Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Experimental dynamic verification of open piled wharf models Yee, Steven W. G. 1996

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1996-0111.pdf [ 28.38MB ]
Metadata
JSON: 831-1.0050354.json
JSON-LD: 831-1.0050354-ld.json
RDF/XML (Pretty): 831-1.0050354-rdf.xml
RDF/JSON: 831-1.0050354-rdf.json
Turtle: 831-1.0050354-turtle.txt
N-Triples: 831-1.0050354-rdf-ntriples.txt
Original Record: 831-1.0050354-source.json
Full Text
831-1.0050354-fulltext.txt
Citation
831-1.0050354.ris

Full Text

E X P E R I M E N T A L D Y N A M I C V E R I F I C A T I O N O F O P E N P I L E D W H A R F M O D E L S by Steven W . G . Y e e B . A . S c , The University of British Co lumb ia , 1988 A T H E S I S S U B M I T T E D IN P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F A P P L I E D S C I E N C E in T H E F A C U L T Y O F G R A D U A T E S T U D I E S Department of Civi l Engineer ing W e accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F BRIT ISH C O L U M B I A January 1996 © Steven W . G . Y e e , 1996 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada Date DE-6 (2/88) II ABSTRACT Designers of wharf structures have historically had little guidance from published standards from which to base their seismic designs. Many of the methods used to develop models for seismic analysis are taken from building or bridge codes, but fundamental differences between the support systems in buildings/bridges and wharves may make these codes unsuitable for analysing wharf structures. Little research has been done to verify the applicability of building and bridge code recommendations for the seismic analysis of open-piled wharves. The aim of this study was to assess the relative importance of different modelling assumptions, taken from guidelines set out by building and bridge codes, to determine the inputs for the dynamic modelling of wharves. In this study, two wharves, a vertical piled wharf and a batter piled wharf, were analysed to determine their dynamic characteristics. The vertical piled wharf was monitored by means of an ambient vibration survey. The ambient vibration measurements provided data from which to calculate the low level vibration characteristics of the wharf. The dynamic characteristics determined from the data were used to calibrate a computer model that was, in turn, used to conduct a parametric study of the major variables used to model wharf structures. The variables that were studied were as follows: non-structural restraints; length to fixity of the piles; structure mass; and stiffness of the framing members. A B S T R A C T It was found that the first two parameters can make a large difference in the dynamic properties of the structure, while the influence of the last two parameters is relatively insignificant. The batter piled wharf was subjected to strong motion shaking during the 1989 Loma Prieta earthquake. Because of the limited availability of strong motion data for wharves, the data recorded during the earthquake provided a unique opportunity to determine the behaviour of the wharf under strong shaking as well as to evaluate its dynamic characteristics. Analysis showed that the individual sections of the wharf moved independently of one another during the earthquake, although these motions were highly influenced by the ground movements in the free field. Frequency and time domain analyses resulted in the determination of one transverse natural frequency at approximately 3.0 Hz for one of the independent sections of the structure. iv TABLE OF CONTENTS A B S T R A C T ii T A B L E OF C O N T E N T S iv LIST OF T A B L E S viii LIST OF FIGURES ix A C K N O W L E D G M E N T S xii DEDICATION xiii C H A P T E R 1 OVERVIEW 1 1.1 B A C K G R O U N D ON WHARF S T R U C T U R E S 1 1.2 O P E N PILE W H A R V E S 2 1.2.1 PILE C A P AND D E C K S Y S T E M 2 1.2.2 VERTICAL LOAD S U P P O R T S Y S T E M 3 1.2.3 LATERAL LOAD S U P P O R T S Y S T E M 3 1.3 MODELLING OF T H E LATERAL S U P P O R T S Y S T E M 4 1.4 SOIL AND S L O P E CONDITIONS AT WHARF FOUNDATIONS 5 1.5 P R O B L E M DEFINITION 7 1.6 OBJECTIVES AND S C O P E 7 C H A P T E R 2 THEORETICAL B A C K G R O U N D AND LITERATURE REVIEW 11 2.1 C O N C R E T E MATERIAL PROPERTIES 11 2.2 PILE TO SOIL INTERFACE MODELLING 12 2.2.1 EQUIVALENT SOIL SPRING M O D E L 13 2.2.2 EQUIVALENT B A S E SPRING M O D E L 14 2.2.3 EQUIVALENT CANTILEVER M O D E L 14 2.3 AMBIENT VIBRATION T H E O R Y 16 2.4 MODAL A S S U R A N C E CRITERION 20 2.5 S T R O N G MOTION DATA ANALYSIS 21 C H A P T E R 3 S T R U C T U R A L MODELLING - VERTICAL PILED WHARF 26 3.1 B A C K G R O U N D INFORMATION 26 3.2 SITE CONDITIONS 27 3.3 S T R U C T U R E DESCRIPTION 27 3.4 MODELLING ASSUMPTIONS 28 TABLE OF CONTENTS V 3.4.1 PILE PROPERTIES 30 3.4.2 PILE C A P PROPERTIES 31 3.4.3 D E C K PROPERTIES 31 3.4.4 MASS E L E M E N T S 32 3.4.5 DYNAMIC ANALYSIS 33 C H A P T E R 4 FIELD TESTING - VERTICAL PILED WHARF 43 4.1 T E S T OBJECTIVES 43 4.2 T E S T EQUIPMENT 44 4.2.1 S Y S T E M #1 - SE ISMOMETER B A S E D 44 4.2.2 S Y S T E M #2 - A C C E L E R O M E T E R B A S E D 46 4.2.3 EQUIPMENT SELECTION 50 4.3 FIELD TESTING 52 4.3.1 TESTING RESTRICTIONS 53 4.3.2 T E S T CONDITIONS 54 4.3.3 DATA COLLECTION S O F T W A R E / H A R D W A R E S E T U P 55 4.3.4 AMBIENT VIBRATION T E S T S 55 4.3.5 F R E E VIBRATION T E S T S 57 4.4 M E A S U R E D DATA 57 4.4.1 AMBIENT VIBRATION DATA . . . .'. i . . . 57 4.4.2 F R E E VIBRATION DATA 58 4.5 DATA ANALYSIS 58 4.5.1 C O M P U T E R P R O G R A M S 59 4.5.2 AMBIENT VIBRATION DATA ANALYSIS 60 4.5.3 F R E E VIBRATION DATA 62 4.6 ANALYSIS R E S U L T S 62 4.6.1 AMBIENT VIBRATION T E S T R E S U L T S 62 4.6.1.1 O V E R A L L SITE M O D E . 63 4.6.1.2 RIGID BODY M O D E S 63 4.6.1.3 FLEXURAL M O D E S 64 4.6.2 F R E E VIBRATION T E S T R E S U L T S 64 4.7 DISCUSSION OF R E S U L T S 64 4.8 SUMMARY 66 C H A P T E R 5 CORRELATION ANALYSIS AND PARAMETRIC STUDY - VERTICAL PILE WHARF . . . . 87 5.1 CORRELATIVE ANALYSIS 87 5.1.1 ANALYTICAL R E S U L T S VS EXPERIMENTAL R E S U L T S 87 5.1.2 CORRELATION O F EXPERIMENTAL R E S U L T S T O T H E ANALYTICAL MODEL 91 5.1.2.1 PILE FIXITY LOCATION 91 5.1.2.2 D E C K STIFFNESS 92 5.1.2.3 S T R U C T U R E MASS AND MASS DISTRIBUTION . . . 92 5.1.2.4 SLIDING BEARING RESTRAINT 93 TABLE OF CONTENTS VJ 5.2 PARAMETRIC S T U D Y 94 5.2.1 OBJECTIVES 95 5.2.2 STUDY P A R A M E T E R S 95 5.2.2.1 PILE LENGTH T O FIXITY 95 5.2.2.2 PILE AND PILE C A P ST IFFNESS 96 5.2.2.3 S T R U C T U R E MASS 97 5.2.2.4 BEARING PAD SPRING ST IFFNESS 98 5.2.3 STUDY COMPARISON BASIS 98 5.2.4 STUDY R E S U L T S 99 5.2.4.1 PILE LENGTH TO FIXITY 99 5.2.4.2 PILE AND PILE C A P ST IFFNESS 100 5.2.4.3 S T R U C T U R E MASS 101 5.2.4.4 BEARING PAD SPRING ST IFFNESS 101 5.2.5 SUMMARY 102 C H A P T E R 6 S T R U C T U R A L MODELLING - B A T T E R PILED WHARF 118 6.1 SITE CONDITIONS 118 6.2 S T R U C T U R E DESCRIPTION 119 6.3 MODELLING ASSUMPTIONS 120 6.3.1 PILE P R O P E R T I E S 122 6.3.2 PILE C A P PROPERTIES 123 6.3.3 D E C K PROPERTIES 123 6.3.4 DYNAMIC ANALYSIS 124 C H A P T E R 7 SEISMIC BEHAVIOUR - BATTER PILED WHARF 136 7.1 M E A S U R E M E N T EQUIPMENT AND EQUIPMENT L A Y O U T 137 7.2 SEISMIC DATA 138 7.2.1 LOMA PRIETA E A R T H Q U A K E - B A C K G R O U N D : 138 7.2.2 ORIGIN OF DATA 139 7.2.3 P R O C E S S E D S T R O N G MOTION DATA 140 7.2.3.1 C O R R E C T E D A C C E L E R A T I O N R E C O R D S 141 7.2.3.2 C O R R E C T E D VELOCITY R E C O R D S 141 7.2.3.3 C O R R E C T E D DISPLACEMENT R E C O R D S 142 7.2.4 DISCUSSION OF DATA 142 7.2.4.1 LONGITUDINAL DIRECTION 142 7.2.4.2 T R A N S V E R S E DIRECTION 145 7.2.4.3 VERTICAL DIRECTION 148 7.3 DATA ANALYSIS 149 7.3.1 C O M P U T E R P R O G R A M S 149 7.3.2 RELATIVE MOTION OF WHARF SECTIONS 150 7.3.3 DYNAMIC CHARACTERISTICS ANALYSIS 151 7.3.3.1 F R E Q U E N C Y DOMAIN ANALYSIS 151 7.3.3.2 R E S P O N S E S P E C T R U M ANALYSIS 157 7.4 SUMMARY 161 TABLE OF CONTENTS VII C H A P T E R 8 SUMMARY AND CONCLUSIONS 203 8.1 SUMMARY 204 8.2 CONCLUSIONS 207 8.3 FURTHER R E S E A R C H 208 R E F E R E N C E S 209 APPENDIX A S A M P L E S E T OF AMBIENT VIBRATION TIME HISTORIES 214 APPENDIX B S A M P L E S E T OF F R E E VIBRATION TIME HISTORIES 223 APPENDIX C OAKLAND O U T E R HARBOUR WHARF - U N C O R R E C T E D A C C E L E R A T I O N TIME HISTORIES 229 APPENDIX D OAKLAND O U T E R HARBOUR WHARF - P S D FUNCTIONS 231 APPENDIX E OAKLAND O U T E R HARBOUR WHARF - A C C E L E R A T I O N R E S P O N S E S P E C T R A 243 LIST OF TABLES C H A P T E R 3 Page Table 3.1 Preliminary Analytical Model: Section and Material Properties 35 Table 3.2 Preliminary Analytical Model: First Ten Natural Frequencies 35 C H A P T E R 4 Table 4.1 Comparison of Significant Features of the Equipment Systems 67 Table 4.2 Summary of Measurement System Performance Under Testing . . . . 68 Table 4.3 Ambient Vibration Test Setup Information 69 Table 4.4 Forced Vibration Test Setup Information 69 C H A P T E R 5 Table 5.1 MAC Results: Measured Modes vs Preliminary Model Modes . . . . 105 Table 5.2 Natural Frequencies: Measured Modes vs Preliminary Model Modes 105 Table 5.3 MAC Results: Measured Modes vs Calibrated Model Modes 106 Table 5.4 Natural Frequencies: Measured Modes vs Calibrated Model Modes 106 Table 5.5 Calibrated Analytical Model: Section and Material Properties 107 Table 5.6 Parametric Study Results: Length To Fixity - Natural Frequencies . 108 Table 5.7 Parametric Study Results: Length To Fixity - Participating Mass. . . 109 Table 5.8 Parametric Study Results: Framing Member Stiffness -Natural Frequencies 110 Table 5.9 Parametric Study Results: Framing Member Stiffness -Participating Mass 111 Table 5.10 Parametric Study Results: Structure Mass - Natural Frequencies . . 112 Table 5.11 Parametric Study Results: Structure Mass - Participating Mass . . . 113 Table 5.12 Parametric Study Results: Bearing Pad Stiffness -Natural Frequencies 114 Table 5.13 Parametric Study Results: Bearing Pad Stiffness -Participating Mass 115 C H A P T E R 6 Table 6.1 Preliminary Analytical Model: Section and Material Properties 125 Table 6.2 Preliminary Analytical Model: First Ten Natural Frequencies 126 Table 6.3 Preliminary Analytical Model: First Participating Mass Factors . . . . 126 ix LIST OF FIGURES CHAPTER 2 Page Fig. 2.1 Soil-Structure Interface Models 23 Fig. 2.2 Assumed Ground Line Location For Steep Waterfront Slopes 24 Fig. 2.3 Depth To Fixity Using Davisson's Method 25 CHAPTER 3 Fig. 3.1 Photograph Of Vertical Piled Wharf Site 36 Fig. 3.2 Vertical Piled Wharf - Plan and Elevation Views 37 Fig. 3.3 Vertical Piled Wharf Computer Model - Isometric View 38 Fig. 3.4 Hollow Octagonal Pile - Typical Section 39 Fig. 3.5 Concrete Pile Cap - Typical Section 40 Fig. 3.6 Deck Unit - Typical Section 41 Fig. 3.7 Ten Analytical Mode Shapes - Plan View 42 CHAPTER 4 Fig. 4.1 Pictoral Diagram Of System No. 1 Equipment Setup 70 Fig. 4.2 Pictoral Diagram Of System No. 2 Equipment Setup 71 Fig. 4.3 Drifting of Ambient Vibration Measurement Signal 72 Fig. 4.4 Ambient Vibration Survey - Sensor Locations and Directions 73 Fig. 4.5 Forced Vibration Survey - Sensor and Force Input Locations 74 Fig. 4.6 Typical Ambient Vibration Survey Time History 75 Fig. 4.7 Forced Vibration Response Filtered At 50 Hz 76 Fig. 4.8 Forced Vibration Response Filtered At 10 Hz 77 Fig. 4.9 Averaged Normalized Power Spectral Density -Transverse Direction 78 Fig. 4.10 Averaged Normalized Power Spectral Density -Longitudinal Direction 79 Fig. 4.11 A PSD's For Wharf A V S Records - Transverse Stations 3 - 6 80 Fig. 4.11B PSD's For Wharf A V S Records - Transverse Stations 7 -11 81 Fig. 4.11C PSD's For Wharf A V S Records - Longitudinal Signals 82 Fig. 4.12 Ambient Vibration Survey - Overall Site Mode 83 Fig. 4.13 Ambient Vibration Survey - Rigid Body Modes 84 Fig. 4.14 Ambient Vibration Survey - Flexural Modes 85 Fig. 4.15 Typical Forced Vibration PSD 86 LIST OF FIGURES x CHAPTER 5 Fig. 5.1 MAC Analysis Results: Preliminary Analytical Model vs Measured Mode Shapes 116 Fig. 5.2 MAC Analysis Results: Calibrated Analytical Model vs Measured Mode Shapes 117 CHAPTER 6 Fig. 6.1 Average Soil Profile At Oakland Outer Harbour Wharf 127 Fig. 6.2 Geology At Oakland Harbour 128 Fig. 6.3 Photograph Of Oakland Harbour 129 Fig. 6.4 Plan View Of Oakland Outer Harbour Wharf 130 Fig. 6.5 Elevation View Of Oakland Outer Harbour Wharf 130 Fig. 6.6 Isometric View Of Oakland Outer Harbour Wharf 131 Fig. 6.7 Typical Concrete Pile At Oakland Outer Harbour Wharf 132 Fig. 6.8 Geometry Of Steel Sheet Piling At Oakland Outer Harbour Wharf 133 Fig. 6.9 Typical Pile Cap At Oakland Outer Harbour Wharf 134 Fig. 6.10 First Ten Analytical Mode Shapes - Plan View 135 CHAPTER 7 Fig. 7.1 Location Of Strong Motion Sensors At The Oakland Outer Harbour Wharf 163 Fig. 7.2 Location Of Known Faults In The Vicinity Of The Loma Prieta Earthquake (Inset Shows Radial Propagation Of Ground Waves Moving Past The Wharf Site) 164 Fig. 7.3 Correlated Acceleration Records From Strong Motion Sensors Located At O O H W During the Loma Prieta Earthquake 165 Fig. 7.4 Correlated Velocity Records From Strong Motion Sensors Located At O O H W During the Loma Prieta Earthquake 166 Fig. 7.5 Correlated Displacement Records From Strong Motion Sensors Located At O O H W During the Loma Prieta Earthquake 167 Fig. 7.6 Corrected Acceleration Time Histories - Longitudinal Structure Sensors 168 Fig. 7.7 Corrected Acceleration Time Histories - Longitudinal Free Field Sensors 169 Fig. 7.8 Corrected Acceleration Time Histories - Longitudinal Free Field And Structure Records at South End of The Wharf 170 Fig. 7.9 Corrected Acceleration Time Histories - Longitudinal Free Field And Structure Records at North End of The Wharf 171 Fig. 7.10 Acceleration Time History Records For Sensors 4 and 5 172 Fig. 7.11 Acceleration Time History Records For Sensors 4 and 6 173 Fig. 7.12 Acceleration Time History Records For Sensors 4 and 9 174 LIST OF FIGURES x j Fig. 7.13 Acceleration Time History Records For Sensors 5 and 6 175 Fig. 7.14 Acceleration Time History Records For Sensors 5 and 9 176 Fig. 7.15 Acceleration Time History Records For Sensors 6 and 9 177 Fig. 7.16 Acceleration Time History Records - Transverse Free Field Sensors 178 Fig. 7.17 Acceleration Time History Records - Transverse Free Field And Structure Sensors at The North End of The Wharf 179 Fig. 7.18 Acceleration Time History Records - Transverse Free Field And Structure Sensors at The South End of The Wharf 180 Fig. 7.19 Typical Real Time Displacements During The Earthquake 181 Fig. 7.20 Coherence: Sensors 4 and 5 182 Fig. 7.21 Coherence: Sensors 4 and 6 183 Fig. 7.22 Coherence: Sensors 4 and 9 184 Fig. 7.23 PSD For Transverse Structure and Free Field Sensors 185 Fig. 7.24 PSD For Longitudinal Structure and Free Field Sensors at South End of The Wharf 186 Fig. 7.25 PSD For Longitudinal Free Field Sensor at South End of Wharf . . 187 Fig. 7.26 PSD For Longitudinal Free Field Sensor at North End of Wharf . . 188 Fig. 7.27 PSD For Transverse Free Field Sensor at South End of Wharf ... 189 Fig. 7.28 P S D For Transverse Free Field Sensor at North End of Wharf . . . 190 Fig. 7.29 PSD For Longitudinal Free Field Sensors 191 Fig. 7.30 PSD For Transverse Free Field Sensors 192 Fig. 7.31 Coherence: Sensors 4 and 10 193 Fig. 7.32 Coherence: Sensors 5 and 10 194 Fig. 7.33 Averaged PSDs for Relative Acceleration Time Histories: Sensors 4 and 5 195 Fig. 7.34 Frequency Response Function: Sensor 10 (Input) and Sensor 4 (Output) 196 Fig. 7.35 Response Spectra at 5% Damping: Longitudinal Free Field Records 197 Fig. 7.36 Response Spectra at 5% Damping: Longitudinal Free Field And Structure Records at North End of The Wharf 198 Fig. 7.37 Response Spectra at 5% Damping: Longitudinal Free Field And Structure Records at South End of The Wharf 199 Fig. 7.38 Response Spectra at 5% Damping: Transverse Free Field Records 200 Fig. 7.39 Response Spectra at 5% Damping: Transverse Free Field And Structure Records at North End of The Wharf 201 Fig. 7.40 Response Spectra at 5% Damping: Transverse Free Field And Structure Records at South End of The Wharf 202 ACKNOWLEDGEMENTS xii I would like to thank my parents Sing and Susie, my brothers Perry and Doug, my employer, Westmar Consultants Inc. and all of my friends who have been understanding, supportive and patient with me through my extended period as a graduate student. Without their support and the occasional push, this thing may never have been completed. I gratefully acknowledge the support I received from my thesis advisors Dr. Carlos E. Ventura and Dr. Sheldon Cherry. Their guidance and input at all levels of this project made what seemed like a mountain of a task appear much easier. Financial support for this project was provided by the Natural Sciences and Engineering Research Council of Canada (NSERC) in the form of a postgraduate scholarship and through Research Grants awarded to both Dr. Ventura and Dr. Cherry. I wish to thank Mr. Howard Nichol, U B C Earthquake Laboratory technician, for his assistance in dealing with all the equipment used in this project and for his help with the occaisional field assignment. I also wish to thank Mr. Jim Grieg and Mr. Thomas Wong of the Civil Engineering Graphics Department for their assistance with all of my computer related problems and their odd lending of printers and other computer things. I save my deepest gratitude for the two people most instrumental in helping me complete this achievement, my good friend Dr. Andreas J . Felber and my friend and advisor Dr. Carlos E. Ventura. Andreas allowed me to tag along on his wagon as he developed the ambient vibration system and give me both technical and psychological support throughout the project. Carlos, whom I have already thanked several times as a friend, an advisor and a financial supporter, has continually supported me even when I pushed him to the limit of his patience. XIII To Sing, Susie, Perry and Doug 1 CHAPTER 1 OVERVIEW In the past seismic analysis of open-piled wharves has been carried out based on the provisions of building codes. More recently, such analysis has been based on the provisions of bridge codes. Current State-of-the-Art designs of wharves in North America use the recommendations of the American Association of State Highway and Transportation Officials (AASHTO) standard specifications for highway bridges, which in turn are taken directly from the recommendations of the Applied Technology Council's ATC-6 standard. However, fundamental differences between the support systems in bridges and wharves may make the A A S H T O recommendations unsuitable for analysing wharf structures. Little research has been done to verify the applicability of building and bridge code recommendations for the seismic analysis of open-piled wharves. The aim of this study was to determine the accuracy of using different modelling assumptions, taken from guidelines set out by building and bridge codes, to determine the important parameters for the dynamic modelling of wharves. 