gage length (Lg) Fabry-Perot cavity length (0 to few tens of microns) Opticalfiber Capillary tube 200 m mirror Fusion spots Figure 1.7 Fabry-Perot Sensor The light source can be either a broadband light emitting diode or a tunable laser over a specified wavelength rang e. Bragg gratings are supplied with a section of the coating around the grating removed to allow for installati on and bonding. The grating itself may appear as a barely-perceptible optical fiber difficult to see with a naked eye. To create the Bragg grating sensor, ultraviolet (UV) light is directed perpendicular to the core of the fiber periodi cally, along a defined section of the fiber optic cable. The process is referred to as wiring FO S gratings (ISIS Design Manual 1, 2001. 20
21 Photonics Research Ontario (PRO) Center of Excellence and E-TEK Electro Photonics Solutions are the manuf acturer of some Fiber Bragg Grating fiber optic sensors. Fiber Bragg Grat ing (FBG) Sensors have been installed on several structures such as the Commodore Barry Bridge in Philadelphia (Aktan et al. 2000) and the Taylor bridge in M anitoba (ISIS design manual I, 2001). Figures 1.8 and 1.9 show a general worki ng principle of Fi ber Bragg Grating (FBG) Sensors. Fiber Bragg Grati ng sensor has the following unique characteristic. Sensed information is encoded directly into optical wavelength, which is an absolute parameter. Therefore, t he output does not depend directly on the total light intensity levels, losses in the connecting fibers and couplers, or recalibration or re-initialization of the system. Fiber Bragg Grat ing sensor is also capable of handling wavelength division mult iplexing by the fabrication of each grating at a slightly different frequ ency within the broadband s ource spectrum on a single fiber. In FBG, mirrors are inscribed inside the fibers. The Bragg Wavelength (IB) is function of the spacing ( ) and the refractive index n of the core.
nIB 2 Output Signal Input Signal Reflected SignalI B I Fiber Core I Output Signal Input Signal Reflected SignalI B I Fiber Core I Input Signal Reflected SignalI B I I Fiber Core I I Figure 1.8 Fiber Bragg Grating Sensor 126.96.36.199 Long Gauge Fiber Optic Sensor Long gauge sensing system comes in tw o types. One method involves using conv entional telecom optical fibers of arbitrary length configured from 2 inches to about 300 feet. This type of long-gauge can be bonded to a structure or embedded in concrete. The distance between two mirrors on the fiber optic leads defines the gauge length of the system. This type of sensor measures the change in path distance between the mirrors while bonded to the host structure or material. The system demodulates the light signals returning from the mirrors by the principle of low coherence interf erometery. The obtained deformation is the average values taken over the gauge length. 22
Figure 1.9 Fiber Bragg Grating Sensor System Components The second method, Brillouin scatteri ng also involves using conventional telecom optical fibers and can be used to measure strains due to thermal or mechanical loading. Brillouin scattering is a distributed sensor that can take readings at various points along the optic al fiber over a large distance in magnitude of ~ 1000s feet. The resolution of this system can be abut 4 to 8 inches and strain values are the aver age values taken over the gauge length. Although expensive, long gauge sensors were used to monitor the behavior of several bridges such as the Rio Puerco bridge in Ne w Mexico (Idriss, Kerseyand and Davis 1997), Highway 401 in Toronto (ISIS 2001) and Lutrive twin bridges between Lausanne and Vevey in Switzerland (Inaudi et al. to be published). Either types of long-gauge syst em is suitable for the applications 23
24 where deformation or strain is required in small or very large diameter cylinders or bridge piers due to thermal or mec hanical loading is required. Another application of these long-gauge systems is to find the st rain in deteriorating bridge pilings and piers wrapped with co mposite sheets of fiber-reinforced polymer, FRP. Long gauge sensors are not capable of high frequency monitoring and thus are not suitable of monitoring structures subject to high frequency dynamic loads. System of long-gage, strain sensors, SOFO (surveillance dOuvrages par Fiber Optiques or monitoring of Structur es by Fiber Optic Sensors) (Inaudi and Vurpillot 1998) have been used in several bridges in Switzerland for monitoring the effect of temperature fluctuati on and stresses due to static and dynamic loading on the structures. This system is best suited to determine the deflection profile of a beam type structures such as bridge, frame, etc. The sensor consists of a pair of single mode fibers installed in the structure. One of the fibers, the measurement fi ber would be in mechanical contact with structural member subj ect to measurement and the other, the reference fiber, is placed loose nearby the structure. Deformation of the structure will then result in a change of the length difference between these two fibers (Inaudi et al. 1977) SMARTEC, a Swiss company install ed 30six-meter long sensors along the length of fourth span on Lutrive Highw ay Bridge, a box girder bridge in Switzerland [Inaudi, 1999]. These sensors were used to monitor the effect of temperature variation on curvature. A double integration of curvature will result
25 in deflection. SOFO system is not capable of high frequency strain monitoring therefore, it is not recommended fo r seismic monitoring application. 1.4.5 Sensing Systems Miniature sensors represent anot her technology used for remote monitoring of structures. An attempt to apply this technology for monitoring civil structural systems was performed at t he John A. Blume Earthquake Engineering Center at Stanford University in co llaboration with Los Alamos National Laboratory. The team devel oped the so-called WiMMS (Wireless, Modular Monitoring System) for remote damage det ection. The data acquisition system is moved to the sensor unit, where the computation is performed. Sensors located at different locations in the structur e, send the information wirelessly to a centralized data storage system. WiMMs sensors are battery-operated accelerometers aimed at monitoring the vi bration characteristics of structural elements. Advanced micro-electro-me chanical (MEMS) wireless accelerometers have also been used for structural monito ring. These devices, also called Macro Motes, have been developed at the Berk eley Sensor and Actuator Center (BSAC). These devices incorporate commu nication, processing, sensing, and batteries into a package about a cubic inch in size.
26 1.5 Literature Review on Application of Fiber Optic Sensors 1.5.1 Low Coherence Fiber Optic Deformation Sensors This system was used in Versoix Bridge near Geneva, Switzerland to measure the displacements of the fresh concrete during the setting phase and to monitor its long-term deformations. The measurement technique relies on an array of standard telecommunication optic al fibers in mechanical contact with concrete. Any deformation of the host struct ure results in a change in the optical length of the fibers. Each sensor line c onsists of two singlemode fibers. One of the fibers, the measurement fiber would be in mechanical contact with structural member and the reference fiber, is plac ed loose near the other one. Deformation of the structure will then result in a c hange of the length difference between these two fibers. 1.5.2 Long-Gauge Structural Moni toring of Civil Structures The fourth span of Lutrive, a 2800 feet twin bridge was fitted with Thirty, 18 feet long SOFO sensors. The sensors were installed in pairs on interior surface of box girder near the top and botto m of bridge web. A series of strains data result in bridge curvature and a double in tegration of curvature would lead to the vertical displacement. The sensors were used to collect data for quasi-static test under thermal loading and under static load as well as for statistical characterization of the dynamic behavior of the bridge. The verification of static and dynamic values and their comparison with the analytical model and computation were not presented. A table presenting an
27 organized collected data is lacking. Conc lusions present application and benefits of SOFO monitoring system but the results of test were not clearly conclusive. A system for protecting the sensors was presented. The program associated (material cost of strain measurement system and the labor) cost with respect to total construction cost was not been presented. 1.5.3 Use of Fiber Reinforced Poly mer Reinforcement Integrated with Fiber Optic Sensors for Concrete Bridge Deck Slab Construction The bridge concrete deck and girders in Joffre Bridge, built in 1950, in Sherbrooke, Quebec, Canada over the St Francois River were severely deteriorated due to heavy corrosion activity (Inaudi et al. 1988)]. The Ministry of Transportation of Quebec, determined to replace the bridge deck and girders to satisfy the serviceability r equirements. A part of concrete deck, a section traffic barrier and sidewalk were reinforced with fiber-reinforced polymer (FRP), carbon fiber reinforced polyme r (CFRP) and Glass fiber reinforced polymer (GFRP). During the construction, in addition to a variety of different sensors, some Fiber optic sensors were installed within these elements. Strain values were recorded and mentioned however, there was no comparison between the sensors strain values and analytical results to indicate whether the strain values were high, low or in agr eement. Without such an indicator, the accuracy and reliability of strain values may be questionable. A table depicting these analytical and experimental result s for the purpose of comparison was lacking. The number of sensors and met hod of installation is not described.
28 There was no explanation as to how the sensors were attached to FRP, CFRP and (GFRP). Were they bonded, loosely a ttached or just placed next to the member? A system for prot ecting the sensors was not presented. The strain values from the sensors were not com pared with the analytical results derived from the code for the verification. No long term remote monitoring was presented. Literature has ignored to present the program associated (cost of strain measurement syst em and the labor) cost of equipment and labor. 1.5.4 Test Model for the First Canadian Smart Highway Bridge Carbon fiber reinforced plastic tendons (CFRP) were used for the first time in to pretension six girders of a concrete highway bridge, built in the City of Calgary, Alberta (Reference). This paper summarizes an experimental program conducted at the university of Manitoba to examine the behavior of four pretension concrete beams similar to the bridge girders pre-st ressed with CFRP tendons. Four prestressed concrete T-beams were exami ne for various limit state behaviors, ultimate capacities, and failure modes. The experimental pre-stressed concrete T-beams were 21 feet long and 13 inch deep with overall s pan-depth ratio as similar to the Calgary bridge girders and sca le of 1:3.3. These beams were fitted with fiber optic sensor for monitoring t he strain induced due to static and dynamic loads. The experiment concluded that the secti on curvature at failure of beams prestressed by CFRP was much less than that for beams pre-stressed by steel
29 strands. However, by increasing the reinforcement ratio beyond 0.56 percent, the section curvature of the beams prestressed with CFRP fairly matched the behavior of the beams pre-stressed with steel. No comparison is made between the sensors strains values and analytical results to indicate whether the experimental strain values were high, low or in agreement with strain readings of sensor s. Without such an indicator, the accuracy and reliability of strain values may by questionable. A table depicting these analytical and experimental result s for the purpose of comparison was lacking. The type and number of sensors and method of installation were not presented. There is no explanation as to how the sensors were attached to CFRP. A system for protecting the sensors was not presented. The experimental strain val ues from the sensors we re not compared with the analytical results derived from the code fo r the verification to discuss the code values. The data acquisition system and analysis software were not presented. No long term remote monito ring was presented. The Literature has ignored to present the program estimated associat ed (cost of strain measurement system and the labor) cost of equipment and labor.
30 1.5.5 Using Fiber Bragg Gratin g Sensors to Monitor Pavement Structures The objective of this research wa s to develop an innovative fiber-optic sensing system to evaluate pavement materials or monitor pavement infrastructure. The sensor was developed and designed to measure simultaneously pavement temperatures and strains (Wang and Tang et al. 2005). The reliability and long-term stability te sts for this sensor were examined by mounting it on the surface of two types of specimens, asphalt and concrete, The paper mentions the shortcoming of simultaneous measurement of strain and temperature and suggests a possible so lution. Experiment was conducted on two specimens, one concrete and an asphalt pavement in a laboratory setting. Sensors were surface mounted. The results of readings between the two specimens theoretical values were co mpared. The application of FBG for pavement condition assessm ent was verified. No reference was made to any field experiment on asphalt, either surface mount or embedded The asphalt and c oncrete pavements surface mount sensors would not be able to resist the impa ct of vehicular traffic. The Literature has failed to present the program associat ed cost of strain measurement system and the labor and overall cost of laboratory and fieldwork. 1.5.6 Using Sensors fo r Remote Field Test Fabry-Perot sensors were used to perform field testing of University Drive bridge in Jacksonville, Florida for FDOT in collaboration with Univ ersity of Florida
31 was not successful. A laptop was used to remotely collect data from surface mount installed data. The program did not work and it was abandoned. 1.6 Summary of Research Work and Implementation of the Objectives The study is related to the application of Fiber Optic Sensors (FOS) to investigate the AASHTO LRFD bridge specifications to determine its level of reliability. A total of sixteen Fabry-Perot FOS sensors were installed on the East Bay bridge, in Hillsborough County, Florida. The bridge is a 4-span continuous reinforced concrete deck-type structure. Th e bridge is considered the first smart structure in the State of Florida. The FP sensor s were both bonded to the longitudinal reinforcing bars and surface-mo unted to the concrete deck. Detailed step-by-step description of the installation process is presented. Static and dynamic tests of the bridge under SU4 trucks were conducted. A finite element model was developed, and its output was compared to the experimental data obtained from the truck tests. The results confirmed the accuracy of FP sensors in evaluating the bridge behavior under traffic loads. A remote communication system was established through phone lines in order to connect the acquisition system to the Internet. This technique enables live traffic monitoring from a central station located in the county ma intenance office. Live traffic data are currently being collected and stored on PC hard drive and CD. These data will be used to (a), evaluate current AASHTO specifications for deck type bridges and (b), facilitate the br idge maintenance process, receive early warnings regarding possible structural deficienc ies, and assist in decision-making
32 processes regarding functionality of bridges. The proposed remote health monitoring technique with FOS sensors proved to be practical, cost-effective, and efficient, providing skillful installation. 1.7 An Overview of Dissertation The use of fiber optic sensors to investigate AASHTO LRFD Design Specifications, AASHTO LFD Standard Specif ications, LFD Bridge Rating, Real time Remote monitoring of bridge condi tion and literature review on needs, development and use of fiber optic sensor s to investigate the structural behavior of bridges have been presented in this chapter. Chapter 2 describes the experimental program porti on of the dissertation. The experimental program consists of laboratory exam ination of two concrete beams using FOS to verify the beams cracking state subject to four point static load, field test depicting project tasks (construction sequences) coordination and installation of sensors and monitoring system. The chal lenge and duration for insta llation of electricity and telephone at the bridge site is mentioned. Chapter 3 presents a brief description of the old (replaced) and the new bridge, the summary of Design Code formulas and calculations for the new bridge and app lication of fiber optic sensors for monitoring of structural behavior. De sign and analysis of the new bridge by FDOT software programs using LRFD and LF D are presented in chapter 3. Also, presented in chapter 3 are LFD rating of br idge subject to Florida legal loads and finite element modeling for verification of experimental collected data. Chapter 4 illustrates methodology for data colle ction, truck load testing data and
33 organization of plots, graphs and table of maximum strains for a duration of one year. In Chapter 5, the current de sign specification is compared with the analytical results of Chapter 3 and the r eal time data collected in Chapter 4. Chapter 6 provides a summary of research findings, conclusions and recommendations.
34 CHAPTER 2. EXPERIMENTAL PROGRAM This chapter illustrates the installa tion of fiber optic strain sensors embedded in the concrete deck and on t he underside surface of deck on East Bay Road Bridge to be used as an important t ool to satisfy the objectives of this research. There are two objectives of this resear ch. The first objective is a shortterm application of Fiber Optic Sensor (FOS) for evaluation of AASHTO bridge design guidelines to investigate the new bridge design method of LRFD and old design method of LFD, evaluate and ve rify the assumptions and parameters considered in design of East Bay Road Br idge. We then compare the design of the bridge with the data obtained from Fiber Optic Sensors installed in the bridge during the construction. This data is also used to investigate and verify the results of LFD method of bridge load capacity rating. The second objective is development of a new methodology for damage detection and life cycle evaluatio n of bridges. To achieve these goals, the longterm application of FOS is essential to bu ild strain history of the bridge by continuous or periodic monitoring of t he bridge and evaluation of collected data for monitoring of structural behavior. Li terature on the related topic emphasizes on expert installation of sensors for gat hering useful and accurate data. Some literature has shown photographs of installed fiber optic sensors but the process
35 and procedure of installation hav e not been clearly presented. The instruction manual sensoptic fiber-optic sensors Fabry-Perot strain Gauge FOS series by RocTest (2000) is the only av ailable source to be considered during installation of fiber optic sensors. D ue to the absence of such vital information and difficult encounters during installation of sensors, a great deal of emphasis has been placed in the step-by-step process of embedded strain sensors installation in concrete media and surface of structural elements. This chapter describes material type, property, vari ables and factors in determining the procedure for installation of fiber optic strain sens ors in laboratory and filed experiment settings. 2.1 Beams Fabrication for Laboratory Test Two reinforced concrete beams were fabricated for laboratory experiment. Beam (1) was a 3.5 x 3.5 x 36 with 4 #3 deformed grade 60 steel (yield strength of 60 ksi) placed one at each corner, as indicated in Figure 2.1. Stirrups were #2 grade 40, smooth steel placed at 4 inches on center. The steel clear cover was 3/4 inches. The concrete compressive strength was 5000 psi. The wood forms were lightly covered with oil to provide easy form removal and prevent damage to the beam. Tapping on t he sides of forms with rubber mallet consolidated the concrete in the forms. Beam (2), was a 3.5 x 3.5 x 36 specimen, had 1 # 3 rebar placed at the bottom middle of the form to simulate 50% steel section loss in flexure, Figure 2.2.
Figure 2.1 Beam (1), 4 # 3 Rebars Figure 2.2 Beam (2), 3 # 3 Rebar 2.2 Laboratory Test Setup The purpose of the laboratory experim ent was to evaluate a new testing system, known as surface mount sensor s (blade). Based on the results of laboratory tests, it will be det ermined to use this system in field experiment or investigate other types of st rain measuring sensors. Due to the budget restraint, we did not purchase equipments and material for this experiment. Some of the material and equipment were available in the laboratory and were fabricated or modified to meet the testi ng requirements. Two strain sensors were purchased and a data conditioner was rented. The testing framework was a rigid welded frame constructed of 3x 5 steel tubing (Figures 2.3 and 2.4). A hydraulic pump with a pressure gauge and two 30-ton hydraulic jacks were available in the lab (Figures 2.5 and 2.6). 36
Figure 2.3 Testing Framework Figure 2.4 Conditioner Setup Data logger Figure 2.5 Hydraulic Pump System Figure 2.6 Hydraulic Jacks System 2.3 Data Acquisition System Components (Hardware) The rented system consisted of a 32-channel Bus system (data acquisition), two surface-mount Fabry Perot fiber optic sensors with the composite laminates (conveniently call ed Blade) and a desktop computer. A communication serial link cable, RS232 established a link between the Bus system and the computer (Figures 2.7, 2.8 and 2.9). 37
Figure 2.7. RS 232 Communication Cable 38 Figure 2.8 Bus System Figure 2.9 Computer Linked to Bus 2.4 Concrete Surface Preparation The surface of the beam was sanded using a 100 grids sandpaper to plane the surface. All loose material was removed and the surface was sanded again using a 200 grids sand paper to pr ovide a smooth surf ace. The sanded surface was dusted and wiped off with paper tissues, wet with 75% by volume isopropyl alcohol. The surface was wiped several times, each time with a new tissue and in only one direction to avoid surface contamination. A quicker
39 alterna yer tive cleaning was to wash the surface with water, however, this method would require 24 hours for the surface to dry while it may again collect dust and debris. In comparison, a concrete su rface cleaned with al cohol can be used immediately. A straight edge was used to verify the surface flatness. Any gap more than one mm is considered excessi ve and must be filled with putty. Optional bottom CFRP/GFRP sheets may be installed to provide a primary surface for blade sensors installation fo r the surfaces larger recessed areas. The accuracy of collected data from FOS is directly related to proper installation of the sensors. As soon as the surface was prepared, a uniform la of epoxy was placed on concrete surface. The sensor was placed on epoxy and covered with another coat of epo xy (Figures 2.10, 2.11). Figure 2.10 Two Components Epoxy Figure 2.11 FOS-N Installed
40 2.5 Installation of Beam on Reaction Frame fromhe igures 2.12 through 2.15 illustrate positio ning of the concrete beam, reaction ame, brackets supporting the beam and 30tons hydraulic jacks. The hydraulic cks were seated on 3.5x 3.5 x 1/2 steel plates an d dense rubber sheets to rovide a surface for uniform load transfe r over the jacks seats areas (Figure n area for the top part of the jack. The re ble to In addition to the FP sensor described in Se gauge was installed on the beam and a digit placed on the reaction fram e to monitor beam deflecti The beam was suspended t reac tion frame by two brackets. F fr ja p 2.1). The top plates were selected to prov ide a action beam was made from two steel tubes and the jacks were not a push against the reaction beam. T he jacks were positioned at 1/3 points. Figure 2.12 Load Assembly Fi gure 2.13 Suspended Brackets ction 2.4, a regular strain al caliper deflection gauge was on (Figures 2.16 and 2.17).