1.1 B A C K G R O U N D ON W H A R F STRUCTURES A wharf is a port structure built specifically as an extension of the land into deeper water. The purpose of the structure is to facilitate the transfer of cargo between ships and land by allowing ships to berth directly alongside a "land bridge". Cranes and other unloading and loading vehicles can assist in the transfer of cargo once the ships are berthed at the wharf. OVERVIEW 2 Closed cell type wharves consist of a retaining type structure with the land filled out to the retaining structure. Cell type wharves are usually built in locations where soil conditions are stable and water depths are relatively shallow. Open type wharves use a slab and column type construction to extend a platform out into the deeper water. The emphasis of this thesis will be on the modelling of open piled wharves. 1.2 OPEN PILE W H A R V E S Open piled wharves, denoted simply as "wharves" hereafter, generally consist of a timber or concrete deck supported by timber or concrete pile caps. The pile caps are in turn supported above the water by a number of timber, concrete or steel piles driven into the soil. New wharves are now generally constructed of concrete and/or steel piles with concrete pile caps and concrete decks. 1.2.1 PILE CAP A N D D E C K S Y S T E M Concrete pile cap and deck systems in Canada normally consist of heavy precast deck units supported by deep pile caps. Because of heavy gravity live load requirements, upwards of 48 kPa over the entire deck surface, the deck slabs are usually very thick. The thickness of deck units can be in excess of 1 m. This in turn requires that the pile caps be very thick, usually approximately 0.6 m thicker than the deck slabs. Because of these geometrical constraints, the usual seismic design preference of having weak beams and strong columns becomes virtually impossible to achieve economically. This is very similar to the case for bridge piers. OVERVIEW 3 1.2.2 VERTICAL L O A D SUPPORT SYSTEM In a wharf, unlike the case for buildings and bridges, no massive foundations are available to create a somewhat fixed base for supporting columns. Piled foundations for buildings and bridges in general consist of a large number of fully imbedded piles in a closely spaced cluster. This creates a somewhat firmly fixed foundation. In wharf structures, the piles serve two purposes. First the piles act as the foundation elements for the structure, providing the interface between the structure and the soil. Second, the piles act as columns supporting the deck above a free zone in water and air. For economic reasons, the number of piles is minimized and thus the piles are spread out over a large distance. The number of piles is controlled by the requirement for supporting the high gravity loads associated with these types of structures. Loadings on concrete piled wharves vary significantly between wharves, but a typical concrete pile wharf might have about 15 m 2 of tributary area per pile, with an average pile loading of about 900 kN per pile. Individual pile loads would be much higher locally due to heavy vehicle loads. The total number of piles and the configuration of the wharf is dependent on many factors such as the number, type and size of ships using the wharf. 1.2.3 L A T E R A L LOAD SUPPORT SYSTEM Two different classes of lateral support systems provide lateral restraint in wharf systems. The two systems can be broadly categorized as batter pile systems and vertical pile systems. The batter pile system can be viewed simplistically as a braced frame type system in which the lateral O V E R V I E W 4 loads are taken out directly by the piles. In contrast, the vertical pile system is basically a moment resisting frame in which lateral loads are carried by the frame action of the structure. Batter pile systems have been very popular because of the overall rigidity of the structural system. However, current practice favours vertical piled systems in seismically active zones because, in part, of the poor performance of batter pile wharves (BPW) in past earthquakes. Vertical piled systems tend to attract lower forces during earthquake shaking and can easily be detailed to achieve significant ductility levels. The low seismic design forces and greater ease of constructing vertical piled wharves (VPW) make these structures very desirable to designers and contractors. 1.3 M O D E L L I N G OF THE L A T E R A L SUPPORT S Y S T E M Because of the large spacing of the piles of a wharf, each pile in itself can be considered as a single "foundation". For analysis purposes, each pile must be modelled separately as opposed to modelling a large group of closely spaced piles as a single foundation. Regardless of the support system used, one of the key issues in modelling wharves is how to model the connection of the piles to the soil at the soil level. Normal analysis techniques used for bridge piers might involve methods of replacing entire foundations with "equivalent" rotational and translational springs. Springs would be placed in each of the three principal directions at ground level. Springs representing cross coupling terms between rotations and translations are generally neglected since their influence is considered insignificant (Buckle et al 1986). OVERVIEW 5 Replacing the foundation with equivalent springs is feasible in bridge piers because there are relatively fewer foundations than in wharves. Bridge foundations are in very discrete locations and are considered well fixed by a large number of piles. In a piled wharf structure, the number of locations where equivalent springs are required would result in an analytical model of the wharf of unmanageable size. The computer resources required to perform static and dynamic analyses of the model would be tremendous. A very practical method for modelling the lateral stiffness of the piles in the soil for wharves involves the use of an equivalent cantilever type model. In an equivalent cantilever model, the interface between the pile and the soil is modelled as fixed. The "point of fixity" of the pile in the soil is chosen to occur at a level which provides either force or displacement equivalency between the modelled structure and the actual structure. The fixity location used depends on whether force or displacement equivalency is required for analysis purposes. This type of model is especially appealing to designers because of its simplicity and ease of application. Many different theories are available for determining the effective pile length to fixity for use in the equivalent cantilever type models. It is not certain how effective these models are for predicting the seismic behaviour of the wharf structures, but performance of wharves in past earthquakes has shown that the methods are acceptable. Some of these theories are explained further in Chapter 2. 1.4 SOIL A N D SLOPE CONDITIONS A T W H A R F FOUNDATIONS Soil conditions at the foundations for wharf structures are generally poor. Soils are often loose silt, sand or gravel deposits. These soils are submerged in water and are highly susceptible to liquefaction. For cases where the possibility of liquefaction exists, its effects must be considered OVERVIEW 5 when determining the mud line from which fixity is measured. Fixity must be assumed to occur at some calculated distance below the zone of liquefaction. Liquefaction is ignored in this study because, in practice, remedial measures are generally taken to limit liquefaction on wharf sites or else the design is limited by the peak ground acceleration at which liquefaction is expected to occur. In addition to poor soil conditions, a sloping shore front causes further difficulties. The slope of the ground is usually kept as steep as possible, in order to achieve the water depth required at the berth face of the wharf in the shortest possible distance. This ensures the most economical deck surface area for the structure. The ground profile can get as steep as 1V:1.75H, which causes two problems that influence the dynamic behaviour of the structure. The first problem is that the steep slope is susceptible to amplified downslope free field motions under strong motion vibrations. The second problem is that the slope causes the piles to be of widely varying lengths. The first problem is mainly a geotechnical problem and only affects the structure in that it induces displacement controlled loads on the structure. This problem is not addressed in this study because of its complexity and the fact that it is generally ignored in practice. Although this problem is ignored in this study and in practice, it can be a very significant problem and must be studied in the future. The second problem, regarding the varying lengths of pile, has a large influence on the dynamic characteristics of the structure. Varying the length of the piles can cause a large difference in lateral stiffness from pile to pile. The shorter piles close to shore tend to control the dynamic characteristics of the wharf. Because of this discrepancy in member stiffness from pile to pile, OVERVIEW 7 assumptions as to the location of the point of fixity can make large differences on the response of the system, especially for the shorter piles close to shore. 1.5 PROBLEM DEFINITION Assumptions used when creating computer models can significantly affect the number of piles required in a structure. Since a large portion of the cost of these structures is in the piles, and thus extra piles in the system can generally lead to significant additional expense, it is important to determine whether or not pile fixity assumptions are correct. While it is generally acknowledged that the high level of redundancy in these structures leaves little likelihood of collapse under seismic loading, very high costs in repair work make it important to ensure that damage is kept below an acceptable limit without using too many piles. Under these circumstances, the assumptions used in modelling wharves in analysis can result in a major economic issue. 1.6 OBJECTIVES AND SCOPE The goal of this research was to assess the accuracy of modelling assumptions used in determining the dynamic properties of an open piled wharf. In order to accomplish this goal, two objectives were defined. One objective was to review methods used to model the various structural elements and the nature of the soil-structure interface, by observing how accurately properties such as natural frequencies and mode shapes were determined. The second objective was to gain improved understanding of the the nature of the movement of a wharf structure under OVERVIEW 8 strong motion activity. To achieve these two objectives, vibration data were obtained and analysed for two separate wharves. The first wharf, a concrete VPW, was monitored by means of an ambient vibration survey. Additionally, impact induced free vibrations were measured. The vibration measurements from the ambient vibration and free vibration measurements provided information for low level vibrations. A computer model of the V P W was created on the commercially available finite element package SAP90. Various modelling assumptions were used to create a preliminary model which was employed to predict the motion of the wharf prior to conducting field measurements, and thus to assist in identifying mode shapes and natural frequencies in the experimental data. The preliminary model for the V P W was modified to best match the experimentally measured dynamic properties of the structure after the experimental data were analysed and results obtained. Furthermore, a parametric study was conducted using the modified computer model. Results from the parametric study were then used to extrapolate the predictions from low level vibration measurements to predictions for the same structure under strong motion vibrations. The second wharf studied is located in Oakland, California and consists of an all concrete deck supported by a combination of vertical and batter piles. This structure was instrumented in the mid 1970's by the California Strong Motion Instrumentation Program (CSMIP), Division of Mines and Geology, Department of Conservation, with a total of twelve permanent accelerometers. Strong motion data from these accelerometers collected during the Loma Prieta earthquake in 1989 were used for the analysis. This data set provided information on the vibration OVERVIEW 9 characteristics of a BPW subject to strong motion vibrations. As was done for the VPW, a preliminary computer model was also produced for the BPW to assist in determining the mode shapes and natural frequencies of the structure. The work presented in this thesis is organized in the following manner: Chapter 2 presents the most important theoretical aspects of the work, including references for this information. Chapter 3 introduces the vertical piled structure which was studied. It includes the background of the structure, the structure description, and the average soil conditions of the site. Also, the preliminary finite element model is introduced. The assumptions used in creating the model are discussed in detail. Chapter 4 presents the preliminary analysis, the field testing, the experimental results, and the detailed analyses of the vertical pile wharf structure. Included in this section is a detailed discussion of the selection of equipment for testing as well as the specifications of the equipment used. Chapter 5 gives the results of a correlation analysis between the finite element computer model created in Chapter 3 and the experimentally measured results determined in Chapter 4. In addition, the results of a parametric study of the variables used to model the dynamic properties of a wharf are given. OVERVIEW -j 0 Chapter 6 describes the batter piled structure that was studied. This chapter Includes information regarding the history of the structure, the structure description, and the general soil conditions of the site. Also presented is the theoretical uncalibrated finite element model of the structure. The assumptions used in creating the model are discussed in detail. Chapter 7 discusses the analysis of strong motion data for the batter piled wharf system. Included is a description of the measurement system used to obtain strong motion data and discussion of, the form of the data, the general trends of the data, and the data analysis and results. Chapter 8 summarizes the conclusions from each of the previous chapters. Appendix A contains a sample set of ambient vibration time histories. Appendix B contains a sample set of impact vibration time histories. Appendix C gives the uncorrected acceleration signals from the Loma Prieta earthquake. Appendix D contains plots of the power spectral densities calculated from the signals measured on the Oakland Outer Harbour Wharf during the Loma Prieta Earthquake. Appendix E contains the acceleration response spectra calculated from the signals obtained on the Oakland Outer Harbour Wharf during the Loma Prieta Earthquake. 11 CHAPTER 2 THEORETICAL BACKGROUND AND LITERATURE REVIEW This chapter provides the background for the theories and methods used in the following chapters. Topics discussed in this chapter include: concrete material property relationships pile to soil interface modelling ambient vibration theory modal correlation analysis strong motion data analysis An overview of each of these topics is presented here. The reader is directed to the reference material for more in-depth explanations. 2.1 CONCRETE M A T E R I A L PROPERTIES Relationships used to calculate concrete properties were taken from well known sources. Concrete properties which were of importance included: Young's modulus Shear modulus THEORETICAL BACKGROUND AND LITERATURE REVIEW 12 Young's modulus (E) was calculated as the secant modulus and is approximated using the following relationship: where f c is the 28 day compressive strength of concrete (MPa) The Shear modulus (G) was calculated using the following relationship: where v is Poisson's ratio for concrete The above formulae are taken from Collins and Mitchell (1987) and are the static moduli. In a more formal approach, the dynamic moduli for concrete sections should be used for modelling, but the standard of practice is to use static moduli. 2.2 PILE TO SOIL INTERFACE MODELLING Foundation modelling is an important part of overall structural modelling to determine the response of a structure to lateral loads. A wharf structure cannot be modelled independently of the foundation soils because boundary conditions imposed by the soil determine how the wharf behaves dynamically. The effects of the soil on the dynamic behaviour of the structure are commonly referred to as the soil-structure interaction of a system. To determine accurate and realistic response results from a computer model, the soil characteristics and properties must be included when modelling the system, especially when analysing short stiff structures. Because (2.1) THEORETICAL BACKGROUND AND LITERATURE REVIEW 13 there are no tools currently available to create a complete soil-structure model, approximations are used to simplify the problem and make it manageable (Buckle et al 1986). Several approximations for modelling the soil-structure interface at foundations are available, including three of the most commonly used approximation models (Buckle et al 1986): equivalent soil springs model equivalent base spring model equivalent cantilever model These three models are the same types of models used as standard practice by highway bridge designers. The three models are shown pictorially in Fig. 2.1 and are discussed further in the following sections. 