41 Figure 2.16 Digital Caliper Figure 2.17 Digital Caliper 2.6 Laboratory Loading Condition The load through the hydraulic jacks was gradually applied to the beam at 15-psi increments until the first crack appeared on the specimen as shown in figures 20 through 23. the analytical values of strain are calculated as follows: Stress: Figure 2.14 Jacks Bottom Plates Figure 2.15 Jacks Top Plates Open space And top plates Assembly P xP I Mc 68.1 5.3 1275.1124 ksi
42 Uncracked Stress: ksi .5.0 (See Figure 2.18) ksiP 5.068.1 lbs kips 297 297 P .0 68.1 5.0 Given: 5000 psi, concrete compressive strength at 28 days Modulus of Elasticity: cf '57000c cf E ksi psi psi E cUncracked strain 4030 5086,030,45000 57000 : 00012.0 4030 5.0 E 1255 c Consider the beam cross-se ction and cracking load, : Calculate neutral axis, Let, crP 25.7 4000 29000 c sE E n 0.75 typ Bar diameter CL ical 3/8 Area of steel, 222.0 inAs
Y C 3.5 CL # 3 Grade 60 Rebar (typical) Figure 2.18 Experimental Beam Cross Section 3.5 T 43 ksify60 Y AY A Y xYsteel bottom stelltop 5.225.7 125.7 2 5.3 39.259.195.375.12 Y 36.1 75.1 36.22 Y inY 17.1 ment of inertia, and cracking load, Determine cracking mo crI crP AY A Y I 5.225.7 125.7 5.32 3 Ysteel bottom steeltop cr 3 Substitute forsteeltopA and steelbotA in the above equation to obtain crI 2 3188.8.131.522.017.1125.722.05.3 17.1 Icr 4 24.3 3 in
44 4.3 12 CPCMcr cr 25.7 36 Icr y Where, C 3.5-1.17-1 = 1.33 And substitute for C, 96.4 4.3 33.112 25.7 36 crP Thus, lbs kips Pcr060,106.1 33.112 4.369.4 The results of test for investigat ion and verification of FOS readings in laboratory setting will be com pared with analytical values. Two thick plates with 14 in (4 3.5) area were placed on the beam under each hydraulic jacks und base for support and uniform load distribution. The beam was loaded and loading was increased in 15-psi increment until the first crack appeared under 86.5 psi (1060 lbs). At this time, the loading rocess was terminated. It was obse rved that the FOS reading of 260 2 ro p for cracking condition was closer to analytical strain value of 245 than to strain gauge reading of 280 The FOS strain reading of 130 for uncracked in value of 125 condition was closer to anal ytical stra than to gauge reading of 145 The analytical deflection of 0.022 in ches as shown in Figure 2.19 at g condition is close to the gauge reading of 0.026 inches. crackin
45 Figure 2.19 Load and Deformation Graph Figure 2.19 represents cracked and uncracked sections of two experimental specimens with relative yield st rength of steel. Stress is 0.5 ksi for uncracked specimen and 1.06 ksi for cracked section respectively. The beam in the reaction frame was closely examined for the presence and location of any cracks while applicati on of load was in progress. Figures 2.20 through 2.23 illustrate the extent and pattern of cracking under the applied load. The results confirm FOS is more accura te than strain gages. FOS values are closer to analytical values. Strain gauge readings are slightly higher than value of the analytical model. Load-Deformation of R/C Specimen4.00 5.00 6.00 0.00 1.00 2.00 3.00 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14Load (kips) Deformation (in) Sec. with 1 Bar Sec. with 2 Bars
Figure 2.20 Tension Cracks Figure 2.21 Propagation of Cracks 46 Figure 2.22 Cracks Directly Under Figure 2.23 Crack are Directly Under 2.7 Conclusions The entire laboratory testing asse mbly was performed economically (approximately $3,000.00) and successfully. It was determined that the Fabry Perot composite laminate sensor would be used in the field experiment. The strain values of Fabry Perot are close to t he results of analytical strain values of prism. The readings of strain gauge were sl ightly higher than the strain values of the Load and on the Side the Load
FOS and analytical strains values 47 classical and an indication of intended under applied static load. Fabry Perot s 2.8 Proposed Remote Sensing System With the emergence of present day tec hnologies, structures can now be monitored g stati loca his remote capability allows continuous monitoring of structures, a condition needed to conduct this research study. Sensors are placed at several critical locations along the ation to the c thought of as an intelligent or smart system providing warnings before any major failure. The proposed remote sensing system consists of the following as shown in Figure 2.24. (a) Fabry-Perot (FP) Fiber Optic Sensors structure. nnect the FP their signal conditioner system. (c) A signal conditioner system housed in a secured on-site location. (d) A power supply to charge the signal c onditioner provided from nearby power lines. (e) A phone line connection to connect the signal conditioner to the Internet The locations and pattern of cracks were behavior of reinforced concrete beam ensors will be used in field test. remotely from a central monitorin on ted several miles away from the field. T structure, and send structural inform entral station. The structure is thus that is capable of sending information and follows the above mentioned approach and attached to critical locations of the (b) Fiber Optic Cables to co sens ors to
48 SL onnections if available) securely to t he Internet, where data could be retrieved software program like Lab View. tem d on the East Bay Road Bridge over Bullfr og Creek in Hillsborough County, Florida behavior of the bridge under traffic loading The bridge is considered the first uous, consists of an 18 cast-in-place reinfo rced concrete slab, and is supported on P sensors are bonded to the bottom bars in the mid-spans 2 embedded in the slab, The embedded FP sensors transmit t he data to the signal conditioner through Fiber Optic Cables placed in conduits to be protected from the environment. The signal conditioner is connected through a phone line (or D c and processed easily from the office, with a Figure 2.24 Proposed Remote Sensing Sys The proposed system depicted in Figure 2. 24 is currently being installe as a part of this research project funded by Hillsborough County to monitor the smart structure in the State of Florida. The bridge is a four span, contin deck-type structure with a total length of 120and width of 55. The superstructure pre-stressed pile bents, each consisting of 5 piles as shown in Figure 2.25. F World Wide Web Password-Protected Connection Office Access Power and Phone Line FP Sensor w/ LabView Supply Post-Processing Signal Conditioner System
49 whe the ents are e transmitted to the Hills rethe maximum positive bending moments are expected, and bonded to top bars over the pile bent 3, where the maximum negative bending mom expected, as shown in Figures 2.26 and 2.27. The signal conditioner system is housed securely to the side of the bridge on a parapet wall The bridge was opened to traffic in February 2005, and it is expected that live traffic data will b borough County Bridge maintenance office. Figure 2.25 Profile of th e East Bay Road Bridge Figure 2.26 FP Sensors Bonded to Reinforcing Steel Figure 2.27 FP Sensors in Conduits FP Sensors Conduits
50 a rida, quent flooding and its narrow width. Coordination of the field experim Hillsborough County project experiment. The fiel d experiment was allowed only through the construction. Hillsborough Cou nalty le use hedule. field experiment was not considered a ju stifiable cause of construction delay. 2.9 Field Experiment The contractor, All Am erican Concrete Inc., had a construction agreement with Hillsborough Coun ty to replace an existing concrete bridge with cast-in-place reinforced concrete bridge on East Bay Road in Gibsonton, Flo Figure 2.28. This bridge was a low profile concrete structure built in the early 1970s. This bridge was classified as functionally obsolete due to fre Figure 2.28 Elevation Vi ew of Old East B ay Road Bridge ent effort with the contractor and management team was crucial to the success of the if it could stay transparent nt y imposes substantial daily pe on the contractor for any unjustifiab ca of delay in a construction sc A
51 s for the superst ructure (All components of bridge sitting above 2004 el to the process of bridge construction. 10 of 12 lane. Shoulder 3 Schedule to set the form the top of the bent cap) and c oncrete pour was on September 20, with the completion date of November 1, 20 04. The critical time to install the embedded sensors was when plac ement of reinforcing steel was in progress. Time was of the essence for installation of the sensors since pouring concrete would begin as soon as reinforcing steel was in place. The installation of sensors had to take place parall 2.10 Determine Location of Sensors 2.10.1 Transverse Positions of Sensors The embedded strain sensors were placed under the wheels in a transverse direction. Figures 2.29 show the position of truck wheels. SU4 truck was used to test this bridge under service load. This position configuration meets AASHTO section 3.6 (AASHTO 1994) requirement for trucks occupying 1.5 6 4 6 3 Curb sensors 12 Lane Bridge C. L. 27.5 Figure 2.29 Transverse Positions of Wh eels on the Bridge Deck
52 t be uming f Axels on the Bridge Deck ments were bonded to rebars e 4 Tuck The shoulder lane was strategically sele cted for installation of sensors for static load testing of the bridge while open to traffic. The load test of the traffic lane is not practical or safe while the bri dge is open to traffic. The bridge mus closed to traffic during the test. The bridge closur e process requires a detour route determined by The Hillsborough Co unty Traffic Department and approved by Public information Services Department. This process is very time cons and request for bridge closure may not be obtained. 2.10.2 Longitudinal Positions o The embedded strain sensors for positive mo placed at mid-span 2. This is a simplified location very close to th point of maximum positive mo ment. The SU4 truck axle s spacing and weight are shown in Figure 2.30. 13.9 kips 18.7 kips 18. 7 kips 18.7 kips 9.17 4.17 4.17 Figure 2.30 Longitudinal Spacing of Axles in SUr
53 ace of concrete deck. ) Surface mount sensors P1 and P2 = these sensors were bonded to (c) ar on teel mat with epoxy. ollonsf this ent. Three types laced in four categories of installation. Three types of sensors are identified as (a) Surface mount, known as FOS-N (blade), (b) Embedded sensors nd (c) Embedded temperature sens ors. The layout of the sensors identified as shown in Figure 2.31. 2.10.3 Locations of Embedded and Surface Mount Sensors The positions of sensors were dete rmined on the topside of the bridge deck by measurements taken from the inside face of the traffi c barrier. Figure 2.31 illustrates this configuration. Legends: (a) Surface mount sensors ASM, BSM, CSM, DSM (FISO-B, Blade) = Sensors bonded with epoxy to the bo ttom surf (b concrete, 3/4 below the surface of deck. Embedded sensors C, D, E, F = Bonded to the bottom surface of reb bottom reinforcing steel mat. (d) Embedded sensors G, H, I, J = Bonded to the bottom surface of rebar on top reinforcing s Figures 2.32 and 2.33 depict detailed location of sensors within the concrete slab bonded to reinforcing steel, bonded to the surface of concrete and b onded to concrete slightly below the surfac e from the top of the deck. Step-bystep procedure and techniques are outlined in the fwing sectio o chapter illustrating the installation of all sensors in this experim of sensor are p a
55 (Typical) ASM P1 and T2 Figure nt if very congest ed # 9 reinforcing bars were tied together P1&P2 18 2 cl. cover Figure 2.32 Surface Bonded Sensors ASM, BSM, CSM, DSM and P2 Sensors, Slightly Below the Surface T1 18 2.33 Sensors C, D, E, F, T1 (Bottom) and G, H, I, J & T2 (Top) 2.11 Field Readin ess and Planning Coordination of effort with the cont ractor was one of the most importa first steps in field experimentation. C ontract drawings indicate the bridge deck was heavily reinforced. The top and bottom mats consisted of # 9 rebar (1.125 diameter) placed 6 on center. The cl ear space between the bars was 4.87. Placement of the top mat would have made the access to the bottom mat impossible, particularly X-section at Bent 2 Mid span 2 C, D, E, F X-section at G, H, I, J X-section at Bent 3 Mid span 1 X-section at
56 D, # 9 barsre placed untied on the form, it was an opportunity to bond these ors to the top of t he bars and then turn (twist) the bars 180 degrees in place ontractor agreed to begin reinforcing steel placement from span 4 stead ofThe strategy was to place bottom and top mats in span 4 and he author p mat over bent 3. ottomop mats in span 1. This woul d provide an ample time for the author instaensors C, D, E, F and T1 on bottom mat bars at mid span 2. idelines for installing sensors within and on structural members. However, the as knowledge and experience of the installe t with tie wire as shown in preceding section 2.8, Figure 2.27. The sensors (C E, F) for flexural stresses had to be bonded to the bottom of th e bars. Whi le re we sens and tie them together afterward. The c in span 1. span 3 up to bent 3. and place the bottom ma t in bent 2. At this point, t b egan installation of G, H, I, J and T2 s ensors on to The next step was for the contractor to place re inforcing steel for the b and t toll s 2.12 Methodology and Procedure Manufacturer of fiber optic s ensors (RocTest) has recommended gu quality of installation would be as good r. Accuracy and good quality of dat a is directly related to proper installation of sensors. The author has exercised a great deal of patience and care during each step of every sensor installation. Numerous photographs and detailed descriptions are pr esented in every step of the sensors and equipmen installation process.
57 e on #9 grade 60 rebar (deformed) was grinde rs. The area o with The abraded area was wiped with Is opropyl alcohol and rinsed with Miped un idirectional using the wipes, using a new w s s soon 2.13 Surface Preparation of Steel Bars An electric angle grinder connected to an inverter connected to the cars battery (there was no electric power at the bridge site) was used to grind the surface of the steel rebar flat and smooth to install the sensor. An area of approximately 3 long and 3/8 wid d flat. A straight edge was used to verify the flatness of the area. Dry abrading was continued with 200 and 300 grit silicon carbide papers to achieve a flat, smooth surface. It was ri nsed with M-prep Neutralizer 5A (from Measurement Group) and wiped with paper tissue such as kimwipe wipe f the sensor was wet with M-prep conditioner A and abraded the area 400 grit silicon carbide paper. The sensor area was checked frequently with a straight edge for flatness and smoothness. prep Neutralizer 5A. The area was w ipe after each wiping to avoid cont amination of the sensor area for bonding. The sensor was placed on the rebar and held down with electric tape, one inch away from micro capillary. A very small drop of 5 minutes epoxy wa placed on incoming fiber optic, approximately 1/8 from micro c apillary. A as 5 minutes epoxy was cured, t he adhesive was prepared and applied with a linear motion along the entire length of the gage (figures 2.34 to 2.42).
58 r ctive co ating was applied to the sensor of nt tact with surface of rebar. Fi gures 2.37 through 2.39 show the application of 5minutes epoxy to sensitive regi on of gauge and optical fiber. Figure 2.34 Position of Sensor on Bottom Mat Rebar The remainder of adhesive was applied to the optical fiber up to the fibe jacket. For additional protec tion, M-coat Prote (Figure 2.38). After this protecti ve coating dried, a rubberized waterpro sheet, nitrite rubber sheet was wrapped around the sensor as shown on Figure 2.39. All material named in this sect ion were purchased from Measureme Group. Figures 2.33 through 2. 40 are the pictorial present ation of installation of the sensor on reinforcing steel. Thes e figures are used with permission from Roctest Canada. Immediately after application of adhesiv e over sensitive region of sensor a piece of Mylar tape was placed over it to keep the sensor in a good con
59 Figure 2.35 M-Bond 5 Minutes Adhesive Figure 2.36 M-Coating and Neutralizer Figure 2.37 Area of Rebar to Place the Sensor on Figure 2.38 Sensor Secured on Rebar
60 5-Minutes Epoxy Placed on Incoming Optical Fiber gure 2.40 Correct and Incorrect Pro cedure for Sensors with Epoxy Figure 2.39 A Very Small Drop of F i
61 t Figures 2.42 Final Procedural Steps of Sensor Installation Figure 2.41 Mylar Tape was Applied to Sensor to Keep it in Good Contac with Rebar
Figures 2.43 and 2.44, illustrate the actual final steps of filled installation of 62 sensor. Figure 2.43 Placing M Sensor Optic Fiber Rubber and Placed in Conduit s, FTI-10, a single channel data logger was used to test the sensors and verify bonded condition (Figure 2.45). Figure 2.45 Single Channel Data Logger Reads the Strain of Sensor in nm ylar Tape on Figure 2.44 Sensor Wrapped in Nitrite Soon after sensors were bonded to rebar The readings on the data logger are in nanometer (nm) and verify a successful bond interface between the sensor and rebar.
63 auge ing at no load) by the gauge leng th. The gauge length of each sensor is unique to that sensor. In this case, the strain in gauge C is: The strain of sensor in nm is ca lculated by dividing FTI-10 reading (g zero, read length gauge FTI Strain 7456 09.2 5.14993 10 nm The following procedures were proposed by Roctest to interpret the reading of Fabry Parot sensor. The relati onship between the leng th of the cavity ) and the strain ( (cavityL ) is determined by the following formula. gage cavity gageL LL L L0 Where: cavityL Length of Fab ry-Perot cavi ty, in Nanometers and varies between 8000 and 23000 in nm gageL Gage length, the space between fused welding, mm L Initial length of Fabry-Perot cavity, in nanometer 0 Total strain measurement, in The total strain ( ) is the raw strain obtained dire ctly from FOS readings with readout units after the gage factor has been defined in readout memory and selected Therefore: 01 Total strain measurement, in 1 Current strain, in 0 Initial strain, in
64 This to Where: tal strain includes the mechani cal strains and thermal strains in the investigated structure. The real stra in induced by the stress due to thermal change can be computed with the following formula: )(*01TTr r Real strain, in Total strain reading, 1T Temperature reading of structure, in C Initial temperature read ing of structure, in 0T C Thermal expansion factor of structure in al Cmm// on which the sensor is fixed. The therm expansion factor )( can be obtained fr om laboratory test. The factor range for steel is: 10 //16 // mm Cmm C A numerical example of this procedure is presented as fow llos: Given: 0 2002.2 units, Initial strain ( ) reading of FOS with fiber optic readout unit 1 2407.8 current strain ( ) reading of FOS with fiber optic readout re ingof structure 12 ctor of structure unit 0T 20.2 C, initial temperature reading of structu 1T 26.2 C, current temperature read TE Cmm// ,thermal expansion fa
Calculate the strain ( ): 01 2407.8 2200.2 = 207.6 and the real strain (r ) can be calculated as follows: )(*01TTr Therefore, 65 6.135 6.7 2.202.261220 r e bars were turned 180 stlation of ic c nethod to assure thesound co represents planning and installation of this prote 2.13.1 Protecting the Sensors a nd Optical Fibers in the Slab The micro capillary section of the sensors is glass and thus is very sensitive to scratch and impact. Also, optic al fiber is very sensitive to bends, on and during de ck sheet for prot ection against impact and moistu re. Fiber optic cables er 90 e elbows (sweep) were used to avoid sharp bends and kinks in the fiber rebar mats to the trate rubber sheet and caulking, then the conduits were tight wires. Photos in Figures 2.46 thr ough 2.49 illustrate this process. In case of sensors fo r flexural condition, thre degrees to place the sensors facing the fo rms. The successful inal sensors and fiber optables was followed by a well thought protectio m ir ndition in the system. The following section ctive system. kinks, sharp curves and impact during the final steps of installati the bridge construction. All sensors bond to rebars were wrapped in a thi nitrite rubber were inserted into one-inch diamet schedule 40 PVC conduits. Large radius degre optic cables. The conduits were guide d through the crowded edge of the slab, openings in the conduits were sealed with ni ly secured to the rebars with steel
66 Figures 2.48 Secure PVC Conduit to Figure 2.49 Placing Fiber Optic in it Figures 2.46 PVC Conduit Protection Figure 2.47 Seal PVC Conduit Rebar Conduit 2.13.2 Protection of Fiber Optic Cables Out of Slab Conduits containing fiber optic cables were brought unto the forms to the side (edge) of the slab. At this area, s ensors and fiber optic c ables are the most vulnerable to the construction activities such as workers traffic and placing and removing the forms. A 2 diameter hole wa s drilled to allow the conduit to ex
67 the slab (Figures 2.50 through 2. 55). Three small boxes (12 12 12) were es at the exit points. option was to exit from the underside of removal of the slab forms with hea vy equipm severed the cables during the process. fabricated to house the cabl Other alternatives for the cables to ex it the forms were investigated. One the slab at the bottom. However, ents would have damaged or Figures 2.50 G, H, I and J Sensors in Conduit Exiting the Forms in Conduit Exiting the Forms Figure 2.51 C, D, E and F Sensors Figures 2.52 FO Cables in the Box Figure 2.53 Forms are Removed
68 2.13.3 Special Installation Typically, surface mount sensors structural elements subsequent to surface s on top of the concrete deck over installed on the top h e age the sensors floating in freshly poured conc since the position of the sensors coul certainty during concrete placement. The last alternative was to place hardened concrete. The positions of two s e completely cured concrete slab. It was determined to grind the top of the a ni An investigation was carried out to find a s pecialized tool such as a router to cut Figure 2.54 Cables are Safely Out of Bridge Slab Figure 2.55 Box Housing the Cables of FOS-B, P1 and P2 are bonded to the surface of the preparation. In this case, it was determined to install two FOS-B sensor intermediate bent 2. If the sensors were would be exposed to traffic and the har s of the slab, they nvironmental elements and dam to their integrity would be imminent. Another suggested alternative was to leave rete. This method was not acceptable d not be guaranteed with any degree of the sensors below the surface of ensors were marked on the surface of th deck, below the surface to provide u form flat area to bond the sensors.
a 2 18 grove into the concrete deck. 69 As an alternative an electr cut and grind the concrete, below the su a designated area over bent 2. Side to concrete to the desired depth. The surface of the area was wiped with a piec thin layer of epoxy was placed on the dried and cleaned cutout area to provide rea rocedure assured sensor protection agains t traffic and groove cuts in the slab for veh droplane action. Figure Figure nsor This type of tool was not found. ic 4.5 angle grinder and a diamond blade was used to rface. Two deep lines were cut in side motion of angle grinder cut the cutout area was cleaned and dusted. The e of clean cloth and isopropyl alcohol. A a uniform and level bonding surface a fo r the sensors. This installation p icular wheels traction and to avoid the danger of hy 2.56 through 2.68 show the process for this installation. 2.56 Bedding for P1 Sensor Figure 2.57 Bedding for P2 Se
70 and P2 and P2 with PVC P1 and P2 Sensors and Optical Cable Figure 2.58 Edge Bedding for P1 Figure 2.59 Edge Bedding for P1 Figure 2.60 Beddings are Prepared and Read y to Install
71 igure 2.61 Sensor P1 is Installed Figure 2.62 Sensor P2 is Installed F Figures 2.63 P1 and P2 Sensors Fi gure 2.64 P1 and P2 Sensors out of the Slab at the Edge of the Bridge
72 Over Bent 2 Figure 2.66 Protective Box Figur e 2.67 Final Step, P1, P2 Figure 2.65 Material were Used to In stall Sensor P1 and P2 on the Deck Sensors in the Protective Box
Figure 2.68 Sensors Housed 73 2.13.4 Protecting Fiber Opti Accessibility to specific areas of the br sensors and equipment, installation of a pr and data logger and safety of personnel we lanning and construction. The details of different alternatives were carefully e two poi nts of exit for conduits and fiber optic ne point of ex it was the underside of the deck. This n was not practical because, (1) to attach the fiber optic cable protective in the Protective Boxes c Cables in PVC Conduits idge for the installation of fiber optic otection system for fiber optic cables re primary concerns during the p investigated. For example, th s ensors were evaluated. O optio
74 ousing to the bridge underside required an extensive scaffold setup or a cherry tive housing and the fiber optic cables, whic sensors. coordinatio ion ywide Division igure 2.69 and 2.76, this was a very difficult, if not impossible task. The afe installation of onduits, cables and DMI possible. Acce ss to the side of the bridge deck was neither mount scaffold. In Figu h picker. Employment of either techni que was not within the budget limit, (2) the method of removal of the forms from the underside of the bridge deck was by sliding the forms out through a narrow space between the forms support and the concrete deck itself. This motion would shear off the protec h were placed within the housing. The point of exit from the side of bridge deck was a practical alternative since it was accessible and the forms removal would not damage the fiber optic The task of installation of the measuring system required detailed n with the authorities in Hillsborough County, County Wide Divis (Maintenance Headquarter) and the bridge c ontractor. The Count had offered assistance, providing scaffo ld and manpower to install one-inch schedule 40 PVC conduits on the side of th e bridge deck. As it is shown in F invaluable assistance of County Wide Divis ion made the s c safe nor practical without a specia l type of bridge parapet res 2.69 and 2.70 the difficultie s and inaccessibility are shown to the bridge parapet for installation of conduits and the various equipments.