2.2.1 EQUIVALENT SOIL SPRING M O D E L The equivalent soil spring model shown in Fig. 2.1b involves the use of non-linear soil springs (p-y curves) connected to the pile at different levels. These springs represent the stiffness of the soil; since this property can vary with depth, each soil layer may need to be explicitly modelled. In this approach a fairly accurate model can be created, but at a high cost. The size of the computer model could be rather large and difficult to manage. THEORETICAL BACKGROUND AND LITERATURE REVIEW 14 2.2.2 EQUIVALENT BASE SPRING MODEL The equivalent base spring model shown in Fig. 2.1c assumes that the soil behaviour is elastic and that the soil can be modelled at each support using six independent springs. These springs are assumed to act at the ground surface. The equivalent base spring model can be quite accurate, provided that cross coupling terms representing the interaction between the different supports are included in the structure stiffness matrix. The cross coupling terms in the stiffness matrix can only be determined by modelling the entire foundation explicitly, so that the effects of the movement of one support on the rest of the structure can be observed. The explicit modelling encompasses the use of soil springs known as p-y curves in the computer model at several levels along each foundation unit. This can be a very involved process, especially for large structures with numerous supports. 2.2.3 EQUIVALENT CANTILEVER MODEL The simplest and most commonly used approach in practice, which is also the method used for modelling in this thesis, is the equivalent cantilever model shown in Fig. 2.1d. In this model, the supports are represented as fixed supports at a given location (point of fixity). The point of fixity is approximated such that the structure has either the same stiffness at ground level or the same maximum bending moment as the actual soil-structure system. There are several methods of determining the location of the point of fixity. Empirical charts, or methods that consider the relative stiffness between the soil and the pile are commonly used (Buckle et al 1986). Some of the soil-pile models are explained further in the following text. THEORETICAL BACKGROUND AND LITERATURE REVIEW 15 The point of fixity is usually determined with respect to the ground line. If the ground slope is very steep, a virtual ground line is used as a reference (Tuschida 1980). This virtual ground line is a line drawn halfway between the actual soil surface and a line projected from the dredge line at the toe of the slope as shown in Fig. 2.2. For reference, the actual ground line will be referred to as ground line #1 and the virtual ground line will be referred to as ground line #2. The first very loose layer of soil must be neglected. As mentioned above, the depth to fixity can be taken from empirical charts or using models that consider relative stiffness between the soil and the pile. These methods usually result in a depth to fixity of between 3 pile diameters to 8 pile diameters (Gaythwaite 1990). The more competent the soil the shorter the depth to fixity. A very commonly used method to determine the depth to fixity, D, is a relationship proposed by Davisson (1973). In this relationship, depths to fixity are considered to be inversely proportional to the fourth root of the soil horizontal subgrade modulus, Kh, and proportional to the pile stiffness, El. The equation for the depth to fixity in granular soils is as follows depending on whether a stiffness or a force criterion must be met: To match stiffness: El 4 (2.3) THEORETICAL BACKGROUND AND LITERATURE REVIEW 1 6 To match moments: (2.4) The depth to fixity is shown graphically in Fig. 2.3. As can be observed, the actual pile-soil model has a maximum deflection and moment associated with a given lateral force (Fig. 2.3a). These two parameters can be matched in a model by fixing the pile at an elevation which would give equivalent maximum values. Thus to match the deflection criterion, the pile would be modelled with the same free length above the mudline and a depth D s below the mudline to give equivalent stiffness (Fig. 2.3b). To match the force criterion, the pile would be modelled with the same free length above the mudline and a depth D m below the mudline giving equivalent moment (Fig. 2.3c). Note that D s and D m have different values. For further details on the empirical formulae, see Davisson (1973), for further details on points of fixity, see Gaithwaite (1990) and Tuschida (1980). 2.3 AMBIENT VIBRATION THEORY An ambient vibration survey is a process used to determine the vibration characteristics of a structure by measuring the response of the structure to low amplitude, ambient vibrations. This section will discuss ambient vibrations, including the following: their use as a tool to determine the dynamic characteristics of a structure; the conditions to be satisfied for ambient vibrations to produce valid results; and where ambient vibrations have been used in past research. THEORETICAL BACKGROUND AND LITERATURE REVIEW 17 Ambient vibrations are background vibrations caused by cultural noises such as wind, traffic, wave action, and micro-tremors. These vibrations provide very low level vibration input to physical structures. Ambient vibrations are assumed to have white noise characteristics and thus have uniform frequency content over a broad band of frequencies. This is opposed to seismic vibrations and forced vibrations which are of much higher level, but possess high energy over narrow frequency ranges and little or no energy in other frequency ranges. The acceleration response of a structure to the ambient inputs can usually be viewed as a mixed random process, a composite of a random response and a deterministic response. The random response is due to the random nature of the input and other background noise. The deterministic response is due to the structure vibrating in its natural modes of vibration. By measuring a structure response to ambient vibration inputs, its dynamic characteristics can be determined from the deterministic portion of the response. The dynamic characteristics which can be determined are the natural frequency, the mode shapes and the damping of a system. Analysis of ambient vibration data can produce reliable results for modal frequencies and mode shapes if the following criteria are satisfied: the data is stationary and ergodic; the structure behaves as a linear system; the input vibrations are broad band and cover the entire frequency range of interest; the modes of interest are well separated and lightly damped (damping levels up to 5% of critical damping are considered light); and the structure is classically damped. T H E O R E T I C A L B A C K G R O U N D A N D L I T E R A T U R E R E V I E W 18 Ambient vibration measurements have been used to determine the dynamic characteristics of a wide variety of structures such as bridges (Abdel-Ghaffar 1985, Felber 1993), buildings (Trifunac 1972, Ngok 1982, Schuster 1994), offshore structures (Brownjohn 1990) and dams (Paultre 1992 and Kemp 1995). Dynamic characteristics of the structure can be determined by analysing the structure response signals in the frequency domain. Some of the key tools for analysing the ambient vibration data are the following: 1. The Power Spectral Density (PSD) function of a signal is expressed as: G^)=X^)X;^) (2-5) where X(co) is the Fourier Transform of the structure acceleration response at location "i"; X (co) is the complex conjugate of X(co); and co is the circular frequency in radians per second 2. The Cross Spectral Density (CSD) function of signals from two separate locations "i" and "j" is defined as: G^)=X,*(ai)XJi<d) (2.6) T H E O R E T I C A L B A C K G R O U N D A N D L I T E R A T U R E R E V I E W -| 9 3. The Phase Angle ((j)) between two signals "i" and "j" is defined as: ^ c o ) = t a n - 1 ( - ^ ) (2.7) where C^co) and Q^co) are the real and imaginary components of G^co) 4. The Coherence (y2) between two signals is expressed as: T f e ) . '(2-8) 5. The Transfer Function between two signals is expressed as: m^JM^M-Bial <2.i» X / W ) Qjjo) Gj(<o) These functions are equations based on random vibrations and are the basis of the solution to ambient vibration problems. The PSD and C S D of the response signal highlight frequency ranges in the response of the structure that have a high energy content. Since the input vibrations are ambient, and thus are assumed to be of uniform frequency content, frequencies in the PSD or C S D containing higher energy content, possibly correspond to natural frequencies of the structure. The phase angle between two signals indicates the manner in which the two locations physically move relative to one another. When two signals move directly in-phase or directly out-of-phase for a given frequency, the signals are possibly moving in a given mode shape. The coherence between two signals at a given frequency indicates the level of correlation between two signals at that frequency. A coherence of unity corresponds to a perfect correlation THEORETICAL BACKGROUND AND LITERATURE REVIEW 20 between two signals while a coherence of zero corresponds to a very poor correlation. In actual practice, coherence is somewhere between zero and unity. Low values of the coherence function between two signals can occur from one or more of the following conditions: extraneous noise in the measurements; resolution bias errors in the spectral estimates; the system relating the two signals is not linear; and one of the signals is a result of multiple inputs. At a natural frequency, pairs of signals are more likely to exhibit high coherence. For further details on random vibrations refer to Bendat and Piersol (1986). For further details on ambient vibrations the refer to Luz (1987), Diehl (1991) and Felber (1993). For further details on dynamics of structures refer to Clough and Penzien (1975). 2.4 M O D A L A S S U R A N C E CRITERION The modal assurance criterion (MAC) (Ewins 1992) is a method that can be used to correlate mode shapes. Coordinates for mode shapes are combined using the following formula: MAC= | M M (2.10) where cpA and <DX are two vectors defining the structure mode shape amplitudes at discrete locations on the structure. THEORETICAL BACKGROUND AND LITERATURE REVIEW 21 The result from the MAC equation is a scalar value which varies from zero to one. This number indicates the level to which the two mode shapes are similar to each other. Two perfectly matching mode shapes would result in a M A C value of 1.0. The closer the number to 1.0, the better the match between two mode shapes. The MAC concept uses the fact that true mode shapes are orthogonal. By multiplying the mode shapes, non matching mode shapes should go to zero. By dividing by both of the mode shapes squared, M A C is normalized. 2.5 STRONG MOTION D A T A A N A L Y S I S Strong motion data are generally obtained from permanent instrumentation installations set out by strong motion instrumentation agencies. In the past, strong motion data were relatively scarce, but with increased use of permanent installations strong motion data have become more readily available. Recently, the instrumentation program set out by the California Strong Motion Instrumentation Program (CSMIP) has provided large quantities of earthquake strong motion data from the 1989 Loma Prieta earthquake and the 1994 Northridge Earthquake. The analysis of strong motion data uses the same principles employed in the analysis of ambient vibrations. Strong motion data are generally analysed in both the frequency and time domains. Some of the techniques used to analyse strong motion data are discussed in this section. Information which can be obtained by analysing strong motion data in the time domain includes the following: THEORETICAL BACKGROUND AND LITERATURE REVIEW 22 peak acceleration, peak velocity and peak displacement; duration of strong motion shaking; frequency of predominant waves (by a response spectrum analysis); approximate frequency range of the event (ie. are high frequency components present); time delays between signals; and amplitude and correlation between signals in the three principal directions. In general, time domain analysis can provide information regarding the characteristics of the seismic waves or seismic event at the location of the sensor. To obtain information on the response of the structure to the seismic event, the data can be analysed using frequency domain analysis and/or response spectrum analysis. The key functions in the frequency domain analysis include the Fourier Spectrum, the transfer function, phase angle and coherence between pairs of signals. These are essentially the same functions evaluated in the analysis of ambient vibration data. In contrast to ambient vibration analysis, the Fourier spectra of structural response is influenced by the frequency content of both the seismic event and the reaction of the structure to the input. This gives the analyst the added complexity of separating the natural frequency of the seismic event from the natural frequency of the structure. The phase angle between the input signal from a sensor removed from the structure and a sensor on the structure is approximately 90°, as opposed to 0° or 180° for two sensors on the structure. As for ambient vibrations, the coherence is high between two sensor signals at a natural frequency of the structure provided the main source of excitation to the structure is only one input, rather than multiple inputs. S3 M a i A a a a u n ± v a 3 i n Q N V a i \ i n o u o > i o v s - i v o u a u o a m 26 CHAPTER 3 STRUCTURAL MODELLING - VERTICAL PILED WHARF A structural model of the experimentally measured vertical piled wharf (VPW) was developed using the commercially available finite element analysis package SAP90 (Version 5.4). A combination of three dimensional beam elements, shell elements, and mass elements was used to create the model. Modelling techniques and rules applied in general practice were followed in determining the global model framing system and individual member properties. This model was employed to estimate the global dynamic characteristics of the wharf prior to conducting tests on the structure so that testing procedures such as the number and location of measurements could be planned. In addition, preliminary results were used to assist in data interpretation. 3.1 B A C K G R O U N D INFORMATION The details of the model used for the V P W are given below. An explanation of the soil characteristics and the existing structural system is followed by an account of the assumptions introduced to determine individual member section and material properties for the computer model. STRUCTURAL MODELLING - VERTICAL PILED WHARF 27 3.2 SITE CONDITIONS The V P W considered in this study is a structure located at a shipping terminal site at the mouth of the Squamish river delta in the Howe Sound about 50 km north of Vancouver, British Columbia. This wharf, situated on relatively loose soils, was designed and constructed in the late 1980's. A photograph of the overall site is shown in Fig. 3.1. The original ground profile of the site was created by delta deposits from the river to the north. The terminal site itself was built in 1970 when 20 m of dredged granular fill was placed over the original river deposits. Offshore, the soils consist of loose sands with traces of gravel to a depth of 15 m below the dredge line. Onshore, the dredged fill is compact to a level of 4 m and is then loose to a depth of 12 m. A compact granular soil underlies the fill. During construction, dredging of the river deposits was required to bring the slope profile to the final grade. Most of the dredging took place at the north end of the site, with progressively less dredging required towards the south end of the structure. It is assumed that future deposition from the river would affect the north end of the structure to a greater extent than the south end of the structure. 3.3 STRUCTURE DESCRIPTION A plan and elevation views of the vertical piled structure is shown in Fig. 3.2. As can be seen, the main wharf apron is a continuous structure that is 153 m long and 32 m wide. The approximate total mass of the structure is 8,400 tonnes (8,400,000 kg). The support system STRUCTURAL MODELLING - VERTICAL PILED WHARF 28 consists of 335 octagonal hollow core precast prestressed concrete piles spaced at 2.3 m centres in the east-west direction and 7.6 m centres in the north-south direction. The piles are driven into the ground with penetrations of 18.3 m to 20.4 m depending on the location of the pile along the wharf. Free standing lengths of pile through water and air range from 7.3 m at the east side to 18.0 m at the west side. These piles are cast integrally with east-west spanning cast-in-place concrete pile caps. The pile caps have a depth of approximately 1.6 m. The piles and pile caps form 21 pile bents, spaced at 7.6 m centres in the north-south direction. The diaphram spanning between the pile caps is comprised of precast prestressed concrete double tee beams, the ends of which are cast monolithically into the pile caps. The precast deck units are topped with a continuous 150 mm thick cast-in-place concrete slab. The wharf apron is theoretically isolated from shore, but is in reality connected to the backlands by an 8 m long transition span. This transition span is comprised of a precast concrete haunched slab that is fixed to a pile cap in the backlands, but is free to slide on the wharf. A low friction teflon bearing pad is used to achieve the isolation of the main structure from the backlands. This type of structure arrangement is generally representative of the form of piled wharf design along the west coast of Canada and the United States since the early to mid 1980's. 3.4 M O D E L L I N G A S S U M P T I O N S The computer model of the V P W structure was developed using three dimensional beam elements for the piles and pile caps, shell elements for the top deck, and mass elements for additional members not accounted for in the beam and shell elements. Parameters that were required for the modelling, such as depth to fixity and individual member section and material STRUCTURAL MODELLING - VERTICAL PILED WHARF 29 properties were calculated using standard techniques as described in Chapter 2. The modelling assumptions introduced to establish the member properties of the preliminary model of the wharf are discussed in this section. The framing system consisted of the following elements: • piles • pile caps • deck The components of the framing system were connected as follows. The piles, represented by beam elements, were framed into pile caps, represented by a separate grid of beam elements. The pile to pile cap frame arrangement was repeated in the model to create the 21 separate pile bents. These pile bents were subsequently connected by shell elements spanning the pile caps and representing the deck diaphram. Connections between the piles and the pile caps were assumed to be full moment connections. A feature of SAP90, the element rigid offset, was used on the pile elements to model the distance from the model node points (located in the mid-point of intersecting members) to the actual interface between the piles and the pile caps. The element rigid offset is explained in more detail in the SAP90 users manual (Wilson et al 1990). The deck members were taken to act as shear panels only. No transfer of moments from the deck elements to the pile caps was assumed. Additional beam elements, representing the webs of the precast double tee beams and their connection to the pile caps, were also included in the model. The double tee webs were used STRUCTURAL MODELLING - VERTICAL PILED WHARF 30 to represent the bending stiffness caused by the deck units and are fixed at both ends to the pile caps. Below ground, the point of full (pile) fixity was assumed to be five pile diameters below the ground profile, determined using ground line #2, as explained in Chapter 2. Mass elements were added to the model to represent the fenders located on every second bent on the west face of the wharf. Figure 3.3 shows an isometric view of the V P W model. Details of the SAP90 model are presented in the following sections. 3.4.1 PILE PROPERTIES A typical cross section of the hollow octagonal concrete piles is shown in Fig. 3.4. The width of the pile is 610 mm between the flat sides and the hollow core is approximately 350 mm in diameter. Section properties to determine the stiffness of the preliminary model were calculated according to the gross concrete section using the dimensions given in the figure. Additional stiffness due to the reinforcing and prestressing steel was ignored. Also ignored was any reduction in stiffness due to cracking of the section expected to occur under high lateral loads. Material properties for the piles were calculated using the formulae discussed in Chapter 2 and using a concrete compressive strength of 48 MPa. Table 3.1 shows the section and material property values used to model the piles. STRUCTURAL MODELLING - VERTICAL PILED WHARF 31 3.4.2 PILE C A P PROPERTIES The geometry of a typical pile cap is shown in Fig. 3.5. Section properties for the preliminary model were calculated according to the gross concrete section using the dimensions given in the figure. Material properties were calculated using the formulae discussed in Chapter 2 and using a concrete compressive strength of 35 MPa. Table 3.1 shows the section and material property values used to model the pile caps. 3.4.3 D E C K PROPERTIES A cross section of a typical individual deck unit is shown in Fig. 3.6. In the structural model, the deck elements were comprised of two different types of elements: shell and beam elements. The shell elements were used to model the deck slab, while beam elements were used to characterize the double tee webs. The thickness of the deck slab spanning between the individual pile bents was assumed to be the thickness of the flange of the double tee section (100 mm) added to the thickness of the cast-in-place topping (150 mm). The webs of the double tees were assumed to be fixed at either end to the pile caps. Stiffness properties of the beam elements for bending about the longitudinal axis were calculated assuming the stiffness properties of the entire double tee section, less the stiffness properties of the flanges. The stiffness of the flanges was included in the deck shell elements. For modelling purposes, beam elements for the webs were only located between the node points created for the pile locations. This approach required less elements than would have been required if all the webs were modelled as individual elements. The fact that this modelling STRUCTURAL MODELLING - VERTICAL PILED WHARF 32 method resulted in fewer elements than in the true structure was taken into account in the following manner. To represent all the double tee webs, the total section properties of all the webs were smeared over the entire width of the wharf. The stiffness for the tributary width of deck was used to calculate the stiffness of the webs representing a given region. For bending about the vertical axis, the theoretical bending stiffness of the entire deck was calculated. The bending stiffness was then distributed evenly between all the elements representing the webs. Section properties for the slabs and the webs were calculated according to the gross concrete section. Material properties for the deck elements were calculated using the actual concrete compressive strength of the precast deck units of 42 MPa. The cast-in-place topping actually only had a compressive strength of about 30 MPa, but the difference in the compressive strengths was neglected because the error in modelling using the higher strength is minimal. Table 3.1 shows the section and material properties used to model the deck diaphram and the double tee webs. 3.4.4 MASS ELEMENTS Mass elements used to represent the fenders were attached to the nodes on the west side of the wharf model. These elements were placed at deck level on the bent lines containing a fender. A total of eleven translational mass elements of 100 kg each were required. The mass of all other elements were added directly to the model by inputting the density, cross-sectional area and length of each element. Added mass from the water around the structure, as well as from any marine growth clinging to the structure, was neglected as is done in standard practice. STRUCTURAL MODELLING - VERTICAL PILED WHARF 33 3.4.5 D Y N A M I C A N A L Y S I S A dynamic modal analysis was performed using the above framing system, and the section and material properties. A total of ten eigenvalues (natural frequencies) and eigenvectors (mode shapes) were determined in the analysis. Participating mass factors were also calculated for each mode. Visual inspection of the natural frequencies and mode shapes showed that the first three modes of vibration are rigid body modes of the deck. Natural frequencies of the three modes range from 1.8 Hz to 2.2 Hz. The first analytical mode is a longitudinal mode at 1.8 Hz. The second mode is a transverse mode with a natural frequency of 1.95 Hz. The third mode is a torsional mode with a natural frequency of 2.2 Hz. Modes four to eight were predominantly in-plane deck modes, while the remaining two modes were coupled in-plane deck and out-of-plane (torsional) deck modes. Natural frequency results of the first ten modes of vibration are shown in Table 3.2 and plots of the plan view of the first ten mode shapes are shown in Fig. 3.7. The participating mass results are also shown in Table 3.2. The results showed that only the first three modes of vibration contribute significantly to the horizontal vibrational response of the structure. Participating masses are virtually zero for the in-plane vibrational modes after the third mode. Participating masses for the out-of-plane vibrational modes are significant for the higher modes, but vertical vibrations are generally ignored in seismic analyses and thus were ignored for this study. STRUCTURAL MODELLING - VERTICAL PILED WHARF 34 The results of the participating mass study were used to determine the number and location of measurement stations that would most effectively measure the significant modes of vibration. The field testing of the structure is discussed in the following chapter. STRUCTURAL MODELLING - VERTICAL PILED WHARF 35 Element Section Properties Material Properties Area Moment of Inertia Polar Moment Thickness Modulus Shear Ix of Inertia of Elasticity Modulus M)_ (mA4) {itiM} fm A4) ... (mj. (MPa) (MPa} Pile 0.21 0.007 0.007 0.0136 n/a 5360000 2230000 Pile Cap 1.2 0.285 0.11 0.095 n/a 4320000 1630000 Deck Web 0.0256 0.052 65 0.002 n/a 5360000 2230000 Deck n/a n/a n/a n/a 0.25 5360000 n/a Diaphragm Springs n/a n/a n/a n/a n/a n/a n/a T A B L E 3.1 - P R E L I M I N A R Y A N A L Y T I C A L M O D E L : S E C T I O N A N D M A T E R I A L P R O P E R T I E S M o d e Natural Part ic ipat ing M a s s Frequency X-Dtrect ion Y-Di rec t ion Z-Direct ion (HZ) . (%) (%) (%) . 1 1.79 52.28 0.00 0.00 2 1.95 0.00 100.00 0.00 3 2.21 47.71 0.00 0.00 4 10.18 0.00 0.00 0.00 5 11.10 0.00 0.00 0.00 6 16.00 0.00 0.00 0.00 7 20.00 0.00 0.00 0.00 8 22.17 0.00 0.00 1.96 9 22.16 0.00 0.00 29.28 10 22.17 0.00 0.00 3.23 X = L O N G I T U D I N A L D I R E C T I O N Y = T R A N S V E R S E D I R E C T I O N Z = V E R T I C A L D I R E C T I O N T A B L E 3.2 - P R E L I M I N A R Y A N A L Y T I C A L M O D E L - F I R S T T E N N A T U R A L F R E Q U E N C I E S STRUCTURAL MODELLING - VERTICAL PILED WHARF 37 ®-OUTLINE OF "STAR WOLD" BULK CARRIER (41,000 DWT) 20 EQUAL SPACES = 152m WHARF DECK TRANSITION SPAN WHARF PLAN H.H.W.L. L.L.W.L. ® 14 EQUAL SPACES = 32m CAST IN ^ - CAST IN PRECAST PRESTRESSED / PLACE TOPPING / PLACE PILE CAP DOUBLE "T's APPROXIMATE EXISTING GROUND LINE •7.62m CAST IN PLACE SLAB 1 STEEL PIPE FENDER PILES 610mm OCTAGONAL PRECAST CONCRETE PILES GROUND LINE 1 I BULKHEAD WALL TYPICAL SECTION THROUGH WHARF F I G . 3.2 - V E R T I C A L P I L E D W H A R F - P L A N A N D E L E V A T I O N V I E W S STRUCTURAL MODELLING - VERTICAL PILED WHARF 40 43 CHAPTER 4 FIELD TESTING - VERTICAL PILED WHARF Field tests consisting of ambient vibration measurements and free vibration measurements were conducted on the VPW. The objective of the tests, the equipment used, the actual field testing, the data analysis and the measurement results are discussed in the following sections. A summary of the analytical tools for analysing ambient vibration measurements can be found in Chapter 2. 4.1 TEST OBJECTIVES Ambient vibration measurements were performed to determine the principal dynamic characteristics of the wharf. The purpose of the measurement program was to obtain data from which to refine the analytical computer model of the wharf. The specific objectives of the test were to: Determine the natural frequencies and mode shapes of the lower modes of the wharf apron. These frequencies and modes include the first longitudinal mode, the first transverse mode and the first torsional mode. FIELD TESTING - VERTICAL PILED WHARF 44 Identify natural frequencies and mode shapes associated with in-plane flexure of the wharf apron, including all classical beam bending modes which could be detected within the range of frequencies measured. Determine the equivalent modal damping factors of the structure. 4.2 TEST EQUIPMENT Prior to testing, selection of a suitable measuring equipement system was required. Initially, only a seismometer (velocity) based system was available at UBC's Civil Engineering Department, but in early 1992, an accelerometer based system became available. The two systems provided a good range of capabilities. It is generally recognized that seismometers are useful for measuring the ambient response of massive stiff structures with ambient accelerations in the order of 10"5 g to 10 3 g (Diehl 1991). Accelerometers are useful for measuring the ambient response of more flexible structures with ambient accelerations in the order of 10"4 g to 10"2 g (Diehl 1991). 4.2.1 S Y S T E M #1 - SEISMOMETER BASED System #1 included of the following equipment: Four Ranger SD217 Seismometers; Teledyne SC201 Signal Conditioner complete with Amplifiers and Low Pass Displacement, Velocity and Acceleration Filter Cards; Zonic+A/D Model 3625 Spectrum Analyser; and Shielded Cable in various lengths. Field Testing - Vertical Piled Wharf 45 In system #1, the seismometers were connected to the signal conditioner via the shielded cable. Analog vibration signals measured by the seismometers were sent to the signal conditioner where they were amplified and filtered. The conditioned signals were then passed through the spectrum analyser where they were converted to digital signals. The digital signals could then be further processed or saved to memory for future analysis. A schematic diagram of the equipment setup is shown in Fig. 4.1. System Specifications and Capabilities Since the seismometers are velocity based transducers, accelerations were computed from the velocity signals using acceleration filter cards to differentiate the signal prior to storing the data. The signal conditioner was capable of amplifying the original velocity signal by up to 400,000 times. Band Pass filters were set to screen the signal in the range from 0.1 Hz to 100 Hz. The spectrum analyser is capable of collecting data for one or two sensors simultaneously. The Spectrum analyser can sample data from two channels at a selected Nyquist frequency within the range from 1 Hz to 100 kHz, and has built in anti-aliasing filters that were set automatically according to the frequency range selected. The spectrum analyser is capable of sampling continuously up to 8192 data points in a single memory segment, after which the data segment had to be saved to memory. By using a multiple segment data collection option in the spectrum analyser, up to eight segments of 8192 data points (65,536 data points) could be obtained and stored. Data could then be saved to permanent memory in the spectrum analyser or transferred to floppy disks. Field Testing - Vertical Piled Wharf 45 Separate near real time visual displays for each of the two signals were available for the time history, the Fourier spectrum, or the frequency response of the sensor signals. Many more functions for the spectrum analyser were also available (Zonic+A/D Spectrum Analyser Users Manual), but not used in this study. Data collected using the spectrum analyser were in a format specifically for use by the spectrum analyser, and had to be converted to ASCII format using a separate DOS based personal computer program, called ADCNV. This program was supplied by the manufacturer of the spectrum analyser. It must be noted that this entire system previously included an analog data recorder. The recorder was not functional at the time of testing and therefore the recording capabilities of the system was reduced from its initial capability of four channels to the two channel capability of the spectrum analyser. This particular system with the original analog data recorder was used successfully to conduct ambient vibration measurements of six buildings in 1981 (Ngok 1982). Similar seismometers have been used successfully on a wide range of structures such as buildings and dams (Diehl 1991). 