75 Figure Problem at the Bridge Edge o Install Conduits ures 2.71 through 2.76 show the step-by step process of installing scaffol nd Figures 2.71 Scaffold Installation Figure 2.72 Scaffold Installation from s 2.69 Accessibility Figure 2.70 Accessibility Problem t Fig ds to use as a safe platform to install conduits carrying FO cables a equipments on the bridge. from the Top the Bottom
76 ith The following Figures 2.77 and 2.78 clear unprotected fiber optic c ables and sensors. Figures 2.73 Scaffold Installation in Progress Figure 2.74 Cover the Scaffold w Wood Planks Figures 2.75 Scaffold Installation is Completed Figure 2.76 Scaffold Installation is Approved for Use ly show the potential to damage
77 Figures 2.77 Potential for Damage Figure 2.78 Unprotected FO Cables As it has been illustrated in previ successfully (e.g., the signals were tr reception was verified) and routed to t incident. The conduits were attached to the side of t orking condition of sensors is owed to protective measures taken during the stallation process of conduits, DMI and connection of fiber optic ables to DMI are shown in Figures 2.79 through 2.84., and installation of Fiber optic cables were fished thr connectors at the end of the cables inside the DMI unit (Figures 2.83 and 2.84). to Fiber Optic Cables and Sensors were Damaged ous figures, sensors were installed ansmitted from sensors to FTI-10, data he eastside edge of the bridge without any he bridge. The successful w installation. In c telephone and electric power at the bri dge are shown in Figures 2.85 and 2.86. ough conduits to the DMI unit. The were cleaned and connected to the ports
78 Figure 2.79 Attaching Conduit to the Bridge Figure 2.80 Fish ing Cables Through Conduit MI duits are Attached to the Bridge and Connected to the DMI Figure 2.83 FO Cables are Guided System Figure 2.84 FO Cables and Sensors Figure 2.81 Conduits Entering Figure 2.82 Con D Through Conduit into DMI are in DMI Box
79 2.14 Installation of Electric Powe r and Telephone on the Bridge The power lines came down from the electric pole to a hand hole box at the base of the electric pole, about 160 feet from the point of installation on the bridge. The telephone box was also about 160 feet from the bridge located near the electric pole. Two 2 conduits were placed 2 below the ground surface in a trench. Telephone and electric lines were pulled through the conduits and housed on the bridge (Fi gures 2.85 and 2.86). Telephone cable and electric wires were extended from the telephone box The purpose of direct el ectric power to DMI was to have an uninterrupted power supply to DMI. DM I is supplied with rechargeable battery pack, however, recharging the battery was on ly possible while connected to DMI. Figure 2.85 Telephone Line is Secured on the Bridge Figure 2.86 Electric Line is Secured on the Bridge and disconnect box to DMI and were connec receptacle. ted to a telephone jack and a
80 A generator and a power inverter were ta ken to bridge to charge the battery. objective, sound framework for SU4 truck process for this task wa presented below. Six different positions load-test. Figure 2.87 depicts different load cases. Legend: Northbound direction Two-lane truck Centerline of span 1 Case 1 Case 2 ase 3 ase 4 5 ase 6 igure 2.87 Six Cases of Static Truck Load-Test This process was inconvenient and took about 6 hours for recharging the battery. All previously described crucial steps were taken carefully to provide an static and dynamic load-test. The s evaluated in detail prior to commencement and is (cases) were assigned for this static the layout of the plan of action to perform these six of traffic Southbound direction of traffic load-test centerline of span2 C C Case C F
81 Section 2.15 describes and illustra tes the positioning of the trucks coincident with six load cases shown in figure 2.87. 2.15 Truck Static Load-Test The locations of the sensors as described in section 2.10.3 and shown in Figure 2.29, were marked with white paint on the bridge top surface on the eight feet wide shoulder and the twelve feet northbound traffic lane (Figure 2.90). The marked positions of the s ensors on the bridge deck on the through 2.96 show the positions of SU4 Two full capacity (70 kips) SU4 trucks we truck positions for static load were selected. the truck was positioned over the sensors. the bridge deck topside (Figure 2.90) in hown in sub-section 2.10. 3 Figure 2.29. The trucks were driven with craw reading econd delay, reading process was stopped and then the trucks were repositioned for northbound lane matched the six load cases shown in Figure 2.87. Figures 2.88 truck(s) during the static load-test. re selected for the load test. Six The center of the rear three axles of Sensor locations were marked on accordance with the layout planning s speed to the exact positions. The s were taken after a 10 to 15 s the next load case.
82 Figurekin pside The marked positions of the sens k on northbound lane matched the six load cases shown in gures 2.89 through 2.94 show the positions of SU4 trucks during the sta 2.90 Case 2, Span 2, Northbound 2.88 Marg Locations of the Sensors on the Deck To ors on the bridge dec Figure 2.87. Fi tic load-test. Figure 2.89 Case 1, Span 1, Figure Northbound
83 Figure Two Trucks Side-By-Side, Northbound and Southbound llect the strain r eadings from DMI data onditioner via RS-232 communication cable (Figure 2.95). 2.91 Case 3, Span 1 and 2, Trucks in Tandem, Northbound Figure 2.92 Case 4, Span 2 Trucks Side-By-Side, Northbound Figure 2.93 Case 5, Span2, Two Trucks, Side-By-Side, Northbound and Southbound Figure 2.94 Case 6, Span1, Two An on site laptop computer was used to co c
84 2.16 Conclusions Collected data from all six-load case s were stored in the computer. The strain values of all 16 sensors for each one of the six load cases are tabulated in Table 3.5 and graphical representation of th is data is given in Chapter 3, section 1, 2 3, 4, 5, and 6 and finite element model analysis for case 1, 2, and 3 ar Figure 2.95 On Site Direct Data Collection via RS-232 3.0. The contour of strain values at the lo cations of all 16 sensors for load case model for load case I, 2, and 3 and beam e also presented in Chapter 3.
85 CHAPTER 3. DESIGN AND ANALYSIS 3.1 Introduction This chapter describes the applic ation of LRFD and LFD design and bridge rating by LFD method. The fo rmulation of LRFD and LFD methods are compared to initiate discussion and c onclusions about relevance and benefits of using each method in design. The design steps and formulas for East Bay Road Bridge using LRFD and LFD design methods are presented and the re sults are compared with the results of field static load test. The means of comparison between LRFD and LFD designs and actual bridge load test are based on the strain values obtained from DMI (signal conditioner) readings through use of fiber optic sensors. The use of SAP computer software for modeling and verifi cation of results is also presented. In addition, frame analysis using the pr ogram MASTAN was also performed. 3.1.1 LRFD Code vs. AASHTO Standard Specifications An extensive laboratory-testing progr am was conducted to investigate the shear strength of the pr estressed concrete girders and published in an article titled Shear Behavior of Full Scale Prestr essed Concrete Girders: Comparison Between AASHTO Specific ation and LRFD Code, Shahawy and Barrington, (1996). The test shear strengths are compared with predictions based on the
86 1989 AASHTO Standard Specifications for Highway Bridges and the application of the 1994 AASHTO LRFD Specifications. The results show that the application of the 1989 AASHTO Specifications gives a much better prediction of shear strength than the LRFD prov isions. The average value of V test / V LRFD was 1.37 vs. V test / V AASHTO at 1.2. The shear strength perdition (V n ) of 1989 AASHTO Specifications (LFD) and LRFD Code were compared with the resu lt of tests. The shear strength perdition (V n ) of 1989 AASHTO Specifications is more in agreement with the result of tests than LRFD Code. Kulicki et al. (1996), conducted LRFD method calculations for shear capacity of a test beam and AASHTO Method Calculations for shear capacity of an identical test beam. The LRFD values calculated by Kulicki were different than the values shown by Shahawy. Howe ver, Kulickis calculated values for LFD (AASHTO Specifications) were nearly similar values to those of Shahawys calculations. At the conclusion of a discussion paper, Kulicki et al, implies that the LRFD Values are more variable than those of 1986 AASHTO Code. Further more, Shahawy and Kulicki demonstrated that the ratio of value of the LRFD Code to the value of test is at 52 percentile prediction, while the ratio of value of the LFD code to the value of test is at 21 percentile. Due to the variability of values of LRFD Code calculations to ve rify the test results, one could conclude that there might be some degrees of inc onsistency in interpretation of LRFD Code provisions by different individuals. This perception of inconsistency would
adversely affect the level of reliability in LRFD code calculations. LRFD method is said (Shahawy and Kulicki, 1996) to be more conservative and would take more computation time than LFD method. The LRFD specifications were approved by AASHTO for use as alternat ive specifications to the AASHTO Standard Specifications for highway Br idges, LFD. Technically, LRFD was meant to be a parallel design method wit h LFD but inadvertently has taken a sharp turn away from LFD The variabi lity in design parameter and formulation between LFD and LRFD codes are illustrated in the next section. 3.1.2 LFD Design Method The design live load for LFD method is ei ther truck load or lane load. The live load design is either HS20-44 truck (44 denotes the publication of the 1944 edition of AASHTO Specification), Figure 3.1, or alternate military loading of two axles four feet apart with each axle weighing 24,000 pounds, Figure 3.2. Lane load on continuous span as in the case of the East Bay Road Bridge is the combination of 0.64 kips per foot uniformly distributed load over the span and 18 kips concentrated load at the cent er of span. For maximum positive moment, only one concentrated load is used per lane as shown in Figure 3.3. In case of maximum negative moment, the second concentrated load is placed in series in the adjacent lane as shown in figure 3.4. The maximum live load moment calculated for the Loading condition is shown below: pact LL DL Load Design Im117.23.1. 87
Where The impact factor is given by 125 50 L I for the East Bay Road Bridge, the span lengt h is either 27 or 33 for the positive moment and for the negative moment, is the average of two adjacent spans (e.g. 2 3327 ). The distribution width, is given by: DE 206.04 LftED In the case of the East Bay Road Bridge, ft xftED6.1123006.04 The number of Design Lanes is determined by taking the integer part of ratio of 12W where, W is the clear roadway width in feet between the curbs or traffic barriers. Therefore the number of Design Live Lane for the East Bay Road Bridge would be 12 40 3.3 or 3. 88
Figure 3.1 Notional HS20-44 Truck, Axles, Wheel Spacing and Weights of Each Axle 24 kips 24 kips 89 4 Figure 3.2 Alternate Military Loading 18 kips for moment 26 ki ps for shear 0.64 kips/ft uniform load Figure 3.3 Lane Load on Continuous Span for Positive Moment
18 kips for moment 26 kips for shear 0.64 kips/ft uniform load Figure 3.4 Lane Load on Continuous Span for Negative Moment 3.2 LRFD Design Method The design live load for LRFD method is HL-93 loading and that is the combination of design truck or design tandem with design uniform lane load. The Design truck load is HS20-44 truck as shown in Figure 3.1, section 3.1.1 and Design Tandem is as shown in Figu re 3.5 with each axle weighing 25,000 pounds. The design live load configuration for LRFD method is illustrated in Figure 3.6. Design truck load is increased by 1.33, which is the dynamic impact load allowance. 8 kips 32 kips 32 kips OR 25 kips 25 kips 14 14 4 Plus 0.64 kips per linear foot, Design Lane Load Figure 3.5 LRFD Design Load Combinat ions (HL-93), Positive Moment In form of a formula, Figure 3.5 appears as: 90
kDesigntruc DesignLane DL Load Design Factored 33.1 75.125.1.. The interpretation of LRFD Code provis ion for Design Load to produce maximum loading condition for negative moment is illustrated by Figure 3.5. The code allows for a 10% reduction for this case. The formulated form of Figure 3.5 is presented as: 9.0 33.1 75.125.1.. kDesigntruc DesignLane DL LoadDesig Factored 14 14 50 14 14 32 kips 32 kips 8 kips 32 kips 32 kips 8 kips 0.64 kips per foot Figure 3.6 LRFD Code, Design Load to Produce Maximum Negative Moment Figure 3.7 represents cross sect ion of East Bay Road Bridge superstructure. 91 DE Figure 3.7 Distri butions width this Figure Shows Actual Cross Section of East Bay Road Bridge Superstructure DE The distribution width of slab ( ) under the design live load based on LRFD and LFD Codes are defined as follows DE 12 44.1 "84 LW ELRFD D and
206.04 Lfeet ELFD D Where L = Actual span length in feet W = Physical edge-to-edge of bridge in feet Considering the East Ba y Road Bridge with W = 55 feet and L = 30 feet, the distribution width in LRFD method can be calculated as: '874.11 12 553044.1 "84 x ELRFD D And similarly, distribution width in LFD method can be calculated as: '6.1123006.04 feet ELFD D The comparison between and indicate that LRFD method is more conservative than LFD method, although by a small margin. LRFD DE LFD DE 3.3 Bridge Load Rating Using Load Factor Method The Load Factor Design (LFD) method has been predominantly used in analysis of Bridge Load Rating. LFD met hod was presented in the first edition, first printing of AASHTO Manual for Main tenance Inspection of Bridges in July 1970. Since then, several editions have been printed. Sec ond edition, first printing was out in June 1974 and the fist printing of third edition was in January 1979. There has not been any significant modification in formulation and application of the LFD met hod during this period. 92
The AASHTO manual for maintenance in spection of bridges requires highway bridges to be rated at two load levels, either by load factor or by working stress methods. 3.3.1 Operating Rating At the first or upper level, rating is referred to as Operating Rating. The operating rating will result in the absolute maximum permissible load level to which the structure may be subjected. Based on the 1979 AASHTO manual for maintenance inspection of bridges, t he following expressions has been used to determine the operating ra ting of structures. Operating Strength Anal ysis General expression: ISRFSSL D u 3.1 3.3.2 Inventory Rating At the second or lower level, rating is referred to as inv entory rating. The inventory rating will result in a load level, which can safely utilize an existing structure of an indefinite period of time. Based on the 1979 AASHTO Manual for maintenance inspection of bridges, t he following expressions has been used to determine the inventory rating of structures. Inventory Strength Analysis General expression: ISRF SSL D u 3/53.1 where capacity reduction factor as per standard specification for highway bridges 93
uS ultimate theoretical strength effect of dead load DS effect of live load plus impact from the rating vehicle ISL R F rating factor For concrete members, the code specif ies for strength and serviceability, the area of tension steel at yield to be us ed in computing the ultimate moment capacity not to exceed 75 per cent of the required steel for balanced condition. The Code, further specified the yield strength of steel shown in Table 3.1. Table 3.1 Yield Strength of Different Grades of Steel Reinforcing Steel Yield Point F y (psi) Unknown steel prior to 1954 Structural Grade Intermediate Grade and unknown after 1954 Hard Grade (Grade 50) Grade 33,000 36,000 40,000 50,000 60,000 The AASHTO manual for maintenance inspection of bridges (17 th edition), specify the following procedure for Bridge Rating using LFD. The moment live load, is due to application of the load combination, LLM pact LL DL Im117.23.1 To determine the Rating Factor, R F for the Operating Le vel, all six Florida legal trucks, SU2, SU3, SU 4, C3, C4, C5 and two design vehicles, described in 94
table 3.2, HS20-44, HL93 and military loading must be investigated and the rating factor, R F for all nine cases must satisfy, 1 )Im1(3.1 3.1 pact M MM RFLL DL u In determination of rating factor for the inventory level, only design truck, HS20 must be investigated and R F for this case must satisfy, 1 )Im1(17.2 3.1 pact M MM RFLL DL u In a simplified form, R F for Inventory level can be obtained in one step, multiply R F of operating level by 67.1 1 Table 3.2 Florida Legal Loa d and Design Live Load Trucks Truck Description of trucks Gross Vehicle Weight SU2 Single Unit 2 axles, gro ss vehicle weight GVW = 34.0 kips SU3 Single Unit 3 axles, gro ss vehicle weight GVW = 66.0 kips SU4 Single Unit 4 axles, gro ss vehicle weight GVW = 70.0 kips C3 Combination, tractor and tr ailer, 3 axles GVW = 56 kips C4 Combination, tractor and tra iler, 4 axles GVW = 73.3 kips C5 Combination, tractor and tra iler, 3 axles GVW = 73.21 kips HS20 A notional Design Truck GVW = 72.0 kips ST5 Tractor pulling Tandem Trailers GVW = 80 kips HL-93 A notional Design Truck HS20/Tandem + lane military Two axles four feet apart 24 kips each axle 95
96 3.4 Steps in Designing of East Bay Road Bridge The Florida Department of Transporta tion dictates the body of codes and specifications to be used in design of a bridge structure. The required codes and specifications are described in the follo wing sub-sections. Two vital information are necessary to guide the structural engi neer in determining the specifics of the design of a bridge. First, bridge hydraulic recommendation is the information required to establish the bridge vertical a lignment. Second, report of core boring is required to establish the foundation type. (e.g. Piling, pier, Spread Footing etc.) 3.4.1 General Specifications The Florida Department of Transpor tation standard specification for road and bridge construction, 2000 edition and supplemental thereafter, was used for design. 3.4.2 Design Specifications The following are the codes used in the design of the East Bay Road Bridge. The American Associ ation of State Highway and Transportation Officials, (AASHTO) and LRFD Bridge Design Specific ations Second Edition with Interim Revisions thru 2002. FDOT Structures Design Guidelines for Load and Resistance Factor Design, Edition 2002.