4.2.2 S Y S T E M U2 - A C C E L E R O M E T E R BASED This system is currently being used at UBC for ambient vibration testing and is continuously being improved and expanded. At the time of the wharf testing, system #2 consisted of the following equipment: Field Testing - Vertical Piled Wharf 47 6 - Kinemetrics Force-Balanced Accelerometers, FBA-11, ±0.25 g Accelerometers Kinemetrics Signal Conditioner with 4 - Kinemetrics AMI-3 Signal Conditioning Cards Keithley KDAC500 Analog to Digital (A/D) converter model 575 with an AMM2 board Compaq 286 Portable Data Collection Computer 486 Personal Computer for Data Analysis Zonic+A/D Model 3625 Spectrum Analyser 425 m of Shielded Cable (in different lengths) A V T E S T - Custom Written Data Collection Software for the A/D Converter An analog signal from the accelerometers, which is measured in volts and later converted to acceleration using a conversion factor, is sent to a signal conditioning card via the shielded cable. At the signal conditioner, the signal conditioning card amplifies and filters the signal which is then sent to the Analog to Digital (A/D) converter. Only one accelerometer could be connected to a given signal conditioning card at any one time. Therefore only four of the six accelerometers could be used for collecting data at any one time. At the time of this writing the system had been expanded to 10 accelerometers and 8 signal conditioning cards. At the A/D converter, the analog data are digitized and then sent to the data collection computer. The parameters for the digitization of the data are communicated to the A/D converter by the data collection software (AVTEST) (Felber 1993) in the data collection computer. The A/D converter is linked to the data collection computer through a special communication cable. The data are only temporarily stored in the data collection computer and, when possible, are transferred to a data storage/analysis computer that is networked to the data collection computer. At the data storage/analysis computer, data can be permanently stored and then analysed or saved to disk. Field Testing - Vertical Piled Wharf 48 In addition to the main data collection system explained above, the spectrum analyser can be connected to the system. The spectrum analyser is connected to the system in parallel at the output end of the signal conditioner. This connection can be made at the onset of testing to assist in determining the optimal amplification level at which to conduct the measurements. Once testing commences, the spectrum analyser can be used to periodically monitor the amplified and filtered signals or to conduct initial spectral analyses of the incoming signal. A schematic diagram of the equipment setup is shown in Fig. 4.2. System Specifications and Capabilities The signal conditioning cards housed in a self contained unit (signal conditioner) are capable of amplifying the original signals from the accelerometers by up to 2,000 times. Furthermore, the program A V T E S T allows the signals to be amplified by up to ten times for a total possible amplification of 20,000 times. As a result of the amplification capabilities and the accelerometer sensitivity, accelerations as low as 10~5 g can be measured with full resolution of the data. Anti-aliasing low pass filters on the signal conditioning cards allow the signal to be filtered at various discrete frequencies ranging from 0 Hz to 50 Hz. A switch on the signal conditioning card also allows the signal to be high pass filtered at either 0 Hz, 0.1 Hz or 1.0 Hz. The rate at which the data can be collected and stored in the field computer is governed by the Keithley A/D converter capabilities and its controlling software A V T E S T . The sampling rate is controlled by the user, through A V T E S T , by selecting a Nyquist frequency ranging anywhere from 1 Hz to 100 Hz. A sampling Nyquist frequency outside of the 1 Hz to 100 Hz range would require modification to A V T E S T . Fie Id Testing-Vertical Piled Wharf 49 The Keithley converter and A V T E S T software allow data to be sampled continuously at the frequency chosen, in 4096 data point segments for up to sixteen sensors concurrently (only four channels were employed in the test). Greater amounts of data can be accumulated by collecting data over a greater number of segments. The number of data segments, and thus the total number of readings, that can be measured is limited only by the storage capacity of the data collection computer. Data is collected in binary form in order to reduce memory requirements in the data collection computer. During data collection, near real time display of the time histories for all of the sensor signals are shown simultaneously on the same set of axes in the data collection computer monitor. While it is difficult to differentiate the signals for each of the sensors on the computer screen, the signals can be monitored for any peculiar motions. Further monitoring can be provided by connecting the spectrum analyser in parallel to any two sensors in the system. The capabilities of the spectrum analyser were discussed previously. The final form of the collected data is in a format that can be used by DOS based software specifically written for the purpose of analyzing data output from A V T E S T . The analysis software, and some of its capabilities, is explained briefly in Section 4.5.1. The analysis software has the ability to convert the binary data files from A V T E S T to ASCII format, so that the user can view the data or use it with other data analysis software. The data collection computer contains an 80286 - 12 MHz micro-processor and has an 80 Mb hard disk drive. The only restrictions placed on the capabilities of this computer were that the micro-processor must be at least an 80286 to run A V T E S T and that the available memory must Field Testing - Vertical Piled Wharf 50 be high enough to temporarily store enough data for one set of tests. The data analysis computer contains an 80486 - 33 MHz micro-processor and a 120 Mb hard disk drive. No restriction, other than available memory restrictions, were placed on the capabilities of this computer. To be most effective, the data analysis computer must be as fast as possible because analysis of the data is very computationally oriented. By using separate computers for data collection and data analysis, the system is capable of conducting data analysis concurrently with data collection and thereby producing significant results almost immediately. The ability to conduct on-site analysis is considered to be a significant advantage of this system. This system has been used successfully to carry out ambient vibration surveys (AVS) on several bridge structures (Felber 1992, Felber 1993 and Horyna 1995) and on a 30 storey high rise building (Schuster 1994). It has also been used to conduct an AVS on a dam (Kemp 1994). Similar accelerometer based systems have been used successfully by other investigators to conduct ambient vibration measurements of several bridges (Douglas (1986) and Nigbor). 4.2.3 EQUIPMENT SELECTION A comprehensive evaluation and testing program was carried out on both the seismometer based and accelerometer based systems to determine which one would be most suitable for the field testing. This evaluation was based on the specifications and capabilities of each system, as well as on testing of the two systems in the field and in the laboratory. The two systems were tested both concurrently and individually. Field Testing - Vertical Piled Wharf 51 Instruments were tested concurrently using two separate processes. The first process involved placing the sensors from both systems on a swing specifically designed and constructed for the purpose of testing the two systems. By exciting the swing to a known maximum displacement and period of motion, the accuracy of the measured results of each system could be compared. In addition, the clarity of the measured signals of the two systems under the same excitation conditions could be observed. Instruments were also tested simultaneously under strong motion excitation by using the laboratory earthquake shake table. Non-concurrent tests of the measuring systems were done using two different methods. The first method involved measuring the system response to input signals generated by a W A V E T E K III electronic wave generator. The second method involved conducting preliminary ambient vibration measurements on the actual test structure. Preliminary measurements on the structure were carried out in a total of four outings, twice for each system. Comparison of significant features of the two systems from the specifications and capabilities discussed above are shown in Table 4.1 below. Performance of the two systems under testing is summarized in Table 4.2 below. As can be inferred from the tables it is apparent that the performance and capabilities of system #2 are far superior to the performance and capabilities of system #1 under very low amplitude excitation. Under higher amplitude excitation, system #2 is still superior to that of system #1, but the difference in performance level is not quite as marked. The results under low amplitude excitation was somewhat unexpected since seismometers are generally regarded as being more suitable than accelerometers under very low excitation because of their higher sensitivity. Further Field Testing - Vertical Piled Wharf 52 research into the origin of system #1 revealed that the seismometers and signal conditioner are over 20 years old. Extensive testing in the laboratory showed that the signal conditioning cards have since deteriorated to the point where the phase relationship between the separate cards is excessive and therefore unreliable. Additionally, the filters in the conditioning cards are no longer effective at high gains. The noise in the signal at high gains is also very significant, even at relatively low gains. Repairing the system would have required a complete reconditioning of the seismometers and signal conditioning cards at a significant cost. In the final analysis, the advantages of system #2 far outweighed the advantages of system #1. Of greatest influence was the fact that age had deteriorated the performance of system #1 to the point that its advantages were lost. Also, the ability of system #2 to carry out measurements simultaneously with four sensors rather than only two, and the ability to carry out analysis simultaneously with data collection was an important advantage. System #2 was chosen as the instrumentation system for further testing. It should be emphasized that the results of the system evaluation may have been different if system #1 had been in better condition. One concern in choosing system #2 for testing was that the accelerometers might not be suitable for low amplitude measurements of very stiff structures such as the V P W structure. 4.3 FIELD TESTING Field tests using the accelerometer based system were conducted on April 24, 1992, June 11, 1992 and October 7, 1992. The first two outings were part of the preliminary field tests used to evaluate the system. The last outing produced ambient vibration and free vibration data from Field Testing - Vertical Piled Wharf 53 which detailed analyses were performed. The particulars of the field testing are explained in the following sections. 4.3.1 TESTING RESTRICTIONS Few restrictions were placed on the testing of the structure by the wharf owners. The only restrictions, other than normal safety regulations, were that operations at the terminal must not be obstructed in any way. Additionally, absolutely no damage to the structure was permitted. Operations at the wharf are such that all activity occurs on the wharf when a ship is berthed. When a ship is in the berth, loading and unloading of cargo is continuous until the ship leaves. The restriction of not impeding operations could easily be met by testing on days where no ship was present at the wharf. There is a high berth occupancy at this particular wharf and thus, to avoid impeding operations, close contact was maintained with the wharf owners to determine available windows for testing. Testing required a window of one full day, but one day prior notice was necessary to mobilize the test equipment. The nature of the equipment used for the ambient and free vibration measurements is such that little damage is possible to the test structure. The only problem encountered due to the no damage restriction was that the method normally used to attach the accelerometers to the structure could not be used. In previous testing on other structures, sensors were bolted onto anchor plates which were in turn attached to the structure using concrete anchor bolts. In this case, the anchor bolts would potentially damage the concrete deck surface and thus this method Field Testing - Vertical Piled Wharf 54 of attaching the accelerometers was abandoned. The anchor plates were instead attached to the structure using quick setting plaster. 4.3.2 TEST CONDITIONS Factors affecting the test conditions at the wharf fall into two categories, those affecting the process of physically obtaining the data and those affecting the quality of the data. The test conditions were very good for physically obtaining the data. The nature of the use of the wharf meant that it was completely clear of obstructions for the duration of the testing. No restrictions or difficulties were encountered in laying out the instrumentation equipment. Test conditions affecting the quality of the data were not as favourable. On the day of final testing weather was cold and very windy. While the wind was beneficial in exciting the wharf, it could also have potentially corrupted the data. In earlier tests, drifting of the measured signal was encountered. This drifting can be seen in the sample data shown in Fig. 4.3. The drifting is thought to be due to the wind moving the entire plate assembly because the plaster had not set properly. While the same drifting was not encountered during final testing (due to improved quality control in placing the instruments), there is the possibility that the wind may have affected the readings. This is especially true at the high gains used for the tests. Another factor affecting the quality of the data was that there was considerable railcar and forklift activity in the warehouse area of the terminal some 150 m away from the wharf. The activity at the warehouse caused the sensors to saturate occasionally, thus corrupting the signals. An additional potential problem was that the transducers were measuring acceleration levels at the limit of their capabilities. Field Testing - Vertical Piled Wharf 55 4.3.3 D A T A COLLECTION SOFTWARE/HARDWARE SETUP The data collection software and hardware were set up to obtain and save data for four accelerometers. The amplification levels and filters were fixed on the hardware to provide the maximum possible resolution on the measured data. A sufficient frequency range to capture all the desired modes of vibration in a reasonable total measurement time interval was established in the software. To satisfy the above requirements for the ambient vibration measurements, data were collected with a Nyquist frequency of 20 Hz. A total of 32,768 points was gathered in 8 - 4096 point segments over a period of approximately 14 minutes. Anti-aliasing, low pass filters were set at 12.5 Hz. Amplifier cards were set at their maximum amplification. 4.3.4 A M B I E N T VIBRATION TESTS The data collection system consisting of the AID converter, the signal conditioner, the data collection computer and the spectrum analyser were located at a control station in a vehicle situated in a central location on the wharf. Ideally the control station would have been located off the structure, but the lengths of available shielded cable required to connect the sensors to the data collection system were too short to allow for this. The number of test setups and the corresponding measurement station locations were chosen to measure best the natural frequencies and mode shapes mentioned in the objectives. In determining the number of test setups, the maximum feasible number of test setups was Field Testing - Vertical Piled Wharf 55 scheduled while accounting for the requirement of completing all the measurements in one day. It is important to note that analysis of results is increasingly simplified and results increasingly more reliable with increased numbers of measurement locations. Final sensor placements on the structure were chosen to be at the 1/8 points, the 1/4 points, the 3/8 points and the mid point along the west side of the wharf. The four comers were also measured along with an additional station at the 1/3 point along the south edge of the wharf. One free field station was located 30 m from the east edge of the wharf intersecting a line extended from the middle of the wharf in the longitudinal (north-south) direction. Both principal directions, transverse (east-west) and longitudinal were represented in the sensor lay out. Measurements in the vertical direction were not taken. All of the sensor placement locations and measurement directions are shown in Fig. 4.4. A total of eight test setups were used to complete all of the desired measurements. One pair of sensors was located at the southwest corner of the wharf to act as a reference station. This reference station was kept in place for each set of measurements to act as a constant between tests. One reference sensor was oriented to measure in the longitudinal direction and the other reference station was oriented to measure in the transverse direction. The southwest corner of the wharf was chosen as the reference station because preliminary analytical modelling showed that this location had significant motion in all the modes of vibration of interest. Two of the remaining four sensors were used to measure vibrations at two of the desired measurement locations. Meanwhile, the final two sensors were positioned in preparation for the next set of measurements. This "leap frogging" of sensors helped to reduce down time between test setups. Table 4.3 identifies the sensor number and direction measured in each setup. Field Testing - Vertical Piled Wharf 57 4.3.5 FREE VIBRATION TESTS Free vibration tests were carried out using the same measurement system used for the ambient vibration tests. The control station was not moved from the original position and the sensor placement stations were the same as for the ambient vibration tests. Free vibrations were imposed on the structure by a tug boat pushing against the structure and then letting go. Numerous free vibration tests were carried out with the tug pushing at different locations and with sensors located at different stations along the wharf. These forced inputs were applied at locations chosen to induce vibration in each of the first three principal modes. These locations are shown in Fig. 4.5. Table 4.4 defines the arrangement of the sensors for each setup. 4.4 M E A S U R E D D A T A The ambient vibration and impact vibration data are discussed in the following sections. 4.4.1 A M B I E N T VIBRATION D A T A The data measured under the ambient vibration tests was saved to data files in binary format. A plot of a portion of a typical time history is shown in Fig. 4.6. A larger sample of the time histories can be found in Appendix A. Visual inspection of each of the time histories does not show any obvious trends; abnormal characteristics also are not evident. It can be noted, however, that the maximum acceleration level is of the order of 0.05 mg, and the vibration frequency is quite high. Although, the maximum acceleration recorded is lower than the Field Testing - Vertical Piled Wharf 58 acceleration level normally considered acceptable for the accelerometers used, the measured signals appear to be clear and free of noise. 4.4.2 FREE VIBRATION D A T A A typical free vibration (push/release) time history low pass filtered at 50 Hz is shown in Fig. 4.7. The same signal low pass filtered at 10 Hz is shown in Fig. 4.8. The plotted time histories of a set of the vibrations low pass filtered at 50 Hz are shown in Appendix B. Unlike the ambient vibration time histories, the free vibration time histories do show obvious trends. As can be seen in Appendix B, the natural frequency excited by the free vibrations seems to be the same frequency in all cases and the damping seems to be quite high in all cases. Note that the maximum acceleration level is in the order of 0.25 g. The data obtained in the measurements were further analysed in the frequency domain to obtain more information. 4.5 D A T A A N A L Y S I S Detailed analysis of the data obtained in the field tests is discussed in the following sections. The analysis packages employed and the manner in which they were used to obtain results are summarized . The natural frequencies and mode shapes of the structure were determined from the ambient vibration data. The free vibration data provided information on the frequencies of the excited modes and the modal damping values of the structure. Field Testing - Vertical Piled Wharf 59 4.5.1 COMPUTER PROGRAMS The computer programs used to analyse and animate the vibration data were ULTRA and VISUAL (Felber 1993). ULTRA is an analysis package that includes the capability of generating transfer functions and which was written specifically for analysing ambient vibration data. VISUAL is a 3-D animation package which accepts input files from analysis packages, such as ULTRA. ULTRA uses the information from pairs of independent data files and is capable of conducting frequency domain analysis on the data. Some of the capabilities of the program include the following: signal conditioning filtering providing base line adjustments adding and subtracting multiples of two signals calculating power spectral density of an individual signal calculating the frequency response between two signals calculating the cross spectral density of two signals calculating the phase angle between two signals calculating the coherence between two signals calculating the modal ratios between two signals, storing results to file plotting results Field Testing - Vertical Piled Wharf 5Q Detailed explanations of each of these capabilities, as well as further capabilities of the program, can be found in the ULTRA users manual (Felber 1993). As mentioned above, VISUAL accepts input files from analysis packages which generate transfer functions. In this study, input files to VISUAL were "modal ratio" files created using ULTRA. The data points in each modal ratio file represent the frequency response function between a station sensor signal and a reference sensor signal, with the frequency response function modified using both a phase and a coherence filter. The phase and coherence filters multiply the frequency response function by a factor of 1 when two, user defined, criteria (one for phase and one for coherence) are met and by a factor of 0 when either the phase or coherence criteria are not met. These files are loaded in conjunction with a user created structure geometry file and a user created file noting the locations and orientation for each one of the modal ratio files. With the loaded data, VISUAL animates the structure at discrete frequency steps. This program is extremely useful to visually identify frequencies which are associated with a mode shape. A description of the techniques used to determine the required structure dynamic properties for both the ambient vibration data and the free vibration data is given below. 4.5.2 A M B I E N T VIBRATION D A T A A N A L Y S I S ULTRA was used to make preliminary estimates of natural frequencies of the structure. Power spectral densities (PSD) for the various acceleration time histories were calculated, exported to a spreadsheet, averaged and normalized. Averaging was accomplished by summing the PSD's calculated from each of the measured signals at each frequency and dividing by the total number Field Testing-Vertical Piled Wharf 51 of signals, while normalizing was accomplished by dividing the averaged PSD by the maximum ordinate of the averaged PSD. The resulting plot, an averaged normalized power spectral density (ANPSD), was inspected for potential resonance frequencies. Peaks in the averaged spectra corresponded to potential resonance frequencies, which in turn could be related to the natural frequencies of the structure. Once potential natural frequencies of the structure were identified, a detailed analysis was conducted. The detailed analysis was used to determine which spectral peaks were most likely to correspond to natural frequencies of the structure. Final determination of natural frequencies involved comparing the phase angle and coherence functions between the signals of the various sensors and the reference sensor. The criteria for final determination of natural frequencies for the structure were as follows: coherence value close to unity for all signals in relation to the reference signal; and phase angle close to 0° for relative movement in-phase and 180° for relative movement out-of-phase between a station sensor signal and the reference sensor signal. The data reduced using ULTRA were loaded into VISUAL. Using VISUAL, relative motions at potential natural frequencies were viewed to note if motions corresponded to motions from expected mode shapes. Frequencies which corresponded to peaks in the averaged Fourier spectra and to well defined mode shapes were deemed to represent natural frequencies of the structure. Field Testing - Vertical Piled Wharf 62 4.5.3 FREE VIBRATION D A T A Free vibration data were also processed and analysed with the aid of ULTRA. The unprocessed time histories were conditioned to obtain smooth free vibration signals and a PSD of each individual signal was calculated and used to determined the natural frequency of vibration of the structure. Damping for the measured natural frequency was determined in the analysis by observing the magnitude of adjacent peaks in the time history and calculating the damping using the log decrement method (Clough and Penzien 1975). The results of the analyses are given in the following sections. 4.6 A N A L Y S I S RESULTS Using the test measurement data and data analysis techniques explained above, one overall site frequency and six structure natural frequencies and mode shapes were determined. In addition, damping, for what is believed to be the measured third mode (east-west translation), was determined. 4.6.1 A M B I E N T VIBRATION TEST RESULTS The first six mode shapes of the structure corresponded closely to mode shapes determined from previous computer analysis. The measured natural frequencies differed significantly from those initially predicted by the finite element model. Field Testing - Vertical Piled Wharf 53 Fig. 4.9 shows the A N P S D for the experimentally measured transverse signals. Significant peaks in the A N P S D correspond to the natural frequencies of the structure. Fig. 4.10 shows the A N P S D for the experimentally measured longitudinal signals. The PSD for all the signals along the length of the wharf are plotted in Fig. 4.11. The experimental results identified structure natural frequencies ranging from the longitudinal rigid body mode at a natural frequency of 2.6 Hz to the third deck bending mode at a frequency of 18.9 Hz. A more complete discussion of the results of the analysis of the test data follows. 4.6.1.1 O V E R A L L SITE M O D E The overall motion of the wharf is shown in Fig. 4.12. As can be seen, the entire wharf structure moves in-phase with the free field station and with a similar amplitude of motion. 4.6.1.2 RIGID BODY MODES Three rigid body modes of the wharf apron identified by the test measurements were longitudinal (north-south) translation at 2.6 Hz, torsion at 3.8 Hz, and transverse (east-west) translation at 3.9 Hz. The longitudinal translational mode exhibits torsional coupling due to the difference in pile lengths in the wharf cross section from east to west. In addition to the rigid body motion in the first three modes, some flexing in the deck can be detected. In particular, the east-west translation mode shows considerable flexing of the deck. This deck flexing can be attributed to the relatively large stiffness of the piled substructure in relation to the stiffness of the wharf deck. The three rigid body modes of the wharf apron are shown in Fig. 4.13. Field Testing - Vertical Piled Wharf 54 4.6.1.3 F L E X U R A L MODES The first three flexural modes of the wharf identified through the test measurements corresponded to natural frequencies of 7.9 Hz, 11.3 Hz and 18.9 Hz. The mode shapes, shown in Fig. 4.14, compared very well to the classical mode shapes of a flexural beam member. 4.6.2 FREE VIBRATION TEST RESULTS Power spectral densities (PSD) for the forced vibration acceleration time histories show that the principal mode of vibration excited in the free vibration tests was the east-west translation mode. A typical forced vibration PSD is shown in Fig. 4.15. The average damping value determined from several tests using the log decrement method indicated damping to be in the order of 8% for this mode. Damping for the longitudinal mode of vibration was not determined because the tug boat that provided the excitation force could not access locations where it could provide a longitudinal input motion without damaging the structure. 4.7 DISCUSSION OF RESULTS The initial concern that the sensitivity of the accelerometer based system might not be suitable for adequately detecting the ambient vibration measurements did not seem to affect the results of the study. It should be noted, however, that the data were difficult to interpret. This may have been associated with the unsuitability of the accelerometers for measuring this particular structure Field Testing - Vertical Piled Wharf 65 or the inappropriateness of the method of analysis employed. Regardless, the final results from the testing program were not initially foreseen, but upon hindsight could be expected. Measured natural frequencies of the structure differed significantly from the natural frequencies determined using the analytical model. The reasons for the differences became obvious once the results were obtained. As expected, the test frequencies resulting from low amplitude vibrations were far higher than the analytical frequencies resulting from assumptions of high amplitude vibrations. The results from the ambient vibration testing gives an upper bound for the natural frequencies of the structure. The measured mode shapes of the structure compared quite well with the analytical mode shapes, except that the first torsional mode and the first translational mode were unexpectedly reversed in order. This result may have been due to an unexpected distribution of mass or stiffness in the structural system. Damping for the translational mode of vibration is reasonably consistent with, but slightly higher than, published values for the third mode of vibration in different concrete structures. 4.8 SUMMARY Field testing was conducted on a vertical piled wharf. The objectives of the testing were to determine the principal dynamic characteristics of the wharf. These characteristics included the natural frequencies and mode shapes of the structure as well as the damping of the structure. Field Testing - Vertical Piled Wharf 66 An accelerometer based instrumentation system was chosen for use in conducting the test. The tests consisted of ambient vibration and free vibration tests. From the tests, seven natural frequencies and mode shapes of the site and the structure were measured. The natural frequencies and mode shapes corresponded to the following: overall site mode 0.3 Hz first longitudinal mode 2.6 Hz first torsional mode 3.8 Hz first transverse mode 3.9 Hz first deck bending mode 7.9 Hz second deck bending mode 11.3 Hz third deck bending mode 18.9 Hz Damping of the first transverse mode of vibration was determined to be 8 % by the log decrement method (Clough and Penzien 1975). Field Testing - Vertical Piled Wharf 57 TABLE 4.1 - COMPARISON OF SIGNIFICANT F E A T U R E S OF THE EQUIPMENT S Y S T E M S SYSTEM CAPABILITIES AND SPECIFICATIONS #1 - Seismometer Based velocity based transducers amplification of signal by up to 400,000 times data sampling range from 1 Hz to 100 Hz automatic setting of anti-aliasing filters four transducers available sampling with two sensors concurrently for up to 65,536 data points per test setup data transferable to floppy disk data transferrable to ASCII format near real time display for time history, fourier spectrum, and frequency response spectrum for two channels simultaneously #2 - Accelerometer Based force balanced accelerometers amplification of signal by up to 20,000 times data sampling range from 1 Hz to 100 Hz manual setting of anti-aliasing filters at various discrete frequencies ranging from 0 Hz to 50 Hz six transducers available sampling with four sensors concurrently number of data points limited only by storage capacity of the data collection computer data transferrable to floppy disk data transferrable to ASCII format near real time display for time histories of all four sensors simultaneously connection of spectrum analyser to system in parallel to receive all the associated capabilities of.the^spectrum analyser on-site data analysis capabilities within minutes after receiving data Field Testing - Vertical Piled Wharf 68 T A B L E 4.2 - S U M M A R Y O F M E A S U R E M E N T S Y S T E M P E R F O R M A N C E U N D E R T E S T I N G T E S T UNIT S Y S T E M #1 S Y S T E M #2 C lean l iness of Signal L o w Ampl i tude High Appl i tude high signal to no ise ratio low signal to no ise ratio low s ignal to noise ratio low signal to no ise ratio P h a s e Be tween S igna ls Measu red by Two Transducers Concurrent ly unrel iable reliable C o h e r e n c e Between S igna ls M e a s u r e d by T w o Transducers Concurrent ly unrel iable reliable E a s e of Operat ion of Sys tem easy easy Portability of Sys tem extremely portable extremely portable Field Testing - Vertical Piled Wharf QQ Filter (Hz) o o o o o o o o Attenuation Factor (dB) ! o o o o o o o o o - " 1 - 1 — 1 — 1 — - I - 1 - 1 6 - l r - L _ | _ i _ _ t r - r -5 1 1 1 1 1 1 1 1 c o c o c o c o n c o c o N o r - r - r - r - r - r - r - r -n c o r o r o c o c o c O N File Name v - C N C O - ^ - L n c D N - O O o o o o o o o o <<<<<<<< oooooooo c o c o c o c o c o c o c o c o Set Up Number T - C N C O ^ i n c O N C O o on o Q. ZD h-LU CO h-CO UJ 2 gg > Q UJ 0 or: O u_ 1 LU _J OQ Force Input Location CO if) lO N N Filter (Hz) o o o o o L O I O I O I O LO Attenuation Factor (dB) ! M-L O L O L O LO LO o ZZ CD ZZ 0 ) CO 5 h - 1— h - 1— r— L O 1^- CO O 1— I— i— 1— i— L O L O "^T L O I— 6 1— 1— —1 1— I— CO CO CO CO ^ " CD £ Z — T - CM co -^ r L O o o o o o O O O O O CO CO CO CO CO T - c\i co ^  m Field Testing - Vertical Piled Wharf 70 01 o O CO CO z Z 111 UJ CO CO CZD UJ CO CO 0_ H UJ CO I— z UJ z> o UJ UJ I-CO >-CO 2 1 G _ J o I-o CD Field Testing - Vertical Piled Wharf 71 or o O CO CO z z UJ UJ CO CO o b or co w 0 o_ 5 ° 1 o I - * UJ co CO 0_ Z) r— UJ CO r-Z UJ Q_ Z> o UJ CN UJ r— CO >-CO u_ 0 § Q _i O \-g o_ 1 CM CD Field Testing - Vertical Piled Wharf 72 [6iu] uoiiBjaieoov Field Testing - Vertical Piled Wharf 75 [6lU] U0!1BJ9 |800V Field Testing - Vertical Piled Wharf 75 CM Field Testing - Vertical Piled Wharf JJ CM Field Testing - Vertical Piled Wharf 78 O CM Field Testing - Vertical Piled Wharf 79 FIELD TESTING - VERTICAL PILED WHARF 80 120 100 80 60 40 20 0 STATION 3 North i 1-"- -wmm J " . IBM | | STA TION4 North p>-g|g|||j | STATION 6 North i * 1 j — —j J Jl — 1.S 10 12 14 16 Frequency [Hz] 18 20 FIG. 4.11A - PSD's FOR WHARF AVS RECORDS - TRANSVERSE STATIONS 3 - 6 Field Testing - Vertical Piled Wharf 81 STA" riON 7 North mm ! 120 100 80 60 40 20 0 STATION 10 North 1 ! *i I I i i 1.1-. 1 L 10 12 Frequency [Hz] F I G . 4.11 B - P S D ' s F O R W H A R F A V S R E C O R D S - T R A N S V E R S E S T A T I O N S 7 - 1 1 Field Testing - Vertical Piled Wharf 82 S T ; M I O N 2 111111 1 1 North i Uj ,1 S T A T I O N 3 1 FT7 North l i — i V i y (A C <D Q I o Q. CO S T A T I O N 11 North £ / i 1 V i MM0*-* y est, I S T £ J I O N 11 > North 1 12 20 Frequency [Hz] FIG. 4 .11C - P S D ' s F O R W H A R F A V S R E C O R D S - L O N G I T U D I N A L S I G N A L S Field Testing - Vertical Piled Wharf 83 E23 F R E E F I E L D S T A T I O N F IG. 4.12 - A M B I E N T V I B R A T I O N S U R V E Y - O V E R A L L S I T E M O D E Field Testing - Vertical Piled Wharf 84 E x p e r i m e n t a l M o d e 1 F r e q u e n c y = 2 . 6 3 H z MODE SHAPE UNDEFORMED SHAPE ^^ ^^ ^^ ^^ ^^ ^^ ^^  ^ ^^^^^^^ FREE FIELD STATION E x p e r i m e n t a l M o d e 2 F r e q u e n c y = 3 . 8 0 H z MODE SHAPE UNDEFORMED SHAPE FREE FIELD STATION E x p e r i m e n t a l M o d e 3 F r e q u e n c y = 3 . 9 0 H z v UNDEFORMED SHAPE m FREE FIELD STATION FIGURE 4.13 - AMBIENT VIBRATION - RIGID BODY MODES Field Testing - Vertical Piled Wharf 85 E x p e r i m e n t a l M o d e 4 FREE FIELD STATION E x p e r i m e n t a l M o d e 5 F r e q u e n c y = 11.3 H z FREE FIELD STATION E x p e r i m e n t a l M o d e 6 I FREE FIELD STATION FIGURE 4.14 - AMBIENT VIBRATION - FLEXURAL MODES 98 peg/ft paild |BO!^3A - Bursal R a j 87 CHAPTER 5 CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE A correlative analysis was performed on the test structure to match computer analytical model results with experimentally obtained results. The correlated model was used to conduct a parametric study to determine the sensitivity of the dynamic characteristics of the wharf to key parameters in the structural system. 5.1 CORRELATIVE ANALYSIS Results from the field data discussed in Chapter 4 were used to modify the analytical model. Key parameters such as pile lengths, pile and deck stiffness, bearing restraint stiffness and structure mass were varied in the model such that a new model matching the experimental data could best be developed. Important dynamic characteristics which were matched were the natural frequencies and mode shapes of the system. Natural frequencies values were matched numerically, while mode shapes were matched using the Modal Assurance Criterion (MAC) analysis first discussed in Chapter 2. 