3.4.3 Design Method Load and Resistance Factor Design Method (LRFD) 3.4.4 Design Loading (a) Dead Load, Unit weight of reinforced concrete 0.150 kcf (b) Future wearing surface 0.015 kcf (c) Traffic railing barrier 0.418 klf each (d) Live load HL-93 Loading 3.4.5 Material Property Cast-in-place Deck, 4500 PSI minimum compressive strength at 28-days. All Reinforcing Steel are ASTM A615 Grade 60. 3.4.6 Code Distribution Width, DE Equivalent design distribution width per lane 11.874 ft 3.5 Analysis FDOT live load generator was used to obtain positive and negative moments for service live load. Dead load moment was obtained from beam analysis. The deflection and moment envelopes for the different truck configuration are shown in Figures 3.8 thru 3.16. The results of positive and negative live load moments and dead load moments for Florida legal load 97
configurations, SU2, SU3, SU4, C3 C4, C5, and notional design live load configurations, HS-20 and HL-93 are tabulated in Table 3.3. 3.5.1 Service Moments The following live load positive and negative moments w ere obtained from Live Load Generator. Dead load moment was obtained from beam analysis. 135 8.170 6.250posM ftkip moment loadDead moment loadlive Negative moment loadlivePositive .. ... ... 3.5.2 Cracked Section Analysis The major assumption made in designing the East Bay Road Bridge was that the deck would crack under service load. Later in this chapter, the results of truck load test would indicate that t he cracked section assumption is very conservative. Given design parameters ft ED.874.11 Tributary width for a single truck in spacingpos.6 Spacing of rebars for top and bottom mat, center to center of the bars inder.2cov Clear Cover for reinforcing steel 2 1 8 9 18 indin dc posc ft dposc.286.1 Depth to C.L. of steel pos barspacing b n 475.23 barn Number of bars per design width of slab 98
2.1 innAbar poss Total steel area in tributary width. 2 2.24.124 inin Aposs Calculate effective tension area of concrete around the flexural reinforcement bar cn db A 2 Substitute for and the tension area of concrete is: bcdbarn224 inA. Knowing the values of and the service limit state stress for reinforcing steel is given by zcdbarn y c saxf xAd z f 6.0 min3 1 thus: 0.36 saf ksi Calculate the neutral axis of the sect ion to determine the actual stress in reinforcing steel. There is an iterative pr ocess, therefore assume an initia l value of in 8.4 NAXGiven: XdA E E bXpossposs slabc s 22 1 and, the result is in 3.5 posNAX Calculate the tensile force in reinforc ing steel due to the service limit state moment. 99
Given: 3posNA poss pos sX d M T and the tensile force, ksi Ts 4.119 150 220Calculate actual stress in reinforcing steel due to the service limit state moment. Given: poss s actualsA T f and the stresses are, ksi factuals 0.5 3.6 3.9 Given: Modulus of Elasticity of reinforcing steel ksi Es29000 Modulus of Elasticity of concrete ksi Eslabc 310475.3 Substitute the values of stress and the modul us of elasticity in strain formulas: steel steelE f 376.173 74.217 471.319moment loadDead moment loadlive Negative moment loadlive Positivel .. ... ... 100
3.5.3 Uncracked Section Analysis It is necessary to obtain the steel stra in of uncracked section to compare with the actual steel strain under s ervice l oad. The first step is to determine the center of gravity, GCy poss slabc s poss c slabc s GCA E E in A in d E E inb y2.189 28 9 2 .18 in yGC.358.8 C.G. of Uncracked Section Calculate moment of inerti a of uncracked section as 2 2 328 9 2 .18 .18 12 .18 in dy E E Ay in inb in bIc slabc s poss Calculate steel stresses for uncracked section as slabc s c posE E I in dy M 28 9 ksi 059.1 33.1 591.1 Given steel stresses, and steel modulus of elasticity, calculate steel strains, sE sE610. 51.36 852.45 275.67ftkip moment loadDead moment loadlive Negative moment loadlivePositive .. ... ... 3.6 Application of Florida Legal Loads Florida legal load consists of six known truck configurations with the maximum allowable gross vehicle weight, defined in T able 3.2. 101
102 The positive live load moment (LL pos ), Negative Live Load Moments (LL neg ), and Dead load Moments are tabulat ed in Table 3.3. The detail calculations of values shown in Table 3. 3 are shown in Appendix A. The values in Table 3.3 thru 3.8 have been extracted from Figure 3.8 to 3.16. Table 3.3 Moments due to Design Truck Loading, kip-ft Moments kip-in SU2 SU3 SU4 C3 C4 C5 HS20 HL-93 LL pos 1563.8 2774.9 3019.4 1685.7 2467 1900.8 2830.4 3007.2 LL neg 989.7 1893.3 2050.1 1752.8 2350 2268.8 2262.5 2049.6 DL 1632 1632 1632 1632 1632 1632 1632 1632 The Tensile forces, T s in the reinforcing steel due to service limit state moment are shown in Table 3.4. Table 3.4 Tensile Forces T s, due to Service Limit State Moment, kips T s kips SU2 SU3 SU4 C3 C4 C5 HS20 HL-93 LL pos 114.7 203 220.9 123.3 180.5 139.1 207.1 220 LL neg 72.4 138.5 150 128.2 171.9 166 165.5 150 DL 119.4 119.4 119.4 119.4 119.4 119.4 119.4 119.4 Cracked section analysis has resulted in the values of actual steel stress and strain shown in following Tables 3.5 and 3.6.
Table 3.5 Actual Stress, f s in Reinforcing Steel, ksi f s, ksi SU2 SU3 SU4 C3 C4 C5 HS20 HL-93 LL pos 4.8 8.5 9.3 5.2 7.6 5.9 8.7 9.3 LL neg 3.0 5.8 6.3 5.4 7.2 7.0 7.0 6.3 DL 5.0 5.0 5.0 5. 0 5.0 5.0 5.0 5.0 Steel strain of cracked section due to service limit state moment is shown in Table 3.6. Table 3.6 Steel Strain, of Cracked Section, SU2 SU3 SU4 C3 C4 C5 HS20 HL-93 LL pos 166.60 249.79 320 179.08 262.08 201.93 300.68 319.47 LL neg 105.14 201.13 217.79 186.20 249. 65 241.02 240.35 217.7 DL 173.37 173.37 173.37 173.37 173.37 173.37 173.37 173.37 The stress and strain due to servic e limit state moments of uncracked section are tabulated in Tables 3.7 and 3.8 for all eight configurations of design trucks. Table 3.7 Stress, of Uncracked Section, ksi SU2 SU3 SU4 C3 C4 C5 HS20 HL-93 LL pos 1.017 1.8 1.959 1.094 1. 601 1.233 1.536 1.951 LL neg 0.642 1.228 1.33 1.037 1.525 1.472 1.468 1.33 DL 1.059 1.059 1.059 1.059 1.059 1.059 1.059 1.059 103
Table 3.8 Strain, of Uncracked Section, SU2 SU3 SU4 C3 C4 C5 HS20 HL-93 LLpos35.085 62.078 67.548 37.711 55. 19 42.523 63.32 67.275 LLneg22.141 42.356 45.863 39.212 52.573 50.756 50.615 45.852 DL 36.51 36.51 36.51 36.51 36.51 36.51 36.51 36.51 The values of positive and negative moments shown in captions of the following Figures 3.9 through 3.16 are tak en from the preceding Table 3.1. These values are the results of static load analysis by Flor ida Department of transportation, FDOT MathCAD software program (Mathsoft 2002). 01 02 03 04 05 06 07 08 0 003 002 001 001 Envelope of Truck DeflectionsGRID POINTinches00087 00225 posdispdqnegdispdq80 0 dq Figure 3.8 Graph of (+) 0. 009 and (-) 0.023 inches of Deflections due to HS-20 Design Truck Loading 104
0510152025303540 1000 500 500 1000 Lane Load Moment EnvelopeMoment (in*kips)631.7522 882.2549 Apos_ln_adj vtAneg_ln_adj vt40 0 vt Figure 3.9 Graph of (+) 631.8 and (-) 882.3 Moments due to 0.64-kip/in Uniform Lane Load (Moment Values are Shown in Table 3.3) The Code values will be evaluated through static and dynamic testing of the bridge and finite element modelin g. The details are given next. 0510152025303540 1000 1000 2000 Permit Trk Pos & Neg Moment Envelope GRID POINTMOMENT (in*kips)1568.3448 989.6971 Apos_trkp vtAneg_trkp vt40 0 vt Figure 3.10 Graph of (+) 1568.3 and () 989.7 Moments due to SU2 Truck Loading (Moment Values are Shown in Table 3.3) 105
0510152025303540 2000 1000 1000 2000 3000 Permit Trk Pos & Neg Moment Envelope GRID POINTMOMENT (in*kips)2774.9048 1893.304 Apos_trkp vtAneg_trkp vt40 0 vt Figure 3.11 Graph of (+) 2774.9 and () 1893.3 Moments due to SU3 Truck Loading (Moment Values are Shown in Table 3.3) 0510152025303540 4000 2000 2000 4000 Permit Trk Pos & Neg Moment Envelope GRID POINTMOMENT (in*kips)3019.4413 2050.0741 Apos_trkp vtAneg_trkp vt40 0 vt Figure3.12 Graph of (+) 3019.4 and (-) 2050.1 Moments due to SU4 Truck Loading (Moment Values are Shown in Table 3.3) 106
0510152025303540 2000 1000 1000 2000 Permit Trk Pos & Neg Moment Envelope GRID POINTMOMENT (in*kips)1685.7279 1752.8422 Apos_trkp vtAneg_trkp vt40 0 vt Figure 3.13 Graph of (+) 1685.7and (-) 1752.8 Moments due to C3 Truck Loading (Moment Values are Shown in Table 3.3) 0510152025303540 3000 2000 1000 1000 2000 3000 Permit Trk Pos & Neg Moment Envelope GRID POINTMOMENT (in*kips)2467.045 2349.9895 Apos_trkp vtAneg_trkp vt40 0 vt Figure 3.14 Graph of (+) 2467 and (-) 2350 Moments due to C4 Truck Loading (Moment Values are Shown in Table 3.3) 107
0510152025303540 3000 2000 1000 1000 2000 Permit Trk Pos & Neg Moment Envelope GRID POINTMOMENT (in*kips)1900.7791 2268.8251 Apos_trkp vtAneg_trkp vt40 0 vt Figure 3.15 Graph of (+) 1900.8 and () 2268.8 Moments due to C5 Truck Loading (Moment Values are Shown in Table 3.3) 0510152025303540 3000 2000 1000 1000 2000 3000 Truck Max Pos. & Neg. Moment Envelope GRID POINTMOMENT (in*kips)2830.4348 2262.4734 Pmom_trkvtinkip Nmom_trkvtinkip 40 0 vt Figure 3.16 Graph of (+) 2830.4 and (-) 2262.5 Moments due to HS-20 Truck Loading (Moment Values are Shown in Table 3.3) 108
109 3.7 Static and Dynamic Loa d Testing of Bridge To evaluate the previous design ca lculations, static and dynamic load tests were performed on the East Bay Road Bridge using single unit four axles, SU4 trucks as shown in Figure 3.17. SU4 trucks are the most effective due to their short configuration and heavy weight (70 kips). Static tests were performed for six different loading conditi ons. In all six load cases, the third axle of a 70kip SU4 truck was positioned in the middle of the span. Strains were measured using the installed FP sens ors for the six different load cases, respectfully. The strain contour lines of the 16 FP sensors are shown in Figures 3.18 through 3.23 respectively. The results of these te sts will be compared in the next section with a detailed finite element model. Dynamic tests of the bridge under moving trucks with different speeds were also perform ed to confirm the sensors accuracy under dynamic loading.
Figure 3.17 Bridge Load Test with SU4 Trucks In figures 3.18 through 3. 24, the scale at the bo ttom edge of contour graph is the distance along the bridge length in feet and the scale at the right side edge of contour graph is the distance along the bridge width in feet. 110
06.7513.520.252735.2543.551.7560 0 8.5 14.5 18.5 24.5 30.5 36.5 46 21-24.5 17.5-21 14-17.5 10.5-14 7-10.5 35-7 0-3.5 -3.5-0 -7--35 -10.5--7 -14--105 -17.5--14 -21--175 -24.5--21Wheel loads Figure 3.18 Experimental Strain Contour Lines, Load Case 1, Truck Positioned at Mid Span 1. Units in 111
112 0 8.5 14.5 18.5 24.5 30.5 36.5 46 06.7513.520.252735.2543.551.7560 Distance along bridge length (ft)Distance along bridge width (ft) 21-24.5 17.5-21 14-17.5 10.5-14 7-10.5 3.5-7 0-3.5 -3.5-0 -7--3.5 -10.5--7 -14--10.5 -17.5--14 -21--17.5 wheel loadsFigure 3.19 Experimental Strain Contour Lines, Load Case 2, Truck Positioned at Mid Span 2. Units in
06.7513.520.252735.2543.551.7560 0 8.5 145 185 245 305 365 18-215 14.5-18 11-145 7.5-11 4-75 0.5-4 -3-05 -6.5--3 -10--6.5 -135--10 -17--13.5 -205--17 -24--20.5 Wheel loads Figure 3.20 Experimental Strain C ontour Lines, Load Case 3, Trucks Positioned at Mid Span 1 a nd Mid Span 2. Units in 113
Southbound Direction of traffic Direction of traffic Northbound 06.7513.520.252735.2543.551.7560 0 8.5 145 185 245 305 365 21-245 17.5-21 14-175 10.5-14 7-10.5 3.5-7 0-35 -3.5-0 -7--3.5 -105--7 -14--10.5 -175--14 -21--17.5 -245--21Wheel loads Figure 3.21 Experimental Strain C ontour Lines, Load Case 4, Trucks Positioned at Mid Span 2, Both Trucks are in North D irection. Units in 114
Wheel loads Southbound Direction of traffic Direction of traffic Northbound 115 06.7513520.252735.2543551.7560 0 8.5 145 185 245 305 365 17.5-21 14-17.5 10.5-14 7-105 3.5-7 0-3.5 -35-0 -7--3.5 -10.5--7 -14--10.5 -17.5--14 -21--17.5 Figure 3.22 Experimental Strain C ontour Lines, Load Case 5, Trucks Positioned at Mid Span 2, No rthbound and Southbound. Units in
Southbound Direction of traffic 116 Direction of traffic Northbound 06.7513.520.252735.2543.551.7560 0 8.5 145 185 245 305 365 6 17.5-21 14-17.5 10.5-14 7-105 3.5-7 0-3.5 -35-0 -7--3.5 -10.5--7 -14--10.5 -17.5--14 -21--17.5 Wheel loads Figure 3.23 Experimental Strain C ontour Lines, Load Case 6, Trucks Positioned at Mid Span 1, No rthbound and Southbound. Units in Figure 3.24 illustrates the dynamic re sponse of the bridge subject to two SU4 truck in tandem, each weighing 70 kips GVW. The truck traversed over the bridge at the speed of 10 mph. The fi rst truck goes on the bridge and the DSM sensor take the reading at 17 At two seconds later, the sensor reads 15
due to the effect of second truck on the bridge. More detailed descr iption if the dynamic behavior of trucks is discussed in Chapter 4. -2 0 2 4 6 8 10 12 14 16 18 0 5 10 15 20 25 30 35 40Time (sec)Strain ( ) Figure 3.24 Dynamic Strain Tables 3.9 through 3.14 contain the strain values of different sets of sensors subject to SU4 truck load at diffe rent constant speed. The evolution of strain values indicate that the change in strain reading of one sensor subject to a single SU4 truck is not significantly sma ller than the reading of the sensor subject to SU4 truck load in tandem at the same speed. In Table 3.9, CSM sensor is surface mounted to the bottom surface of slab at the center of span 1. These tables were generated during data collect ion process in chapter 4 and shown here for analytical discussions. 117
Table 3.9 Dynamic Response of One Surface Mount Sensor 118 Table 3.10 Dynamic Response of One Embedded Sensor In Table 3.10, sensor F is bonded to bottom mat reinforcing steel. One channel was on, Readings are in Speed MPH Load SU4 CSM Remarks 10 Single 21 10 Tandem 16.5, 18.5 20 Single 13.5 close to that of tandem, 13.5 20 Tandem 12, 13.5 13.5 30 Single 13 30 Tandem 13.5, 15.5 24,25 sec to first and second peaks 35 Single 13 35 Tandem 12, 13.5 40 Single 13 One channel on, Readings are in Speed MPH Load SU4 F Remarks 10 Single 19 10 Tandem 15, 16 20 Single 17 20 Tandem 12.5, 14 30 Single 14 30 Tandem 17, 18.5 24,26 sec to first and second peaks 35 Single 16 35 Tandem 14, 15 40 Single 12
Table 3.11 Dynamic Response of Two Sensor 2 Channels were on, Readings are in Remarks Speed MPH Load SU4 CSM F 10 Single 17.5 11 10 Tandem 4.5, 14 13.5, 15 Anom alies, CSM reading (4.5) 20 Single 12 11.5 20 Tandem 11.5, 12 9, 13 30 Single 13 15 30 Tandem 7, 11 8, 14 35 Single 11 12 35 Tandem 6.75, 9 7, 10 40 Single 8.5 8 40 Tandem -------------unsafe to maintain distance 45 Single 7.5 6.5 45 Tandem -------------Table 3.12 Dynamic Response of Four Sensors 119 unsafe to maintain distance 4 Channels were on, Readings are in Speed MPH Load SU4 H CSM P2 F 10 Single 19 13 11 14 10 Tandem 18.5,14.5 11.5, 6.5 13, 15 12, 14 20 Single 17 10 11 12 20 Tandem 15, 14.5 10/805 9/12 11 30 Single 7 9 6.5 10.5 30 Tandem 11.5, 11.5 4.5, 5.5 5, 11.5 5, 10.5 35 Single 10 8 6 11 35 Tandem 7, 10.5 4.5, 5 5.5, 8 10, 11.5 In Table 3.12, sensor H is bonded to the top mat reinforcing steel. Sensor P2 is surface mount to the deck topside, below the concrete surface.
3.8 Finite Element Modeling A finite element model for the bridge was developed usi ng the commercial software SAP2000 (Computers and Structures Inc., 2004). The bridge deck was modeled using 4-node shell elements. The nodes along the bent lines were assumed fixed in the vertical directi on only. The nodes along the central bent #3 were assumed fixed for both displacements and rotations due to the presence of hook anchorages extending from the bent. Only half of the deck was modeled, as the presence of the fixed supports along t he central bent prevents any forces to be transferred from one side of the deck to the other. The model was used to study the behavior of the bridge under the same loading condition of the static test described in the previous secti on. In this case eight point loads were used to represent the wheel loads. The analytical strain contour lines for load case two are shown in Figure 3.25 is compared to the experimental contours of Figure 3.19. The maximum strain value obtained under the wheels is 35 .. The corresponding recorded ex perimental value was 32 .. From these results and Figures 3.18 and 3.23, it can be concluded that the FP sensors are capable of providing a high degree of accuracy in s ensing the response of the bridge under the truck loads. Figure 3.19 below, is repeated as Figur e 3.26 for a closer viewing and for comparison with Figure 3.25. 122
FIGURE 3.25 Analytical Strain Contour Lines for SU4 Truck on Span 2 Units in 0 8.5 14.5 18.5 24.5 30.5 36.5 46 55 06.7513.520.252735.2543.551.7560 Distance along bridge length (ft)Distance along bridge width (ft) 21-24.5 17.5-21 14-17.5 10.5-14 7-10.5 3.5-7 0-3.5 -3.5-0 -7--3.5 -10.5--7 -14--10.5 -17.5--14 -21--17.5 Figure 3.26 Experimental Strain Contour Lines, Load Case 2, 123
In Tables 3.4 and 3.6, the values of strain for cracked and uncracked section due to SU4 are list ed as 320 and 67.548 micro strains, .. respectively. It would be difficult not to notic e that the value of strain, 320 .. for cracked section, in case of SU4 truck loading, is excessively high and conservative. This comparison simply invalidates the assu mption of cracked section in designing the bridge. On the other and, the assumpti on of Uncracked section which yields in strain value of 67.548 .. is somewhat more in agr eement with the results of field truck load test. These strain values, as shown in Figure 3.18 thru 3.23 are in range of 21 to 24.5 .. Even though these values are for stat istically determinant structures, the restraining effects of indeterminate st ructure will only s lightly reduce it. The following section describes the beam model representing the bridge deck subject to SU4 truck static load for three different load cases as shown in section 3.9, Figures 3.9, 3.10 and 3.11. 3.9 Beam Model Analysis Subject to Static Load The finite element modeling is used to evaluate single beam models. The program MASTAN was used to conduct t he beam analysis. Six different load positions (load cases) were considered in the analysis. Section 2.14, Figure 2.87 depicts the layout plan of action to perform these six different load cases. The process of performing this task was evaluat ed in detail prior to commencement and is presented in Section 2.15, Figures 2.88 through 2.95. 124
Figures 3.26, thru 3.28 show the mo ment diagrams of the beam model for three load cases. In load case 1, a single SU4 truck wa s placed on center of span 1. The corresponding maximum positive moment at Span 1 was +2783 kip in and the corresponding maximum negative moment at Bent 2 was -1500 kip in, (Figure 3.26). The corresponding strain values for case 1 loading is 61 in span 1 in tension (maximum positive moment was 2783 kip in) and 35 over Bent 2 in compression (maximum negative moment was 1500 kip in). Figure 3.27 Moment Diagram for Beam Model for Case 1 Static Load Test In load case 2, a single SU4 truck wa s placed on center of span 2. The corresponding maximum positive moment at Span 2 was +2338 kip-in and the corresponding maximum negative moment at Bent 3 was -3796 kip-in (Figure 3.27), 125
corresponding strain values for case 2 loading is 56 in span 2 in tension (maximum positive moment was 2338 kip-in) and 90 over Bent 3 in compression (maximum negative moment was 3796 kip-in). Figure 3.28 Moment Diagram for Beam Model for Case 2 Static Load Test In load case 3, two SU4 trucks were placed in tandem on center of span 1 and span 2. The corresponding maximum pos itive moment at Span 1 was +2136 kip-in and maximum negative moment at Bent 2 was -2973 kip-in. The corresponding maximum positive moment at Span 2 was +1918 kip-in and maximum negative moment at Bent 3 was -2956 kip-in (Figure 3.28). corresponding strain values for case 3 loading is 51 in span 1 in tension (maximum positive moment was 2136 kip in) and 71 over Bent 2 in compression (maximum negative moment was 2973 kip-in). 126
Figure 3.29 Moment Diagram for Beam Model for Case 3 Static Load Test Strains at the cracking condition d ue to the positive and negative moments (+3019.4, -2050.1 kip-in) are +320 and -217.79 respectively. Strains at the uncracked section are +67.54 and -45.8 respectively. In comparison, the strain value of finite element model is 35 and strain value due to the static load test (SU4 truck) was 24.5 This strain value (24.5 ) is less than half strain value of 61 calculated for the beam model fo r case 1 loading on span 1 over sensors CSM and DSM. The program that was used to ca lculate the moments for beam model does not calculate the strain. The fo llowing steps are used to determine the strain values of beam model fo r three different load cases. Given: Area of concrete section is, dhAc 127
Where: h concrete deck thickness d distribution width, 11.87 feet Therefore 225641287.11"18 A The moment of inertia can be calculated as: 312 1 bhI 4 384.225,69181287.11 12 1 in I Stress is: I Mc the moment for load case 1 is 2783 kip-in Substitute for M, c and I to get the strain ksi259.0 84.225,69 72783 And strain: cE where cE is modulus of elasticity of concrete ksi psi psi Ec4227 13.233,227,45500 57000 128 000061.0 4227 259.0 or 61 This value of strain is relatively hi gh indicating that the distribution width, is possibly too small. This assumption can be investigated and verified by recalculating the strain values with a larger distribution width, DE DE
129 CHAPTER 4. DATA COLLECTION This chapter describes the methodology for collecting data from the bridge under service load. This data was us ed to check the East Bay Road Bridge designed by LRFD code, and compare the values of strains obtained by using The Florida Department of Transpor tation software program with the experimental values of strains obtained fr om static load test of bridge with SU4 truck. Details of this evaluat ion are described in Chapter 3. The static load was performed under a controlled weight and speed condition. The bridge was closed to traf fic while the locations of sensors were marked on the deck top surface (Figures 4.1 and 4.2). With no truck on the bridge, the sensors readings were recor ded. These readings are the baseline or the zero readings of the sensors (Table 4. 1). A SU4 truck with gross weight of 67,360 pounds was placed on the marks on t he bridge. The tires of the middle rear axle were placed directly over the sensors, (F igure 4.3). Four sensors designated as C, D, E, and F were bonded to the primary reinforcing steel of the bottom mat in span 2 for m onitoring strain due to the positive moment. Two sensors designat ed as P1 and P2 were placed on the deck below the surface over bent 1. Four sensors designated as G, H, I and J were bonded to the primary reinforcing steel top mat over bent 3 (Refer to Chapter 2, experimental Section 2.10.3, Figures 2. 31, 2.32 and 2.33).