5.1.1 ANALYTICAL RESULTS VS EXPERIMENTAL RESULTS Comparison between the analytical results first introduced in Chapter 3 and the experimental results show that the mode shapes determined experimentally match the analytical mode shapes CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 88 quite well. The first ten analytical mode shapes were shown previously in Fig. 3.7 and the first six experimental mode shapes (from ambient vibrations) were shown previously in Fig. 4.13 and Fig. 4.14. The first three analytical and measured mode shapes are rigid body modes of the structure, while the remaining analytical and measured modes are deck bending modes. From these figures it should be noted that, analytical modes 5 and 7 were not detected in the field measurements, modes 2 and 3 (experimental) are interchanged relative to their analytical order, and analytical modes 8, 9 and 10 are all similar in the in-plane (plan) view. It was expected that some analytical modes would not be detected in the field measurements: Factors such as the actual frequency content of the excitation and its spatial distribution, the location of the sensors, the sampling rate, and limitations in data analysis capabilities restrict the modes that can be detected with confidence. The above noted order interchange of modes 2 and 3 can possibly be attributed to unaccounted for distributions in mass and/or stiffness of the structure. It has been noted in Chapter 3 that soil deposition from the river to the north end of the structure favours the north end of the wharf and that dredging was required, at that end, prior to construction. It is possible that the soil to the north is more competent or at a higher elevation than the theoretical design elevation. This may cause uneven soil stiffness from one end of the structure to the other, favouring torsional modes of vibration and thus potentially interchanging the order of the torsional and translational modes. The similarity of modes 8, 9 and 10 (by computer analysis) in plan view can be attributed to the fact that the vertical component in each of the mode shapes is not displayed. Each of the three modes has a different vertical movement. Because the vertical vibrations of the structure were not measured, there is no way of differentiating between the three mode shapes in the measured results. These three mode CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 89 shapes also have similar natural frequencies; thus, even if each of the mode shapes was measured, analysis capabilities and confidence levels make it difficult to differentiate between modes. As discussed above, not all of the analytical and measured mode shapes match in the same order. To ensure proper comparisons of mode shapes, an analysis was performed to determine which analytical mode shape best corresponded to each experimental mode. Visual inspection of the mode shapes was sufficient to make preliminary matches for the first five mode shapes, but further techniques were required to match the sixth mode. A MAC analysis was used to make final matches for the mode shapes. The MAC analysis was carried out using the computer program MAC developed at UBC by Dr. Carlos Ventura (Ventura 1993). M A C utilizes displacement output from VISUAL and SAP90 for the direct analysis of mode shapes. The computer program MAC extracts the pertinent information from the VISUAL output files and the SAP90 output files and performs a correlative analysis between mode shapes. Output from the program MAC is in the form of a matrix of numbers which ranks the correlation between each analytically obtained mode shape to each experimentally obtained mode shape. The larger the MAC value between the analytical and experimental mode shapes, the higher the correlation between the two mode shapes. A M A C value of unity indicates a perfect match between two mode shapes. C O R R E L A T I O N A N A L Y S I S A N D P A R A M E T R I C S T U D Y O F T H E T E S T S T R U C T U R E 90 The results of the MAC analysis between the measured mode shapes and the preliminary analytical model are shown in Table 5.1. The largest MAC value for each experimental mode is highlighted in the table. As can be seen in Table 5.1, the MAC analysis shows results consistent with the preliminary visual inspection of the first five modes (Figs. 3.7 and 4.13 - 4.14); measured modes 1, 2, 3, 4 and 5 match analytical modes 1, 3, 2, 4 and 6. Experimentally measured mode 6 is best matched by analytical mode 8, although modes 9 and 10 are visually similar to mode 8. The results of Table 5.1 are shown graphically in a bar graph in Fig. 5.1. The natural frequencies to be compared were those for the mode shapes determined from the MAC analysis. Table 5.2 summarizes the analytical mode which best matches each experimental mode, indicates the value of the natural frequencies for each mode and the percentage difference with respect to the experimental mode. Only the modes that were detected experimentally are presented in the table. The results of the natural frequency comparison showed that the preliminary results did not match the experimental results very well, with analytical natural frequencies in error, anywhere from 13% to 51%. The first three measured frequencies were higher than the frequencies predicted analytically, indicating that the modelled frame system was either too soft or the modelled mass too high. The last three measured modes were lower than those predicted using the analytical model, indicating that the deck was modelled as being too stiff. The differences in the analytical and measured natural frequencies, whether the frequencies were too high or too low, were considerations for modifying the analytical model to correlate it with the measured results. C O R R E L A T I O N A N A L Y S I S A N D P A R A M E T R I C S T U D Y O F T H E T E S T S T R U C T U R E 91 5.1.2 C O R R E L A T I O N O F E X P E R I M E N T A L R E S U L T S T O T H E A N A L Y T I C A L M O D E L Several key parameters were modified to create an analytical model which best fitted the measured data. Parameters deemed necessary to modify were the individual member stiffness, the pile lengths and the structure mass and mass distribution. In addition, springs were added to the model at the east side of the wharf to represent the frictional restraint caused by the transition section of the wharf prior to the sliding bearings taking effect. Several combinations of modifications were performed before developing a best match model. A MAC analysis using the final calibrated model is presented in Table 5.3. Again the value corresponding to the largest MAC value is highlighted. The results from Table 5.3 are shown graphically in a bar graph in Fig. 5.2. Note that the analytical modes that best match the measured modes have not changed. A comparison of the measured natural frequencies and the modified analytical model natural frequencies is presented in Table 5.4. The agreement is much better, with the differences ranging from a low of 1% to a high of 26%. Table 5.5 gives a summary of the key analytical model parameter values for the calibrated model. Parameters which were modified are highlighted in the table. A more detailed discussion of each of the model parameters is given in the following sections. 5.1.2.1 P I L E F I X I T Y L O C A T I O N The preliminary model used a pile fixity location determined using the first method for ground line determination in Chapter 2 (ground line #1). Ground line #1 is better suited for a strong motion CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 9 2 model than ground line #2; therefore, the fixity depth was likely over-estimated for the low vibration levels measured. Because the fixity depth was anticipated to be longer than the actual condition, the measured natural frequencies for the rigid body modes were lower than those determined analytically. For the low level of vibrations encountered on the ambient vibration survey (0.025 mg), the level of apparent fixity was likely very close to the mudline. The final fixity point for the piles was considered to be at the assumed mud line. 5.1.2.2 D E C K STIFFNESS The in-plane deck stiffness in the preliminary model assumed that the deck bends as a single supported beam. This is a very unlikely condition since some buckling out-of-plane can be expected for a 30 m wide by 500 mm thick deck. Out-of-plane buckling of the deck would result in a lower effective in-plane deck stiffness which in turn would result in lower natural frequencies for the deck bending modes. The in-plane deck stiffness in the final model was adjusted to best match the deck bending modes computed from experimental data. 5.1.2.3 S T R U C T U R E M A S S A N D M A S S DISTRIBUTION The structure mass in the preliminary model was calculated assuming the member dimensions given in the construction drawings for the structure. The mass of cast-in-place concrete was assumed to be 24 kg/m 3 while the mass of precast concrete was assumed to be 25.7 kg/m 3. The mass of all the deck level members, including the topping, precast double tee units and pile caps CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 93 was assumed to be uniformly distributed on a plane at the centre of mass of the deck. The mass of each pile was taken to be uniformly distributed along the entire length of the pile. The total mass of the structure was unchanged in the final model, but the distribution of the mass at deck level was changed to better match the actual distribution of mass on the actual structure. No change was made to the total mass of the structure because, the construction drawings would be the best guess for the member dimensions. Also the assumed mass densities of the concrete are well established numbers. The assumption of the mass being uniformly distributed over the entire deck was initially considered to be adequate because the stiffness of the deck would serve to distribute any uneven mass throughout the deck. This assumption is acceptable for the rigid body modes of vibration, but the deck bending modes are dependent on the distribution of mass on the deck. The final model used a more accurate distribution of mass at deck level with more mass applied at the perimeter of the structure. 5.1.2.4 SLIDING BEARING RESTRAINT The low friction sliding rubber bearing under the transition slab at the east side of the wharf, is meant to isolate the wharf from shore. Under strong motion conditions, the bearing pads slide and thus effectively separate the structure from shore. Any friction that may occur between the bearings creates an effective restraint on the wharf and thus create a stiffer system at low amplitudes of vibration. The initial model considered the wharf to be isolated from the shore. No effect was considered for any friction. In reality, there would be some friction between the bearing pads. Adding friction J CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 94 to the model would result in higher natural frequencies for the analytical model. This would be consistent with the requirements for matching the analytical model to the experimental results. The manufacturers specifications show that a static coefficient of friction between the surfaces of the bearing pads of 0.2 is expected. The final model incorporates some stiffness consideration for the sliding bearings. This consideration is in the form of a spring constant along the shore edge of the wharf. The initial spring constant was estimated assuming that the sliding bearings are firmly fixed and any movement in the bearings are due to shear deformations in the rubber making up the bearings. Linear strain deformations of the bearing pad over the thickness of the pad were assumed. An initial estimate for the spring constant was 74 kN/m for a 13 mm thick pad. The spring stiffness used in the final model was determined by adjusting the initial spring constant estimate until the measured results could be matched. 5.2 PARAMETRIC STUDY A parametric study was performed using the model best matching the measured data as a reference (control) model. The results that were monitored were the natural frequency and the participating mass of the structural system. The natural frequency is the basis for all dynamic calculations and the participating mass is an indicator of how important a mode of vibration is to the total vibration response. CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 95 5.2.1 OBJECTIVES The objective of the parametric study was to determine the significance of the different parameters in modelling the dynamic characteristics of the wharf. 5.2.2 STUDY PARAMETERS Parameters deemed to be significant in modelling the structure were the following: pile length to fixity; pile section stiffness properties; structure mass; and bearing pad spring stiffness. A wide range of values were used for each parameter to ensure that all possible values of each parameter were bounded. Each parameter is discussed in greater detail in the following sections. 5.2.2.1 PILE L E N G T H TO FIXITY Many different methods are used in practice to determine the depth at which the piles can be assumed to be fixed in the soil. The final modified model uses a depth to fixity at the design mudline (ground line #1). This corresponds to a lower bound value that would be used for small amplitude vibration as found in ambient vibrations. Four different models for depth to fixity (see CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 96 Chapter 2) were studied to determine variations in the dynamic characteristics. These models included: Model F1 - fixity at ground line #1 plus five pile diameters; Model F2 - fixity at the mud line assumed for steep slopes (ground line #2); Model F3 - fixity at ground line #2 plus five pile diameters; and Model F4 - fixity at the depth calculated using Davisson's method. All of the above four models are suitable for high levels of vibration. Five pile diameters was chosen as the depth to fixity below the ground line because it was assumed to be an average value for soils of the type found at the field measurement site. In using model F4, a modulus of horizontal subgrade reaction of ten was chosen to represent a soil of average stiffness. 5.2.2.2 PILE A N D PILE C A P STIFFNESS The stiffness of the framing members were varied to represent the difference in member stiffness under cracked and uncracked conditions. The base model assumed an uncracked condition thus using gross section properties to model framing members. This corresponds to a low level of vibration condition as found in ambient vibrations. Four different pile and pile cap stiffness were studied. These values were calculated by changing the effective moment of inertia (leff) as follows: Model S1 - l e t f = % lg; Model S2 - l e f f = 54 I • CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 97 Model S3 - l e f f = 11/2 l g ; and Model S4 - l e f f = 2 lg. where l g is the gross member moment of inertia. The first two variations in stiffness account for reduced stiffness under high amplitude motions while the final two variations account for unexpected additional member stiffness. 5.2.2.3 S T R U C T U R E M A S S The mass of the structure was varied to show potential differences between the as-built structure and the construction drawings, and deviations of the material densities from recognized values. The initial matched model represented the condition where the dimensions were those in the construction drawings and the material densities those published as standard values. The two different structure mass values below were studied: Model M1 - Mass = 1 / 2 M< Model M2 - Mass = 2 M; where M; is the initial estimate of the mass. The two different values represent extreme upper and lower bounds for the structure mass respectively. CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 98 5.2.2.4 BEARING PAD SPRING STIFFNESS The spring stiffness used to represent the restraint caused by the rubber sliding bearings was varied to represent the possible values for the properties of the rubber. The properties of the rubber used in the pads are highly variable under different environmental conditions and over time. The initial model assumed the shear modulus of the rubber to be 36,000 MPa. Five different rubber shear modulus (KJ values were studied. These variations as a proportion of the initial shear modulus (K) were as follows: Model K1- K s = 0 Model K2 - % K, Model K3 - K^= Vz K; Model K4 - VA K; Model K5 - 1^ = 2 K; The first variation represents the case of an ideal frictionless sliding bearing, while the second and third variations represent a softer rubber than that assumed in the initial model. The final two variations represent a bearing pad that may be stiffer than that assumed. Stiffening of a bearing pad could result from the compression of the pad by the weight of the dead load being supported. 5.2.3 STUDY COMPARISON BASIS Comparisons for the parametric study were based on the natural frequencies of the structural system and on the participating mass in the horizontal plane for each mode of vibration. Natural CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 99 frequencies and participating mass were compared on the basis of absolute change in value and by percentage deviation from the initial base model values. 5.2.4 STUDY RESULTS Results of the parametric study are discussed in the following sections. 