130 FISO Commander standard software version 2 developed by FISO Technologies Inc was used for data collectio n. A laptop computer was directly connected to DMI-16, data logger thr ough RS-232 communication cable on site (Figure 4.4). Table 4.1 repr esents the strain values of static load test. Sensors G, J, C and D are located at the exterior lane (8 feet shoulder, the emergency lane). Sensors G and J are bonded to the to p of the reinforcing mat over bent 3 and sensors C and D are bonded to the botto m reinforcing mat at the mid-span 2. The alphabetically out of order position of sensor J was due to the shorter length of the fiber optic cable. Howe ver, the ascending numbers of channels were assigned to the sensors in alphabetical order (e.g., G = Channel 1, H = 2 I = 3 and J = 4). The strain readings of sensors G, J, C and D are due to the negative and positive mom ents. The higher values of strain readings for sensors H, I, E and F are relevant to the positions of the sensor s located in the travel lane (Chapter 2, Section 2.10.3, Figure 2.31 through 2.33). Sensors P1 and P2 are embedded in the concrete below the surf ace over bent 2 to capture strain due to the negative moment. Tables 4.2 and 4.3 represent remotely collected data from the static load. The connection between the desktop com puter and DMI took place through the modem. The bridge was closed to traffic and the office was notified to record the zero readings. The rear middle axle of the truck was positioned on the bridge deck over the sensors designated as E and F, for the flexural condition. All channels were turned on and signals were transmitted through the modem to the computer.
Table 4.1 On Site Data Co llection with a Laptop Com puter from Static Load Test with SU4 Truc k, GVW= 67,360 lbs Channel Sensor Zero Readings ( ) No traffic on bridge Truck on the bridge ( ) Resultant readings ( ) 1 G -0.6 3.8 4.4 2 H 0.4 9.4 9.0 3 I 1.2 9.4 8.2 4 J 0.6 8.8 8.2 9 P1 2.2 8.4 6.2 10 P2 2.4 6.6 4.2 11 T1 34.60 C0 34.60 C0 34.60 C0 12 T2 24.78 C0 24.78 C0 24.78 C0 13 C 2.0 8.2 6.2 14 D 2.6 13.2 10.6 15 E 3.0 28.4 25.4 16 F 3.0 28.2 25.2 131
Table 4.2 Remotely Colleted Data wi th a Desktop Computer from Static Load Test with SU4 Truck, GVW= 67,360 lbs Channel Sensor Zero Readings ( ) No traffic on bridge Truck on the bridge ( ) Resultant readings ( ) 1 G -0.2 3.6 3.8 2 H -0.2 9.4 9.6 3 I 0.0 8.4 8.4 4 J 0.2 7.8 7.6 9 P1 -2.2 9.0 11.2 10 P2 -1.8 10.0 11.8 11 T1 34.60 C0 34.60 C0 34.60 C0 12 T2 24.78 C0 24.78 C0 24.78 C0 13 C 1.2 8.0 7.8 14 D 1.8 13.0 11.2 15 E -2.4 23.6 25.6 16 F -2.4 25.4 27.8 132
Table 4.3 Verification of Remotely Collected Data with a Desktop Computer from Static Load Test with SU4 Truck, GVW= 67,360 lbs Channel Sensor Zero Readings ( ) No traffic on bridge Truck on the bridge ( ) Resultant readings ( ) 1 G 0.0 3.6 3.6 2 H -0.5 9.4 9.9 3 I 1.0 8.4 7.4 4 J 1.0 7.8 6.8 9 P1 -0.5 9.2 9.7 10 P2 0.5 10.0 9.5 11 T1 34.60 C0 34.60 C0 34.60 C0 12 T2 24.78 C0 24.78 C0 24.78 C0 13 C -0.5 7.8 8.3 14 D 0.5 13.2 12.7 15 E 0.0 23.6 23.6 16 F 0.0 25.4 25.4 133
Figure 4.1 Locating the Sensors on Deck Topside Figure 4.2 Marking the Locations of the Sensors on Deck Topside Figure 4.3 SU4 Truck was Placed with the Tires of the Middle Rear Axle Placed on the Mark Over the Sensors on Span 2 4.1 Remote Monitoring of th e East Bay Road Bridge The data logger (DMI) loca ted 35 miles from the office was remotely connected to a desktop computer via modem. DMI was set to the direct data acquisition mode (Figure 4.10). This mode utilizes the graph of strain as a raw 134
form during the scanning. This data was saved on the computers hard drive for later analysis. Screen shots of the dynam ic strain profiles were taken in the office are shown (Figures 4.5 and 4.6). The desired channels were turned on. Speeds of the trucks or cars going over the bridge were unknown. The large spikes indicate the passage of large tr ucks traveling over the bridge. Small spikes in the graph indicate the passage of cars over the bridge. The traffic data are currently being continuously collected and analyzed for the purpose of investigating the long term bridge behavio r under traffic loading. These data were used to compare maximum recorded stresses to LRFD design values, and detecting possible future deficiencies thr ough a long term remote monitoring of the bridge. Discussion on this subject will be presented in Chapter 5. Figure 4.4 shows field truck load test and on site live data collection Figure 4.4 Field Truck Load Test and on Site Real Time Data Collection 135
Figure 4.5 Graph of Heav y Trucks on the Bridge FIGURE 4.6 Graph of Cars a nd Light Trucks on the Bridge 136
137 4.2 Running the Software, FISO Commander 4.2.1 Detect Comport for Laptop Communication Port Figures 4.7 through 4.12 show the necessary steps in using the software to connect to the DMI for data collecti on. The first step for using FISO commander Standard Edition v1.9.8 is to determine the comport of the laptop computer that will be used for on site di rect connection. Click on start menu, click on settings, open control panel and select systems icon from control panel. Next, click on Hardware then Device Manager At this point, click on the + sign next to Port (Com & LPT). The co mmunication port will be indicated in the parenthesis, e. g., (Com1). As soon as the comport is determined, execute the software by clicking on the application icon in program files. A screen as shown in Figure 4.7 will come up and show the information on the system. To initialize communication with DMI, scroll down and select communication port and wait. After about 3 to 4 seconds, a dialogue box will come up and prompts for teleph one number input. Input the telephone number and click O K (Figure 4.8)
Figure 4.7 Initializing DMI Figure 4.8 Modem Communicati on Initialization Dialogue Box Figure 4.9 shows the software malfunc tion. During an attempt to connect to DMI data conditioner, the software could not detect DMI but found a portable single channel data logger, FTI-10. This error frequently reoccurred. In this 138
case, quit the program and restart the soft ware until the correct data conditioner is detected. 4.2.2 Detect Comport for Desktop Remote Connection To determine the comport for remote connection at the office located anywhere, the following steps should take place. Click on the start menu, click on settings, click on the control panel, double click on the phone and modem options, click on the modems button and the modem compor t will be indicated. Figure 4.9 Conditioner Initiali zation Error (FTI-10) After communication has been established, a dialogue box as shown in Figure 4.10 will come up. This dialog box has several buttons for various applications. Configure conditioner butt on will allow to change the type of scan 139
from automatic to manual and set the scan rate. Configure transducer Button will allow sensors to be turned on or off. After desired transducers are selected, the save button must be clicked before proceeding to the next function (Figure 4.11). Finally, the Direct acquisition (graph) allows starting, stopping and saving the data acquisition (Figure 4.12) The Delay acquisition function (Figure 4.13), was not used in this experim ent. However, this function allows for setting the time and duration of the acquisition for the conditioner similar to a common timer. 140 Figure 4.10 Configure Conditioner
141 igure 4.12 Direct Ac quisition with Graph Figure 4.11 Configure Transducer/Sensors Assignment F
142 oftware v1.9.8. During the early da ys of the experiment, data (strains) ppeared to be within the expected range. However, remote connection to the ptop was not possible. Soon, the integrity of collected data began to eteriorate. A new version, v1.9.9 was developed for this experiment and used r a short period of time The remote connection with laptop was not achieved nd poor integrity in collected data was noticeable. A newer version, v1.9.9 uild E2) !!! for Ebrahim only !!! was developed and put to use. The same ymptoms as with v1.9.9 were immedi ately apparent when using this version igures 4.14 through 4.17) Figure 4.13 Delay Acquisition 4.3 Using Different Versions of FISO Commander Software The data collection from the field experiment began with FISO commander s a la d fo a (B s (F
143 Figure 4.16 Dynamic Graph Channels Figure 4.17 Dynamic Graph 3Channels FISO Technologies took an approach in developing software compatible with DMI and the laptop computer, as well as the desktop computer. Most of the ftware is comparable with the la st version, v1.9.9 (Build This version, v2 has resolved the laptop remote s opened in MS Excel, (stain values Figure 4.14 Dynamic Graph 4Channels Figure 4.15 Dynamic Graph 1Channel 2 features of this latest so E2)!!! For Ebrahim only!!!. connection, and the processed raw data wa ~26 ) was close to the strain value of ~28 obtained with direct connection to DMI via RS-232 communicati on cable.
144 4.4 Ru d icon in on umber is st ored in the memory for the subsequent this version of FISO commander has takes place, a dialog box shown in in Figure 4.20 were working properly y defau isplay. The preset information did not change during this nning FISO Commander v2 Software Program initialization of FISO co mmander v2 is similar to v1.9.9 an opens from the program files. After t he comport is determined using the steps described for v1.9.9, execute the software by clicking on the application the program files. Figure 4.18 wil l appear and show the information on the system. To initialize comm unication with DMI, place a check mark in the modem box, then scroll down on the connection bo x, click on the desired comport and wait. A few seconds later, a dialog box will appear and the prompts for the telephone number input. Input the telephone number and click on the dial butt (Figure 4.19). The telephone n connections. Every attempt to connect to DMI with been successful. As soon as the connection Figure 4.19 will come up with all buttons being active. All of the buttons on the dialogue box in every connection attempt. A factor lt is preset in the system configuration d experiment. (Figure 4.21)
145 Figure 4.19 Modem Setup Dialogue Box Figure 4.20 Application Selection Dialogue Box Figure 4.18 Connection Initialization
146 igure 4.21 System Conf iguration Information The gage/channel configuration butt on in Figure 4.20 is the most 4.22 will come up. In this the gage lis er must correspond expe nged during the dge. The purpose of the load test condition. For the first time, all gages (sensors) ar e set to the Off po sition. To turn d scroll up or down to select the orresponding gage name. The gage factor will automatically appear next to the gage name and in the next column the gage zero will show up. The gage zero is F frequently used option in this experiment. When this button is clicked, Figure box, the gage information must be accurately input in t column. Gage name, gage factor and channel numb to each other, otherwise e rroneous data will be gathered during the riment. When the gage list is comple ted, it will remain uncha entire experiment and likely th rough the life of the bri gage setting is to turn on or off as many gages as are desired for a particular on a gage, click on the gage number an c
147 a number that is internally calculat ed for each gage based on the gage length and gage factor. This number is the baseline for the gage readings when subject to a load condition. ox. When the on start button is click ed, scanning will begin with pre-determined color contrast problems by transparent colors that are not clearly visible. Figure 4.22 Gauge List and Channel Setting When the desired gages in Figure 4.22 are selected, click on the graphic acquisition button in Figure 4.20 and wait. The dialogue box in Figure 4.23 will come up. In this box, insert the desir ed duration of data acquisition or check the infinite for up to 50,000 bytes of scanning, then either click the on start button to begin scanning or click on the advance conf iguration button to go to next dialog b color of graphs by default. This option may cause
148 on button to change the selection. A graphic configuration box will appear. At this point, click on t he + sign next to the Graphic option in the Graphic visible channel section of the dialogue box seen on Figure 4.24. Select the channel to c heck the pre-determined colo r of channel and click O Figure 4.23 Graphic Acquisition for the Selected Sensors If the color contrast is not acceptabl e, click on the stop button and then the advanced configurati K if the color meets an acceptable contrast. To select another color, click on the color (green in this case). This is s hown in the channel properties section of the box in Figure 4.24. This action will bring up the basic color chart (Figure 4.25). Select the desired color and click O K. The graphic acquisition dialog box of Figure 4.23 will appear. Click on the start button to begin scanning. A warning ialogue box will appear, with prompts of Yes or No opti ons, Figure 4.26. d
149 step will take return to the ialogue box in Figure 4.23. Data will be sa ved in sub-directory DATA of the progra re 4.24 Graphic Conf iguration Dialogue Box Figure 4.25 Basic and Custom Colors Chart Figure 4.26 Warning to Protection Unsaved Data n application ogram. If N Click on Yes button the initiate the scan. This d m for analysis use or deletion. Figu Click on the Start button to initialize the data acquisitio pr o button is clicked, the pr ogram will be aborted. It is
150 s Acquisition recommended to avoid aborting the program si nce restarting the program is time consuming and there is a possibi lity of not restarting. Memory Acquisition application of t he v2 of the software (Figure 4.27) acquires data similar to v1.9.9. Howe ver, version v2 has a more appealing graphic presentation than version v1.9.9 (Figure 4.23). In this application, scanning can be programmed for long term m onitoring at various time interval similar to v1.9.9 (All version preceding v2 are now considered obsolete). Figure 4.27 Memory/Delay File Acquisition application of v2 is similar to the Direct Acquisition (File) application of v1.9.9. This applic ation is in the process for further development (Figure 4.28).
151 Figure 4.28 File Acquisition A pplication of v2 Software motely Collected A desktop computer at the office was remotely connected to the DMI system. In case of static load test, a ll channels can be turned on to see the true load distribution sensed by the gauges plac ed in critical locations (Tables 4.1 through 4.3). However, in dynamic test when the truck goes over a sensor and reaches the next sensor, DMI is still scann ing the first channel and the reading of the other channels are diffe rent than those recorded fr om static load with the same truck load. These differences also vary for different vehicle speeds of travel. The manufacturer does not re commended DMI-16 data conditioner more than eight channels for dynamic load test. The following tables and graph indica te that when four channels are turned on, a significant drop in the two last channels are obs erved. Questions may arise concerning the other data conditioners available capability of dynamic scanning of more than two channels for fast moving vehicles. The answer is yes 4.5 Presentation of Re Dynamic Data
152 uch a system is available (BUS Comm ander by FISO) however; the price was eyond the research budget for this study. he following figures are the depiction of real-time truck activities over the bridge. The data acquisition is complete ly uncontrolled, that is, the position, nown. All the data was remotely acquired via telephone line through the m odem. The on and off modes of the rs were ly acqired data by DMI was stored in a directory on e local hard-drive for later analysis Figure 4.29 shows the truck load and mperature response by four sensors, H, I, T1, and T2. The sensors H and I re those bonded to the top reinforcem ent for negative moment. Figure 4.30 lso shows the truck load and temperature response by four other sensors, E, I, 1, and T2. The sensor E is for pos itive movement and I is for negative moment, respectively. Unlike version v1.9.9, the v2 versi on of this software does not display the strains values as they are detected by the sensors. In version v2 of the software, the raw data has to be decoded (processed) in Microsoft Excel through comadelimited feature of the program. When processing data in MS Excel, a channels have to be set at zero otherwise the graph will show different points of origins for the channels. On the other hand, v1.9.9 had an advantage for directly us Figures 4.14 rough 4.17. s b T number of trucks and their speeds are unk sensors were controlled remotely by a desktop and a laptop computer. The caption for each figure indicates which s ensor was on while the off senso not of any concern. Remoteu th te a a T ll displaying the strain values on the graph as shown in previo th
153 ). tom of Slab Figure 4.29 Channels H (Green), I (Blue) T1 (Red) & T2 (Y ellow); Strain and Temperature Graphs. (Speed is Unknown) Figure 4.30 Graph of Strain Values fo r Channels E (Yello w) and I (Blue Where T1 (Red) & T2 (Pink) are Gra phs of Temperature for Top and Bot
154 Figure 4.32 Response of E and P2 Sensors Figure 4.31 Graph of Response from P1 and P2 Sensors
155 Figure 4.33 Response of D, E and F Sensors to a Tandem Condition e following tables illustrate the collected data under a controlled tigate the tion). Also, the effect of the same load with different s peeds was investigated. For location of sensors, refer to Chapter 2, section 2.10, and Figur es 2.29 through 2.33. Th condition for the load and the speed of v ehicles. Two SU4 trucks were employed for this on site dynamic load test. The purpose of this test was to determine the maximum strain values at the critical location of the bridge and inves effect of moving vehicle on other parts of the bridge (load distri bu
156 eds One channel was on, Readings are in Table 4.4 Strain Values of Sensor CSM for Different Spe Speed MPH Load SU4 CSM Remarks 10 Single 21 10 Tandem 16.5, 18.5 20 Single 13.5 Close to Tandem at 20 mph 20 Tandem 12, 13.5 13.5 30 Single 13 30 Tandem 13.5, 15.5One second delay between the trucks 35 Single 11 35 Tandem 12, 13.5 40 Single 13 Table 4.5 Strain Values of Se nsor F for Different Speeds speed MPH Load SU4 One channel turned on, Readings are in F Remarks 10 Single 19 10 Tandem 15, 16 20 Single 17 20 Tandem 12.5, 14 30 Single 14 30 Tandem 17, 18.5 Two seconds between the trucks 35 Single 16 35 Tandem 14, 15 40 Single 12 Table 4.6 Strain Values of Sensors CSM and F for Different Speeds 2 Channels were on, Readings are in Speed MPH Load SU4 CSM F Remarks 10 Single 17.5 11 10 Tandem 4.5, 14 13.5, 15 Anom alies, CSM Readings, (4.5) 20 Single 12 11.5 20 Tandem 11.5, 12 9, 13 Truck were close to each other 30 Single 11 10 30 Tandem 7, 11 8, 14 35 Single 9 8 35 Tandem 6.75, 9 7, 10 40 Single 8.5 8 40 Tandem -----------Unsa fe for this condition 45 Single 7.5 6.5 45 Tandem -----------Unsa fe for this condition
157 s of 40 nd 45 MPH. Driving two fully loaded SU 4 trucks in Tandem position at 40 and Table 4.7 Strain Values of Sensors H, CSM, P2 and F for Different Speeds At 30 mph and higher speeds, the strain values may not be considered at andemhese speed s cse to ther. Tfference between the trucks to the point of load varied from 2 to 4 seconds. The following graphs (Figures 4.43 thru 4.40) indicate strain values of sensors at different locations for singl speed. A close observation of values in t he tables reveals that as more channels re turned on, drop in strain readings are noticed. The drop in strain readings aso obss the speedcrease. In Table 4.6, there are no strain val ues for Tandem trucks at speed a 45 MPH were not performed to pr event rear end collisions. 4 Channels were on, Readings are in Speed MPH Load SU4 H CSM P2 F 10 Single 19 13 11 14 10 Tandem 18.5, 14.5 11.5, 6.5 13, 15 12, 14 20 Single 13 10.5, 12 12, 13 11.5, 10 20 Tandem 9.5, 8 10, 8.5 11, 10.5 10.5, 12 30 Single 7 9 6.5 11 30 Tandem 11.5, 11.5 4.5, 5.5 5, 11.5 5, 10.5 35 Single 10 8 6 11 35 Tandem 7, 10.5 4.5, 5 5.5, 8 10, 11.5 a true t At t s, it was very difficult to drive the trucklo each o he di in time e and tandem truck load test at 10 mph a re al erved a is in d
160 Single Sensor, CSM=16.5 and 18.5 Dynamic load t est, two SU4 trucks in tandem at 10 mph CSM=16.5 and 18.5 me -2 3 8i 13ain ( 18) 23 ch6-CSM Dy nam c Str me 0102030405060708090 Time (sec.) Figure 4.34 Dynamic Load Test, Two SU4 Trucks in Tandem at 10 mph Over a Figure 4.35 Dynamic Load Test, Two SU4 Trucks in Tandem at 10 mph, F=15 &16 Dynamic Load test, two SU4 trucks in tandem at 10 mph F=15 &16 me -2 0Time (sec.) 2 4Dy 6nam 8 Stra 10in ( 12me ) 14 16 18 Ch 16F ic 050100150200250300350400450500
Dynamic Load test, sigle SU4 truck at 10 mph me 13 18 23me CSM=17.5, F=11 -2 3 8 01020304050607080Time (sec.)Dynamic Strain ( ) ch 6-CSM ch 16-F 161 Figure 4.36 Dynamic Load Test, Single SU4 Truck at 10 mph, CSM=17.5 and F=11 Figure 4.37 Dynamic Load Test, Two SU4 Trucks in Tandem at 10mph, CSM=14&9.4 and F=15&13.5 Dynamic Load test, two SU4 trucks in tandem at 10mph CSM=14&9.4, F=15&13.5 me -2 0 2 4 6 8 10 12 14 16 0102030405060708090Time (sec.)Dynamic Strain ( me ) ch6-CSM ch16-F
162 Figure 4.38 Dynamic Load Test, Two SU4 Trucks in Tandem at 10 mph H=18.5 & 14.5, CSM=11.5 & 6. 5, P2=13 & 15, F=12 & 14 Dynamic Load test,two SU4 trucks in tandem at 10 mph H=18.5 & 14.5, CSM=11.5 & 6.5, P2=13 & 15, F=12 & 14 me -2 3 8 13 18 23 0102030405060708090Time (sec.)Dynamic Strain ( me ) ch2-H ch6-CSM ch10-P2 ch16-F =14&10, 5, F=15&8.5 Figure 4.39 Dynamic Load Test, Single SU4 Truck at 10mph, H I=14&6.5, DSM=18&9, CSM=19&5, P1=8.5, P2=10.5, E=14.5&6. Dynamic Load test, two SU4 trucks in tandem at 10 mph H=14 & 10, I=14 & 6.5, DSM=18 & 9, CSM=19 & 5 P1=8.5, P2=10.5, E=14.5 & 6.5, F=18.5 23 5 & me -2 3 8 13 18 0 20406080100120140Time (sec.)Dynamic Strain ( me ) Ch 2-H Ch 3-I Ch 5-DSM Ch 6-CSM Ch 9-P1 Ch 10-P2 Ch 15-E Ch 16-F
163 ASM=2.5, BSM=5, CSM=6, DSM=6, Dynamic Load test, two SU4 trucks in tandem at 10mph G=2, H=13, I=9, J=9, P1=5, P2=7.5, ASM=2.5, BSM=5, CSM=6, DSM=6 C=6, D=4, E=4.5, F=0.5 me -2 0 2 4 6 8 10 12 14 0102030405060708090100Time (sec.)Dynamic Strain ( me ) ch1-G ch2-H ch3-I ch4-J ch5-ASM ch6-BSM ch7-CSM ch8-DSM ch9-P1 ch10-P2 ch11-T1 ch12-T2 ch13-C ch14-D ch15-E ch16-F Figure 4.40 Dynamic Load Test, Two SU4 Trucks in Tandem at 10 mph G=2, H=13, I=9, J=9, P1=5, P2=7.5 C =6, D=4, E=4.5, F=0.5 The following graphs (Figures 4.41 thru 4.50) indicate strain values of sensors at different locations for single and tandem truck load test at 20 mph speed. Figure 4.41 Dynamic Load Test, Single SU4 Truck at 20 mph, CSM=13.5 Dynamic Load test, sinble SU4 truck at 20 mph CSM=13.5 me -2 0 2 4 6 8 10 12 14 16 05101520253035404Time (sec.)Dynamic Strain ( me ) Ch 7-CSM 5
164 igure 4.42 Dynamic Load Test, Two SU4 Trucks in Tandem at 20 mph, F CSM=12 & 13.5 Figure 4.41 shows sensor CSM has recorded 13.5 under a single SU4 at 20 mph. Figure 4.42 shows sensor CSM has recorded 13.5 under a the second truck in tandem and Figure 4.43 shows sensor F has recorded 13.5 under second truck in tandem. This com of second truck in tandem is close to t he strain value for a sing truck. This concludes that stress for a tandem truck under the same condition is close to the stress for a single truck. Also, from the same figures, it can be seen that the ded to rebar, for the similar loading condition. parison shows that the strain value value of strain for surface mount sensor is slightly larger than the sensor bon Dynami Load test, two SU4 trucks in tandem at 20 mph CSM=12 & 13.5 me -2 0 2 4 6 8 10 12 14 16 010203040Time (sec.)Dynam ic Strain ( me ) Ch 7-CSM 506070
165 igure 4.43 Dynamic Load Test, Two SU4 Truck in Tandem at 20 mph, = 12.5 and 14 Dynamit Load test, SU4 truck in tanden at 20 mph F = 12.5 & 14 me -2 0 2 4 6 8 10 12 14 16 010203040506070Time (sec.)Dynamic Strain ( me ) Ch 16-F F F Dynamic Load test, single SU4 truck at 20 mph CSM=12, F=11.5 me 14 Ch 6-CSM Ch 16-F 12 -2 0 2 4 10 01 02 03 04 05 06 0Time (sec.)Dynatr ) 6 8mic Sain ( me Figure 4.44 Dynamic Load Test, Singl e SU4 Truck at 20 mph, CSM=12, F=11.5
Dynamic Load test, two SU4 trucks in tandem at 20 mph CSM=11.5&12, F=9&13 me 01020304050607080Time (sec.) -2 0 2 4 6 8 10 12 14Dynamic Strain ( me ) Ch 6-CSM Ch 16-F 166 amc Load Test, Two SU 4 Trucks in Tandem at 20 mph, SM=11.5&12, F=9&13 igure 4.45 Dyni F C mic Load Test, Single SU 4 Truck at 20 mph H=11.5, I=8.5, 5, PI=5.5, P2=5.5, E=5.5, F=3 Dynami Load test, single SU4 truck at 20 mph -2 0 4 6 8 14Time (sec.)Dynaic St H=11.5, I=8.5, DSM=13, CSM=6.5, PI=5.5, P2=5.5, E=5.5, F=3 me 2 10 12 01020304050607080mrain ( me ) Ch 2-H Ch 3-I Ch 5-DSM Ch 6-CSM Ch 9-P1 Ch 10-P2 Ch 15-E Ch 16-F Figure 4.46 Dyna DSM=13, CSM=6.
167 Figure 4.47 Dynamic Load Test, Single SU4 Truck at 20 mph, H= 10, CSM= 6.5, P2= 6.5, F=5 Dynamic Load test, single SU4 truck at 20 mph H=10, CSM=6.5, P2=6.5, F=5 me 0 2 4 6 8 10 12 11520253035404550Time (sec.)Dynamic Strain ( me ) Ch 2-H Ch 6-CSM Ch 10-P2 Ch 16-F -2 050 igure 4.48 Dynamic Load Test, Two SU4 Trucks in Tandem at 20 mph, =3.5 & 11, I=5.5 & 10.5, DSM=13.5 & 1, CSM=12.5 & 3 Dynamic Load test, Two SU4 trucks in tandum at 20 mph H=3.5 & 11, I=5.5 & 10.5, DSM=13.5 & 1, CSM=12.5 & 3 me -2 0 2 8 10 12 14 16 0102030405060708090100Time (sec.)Dy in ( me ) Ch 2-H Ch 3-I Ch 5-DSM Ch 6-CSM 4 6namic Stra F H
Figure 4.49 Dynamic Load Test, Single SU 4 Truck at 20 mph, H=11, I=10.5, Dynamic load test, Single SU4 trucks in tandem at 20 mph H=11, I=10.5, DSM=13.5, CSM=12.5, P1=8, P2=8.5, E=17.5, F=12 me -2 3 8 13 18 23 0102030405060708090100Time (sec.)Dynamic Strain ( me ) Ch 2-H Ch 3-I Ch 5-DSM Ch 6-CSM Ch 9-P1 Ch 10-P2 Ch 15-E Ch 16-F DSM=13.5, CSM=12.5, P1=8, P2=8.5, E=17.5, F=12 Dynamic Load test, single SU4 truck at 20 mph DSM=19.5, H=10 me -2 3 8 13 18 23 0102030405060708090100Time (sec.)Dynamic Strain ( me ) Ch 1-G Ch 2-H Ch 3-I Ch 4-J Ch 5-DSM Ch 6-CSM Ch 7-ASM Ch 8-BSM Ch 9-P1 Ch 10-P2 Ch 11-T1 Ch 12-T2 Ch 13-C Ch 14-D Ch 15-E Ch 16-F Figure 4.50 Dynamic Load Test, Single SU4 Truck at 20 mph, DSM=19.5 and H =10 168
169 In Figure 4.50, two readings are d epicted, other reading were very low test with a speed values of strain. The horizontal line s indicate the temperature readings. The following graphs (Figures 4.51 thru 4.57) indicate strain values of the sensors at different locations for singl e and tandem truck load of 30 mph. Figure 4.51 Dynamic Load Test, Two SU4 Truck at 30 mph, CSM=13.5 and 15.5 CSM is a Surface Mount Sensor Figure 4.51 shows the readings of st rain values for a true tandem truck condition. Normally, this is not a reo ccurring situation except for the field experiment. The graphs shows increase in strain value due to the second truck. Figure 4.52 also shows a true tandem condition, however the strain due to the Dynamic Load test, two SU4 trucks in tandem at30 mph CSM=13.5 &15.5 me -2 0 2 8 10 12 14 16 18 051015202530354045Time (sec.)Dtrain ( me ) Ch 7-CSM 4 6yna mic S lue teel. second truck has decreased. It is stipulat ed that this decrease in the strain va c an be due to the location of the sensor bonded to the reinforcing s
Dynamic load test, two SU4 trucks in tandem at 30 mph F=17 &15 me -2 0 2 4 6 8 10 12 14 16 18 05101520253035404550Time (sec.)Dynamic Strain ( me ) Ch 16-F 170 n Tndem at 30 mph, = 17 and 15 Figure 4.52 Dynamic Load Test, Two SU 4 Trucks ia F igure 4.53 Dynamic Load Test, Two SU4 Truck in Tandem at 30 mph, Dynamic Load test, two SU4 truck in tandem at 30 mph CSM=7 & 11, F=8 & 14 me -2 0 2 4 6 8 10 12 14 16 01 02 03 04 05 0Time (sec.)Dynamic Strain ( me ) 6 0 Ch 6-CS M Ch 16-F F CSM=7 & 11, F=8 & 14 Figure 4.51 indicates that two tr ucks of equal weights were approaching e bridge at the speed of 30 mph. Howeve r, the second truck fell too far behind the first truck. The graph shows that the second truck did have an apparent th
171 d 0 second time lapse between the two trucks. It is possible that the second truck effect on the sensors and incr eased the readings. It is not clear why the secon truck resulted in the sensors readings to increase since there is almost a 4 had reduced its speed to increase the readings by 1.5 to 3 5, Figure 4.54 Dynamic Load Test, Two SU 4 Trucks at 30 mph, CSM=6&7. F=7&10 Figure 4.55 Dynamic Load Test, Singl e SU4 Trucks at 30 mph, CSM=13 Dynamic Load test, single SU4 truck at 30 mph CSM=13 me -2 6 8 14yc Strain ( me ) 0 2 4 051015202530354045Time (sec.)Dnami 10 12 Ch 7-CSM Dynamic Load test, two SU4 tr uck in tandem at 30 mph CSM=6 & 7.5, F=7 & 10 me -2 0 2 4 6 8 10 010203040506070Time (sec.)Dynamic Strain ( me ) 12 Ch 6-CSM Ch 16-F
Dynamic Load test, two SU4 trucks in tandem at 30 mph H=11.5 & 11.5, CSM= 4.5 & 5.5, P2=5 & 11.5, F=5 & 10.5 me -2 0 2 4 6 8 10 12 14 0102030405060708090Time (sec.)Dynamic Strain ( me ) Ch 2-H Ch 6-CSM Ch 10-P2 Ch 16-F 172 1. CSM = 4.5 and 5.5, F= 5 and 10. Figures 4.56 Dynamic Load Test, Two SU4 Trucks in Tandem at 30 mph, H= 11.5 and 15, 5 Dynamic Load test, single SU4 truck at30 mph H=7, CSM=9, P2=6.5, F=3 me -2 0 2 4 6 8 10Dynamic Strain ( me ) Ch 2-H Ch 6-CSM Ch 10-P2 Ch 16-F 01 02 03 04 05 06 0Time (sec.) Figure 4.57 Dynamic Load Test, Single SU4 Truck at 30 mph, H=7, CSM=9, P2=6.5, F=3
Dynamic Load test, single SU4 truck at 30 mph, G=3, H=6. 5, I=9, J=8.5, DSM=15, CSM=1.5, ASM=1.5, BSM=-1.5, P1 =4.5, P2=4, C=2.5, D=4, E=10.5, F=16 me T1=12.55 T2=18.15 celsius 21 -4 1 6 11 16 01020304050607080Time (sec.)Dynamic Strain ( me ) Ch 1-G Ch 2-H Ch 3-I Ch 4-J Ch 5-DSM Ch 6-CSM Ch 7-ASM Ch 8-BSM Ch 9-P1 Ch 10-P2 Ch 11-T1 Ch 12-T2 Ch 13-C Ch 14-D Ch 15-D Ch 16-F 173 Figure 4.58 Dynamic Load Test, Single SU4 Truck at 30 mph, G=3, H=6.5, I=9, J=8.5, DSM=15, CSM= 1.5, ASM=1.5, BSM=-1.5 P1=4.5, P2=4, C=2.5, D=4, E=10.5, F=16 T1=12.55, T2=18.15 0C e xt graphs (Figures 4.59 thru 4.68) show strain values of the ucks since there was one case, sor graph was of the first four These graphs Thne sensors at different locations for si ngle and tandem truck load test at 35 mph speed. It was determined not to run the test with tandem tr a risk of an accident for the trucks due to the proximity of the trucks. In eight sensors were turned on. For the sa ke of clarity, the eight-sen divided into two four-sensor graphs. Fi gure 4.59 is the graph sensors of an eight sensor group from Figure 4.61. Figure 4.60 is the graph of second four sensors of an eight sensor group from Figure 4.61. are illustrated in Figures 4.59, 4.60 and 4.61.
174 Figure 4.59 Dynamic Load Test, Single SU4 Truck 35 mph, H=11, I=8.5, DSM=4, CSM=9.5 Dynamic Load test, Single SU4 truck 35 mph H=11, I=8.5, DSM=4, CSM=9.5 me -2 0 2 4 6 8 10 12 01 02 03 04 05 06 0Time (sec.)Dynamic Strain ( me ) Ch 2-H Ch 3-I Ch 5-DSM Ch 6-CSM Figure 4.60 Dynamic Load Test, Single SU 4 Truck, at 35 mph, P1=7, P2=6, E=6, F=12 Dynamic Load test, single SU4 truck, at 35 mph P1=7, P2=6, E=6, F=12 me -2 0 2 4 6 8 10 12 14 01 02 03 04 05 06 0Time (sec.)Dynamic Strain ( me ) Ch 9-P1 Ch 10-P2 Ch 15-E Ch 16-F
175 SM=4, CSM=9.5, P1=7, P2=6, E=6, F=12 Dynamic Load test, single SU4 truck at 35 mph H=11, I=8.5, DSM=4, CSM=9.5, P1=7, P2=6, E=6, F=12 me -2 0 2 4 6 8 10 12 14 01 02 03 04 05 06 0Time (sec.)Dynamic Strain ( me ) Ch 2-H Ch 3-I Ch 5-DSM Ch 6-CSM Ch 9-P1 Ch 10-P2 Ch 15-E Ch 16-F Figure 4.61 Dynamic Load Test, Single SU 4 Truck at 35 mph, H=11, I=8.5, D All 16-channels were turned on for the fully load S edU4 truck to drive over the bridge with a speed of 40 mph. Figure 4.62 shows that sensors G and P2 did not respond to the truck while the response of others were 1 or 2 maximum J=5 with a =0, H=1, 1, SM=-1, BSM=1, P1=1, P2=0.5, C=1.5, D= 1, E=1, F=1 Dynamic Load test, single SU4 truck at 40 mph G=0, H=1, I=1.5, J=5, DS M=-2, CSM=1, ASM=-1, BSM=1, P1=1, P2=0.5, C=1.5, D= 1, E=1, F=1 me, T1=142.1, T2=18.25 Celsius -2 3 8 13 18 23 02040608010012014016Time (sec.)Dynam ic Strain ( me ) Ch 1-G Ch 2-h Ch 3-I Ch 4-J Ch 5-DSM Ch 6-CSM Ch 7-ASM Ch 8-BSM Ch 9-P1 Ch 10-P2 Ch 11-T1 Ch 12-T2 Ch 13-C Ch 14-D Ch 15-E Ch 16-F 0 d Tst, Single SU4 Truck at 40 mph. G Figure 4.62 Dynamic Loae I=1.5, J=5, DSM=-2, CSM=A T1=142.1, T2=18.25 0C
176 4.6 Dynamic Response of the Bridge Subject to Live Traffic Load tion site near the bridge. time, the bridge repair or es and three subsequent readings recorded after 6 months, 11 mont hs and 13 months from the initial period of time indicate the effect of time and traffic loading on the bridge. Figure 4.63 Dynamic Strain Readi ngs in Real-Time Recorded on February 4, 2005 at 4:30 pm H=5, I=7, E=7, F=5.5 The next group of graphs r epresents vehicles traversing the bridge. The type, size and speeds of these vehicles were not known. The highest peak indicates the heaviest vehicle in this gr oup. This data was collected remotely approximately 35 miles from the bridge. In some case s, the load was observed to be predominantly SU4 trucks hauling dirt to a construc The difference between the values fo r each period can generate predicted increase in strain until it will reach the safe operating value at which management can make an intelligent decisio n with regard to the replacement. Refer to Chapter 3 for design strain values at which the bridge can safely and indefinitely operate. Figures 4. 63 through 4.68 show the strain values for four consecutive intervals. The in itial strain valu readings are presented in these graphs. The increase in strain values after each Dynamic strain readings in real-time, recorded in January 2, 2005, at 430pm, H=5, I=7, E=7, F=5.5 me -2 0 2 4 6 8 00:00.007:12.014:24.021:36.028:48.0 -1 1 3 5 7 36:00.043:12.050:24.0Time (sec.)D ynam ic Strain ( me ) Ch 2-H Ch 3-I Ch 15-E Ch 16-F
Figure 4.64 indicates an increase in stra in values for I and F sensors by as much as 6 and 4 respectively (Figure 4.64) after about six months later from the init ial readings. Dynamic strain readings in real-time, recorded in June 2005 at 1139 am, I=13, F=9.5 me-2 6 10 14 00:00.002:52.805:45.608:38.411:31.214:24.017:16.8Time (sec.)Dynamic Strain ( me ) Ch 3-I Ch 16-F 2 June 14, 2005 at 11:39 am, I=13, F=9.5 Figure 4.64 Dynamic Strain Readings in Real-Time, Recorded on Figure 4.65 depicts the incr ease in strain values fo r H and F sensors by as much as 8 and 7.5 respectively approximately si x months after from initial readings. 177
178 Figure 4.65 Dynamic Strain Readings in Real-Time, Recorded on November 28, 2005 at 2:50 am, H=13, F=13 The following Figure 4.66 shows the increas e in strain values for E and F sensors by as much as 12 and 22.5 respectively. This increase took place approximately one year after the initial readings. Sensor D is not in the traffic lane, therefore, the strain value of 11 is due to the load distribution. Dynamic strain readings in Real-Time, in November 28, 2005 at 250 pm, H=13, F=13 me -2 2 4 6 8 10 12 14 0Time (sec.)Dynamic Strain ( me ) Ch 2-H Ch 1 6-F 0 0:00.002:52.805:45.608:38.411:31.214:24.017:16.8
Dynamic strain readings in real-time, recorded in January 15, 2006 at 928 am, D=11, E=19, F=28 me -2 3 33Time (sec.)namc Strain ( 13 18 23 28i me ) Channel 14-D Channel 15-E Channel 16-F 8 00:00.007:12.014:24.021:36.028:48.036:00.043:12.0Dy 179 Figure 4.66 Dynamic Strain Readings in Real-Time, Recorded on January 15, 2006 at 9:28 am, D=11, E=19 and F=28 sensors) shown in Figure 4.67 are lo wer than those shown in Figure 4.68 by The strain values of sensors E and P2 (the pair sensors to F and P1 magnitude of 6.5 and 2 This indicates a reasonable change in the magnitude of the strain val ues in that period and thus the strains are indeed time dependent variables.
180 Dynamic Strain readings in real-time, recorded in November 30, 05 at 741 am, E=21.5, P2=11 me Figure 4.67 Dynamic Strain Readings in Real-Time, Recorded on November 30, 2005 at 7:41 am, E = 21.5 and P2= 11 -2 3 8 23 00:00.007:12.014:24.021:36.028:48.036:00.0Time (sec.)Dynamic Strain ( me ) Channel 10-P2 Channel 15-E 18 13 Dynamic Strain readings in real-time recorded in January 2, 2006 at 745 am, P1=13, F=28 me-2 3 18 23 28 33 ( me ) Channel 9-P1 Channel 16-F Figure 4.68 Dynamic Strain Readings in Real-Time, Recorded on January 2, 2006 at 7:45 am, F=28, P1=13 8Dyna 13mic Strain 00:00.007:12.014:24.021:36.028:48.036:00.043:12.050:24.057:36.004:48.0Time (sec.)
181 .7 Conclusions 4 Although, there was a serious problem with the software for communication and data collect ion, the end result was a successful experiment. The communication problem was eventually resolved with the development of version v2 software. Credi t for accurate and sensible data collection goes to the meticulous and skillful installation of the sensors, with bot h surface mount and those bonded to the reinforcing steel. Tr uck load test proved to be an effective and accurate for both load tests, static and dynamic conditions. The strain values of P1, P2, E and F sensors from Tables 4.1, 4.2 and 4.3 are a close match to the strain values of the sensor from the graphs of Figures 4.60 to 4.65. The strain values of E and F sensors show an increasing tr end with time from January 2005 to January 2006. The di fference for this period was 22.5 The reading of sensor F remained the same (28 ) fr om January 2 to January 15, 2006. Synchronized cameras and weig hing scales combined with the selected sensors will provide a complete invaluab le data for analysis of bridge behavior in various loading situations.
182 CHAPTER 5. RESULTS AND COMPARISONS 5.1 Introduction In summary, this chapter will discuss the iss ues presented in the contexts of preceding chapters. First, a brief discussion on evolution of the topic of this research is presented. Then, the other items of discussions are presented as follows: (a) Why the fiber optic sensing te chnology was preferred over the nonfiber optic measuring system. (b) Why a particular sensing system from a group of systems within the same technology was selected and how these systems will compare. (c) Site specific implementat ion of measuring system in comparison to other common procedures. (d) The signific ance of objectives of this study and how requirements of these objectives were met. (e) How the analytically predicted strain values and moments co mpared to the experimental results (Evaluation of collected data). 5.2 Evolution of the Topi c of this Dissertation The author has been involved with issues related to design, construction and inspection of existing and new bridges for more than 20 years. During early years of observations, there were questions about weight restrictions imposed on newly constructed bridges that we re designed using LFD-based format
183 presented in AASHTO Standard Specifications for highway Bridges (1944). This issue was soon resolved (RSH 1990s) (Waters Avenue bridge in Tampa, Florida, designed by Reynold Smith and Hills, a structural engineering consultant, a Jacksonville, Florida based office). SU4 truck was used as an alternative design live load truck in additi on to HS20-44. The purpose of this trail was to investigate the effect of SU4 in the design of the br idge and compare the results of analytical stress/ strain pr ediction with those resu lted from HS20-44. The incorporation of SU4, a short bas e four-axle 70,000 pounds truck as live load design proved to be effective and t he newly constructed bridges were no longer subject to weight restrictions due to SU4 trucks. Later, the design was checked for all Florida legal load trucks. The other controversial and contradi ctory issue was about structural integrity of more than ei ghty existing bridges. These bridges were typically reinforced concrete structures. The s ubstructure was consisted of prestress concrete piles and cast-in-place concrete pile caps. The superstructure was composed of prestress concrete channel beams and cast-in-place concrete deck slab (figure 2.28). these bridges were designed and constructed in 1970s. The design and construction of such a large number of bridges was possible due to the simplicity of the design (AASHTO Standard Specifications for highway Bridges, LFD method) whic h had helped to construction such a large number bridges efficient and quick in a short period. Although, these bridges were posted fo r about 1/3 of thei r allowable load carrying capacity, they were in route for and subject to variety of different trucks
in a daily basis. They were routinel y observed and inspected for signs of stress due to heavy truck traffic and overload. A long term monitoring resulted in a conclusion that these bridges were capable of carrying load at their full capacity without jeopardizing their structural int egrity. The good conditions of bridges were verified through a three phases of non-destructive test and structural investigation. The result of this in vestigation concluded that none of those bridges required weight restrictions. The process for removal of weigh restrictions was tedious and time consum ing (more than four years). A nominal effort by the author and minimal budget of about $150,000.00 resolved a monumental issue of budgeting and spendi ng several millions of dollars to replace more than eighty bridges. T he resolution was approved by Hillsborough County (Board of County Commissioners). All weight restriction signs were removed and the bridges were open to all types of trucks includ ing Florida legal loads. These bridges were eventually classified as functionally obsolete and were placed in a long term program for widening or replacement. The point of this discussion is, if the present technology was present at that time, the bridges could have been fitted with smart sensors, they could have be continuously monitored for any sign of stress at the critical locations and managed confidently and effectively. S ensors transmit stresses in quantitative values as micro strain ( ) and these values can be compare with pattern of strain history that has been kept in record. Any abnormal condition can be detected instantly and then a closer obser vation and investigation can follow. This is how the concept for topic of this study was conceived. New bridges can 184
185 be instrumented with a fraction of to tal replacement budget. The bridge engineers and management can sit back, relax and just watch the monitor for any sign of over stress. Figures 4.61 thru 4.65 clearly demonstrate bridge behavior under real live load quantitativ e measure of strain values. 5.3 Selection of Sensing System for Bridge Instrumentation Smart sensing system is fairly a new technology and there is a limited literature available practica lly with regard to the topic of this research. The objectives of the research were to in vestigate addressed through a literature review process. Literatu re review provided sufficient information resulting in selection of sensing system suitable for Ea st Bay Road Bridge instrumentation. Literature review compared the non-fiber optic measuring systems fiber optic sensing technologies. These sensing syst ems are insensitive to electromagnetic interference, they are very small and light, they are ideally suited to be embedded in composite material, they do not affect the mechanical properties of host material, they are insensitive to co rrosive environment thus will not corrode and they are capable of withstanding hi gh temperatures. A laboratory experiment was set up to examine FabryPerot strain gauge and data acquisition system (Chapter 2 Experimental). Two other commonly used sensing systems were researched and evaluated for the final application on the bridge.
186 5.4 Comparison Between the Most Commonly Used Sensors Three different types of fiber opt ic sensors have been commonly used in civil engineering infrastructure. Each type has specific application, advantage and disadvantage with respect to others. The following sections 5.4.1 and 5.4.2 present a brief description of each type. 5.4.1 Fiber Brag grating and Long Gauge Fiber Optic Sensors Fiber Brag grating and long gage fiber Optic Sensors are considered Distributed Optical Sensing system that c an possibly measure (sense) at multiple points with a single optical fiber. Severa l sensors can be attached to the same cable for up to 14 miles long. This process is called multiplexing. However, these systems have multiple disadvantages for being implemented in this study. These dis advantages are listed as: (1) complex techniques that often have to be used for signal processing, (2) require highly stable and expensive laser light source (3) precision depends on wavelength stability and Bragg grating is olation capacities, (4) a ffected by vibration and temperature effects, (5) unique fiber optic cable is brittle and must be handled with caution. Use of these systems are associated with potential risk of loosing all the installed sensors if the cables are damaged or broken, (6) difficult to monitor very small wavelength changes (F BG), Considerable time averaging is often required to assess and map the s patial changes in loss or scattering coefficients along a fiber (Back scatte ring) and finally detection units (Data
187 loggers) are often incompat ible with FBG and unreliabl e because they are built by different manufac turer. 5.4.2 Fabry-Perot Fiber Optic Sensor Except two minor disadvantage (1) can run only two to three miles and (2) can not measure multiple poi nts with one optical fiber (multiplexing), this White Light Interferometry FISO Point M easurement system has many advantages which made it completely suitable for this study. The advantages of this system in clude but not limited to: (1) High sensitivity for multiple quantities such as temperature, strain, pressure, displacement and reflective index with t he same signal conditioner, (2) a simple White Light Interferometry technique for tr eatment of the signal, and that is an optical instrument that allo ws two beams of light derived from a single source (and thus of the same frequency and in pha se at identical distances from the source) to traverse paths whose difference in length determines the nature of the interference pattern obtained when the beams are allowed to interfere. The wavelength of light can be measured if t he path length difference is known, and vice versa, (3) Inexpensive and calibrati on free signal conditioner (data logger), (4) Thirty two channel signal conditioner has the capability of adding and/or replacing one sensor at the ti me for installation and maintenance, (5) it is tolerant to light loss, (6) data can be collected from each individual channel with a portable hand held signal conditioner FTI-10, (7) unlike Distributed Optical Sensing systems, the sensor s and fiber optic cables can be repaired on the site
188 of the structure, (8) it has a full dia gnostic function available at all time and capable of up to 200,000 Hz sa mpling rate, (9) it has a high precision suitable for medical applications. Because of all the above reasons, Fabry-Perot fiber optic sensors were used in this study. 5.3 Site Specific Instrumentation Sensors are normally surface mount or embedded. The surface mount sensors are bonded to the host structure wi th adhesive. On the other hand, the embedded sensors are either we lded to structural steel or bonded to the smooth and properly prepared surface of reinforci ng steel. A specific case required installation of sensors on topside of conc rete slab. This unusual condintion not been practiced before and no instructions were available. The step-by-step process is described in Chapter 2 and illustrated by numerouse photographs. 5.4 The Significance of Ob jectives of this Study A case study for the application of Fi ber Optic Sensors (FOS) for remote health monitoring of bridge structures is presented. A total of sixteen Fabry-Perot FOS sensors were installed on the East Bay Road Bridge, in Hillsborough County, Florida. The bridge is a 4-span continuous reinforced concrete deck-type structure. The bridge is considered the firs t smart structure in t he state of Florida. The Fabry-Perot sensors were both bonded to the longitudina l reinforcing bars and surface-mounted to the concrete deck. Detailed step-by-s tep description of the installation process is presented in C hapter 2. Static and dynamic tests of
189 the bridge under SU4 trucks were conduc ted. A finite element model was developed, and its output was compared to the experimental data obtained from the truck load tests. The results confirmed the accuracy of Fabry-Perot sensors in evaluating the bridge behavior under traffic loads. A remote communication system was established through phone lines in order to connect the acquisition system to the Internet. This technique enables live traffic monitoring from a central station located in the countys maintenance office. Live traffic data are currently being collected and stored on Compact Disk to generate a long term strain history for the bridge. This data will be used to facilitate the bridge maintenance process, receive early wa rnings regarding possible structural deficiencies, and assist in decision-ma king processes regarding functionality of the bridges. The proposed remote health monitoring technique with FOS sensors proved to be practical, cost-effective, and efficient providing its installation is performed in a very careful, accura te and skillful manner. Data analysis and evaluation confirmed current LRFD spec ifications for deck-type bridges are highly conservative.
190 5.5 Discussions Current AASHTO LRFD design width used in analysis produced strain higher than those of fi eld measured corresponds to a conservative design approach. To have more accurate design, the distribution width can be increased. Analysis with distribution wid th is almost twice more than the one used from the code (11.87). Strain resu lted from incorporating twice the code distribution width (20) seems to have decreased and match better with the collected data. This assumption is based on observed uncracked condition. Further continuous monitoring might indicate an increase in collected strain possibly up to the crack. At this point, after several years of monitoring, the refinement of distributi on width might be possible. Continuous monitoring of the bridge subject to traffic is essential to collect data for condition evaluation and damage as sessment. This data can also be used to predict the useful life of the bri dge. Theoretical life time expectancy of East Bay Road Bridge is 75 years. Aut hor has reviewed design and construction documents for East Bay Road Bridge and has not come across any technical information or references verifying the 75 years predicted life expectancy for this structure. The results of SU4 truck tests along wit h the output of the finite element model, as well as the data collected from remote monitoring suggest that the bridge deck did not experience cracking under traffic loads, or experienced only secondary widely spaced cracks not visible to naked eye. Also, close physical inspection and Visual observations of bridge deck underside confirmed this
191 finding. To evaluate the performance of the bridge under service loads, the moment-curvature relationship of 1ft st rip of the bridge section was developed using inelastic fiber beam models. T he fiber constitutive models used for confined and unconfined concrete follow ed the modified Kent and Park (1971) model, and the reinforcing steel stress-st rain behavior was assumed to be elastoplastic. From the moment-c urvature plot in Figure 3.23 it can be concluded that: (a) The bridge is over-reinforced, as the concrete crushing point (ultimate strength level) occurs before steel yielding. (b) The cracking point is higher than the tr affic level point. The bending moment corresponding to traffic level was evaluat ed from finite element analysis of the bridge under SU4 trucks. These val ues also match with the data recorded through remote monitoring. (c) The ultimate strength of the bridge highly exceeds the ultimate moment demand assumed in the LRFD design process. The preceding observations, along wit h the data collected through remote monitoring (Tables 4.1 thru 4.3 and Figur es 4.60 thru 4.65) suggest that the current design specifications for deck-ty pe bridges are highly conservative under service loading. Further studies and dat a collection are needed to confirm this conclusion. In addition, research and data analysis need to be performed at the ultimate stage.
5.6 Evaluation of Collected Data In this section, an evaluation of the design specifications for the East Bay Road Bridge is performed using the collect ed FOS data. It should be emphasized that the current data is c hanging on a daily basis due to the gradual deterioration of the bridge condition, as confirmed by the observed behavior described in Chapter 4. An accurate evaluation of t he design specifications should be based on the maximum recorded strain values over the entire life time of the bridge. Since these strain values can not be pr edicted presently, the current study was based on the values recorded so far. Thes e values are assumed to represent the service condition of the bridge. The in crease in recorded data, which suggests gradual damage of the bridge, will be discussed in the next section. The maximum positive strain recorded since the opening of the bridge was observed on 1/30/06 for sensor F, and is equal to 28 The maximum negative strain recorded equals 18.5 and was recorded on 1/30/06 for sensor H. These values suggest that the bri dge deck did not experience cracking under traffic loads, or experienced only secondary widely spaced cracks, as the cracking strain is 320 Visual observations also c onfirmed this fact. Since the original design was based on cracked sect ion analysis, this design is assumed to be conservative. To evaluate the performance of the bridge under service loads, the moment-curvature relationship of an 11. 87ft strip of the bridge section was developed using inelastic fiber beam models (Ayoub and Filippou 2000; Ayoub 2003). The fiber constitutive models us ed for confined and unconfined concrete 192
followed the modified Kent and Park model (Kent and Park 1971), and the reinforcing steel stress-strain behavior was assumed to be elasto-plastic. The width of the strip was assumed according to AASHTO specifications for tributary widths for design trucks (AASHTO 2004) as discussed previously in Chapter 3. From the moment-curvature plot in Fi gure (5.1), it can be concluded that: (a) The bridge is over-reinforced, as the concrete crushing point (ultimate strength level) occurs before steel yielding. (b) The cracking point is higher than the actual service level point. The bending moment corresponding to actual service level was evaluated from finite element analysis of the bridge under SU4 trucks, which also matches with the data recorded through remote monitoring. These moments are also slightly lower than the value used in the service design process based on dead load plus lane and truck loads, which equals 4600 kip-in. However, since the design process assumed the concrete section to crack under service loads, the corresponding computed maximum concrete strain was 319.4 which highly exceeds the recorded values. (c) The ultimate strength of the bridge highly exceeds the ultimate moment assumed in the design proce ss, which equals 7200 kip-in. The preceding observations, along with the data collected through remote monitoring suggest that the current design specifications for deck-type bridges are highly conservative under service load ing. Additional data are currently being collected in order to confirm this conclusion. 193
0 5000 10000 15000 20000 25000 30000 35000 40000 45000 00.000050.00010.000150.00020.000250.00030.000350.00040.00045Curvature (1/in)Moment (k-in)Ultimate Strength Traffic Level Cracking Point Figure 5.1 Moment-Curvature Rela tionship for Bridge Section To further elaborate on the issue of re corded strains before and after cracking, the following discussion is presented. 5.7 Flexural Cracking in Bridge Concrete Deck Consider a concrete element reinforced with steel bars. In this case tensile stresses are transmitted across the cra ck through the bonded reinforcing steel. The following example serves as illustration of this behavior: A reinforced concrete member is subjected to axial tension in Figure (5.2a). If the axial stress does not exceed the tensile strength of concrete, the member is ideally free of cracks. This state is referred to as stat e 1. The steel and concrete strains, s 1 and c 1 respectively, are compatible along the member. The average strain is: 194
1 11AE N nAAE Ncscc cs (5.1) where N is the value of the axial force, A c and A s are the cross sectional areas of concrete and steel and nEE s c with E s and E c being the moduli of elasticity of steel and concrete. When the concrete stress exceeds the tensile strength, cracks appear. At a crack the stress is completely carried by the reinforcement and the concrete stress is zero. This condi tion will be referred to as state 2. The steel stress and strain are given by: s s N A2 (5.2) s s s N EA2 (5.3) In the portion between two cracks part of th e tensile stress carried by the steel at the crack is transferred to the concrete through bond. The stress and strain are in an intermediate state between states 1 and 2, as depicted in Figure (5.2). Midway between consecutive cracks, the se ction is in state 1 and the steel stress is less than s 2 At a crack the section is in st ate 2 with the steel stress at its maximum value s 2 and with the concrete stress equal to zero. The difference in steel stress is transmitted to the conc rete through bond, so that the member elongates less than the bare steel. Denoting the average strain of the cracked member in Figure (5.2a) as m then mLL (5.4) where L is the original length of the member and L is the member elongation. 195
196 Before cracking, compatibility of strain s is maintained so that Eq. 5.1 holds 11 csm After cracking, the value of m lies for a given stress level between the steel strain in the perfectly bonded case s 1 and the steel strain at the crack s 2 L S a) Cracking of a Tie b) Stress in Reinforcement c) Bond Stress d) Stress in Concrete s 2s 1c 1 Figure 5.2 Stress at Cracking
Denote the reduction in steel strain due to the participation of concrete between cracks by then m s 2 (5.5) Based on experimental evidence, it is assumed that varies inversely with the applied axial load N (CEB 1985): max N N r (5.6) where N r is the cracking load and max is the steel strain difference between states 1 and 2 at the first crack. From the graph in Figure (5.3): N Nr ss 12 max (5.7) Substitution of Eq. 5.6 in Eq. 5.7 give s the average strain value of the member: 2 11ss m (5.8) where is a dimensionless parameter that represents the amount of cracking and is given by: 21 N Nr (5.9) 0 for an uncracked member. The difference between the solid line and the line representing the bare steel in Figure 5.3 is referred to as tension stiffening. It represents the increase in stiffness due to the conc rete contribution between cracks. Tension stiffening can be si gnificant up to the yielding of the 197
reinforcement, but drops considerably near the yield point. After yielding of the reinforcement at the most critical se ction, the member elongates without significant increase in load and the tension carried by the concrete becomes negligible. Average Strain Axial force (state 2)mmaxssN EAs 2 y cN EAs 1 1(state 1)NNcr Figure 5.3 Axial Force vs. Average Strain for an Axially Loaded Reinforced Specimen The installed FOS sensors are either embedded and bonded to the rebars or surface-mounted to the concrete. T he surface-mounted sensors would record the strain 1 c 1 s before cracking. After cracking, if the sensor is exactly located at the crack position, its r eading will drop to zero. It is more likely, however, that the sensor exists between two cracks. In this case the reading of the sensor will drop but to a non-zero value. Gradual decrease of the sensor readings indicate 198
the formation of additional cracks until the deck becomes severely cracked. In this case, the readings will approach zero values. The role of the surfacemounted sensors therefore is to detect the formation of the initial cracks and to monitor the crack propagation with time. After the deck becomes severely cracked, these sensors will not be able to record service strain values. 1991 The role of the embedded sensors on the ot her hand, is to monitor the service strain values in addition to recording t he maximum steel stress values at cracked locations. Before cracking the sensors would record the strain 1 s c When the strain 1 c reaches the concrete cracking strains, this will indicate the formation of the first crack. The steel strain at t he crack location will increase and reach the value of 2 s but the steel strain bet ween cracks will be less than 2 s If the sensor is exactly located at the crack position, it will record the value of 2 s This is however unlikely to happen, and it is a ssumed that the sensor is recording an average value that equals m as defined in Eq. (5.5). In order to extrapolate the value of the steel strain at the crack lo cation, Eq. (5.8) is used to estimate the value of the axial force N resisted by the reinforced concrete section, which is again used with the help of Eq. (5.3) to ev aluate the steel strain at the crack location 2 s A further increase in the value of either m or 2 s under the same loading conditions indicate the formation of additional cracks or a decrease in the value of the crack spacing S identified in Figure (5.2 ). Currently, the recorded FOS strain values indicate no cracking or the existence of minor and widely space cracked. As the bridge deteriorates with time, the sensors readings should increase, and the process described above for both surface-mounted and
embedded sensors will be impl emented to detect the fo rmation and propagation of cracks, as well as the maximum st eel stresses at the crack locations. 5.8 Evaluation of Design Specifications As stated earlier, the maximum reco rded positive strain value was 28 while the maximum negative strain was 13 The corresponding design values are 319 and 218 for positive and negative case s respectively for cracked conditions, and 67 and 48 for uncracked conditions. These values indicate that the design process was highly conservative for the assumed cracked conditions. Even if the section is assumed to be uncracked, the design values exceed the maximum recorded values. The discrepancy between the design and recorded strain values c ould be attributed to the following parameters: (a) The assumed distribution width in the de sign calculations, (b) the inclusion of the barrier wall in the anal ysis, and (c) the overestimation of the actual truck loads acting on the bridge. Ea ch of these items is described in more details herein. (a) Distribution width : The distribution width assumed in the analysis equals 11.87 ft. To evaluate the accuracy of this distribution width, finite element analysis of the bridge deck using shell el ements and under the static load of an SU4 truck is performed and compared to analysis with frame elements and also to experimental results. These analyses we re described in details in Chapter 3 (Figures 3.19, 3.25 and 3.26). The shell fini te element results indicate that the strains within the 11.87 ft strip around t he wheel load are wit hin the range of 70200
100% of the peak strain. The width of the strip around the wheel load with nonzero stresses actually equals 18 ft. This conclusion is also valid from the experimental plots of Figure (3.19). The frame analysis of the bridge deck with an equivalent width of 11.87 ft pr oduced maximum strains of 61 as described at the end of Chapter 3. This value is clearly overestimated as the maximum recorded value was 28 If the frame analysis was repeated with a section width of 11.87 x 61/28 = 25. 85 ft, the maximum resulting strain would equal 40 which matches with the recorded data. In conclusion, it appears that the distribution width of 11.87 ft provided by the code is highly conservative assuming the current uncracked condition of the bridge. A value of 18 ft. seems to better match with recorded data. This conclusion, however, is expected to change as the bridge starts cracking and det eriorates with time. The author will continue to monitor the behavior of the bridge and re-evaluate the distribution width that matches with crack ed conditions. It is the authors belief, however, that (b) Barrier Wall: Traffic barrier walls in solid slab bridges act as upward vertical beams which enhance the moment capacity of the bridge significantly (Shahawy et al., 1999). The effect of traffic barrier walls and br idge sidewalk parapets were observed in East Bay Road Bridge. This effect was sensed by the gauges located near the walls. Sens ors C, D, G, J, ASM and BSM had strain values of 1.5 to 2.5 while at the same time under t he same loading condition, sensors E, F, H, I, CSM and DSM located under t he wheel load had strain values ranging from 19 to 28 Neither AASHTO standard spec ifications (LFD) nor AASHTO LRFD code have considered the effect of barrier walls in design of bridge slab. 201
202 Overall low strain readings of sensors ar e due to slab stiffness attributed by the barrier walls. To further investigate this conclusi on, the finite element analysis of the bridge deck was repeated with the inclusion of the barrier walls. The barrier walls were modeled as additional shell elements acting at the edges of the deck. The strain contour plots for this case ar e shown in Figure 5.4 and are compared to the ones described earlier in Chapter 3 and shown again in Figure 5.5, where the barrier walls were not simulated. From t he figures two conclusions are drawn: (a) The strains near the edge beams were minima l for the case of the model with the barrier walls confirming the observed recorded behavior, and (b) the maximum strains under the wheel loads dropped from a value of 15.63 to 13.2 which accounts for a 16% decrease. From the discussion above and from the recorded strain values, it is the authors belief that there exists a major need to include the effect of barrier walls in the design and analysis of bridge structures.
Figure 5.4 Strain Contours with Inclusion of Barrier Walls Figure 5.5 Strain Contours without Inclusion of Barrier Walls 203
(c) Actual Load: East Bay Road Bridge was designed based on AASHTO LRFD Code with the governing design live load LH93. LH-93 is a not ional non existing truck that has been configured to produce ma ximum critical live load condition. Without application of specia lized equipment such as sc ale and cameras, there is a little information to verify the actual trucks weight and type traversing over the bridge. However, abundant of SU4, C4 and C5 trucks moving over the bridge is evident by frequent field observation. The strain values recorded through remote monitoring is also in conformance with the strains obtained from bridge load test subject to fully loaded SU4 trucks. The small strain values (28 and 19 for positive and negative moments respective ly) sensed by FOS will only confirm the conservative state of LRFD design, cons ervative live load distribution width and effect of traffic barrier walls rather than absence of actual load in motion over the bridge. (d) Bridge Rating: The bridge analysis under Fl orida legal trucks was performed and presented in Section 3.6. As stated earlier, the current practice for this analysis is based on LFD procedures, while the original design is performed in accordance with LRFD procedures. This in compatibility between the design and rating procedures has caused confusi ons between design engineers and has led the Florida Department of Transportation lately to suggest that the rating be performed in accordance with LRFD proce dures. For instance, in several cases bridges designed in accordance with LRFD procedures did not pass the rating test before the opening of the bridge, and the br idge therefore needed to be posted. Current design and analysis tools, however, are still tailored to match 204
with LFD procedures and there exists a need to modify these tools and so they match with the LRFD approach. For the East Bay bridge, the maximum posit ive strains for the Florida legal trucks were presented in Table 3.4 for cracked conditions. The maximum positive strain was that of the SU4 tr uck and is equal to 320 while the maximum negative strain was that of the C4 truck and is equal to 249 While the maximum positive strain is close to the ma ximum positive design strain of 319.4 under the HL-93 truck the maximum negative strain of 249 exceeds the maximum negative design value of 217 This conclusion implies that the C4 truck was more critical for the East Bay bridge t han the design HL-93 tru ck. Considering the fact that the C4 truck is a real truck that is likely to be moving over the bridge, the strain obtained from this truck confirms the conclusion that the assumption of cracked behavior used in the design is conservative. The maximum positive and negative stra ins for uncracked conditions are presented in Table 3.6. The maximum positive values are 67 for both the design and the critical SU4 truck. The maximum negative values are 45 and 52 for the design HL-93 and the C4 trucks respectively. These values confirm the earlier conclusion that the C4 truck is more critic al than the design truck even for uncracked conditions. Comparing the strains due to the legal trucks to the maximum recorded values implies that the current design guidelines are still conservative even assuming uncracked cond itions. The reasons were discussed earlier and are related to the distributi on width, the presence of the edge beam, and the estimation of the real load acting on the bridge. 205
5.9 Damage Identification of the East Bay Road Bridge The installed health monitoring system will be also used as a tool to detect long term damage of the East Bay Road Bridge. The process is described as follow: (a) The readings of all FOS sensors will be collected and stored. The maximum positive and negative strains among all sensors will be identified. (b) A damage index for service conditions (DI) service that represent s the damage condition of the bridge will be evaluated. The damage index is defined as follow: capacity serviceDI max Where max is the maximum recorded strain at time t and capacity represents the maximum strain that an element can re sist. Theoretically, the maximum service capacity should equal the st eel yield strain of 1897 however the current maintenance practice requi res using a value of 0.85 y = 1518 The value of (DI) service is assumed to equal zero at the init ial stage of the bridge. A schematic diagram of the expected shape of the time vs (DI) service plot is shown in Figure (5.6). The values for (DI) service have been computed for the bridge condition so far and are shown in red in Figure (5.4). The dotted line in Figure (5.4) shows the expected behavior over the life time of the bridge. The maximum value obtained at the end of the first y ear, however, is less than 0.03 due to the uncracked condition of the bridge. The author will continue to monitor the behavior of the bridge in collaboration with Hillsborough County, and construct the damage 206
function over the entire life cycle of the bridge. The data corresponding to the behavior of the first year are shown in Table (5.1) below. Table 5.1 Strain Progression with Respect to Time, Sensors readings, Date H I E F 1-2-05 5 7 7 5.5 6-1-05 8 13 12 13 9-2-05 13 15 17 18 1-2-06 17 18 21 23 1-30-06 21 20 25.5 28 (c) A frame finite element model of the bridge will be de veloped and subjected to the AASHTO design truck. The stiffness coefficient ( ES) of the model will be tuned in order for the model to matc h with the maximum recorded strains, E being Youngs modulus and S the section modulus Two values for the stiffness term (ES) will be evaluated, namely a value fo r matching with maximum positive strains ( ES) + and a value for negative strains ( ES) The critical of the two will be used for the calculations to follo w. The initial stiffness value which corresponds to the initial cracking condi tion will be documented. The stiffness coefficient ( at time t will be evaluated and compared to the initial in order to monitor the strength det erioration rate of the bridge. ()oES )tES ()oES 207
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0102030405060708090100 Time (years)DIservice Figure 5.6 Damage Index of the East Bay Road Bridge 208
209 CHAPTER 6. SUMMARY, CONCLUSIONS, AND FUTURE WORK 6.1 Research Planning 6.1.1 Laboratory Test a nd Field Investigation The laboratory test was performed wit h project economy in mind. The literature review resulted in the evaluatio n of three sensors. (1) Fabry-Perot strain gauge, (2) Fiber Bragg grating opt ic strain gages and (3) long-gauge strain gauge. Limited application and lack of suit able data acquisition make Fiber Bragg grating optic strain gauges and long-gauge st rain gauge sensing systems a poor choice for the bridge load test instrumentation at this time. These systems use two fibers, one as a transducer and one as a reference fiber. These systems are capabl e of multiplexing (e.g., sensors are used in series and only one fiber optic cabl e leads to data logger). Using one fiber optic cable presents a serious and potential risk of loosing all installed sensors. On the other hand, Fabry-Perot Inte rferometer was found to be a suitable sensing system for this research. Th is system offered ease and simplicity in installation and operation. Fabry-Perot uses interferometry technique, a unique way of utilizing the light emitted from a white light source.
210 6.2 Laboratory Application of Sensors The laboratory application of FabryPerot sensors was economical and easy. Installation of sensors on te st specimen was easy, quick and clean. Connection to readout unit was simple and successful. The results of load test were closer to analytical values than to digital strain gauge. The versatility, ease of application and data collection and accu racy of collected data with this new smart sensing technology has rendered older conventional instrumentation obsolete. 6.3 Field Application of Sensors The field application of sensors for East bay Road bridge proved to be as simple as the laboratory application. The successful bonding of the sensors was verified by a potable readout unit and the field test results were verified by analytical models as presented in Chapter 3. 6.4 Conclusions and Recommendations 6.4.1 Fine Tuning of LRFD Code The literature review indicates the absence or the lack of much needed research study and field verification of AASHTO LRFD design specifications for concrete bridges. Results of field data, beam modeling and FEM have indicated over conservative design for the East Bay Road bridge. The results of field experiments have indicated that the LRFD design method has been expanded and diverted much beyond its intended purpose and technicality.
211 6.4.2 Cracked Section vs. Uncracked Section The sensors readings as well as the visual observations suggest that the current condition of the East Bay Road bridge corresponds to uncracked behavior. The design of the bridge, howev er, is based on cracked analysis which resulted in over conservative cross sect ions. This conclusion, however, is based on the current observed behavior and ma y change as the bridge deteriorates with time. 6.4.3 Load Distribution Widt h (Code Tributary Width) The recorded data along with finite element modeling confirmed that current specifications for distribution wid ths are conservative. This conclusion, however is based on the observed uncrack ed condition of the bridge, and might change as the bridge deck starts cracking. There exists a need to develop more accurate criteria for distribution wid ths that better matches with observed behavior. 6.4.4 Discuss the Effect of Parapets and Traffic Barriers on Bridge Deck Stiffness Analytical investigations as well as sensors readings showed that the traffic barrier wall has a considerable effe ct on the stiffness and load distribution of the East Bay Road Bridge. The current bridge design has ignored this effect, which contributed to the conservative behavior of the bridge. There is a need to
212 revisit the current design guidelines to acc ount for the increase in stiffness due to the presence of the barrier wall. 6.5 Future Studies 6.5.1 Continuous Monitoring of the East Bay Road Bridge The author will continue to monito r the behavior of the East Bay Road bridge in collaboration with Hillsborough County officials. The conclusions drawn on current data will be re visited based on the new data as they become available. It is expected that the bridge will eventually st art cracking, which will be reflected in an increase of the sensors readings. 6.5.2 Damage Identification of the East Bay Road Bridge The author proposed earlier a methodology in Chapter 5 to evaluate the structural health and damage condition of the East Bay Road Bridge through a damage index function. The author will conti nue to collect data and construct the entire damage function of the bridge. This damage function will serve as basis to accurately evaluate the real life time expectancy of the bridge. This study will help better understand the performance of si milar bridge structures, and improve their maintenance process accordingly. 6.5.3 Weight-In-Motion (WIM) Systems Future studies will aim at accurate ly evaluating the weight of trucks moving on the bridge in addition to the resulting strain readings. This will be
213 made possible through the use of Weight-in-Motion (WIM) systems. Weight-inmotion systems are reliable tools used across the nation to obtain the following information: axle weight of trucks and cars, axle spacing, and speed. The collected truck information will help better evaluate the collected sensors readings, and therefore better understand t he bridge behavior under traffic loads. 6.5.4 Wireless Sensors The technology for health monitoring of bridge structures is moving with a fast pace. While the sensors used for th is project performed adequately, wireless technology offers additional features. In this case, sensors communicate wirelessly, which will eliminate the need for on-site cabling. Installation of such sensors might be more complex though, as they still need to be attached to an electric card, which will require additional care and innovation during construction. Furthermore, most of t hese sensors are battery operated, which renders long-term use impractical. Curr ent research is undergoing, however, to solve this issue, using several innovative techniques. Future research should focus on the use of such advanced sensors and their applicability for bridge monitoring. 6.5.5 Estimate of Bridge Life Expectancy Continuous monitoring of t he bridge subject to traffic is essential to collect data for the condition, eval uation and damage assessment. This data can also be used to predict the useful life of the bridge. Theoretical life time expectancy of
214 the East Bay Road Bridge is 75 years. The author has reviewed design and construction documents for East Bay Road bridge and has not come across any information or references verifying 75 years predicted life expectancy for this structure. Continuous data collecti on, if formulated properly, will provide invaluable tool for societal and ec onomical management of civil engineering infrastructure and will predict its normal tr ue life time expectancy. The suggested formulation consists of the following variabl es: (1) initial material properties and strength at the time of construction, (2) collected data from nondestructive material testing and strength in every five years period, (3) FOS strain readings at the same time line. The differenc e between the values for each period can generate predicted increase in strain until it will reach the safe operating value at which time, the management can make an in telligent decision about the bridge. 6.5.6 Development of New Bridge Management Systems Using Remote Health Monitoring Techniques It is the authors hope that the current study becomes a starting point into development and implementation of new bridge maintenance systems that follows the present technological era. In this case, the new maintenance structure will rely on a centralized bridge management office where data gathering and data evaluation is perform ed. The current system for bridge maintenance requires engineers to make periodical checks to assess bridge damages. With the implementation of the Fi ber Optic sensors, the ultimate goal would be to decrease the frequency of ins pecting for bridge damages. The main
215 objective of the new system is to determine who is in charge of gathering data, analyzing data and taking the proper actions recommended by the data analysis. It is critical that the new system works efficiently to ensure public safety. It is imperative that the channel of communication and the management structure be in line with the new system so that data does not get overlooked or lost. The author hopes that Hillsborough County be the first to employ such an advanced system, and to work closely with their bridge management team to evaluate current procedures, propose new procedures and resolve any issues that might arise due to the implementation of the new technology. Such new methodologies will improve the safety of these bridges, improv e the emergency response following possible failures, and minimize the impact of traffic delay due to possible bridge closure, resulting in millions of dollars of savings to the County.
216 REFERENCES Ayoub, A. S., and Filippou, F. C., Mixed formulation of non-linear steel-concrete composite beam element, Journal of Structural Engineering ASCE, 2000, pp 371-381. A. A. Abdelrahman, G. Tadros, and S. h. Rizkalla, Test model for the First Canadian Smart Highway Bridge , ACI Structural Journal V. 92, N0. 4, JulyAugust 1995. AASHO, Standard Specifications for Highw ay Bridges, American Association of State Highway Officials Washington, D.C., 1973. AASHTO Bridge Design Specifications, third edition, 2004. AASHTO, LRFD Code for t he Design of Highway Brid ges, American Association of State Highway and Transportation Officials Washington, D.C., 1994. AASHTO, Standard Specifications for Highw ay Bridges, American Association of State Highway and Transportation Of ficials Washington, D.C., 1989. AASHTO, Standard Specifications for Highw ay Bridges, American Association of State Highway and Transportation Of ficials Washington, D.C., 1977. Aktan A.E., Grimmelsman, K.A., Barrish, R.A., Catbas F.N., and Tsikos, C.J., 2000, Structural Identification of a Long Span Truss Bridge, TRR No. 1696, National Academy Press, Washington D.C., pp 210-218. Aktan, A.E., et al. (1996), Condition Assessment for Bridge Management, Journal of Infrastructure System, ASCE, 2(3), p108-117. Amer, A., Arockiasamy, M., Shahawy M. Load di stribution of exix ting sokid slab bridges based on field tests, Journal of Bridge Engineering v 4 n 3 1999. p 189193. Ayoub, A. S., Mixed formulation of non-linear beam on foundation elements, Computers and Structures, 81(7), 2003, pp 411-421.
217 Ayoub, A. S., and Filippou, F. C., Mixed formulation of non-linear steel-concrete composite beam element, Journal of Structural Engineering, ASCE, 2000, pp 371-381. Benmokrane, B., EI -Salakawy, E.F., E lRagaby, A., Lackey, T., and Desgagn, G. 2003, Design and Testi ng of Concrete Bridges Reinforced with Glass FRP Bars, ASCE J. of Bridge Engineering Submitted for Publication, 35 p. Brahim Benmokrane, Habi b Rahman, Phalguni Mu khopadhyaya, Radhouane Masmoydi, Mohammed chekired, Jean-Franc ois Nicole, and Adel El-Safty, Use of fiber reinforced polymer reinforcement integrated with fiber optic sensors for concrete bridge deck slab construction, Can. J. Civ. Eng. 27:928-940 (2000). Catbas, F.N., Lenett, M., Ak tan, A.E., Brown, D.L., Helmicki, A.J., Hunt, V. (1998), Damage Detection and Conditi on Assessment of Seymour Bridge, Proceedings of the 16 th International Modal Analysis Conference, Santa Barbara, California. CEB. (1985), "Cracking and Deformations." Bulletin d'information No. 158, Comit Euro-International du Bton, Paris, France. elebi, M., Sanli, A. GPS in pionee ring Dynamic Monitoring of Long-Period Srtucture. Earcthquake Spectra, V. 18, No. 1, 2002, pp. 47-61. Chase, S. B., and Washer, G. (1997), Nondestructive Evaluation for Bridge Management in the next Century, Public Roads, July/Aug, p16-25. Choquet, P., Juneau, F., and Bessette, J., 2000, New Generation of Fabry-Perot Fiber Optic Sensors for Monitoring of Structures, Proceedings of SPIE 7 th Annual International Symposium on Smart Structures and Materials Newport Beach, CA. D. Inaudi, S. Vurpillot, Long-Gage Struct ural Monitoring for Civil Structures, Paciffic Nordwestern Conference on fiber optic, 1988. D. Inaudi, S. Vurpillot, N. Casanova, P. Kronenberg, Structural monitoring by curvature analysis using interferometric fiber optic sensor, to be published in Smart Materials and St ructural Journal. Haque, Mohammed E. (1997), Uniform El ement Identification System for database management for roadway br idges Nov. ASCE, p 183-188. Hearn, George; Shim, Hyung-Seop ( 1997), integration of nondestructive evaluation methods and bridge managem ent systems, ASCE, p46-473.
218 Idriss, R.L., Kersey, A.D., Davis, M. 1997, Highway Bridge Monitoring Using Optical Fiber Sensors, Geotechnical News, Vol. 15, no. 1, pp. 43-45. ISIS Design Manual 1, 2001, Installa tion, Use and Repair of Fiber Optics Sensors, University of Manitoba, Manitoba, Canada. Kent, D.C., and Park, R., "Flexural Members with Confined Concrete." J. Struct Engrg., ASCE, Vol. 97, No. ST7, 1971, pp. 1969-1990. Lee, A., Robertson, I., I nstrumentation and Long-Term Monitoring of the North Halawa Valley Viaduct, Research Repor t UHM/CE/95-08, University of Hawaii, Honolulu, HI, September 1995, pp 149. MathCAD 11, Mathsoft Engineering & Education, Inc., 1986 2002. Modjesky and Masters Inc, Mechanicsburg, Pennsylvania, Kulicki, Scott R. Eshenaur and Andrew L. T homas, Shear Behavior of Full Scale Prestressed Concrete Girders: Comparison Betw een AASHTO Specification and LRFD Code, PCI Journal, May-June 1996, PCI Journal, May-June 1997. Mohsen A. Shahawy and Barrington deV Ba tchelor, Shear Behavior of Full Scale Prestressed Concrete Gird ers: Comparison Between AASHTO Specification and LRFD Code, PCI Journal, May-June 1996. SAP2000, Analysis Reference Manual, Co mputers and Structures, Inc., 2004. Shahawy, Mohsen A. and Batchler Barrington deV, (1996), Shear Behavior of Full-Scale Prestressed Concrete Gi rders: Comparison Between AASHTO Specifications and LRFD Code PCI Journal v 41 n 3. p 48-62. Sikorsky, C., Stubbs, N., and Karbhari, V ., 2003, Automation of Damage Index Method to Evaluate Structural Safety, Proceedings of the Thirteenth International Offshore and Pola r Engineering Conference, Honolulu, Hawaii. Straser, E.K., and Kiremidjian, A. S., 1998, A Modular, Wireless Damage Monitoring System for Structures. Report No. 128, The John A. Blume Earthquake Engineering Cent er, Stanford University. Van Daveer, J.R. (1975), Techniques for Evaluating Reinforced Concrete Bridge Decks, ACI Journal, December, p 697-704. Wang and Tang, et al, Using Fiber Bragg Gr ating Sensors to Monitor pavement Structures, TRB 2005 Annual Meeting.
ABOUT THE AUTHOR Ebrahim Mehrani was born in Abadan, Iran. He received his High school diploma in math major in 1962 from a public school in Abadan and attended school of Architect in Tehran. He wa s later admitted to Abadan College of petroleum engineering. He worked in the design and construction industry for the next few years. In Spring of 1974, he was admitted to the University of South Florida, where he received his B.S. and M.S. degrees in Structural Engineering. He worked in the private sectors befor e accepting the position in Hillsborough County as The County Structural Engine er. His duties within this capacity included the design and construction of the new county bridges, and the inspection of the county existing bridges. He is currently the owner and president of DCI Solutions, Inc. in Tampa, Florida, a structural design-build consulting firm that specializes in design and cons truction of residential and commercial buildings, and inspection of bridges. He has over 30 years of experience in the field of structural design and constructi on of buildings and bridges. He is a registered professional engineer and a special inspector in the state of Florida, and both a state certified general contractor and a state cert ified bridge inspector. He has published M.S. thesis, two conference papers, one to appear in ASCE 2006 and one in ACI Spring Convention 2006. He has submitted two journals to ASCE for publication.