5.2.4.1 PILE L E N G T H TO FIXITY The results of the parametric study for the four different length to fixity levels are shown in Table 5.6 and Table 5.7. All four of the study fixity levels resulted in lower natural frequencies than those of the base model with differences ranging from a low of 2 % to a high of 27 %. As expected, variations in the natural frequencies were greatest for the three rigid body vibration modes. The longitudinal mode followed by the torsional mode and then by the translational mode of vibration was most affected by the change in pile fixity levels. The results indicate that the method that is chosen to determine the depth of fixity of the piles is not significant because the resulting natural frequencies of the structure are all very similar. The difference in the participating mass shown in Table 5.7, between the base model and the four different fixity level models was less than 5 % for the first two modes and 54 % for the third mode. The large change in the participating mass for the third mode is due to the magnitude of the participating mass for this mode. The participating mass for the third mode was very small; therefore, although the actual change in the participating mass for this mode was small, the CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE "| 00 percentage change was high. The participating mass was negligible for all other modes of vibration. 5.2.4.2 PELE A N D PELE C A P STIFFNESS The results of the parametric study for the four different pile and pile cap stiffness levels are shown in Table 5.8 and Table 5.9. The structure natural frequencies increased with higher member stiffness and decreased with lower member stiffness when compared to the base model results. When the member stiffness was increased to twice the base model stiffness, the natural frequency increased by up to 23 %. When the member stiffness was reduced to one quarter the base model stiffness, the natural frequency decreased by up to 21 %. It is unlikely that the stiffness of the members would reach two times the stiffness of the gross section nor is it likely that the entire length of each member would have one quarter of its original gross section stiffness. More reasonable values for member stiffness are 34 l g for reinforced concrete and VA l g for prestressed concrete (Buckle et al 1986). For the more reasonable range of stiffness, the natural frequencies varied by up to 13 % from the results of the base model. The results show that the framing member stiffnesses are not particularly significant when they fall within this more reasonable range. The difference in participating mass shown in Table 5.9, between the base model and the four different member stiffness levels was less than 5 % for the first two modes and 60 % for the third mode. The large difference in the participating mass between the control model and the parametric study models for the third mode can be attributed to the fact that the participating CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 101 mass for the third mode of vibration was small. Thus a small change in magnitude in the participating mass caused a large percentage change. It should also be noted that the difference in participating mass was less than 30 % for the more reasonable range of member stiffnesses. 5.2.4.3 STRUCTURE M A S S The results of the parametric study for the two different structure masses are shown in Table 5.10 and Table 5.11. The results show that the structure natural frequencies increased with decreased mass and decreased with increased mass when compared to the base model results. For a decrease in mass to one half the base model mass, the increase in natural frequency was 84 %. For an increase in mass of two times, the natural frequency increased by 30 %. The range of masses provided in this parametric study was much larger than that which would be experienced in real life. A reasonable value for the mass might be a variation of ± 15 % from the base value. This would imply a variation in the natural frequency of approximately -5 % to 12 % from those determined in the base model. The difference in the participating masses for the study models is 0 % to 7 % of those in the base model for the significant modes of vibration. These results show that variations in the structure mass is a relatively insignificant factor in the determination of the natural frequencies and participating mass factors of a vertical piled wharf structure. 5.2.4.4 BEARING PAD SPRING STIFFNESS The results of the parametric study for the four bearing pad stiffness are shown in Table 5.12 and Table 5.13. Natural frequencies increased with increased bearing pad stiffness. The difference in the frequencies between the base model and the study models ranged from 0 % for some of CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE <| 02 the higher modes to 33 % for the lower modes. The range of possible spring stiffness values is reasonable. The variation in the natural frequencies for the different spring stiffness was greatest for the transverse mode of vibration, followed by the torsional mode and then by the longitudinal mode. Again the effects of the bearing pad spring stiffness on the deck bending modes is insignificant. The effects of the spring stiffness on the participating masses was negligible for the first two modes of vibration. For the third mode, the participating mass varied by up to 162 % when the stiffness of the bearing pad springs was assumed to be 1/4 that of the base model. For the 1/4 stiffness model, the third mode became a much more significant factor in the dynamic analysis of the structure. The participating mass from the remaining modes of vibration were too insignificant to be of concern in a dynamic analysis. 5.2.5 S U M M A R Y The measured natural frequencies and mode shapes were compared to the preliminary analytical model natural frequencies and mode shapes. The measured mode shapes which were detected matched the analytical mode shapes very well except that the second and third mode shapes were interchanged. The natural frequencies matched poorly with discrepancies of up to 51 % in the important three rigid body modes. A correlative study was used to match the analytical model to the results from ambient vibration measurements. Mode shapes were studied using a modal assurance criterion analysis and CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE -| 03 natural frequencies were matched numerically. Key parameters which were modified to match the natural frequencies between the two sets of results were as follows: pile fixity location pile and pile cap stiffness deck stiffness structure mass and mass distribution sliding bearing restraint The final match of the analytical model to the measured results showed only a 5% discrepancy between natural frequency results for the three rigid body vibration modes, the most significant modes. This calibrated model became the base model for a parametric study on the key variables of the structure The parametric study was conducted to determine which parameters were most significant to the natural frequencies and participating masses of the structure. Parameters which were studied were as follows: pile length to fixity pile and pile cap stiffness structure mass sliding bearing restraint CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 1 Q4 The study showed that the level of sensitivity of the model to the different parameters from most sensitive to least sensitive were as follows: bearing restraint pile length to fixity pile and pile cap stiffness structure mass A reasonable range of bearing restraint stiffness showed a change in natural frequency of up to 33 % from that of the base model. Different pile length to fixity models changed the natural frequency by up to 27%. Pile/pile cap stiffness and structure mass had little effect on the natural frequency of the structure. Participating mass changed little for the first two modes of vibration in all cases, while the participating mass for the third mode of vibration changed significantly based on percentages, but not based on actual magnitude in some cases. C O R R E L A T I O N ANALYSIS A N D P A R A M E T R I C S T U D Y O F T H E T E S T S T R U C T U R E 105 r - M t - L D N T - CO CO | ( 0 | | O O O C M T - 0 0 ( O C O C N d d d o o o d d co N c o r o c o c o o o o o ) | uo || O o O O O C O M - O O O d ooQJOcicicio • m i CO CL X UJ t - rt N 1- x - O) r - CD o O o < » o o o o o o o d d c i o ' d c i d d CO h~ CO CO T— CO CO co|| o r - ^ o ^ o o o o o c o O o" o" o" o" d d o" T - T - O N J C O C O O ^ I O C O C D I CM || O O t - t O O O ( N O O O N O O O O d O O O O O OO r-- T - L O O CM C O o O ^ O O ^ ^ O O 1 ^ O d d o" d d C M C O T L O C D l v - C O C T ) S 9 p 0 Jty B O j l A" { e u v CO UJ Q O UJ Q O < UJ DTI £ CO UJ Q 0 Q UJ 1 CO UJ O z UJ D 0 UJ 01 2 • CM LO UJ _ l CO < Difference (%) & 2 7Z CM IO CO C? f CO CM <o c _ "3 -cr 3; Z |> U, 1.7 2.1 1.9 15 20.7 Analytical j Mode i r- CO CM «a- CO CO Natural Frequency (Hz) CO CO OO CO ° ? n . CM CO CO h-' II 2^ Experimental j Mode T - CM CO LO CO CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 106 •r- CN CN j] co || q o o o o o o o to 0 •»-». c 0 CO 0 a UJ CN 0 ) 0 ) 0 0 T - T - O O o* o o o o CO LO C O O N ^ - O O f f l t D r o c M I io|| O O O O O C O C N C N V - O 0 0 0 0 0 0 0 0 ° ° O 1 - o o o o CD 0> T - CN CO o o o CO o o o o o o o o o o O N- O O N - O o d co r--h - CN C O T - N S O o o o o O O O O O O O O o CO £2 00 o o o o CD CN CD \r \r co O i - T - CN T -d d o" d d o N O N T-0 ) 0 0 ) 0 d d o" d d r-- r-- co CN -q-O O CN O O O " d d d d d ( N l f O ^ l f i ( D N 0 0 O ) ° S 9 p O |/\| & o j } i( i B u y co UJ Q O UJ Q O Q UJ m _i S co UJ Q o Q UJ or CO Si CO UJ o z UJ ZD a UJ or U L _ J g m UJ —1 CQ < Difference Natural Frequency (Hz) o> r-~ ™ ^ CN CO CO 0 0 J ° Analytical Mode t - CO CN -^ r CD CO | Natural 1 Frequency (Hz) CO 00 CO CD ^ °? CN co co 1^ °2 Experimental Mode T - CN co m CD CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE -\ QJ CO UJ I-UJ o Cd CL _J < cn UJ o I-o UJ CO l i UJ Q o s < Q UJ cn _i < o LO LO LU _J CO P o rai » <D 3 w: •a T J « Ivu «». ro ~ ro r-e 5 -3 °-pi ««=• •ei i * Q. e igj: u 0) ral rx i l l ~ Ei ro < -=11 co; E LU O O O O O O O O O O O O co co co CN CO CN CN T- CM ro o o o o o o o o o o o o o o o o CD CN CD CO co co co co to TJ- L O in ra ro ro <£> ro S lO CM CT> O -SP ^ o o o o CN CD ro M O N _ o co m O CN O c o d d CN CN -r-' CN > cu ro O cu E cn ro w c a CO CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 108 CO UJ o 2 UJ ZD O* UJ or u. '< 2 I >< u_ O r-X H (D 2 UJ CO Li-ID CO UJ or >-Q D h-CO g or r— UJ $ i C D L O UJ _1 CQ Difference (%) -27% -15% -17% -7% -6% -5% -7% -10% -16% -16% «•—* •xTCnpOCOr-OOOT-CN C D T - c N j - s r c o ^ c o c o c o c o T - " C O C O l< CJ> C O CO CO co" CO Difference -sP s P vp vp >p vO >p <P v P "sp C r 0s* os> 0s* 0s 0s 0s 0s 0s" 0s" COOv-^COCMCNfOOO) T — T — T - 1 1 1 1 1 1 1 1 1 1 to I s T u. X O r - S N O i n O ( D N S CN t ^ ( D t - ( O N O O O C N co co d co d d d d T - T - T - CN CN CN [Difference (%) -25% -14% -16% -6% -5% -4% -4% -6% -13% -13% CN N U_ X c n c o c o o c D o c o c o r — c o CnCNCNLOOOCOCOCOCOCO T - ' co co d co d d d d T - T — T — T - T — 1 Difference (%) -20% -11% -12% -4% -3% -2% -2% -2% -4% -4% x- IS} . U_ X "tf-CDCNCDCOM-O'vrv-v-v - ; c o ^ c p q c q N - c N c o c o c N c o ' c o N i d c o d d v ^ T - ' T - T - T - CN CN CN T3 I I^^Cf>CDO)COCOLOCOf^ 15 <£ c v i c o c o r ^ d ^ d d c N C N J r ^ J ^ i T - x- CN CN CN CM o ii rwmmsmmmmmmm CO CO or or UJ UJ h- H LU UJ < < Q Q UJ UJ _ l _ l CL OL LO LO + + C N C N % UJ UJ UJ 2 2 2 _J _ l _J Q Q Q 2 2 2 Z) Z> z> o o o or or or o o o < < < i t TY X X X U_ u_ u_ II II II X— CM CO Li_ u_ u_ o X f— UJ O CO CO Q CO Z) >< LL II 2 CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE <| Q9 T A B L E 5.7 - P A R A M E T R I C S T U D Y R E S U L T S : L E N G T H T O FIXITY -P A R T I C I P A T I N G M A S S Mode Calibrated Model Fixity M odel F1 Difference Long - Dir. Trans -Dir Long - Dir Trans - Dir Long - Dir. Trans - Dir. m . ( % L _ m i%) «M : (%) 1 92.40 0.00 95.36 0.00 3.21% n/a 0.00 99.98 0.00 99.99 n/a 0.01% 3 7.56 0.00 4.62 0.00 -38.97% n/a l i l i i i i i : 0.00 0.00 0.00 0.00 n/a n/a 5 0.00 0.00 0.00 0.00 n/a n/a 6 0.02 0.00 0.01 0.00 -44.44% n/a 7 0.00 0.00 0.00 0.00 n/a n/a 8 0.00 0.00 0.00 0.00 n/a n/a 9 0.00 0.00 0.00 0.00 n/a n/a 1D 0.00 0.00 0.00 0.00 n/a n/a Mode Cafibrated Model Fixity Model F2 Difference Long. - Dir Trans. - Dir Long -D.r Trans - Dir Long - Dir Trans - Dir m m\ {%) m m ! . _(%) . . 1 92.40 0.00 96.20 0.00 4.11% n/a 2 0.00 99.98 0.00 99.91 n/a -0.07% 3 7.56 0.00 3.79 0.08 -49.95% n/a 4 0.00 0.00 0.00 0.00 n/a n/a 5 0.00 0.00 0.00 0.00 n/a n/a 6 0.02 0.00 0.01 0.00 -55.56% n/a 7 0.00 0.00 0.00 0.00 n/a n/a 8 0.00 0.00 0.00 0.00 n/a n/a 9 0.00 0.00 0.00 0.00 n/a n/a 10 0.00 0.00 0.00 0.00 n/a n/a Mode Calibrated Model Faity M odel F3 Difference Long. - Dir Trans - Dir Long -Dir Trans - Dir Long - Dir Trans; - Dir. <%> m m I I I W 1 1 I m 1 92.40 0.00 96.43 0.00 4.36% n/a 2 0.00 99.98 0.00 99.96 n/a -0.02% 3 7.56 0.00 3.56 0.03 -52.97% n/a 4 0.00 0.00 0.00 0.00 n/a n/a 5 0.00 0.00 0.00 0.00 n/a n/a 6 0.02 0.00 0.01 0.00 -61.11% n/a 7 0.00 0.00 0.00 0.00 n/a n/a 8 0.00 0.00 0.00 0.00 n/a n/a 9 0.00 0.00 0.00 0.00 n/a n/a 1D 0.00 0.00 0.00 0.00 n/a n/a Mode Calibrated Model Fixity M odel F4 Difference Long - Dir Trans - Dir Long - Dir Trans - Dir Long - Dir Trans - Dir <%) m (%) (%} (%) m 1 92.40 0.00 96.48 0.00 4.42% n/a 2 0.00 99.98 0.00 99.99 n/a 0.01% 3 7.56 0.00 3.50 0.00 -53.69% n/a 4 0.00 0.00 0.00 0.00 n/a n/a 5 0.00 0.00 0.00 0.00 n/a n/a 6 0.02 0.00 0.01 0.00 -61.11% n/a 7 0.00 0.00 0.00 0.00 n/a n/a 8 0.00 0.00 0.00 0.00 n/a n/a 9 0.00 0.00 0.00 0.00 n/a n/a 10 0.00 0.00 0.00 0.00 n/a n/a CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE -| -| Q CO LU O z U J a U J cc 3 • CO CO UJ z CO a: U J CD UJ O CO 3 CO UJ or >-a D i -co o cc tu s £ CO LO UJ _ J fJO Difference v O <sp s„p « s O vP *sP <sp -sP NP 0s" 0 s* 0s" 0 s 0 s* ©** 0 s 0 s* 0 s* 0 s •r-T-COCDCNCOOiOOO M r r T-o> www ^ H CO " x ifc~ CNJCDCOCOOlCOOOCOlOCO CM v- CO CM LO CM v CD (N CM C0^^r(lidlOC)T-CM(M T- T- CM CM CM CM Difference (%) sp sp sp s P sp vP vP NP NP VP O"* 0 s* 0 s 0 s* 0 s 0 s* 0 s O N 0 s 0 s* r - ( D N O ) r - m O C O O O leff = 1-1/2 lg COOTLOOTOCOCOCOCNCN C D C O T - C O ^ - C O I - C N C M C N CM co cd d d ^ CM CM t - T- CM CM CM CM Difference (%) sp "sp «sP "sP sp s P sp sP "sp sP o N O** 0 s* 0 s 0 s* 0 s* o"* 0 s* CONNWr-OJCOCOOT-T _ 1 1 T _ I I I I I I 1 1 D) CM CN V N W II £ tfc CMCMv-COCnCDT-COCDLO COLOCDCOCMOJIOOOO CM co cd cd d CM o i d CM CM" T- T— T— CM CM CM Difference (%) sP sP <sP *sP VP "sP *sP S p N p NP 0 s 0 s* 0 s" 0 s 0 s* 0s" 0 s* 0 s 0 s* 0s" T - O T - C O T - N O ) C O T - T -C M T - T - C M 1 T- ' ' ' ' 1 1 1 1 1 S1 leff = 1/4 lg (Hz) c M c n i O T - c o c o ^ r o c o c o T - C O ^ T - C M N C M O O O ) tNCOtOCDdr-CCiOr-T-T- T- -r- CM CM CM Calibrated Model (Hz) NKO)lDO)cCt«)Lf)H)K 0DNC00)C0r-O(0T-r-CM <0 CO ^ O ^  O O CM CM T- T- CM CM CM CM U) CM CM 1 T CM II II 11 II CD CD o CD > > > > o o o "4—* o .CD CD CD *= *= *= cu CD CD CD 1 1 CM 1 CO CO CO CO CO CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE -| -| -| TABLE 5.9 - PARAMETRIC STUDY RESULTS: FRAMING MEMBER STIFFNESS -PARTICIPATING MASS Mode Cafibrated Mode} Stiffness Model S1 Difference Long. - Dir Trans - Dir Long - Dir Trans - Dir Long - Dir. Trans. - Dir <%) i%) m <%) (%) fltt 1 92.40 0.00 95.62 0.00 3.48% n/a 2 0.00 99.98 0.00 99.99 n/a 0.00% 3 7.56 0.00 4.35 0.00 -42.44% n/a 4 0.00 0.00 0.00 0.00 n/a n/a 5 0.00 0.00 0.00 0.00 n/a n/a 6 0.02 0.00 0.02 0.00 -5.56% n/a 7 0.00 0.00 0.00 0.00 n/a n/a 8 0.00 0.00 0.00 0.00 n/a n/a 9 0.00 0.00 0.00 0.00 n/a n/a 1D 0.00 0.00 0.00 0.00 n/a n/a Mode CaBbrated Mode) Stiffness Model S2 Difference Long - Dir Trans - Dir. Long - Dir Trans - Dir Long. - Dir. Trans - Dir (%) m (%) _ i%\ {%\ 1 92.40 0.00 94.59 0.00 2.37% n/a 2 0.00 99.98 0.00 99.98 n/a 0.00% 3 7.56 0.00 5.38 0.00 -28.83% n/a 4 0.00 0.00 0.00 0.00 n/a n/a 5 0.00 0.00 0.00 0.00 n/a n/a 6 0.02 0.00 0.02 0.00 -11.11% n/a 7 0.00 0.00 0.00 0.00 n/a n/a 8 0.00 0.00 0.00 0.00 n/a n/a 9 0.00 0.00 0.00 0.00 n/a n/a 1D 0.00 0.00 0.00 0.00 n/a n/a Mode Cafibrated ModeJ "' Stiffness Modet S3 Difference Long. - Dir. Trans. - Dir Long - Dir Trans - Dir Long. - Dir. Trans - Dir. m mi {%) {%> {%) (%) 1 92.40 0.00 90.14 0.00 -2.45% n/a 2 0.00 99.98 0.00 99.98 n/a -0.00% 3 7.56 0.00 9.81 0.00 29.75% n/a A 0.00 0.00 0.00 0.00 n/a n/a 5 0.00 0.00 0.00 0.00 n/a n/a 6 0.02 0.00 0.02 0.00 22.22% n/a 7 0.00 0.00 0.00 0.00 n/a n/a 8 0.00 0.00 0.00 0.00 n/a n/a 9 0.00 0.00 0.00 0.00 n/a n/a 10 0.00 0.00 0.00 0.00 n/a n/a Mode Cafibrated Model Stiffness Model S4 Difference Long - Dir Trans. - Dir Long - Dir Trans - Dir Long - Dir. Trans - Dir <%) L (%) (%) {%) i%\ . ( % ) 1 92.40 0.00 87.90 0.00 -4.87% n/a 2 0.00 99.98 0.00 99.97 n/a -0.01% 3 7.56 0.00 12.04 0.00 59.16% n/a 4 0.00 0.00 0.00 0.00 n/a n/a 5 0.00 0.00 0.00 0.01 n/a n/a 6 0.02 0.00 0.03 0.00 50.00% n/a 7 0.00 0.00 0.00 0.00 n/a n/a 8 0.00 0.00 0.00 0.00 n/a n/a 9 0.00 0.00 0.00 0.00 n/a n/a 10 0.00 0.00 0.00 0.00 n/a n/a CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE <| <| 2 CO UJ O 2 UJ Difference (%) >P vP s O >^ 0 N O s O >p 0 S 0 S O** 0)OOC0r-0)0)^0)0) CM CO CO 1 CM T - CM CM CM CM • i i i i i i i i CM csl N S O O V T C O T - N O C D O O O C O CO CO 1— CO T- LO T - (O N CO ^ c M C M i ^ c o ^ ^ r i O L O L r i Difference (%) \ P Sp «sP sP yp <sP sP "sP vO ( V 0 S 0 S 0 S 0 S" 0 S 1 0*^ 0 S" o** T - O O ^ N C N T - C M M T -• ^ - - ^ • ^ r o o L O c o ^ L n ^ r ' ^ -2 ^ X S i S C O t D T - C O C n c D C O N C O N C M ^ ( D C O O)COCOCOCO co i o uo co c\i co x— T— T— T- T - CM CM CO CO CO Calibrated Model (Hz) K K C O C D O ) < O C O L O ( O N ( O N C O C a c O r - O f D r - T -CM CO CO | ^ O ^ O C) CM CM T - t - CM CM CM CM CM T — CM II II CO CO CO CO < < CM CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE <| "| 3 T A B L E 5.11 - P A R A M E T R I C S T U D Y R E S U L T S : S T R U C T U R E M A S S -P A R T I C I P A T I N G M A S S M o d e Cal ibrated Mode l M a s s M o d e l M1 Dif ference Long - Dir. Trans - Dir Long. - Dir. T rans . - Dir. Long . - Dir. T rans - Dir (%) (%) (%) (%)_ (%) . 1 92.40 0.00 91.91 0 - 0 . 5 3 % n/a 2 0.00 99.98 0 99 .965 n/a - 0 . 0 2 % 3 7.56 0.00 8.059 0 6 .54% n/a 4 0.00 0.00 0 0 n/a n/a 5 0.00 0.00 0 0 n/a n/a 6 0.02 0.00 0.012 0 - 3 3 . 3 3 % n/a 7 0.00 0.00 0 0 n/a n/a 8 0.00 0.00 0 0 n/a n/a 9 0.00 0.00 0 0 n/a n/a 10 0.00 0.00 0 0 n/a n/a M o d e Cal ibrated Mode l M a s s M o d e l M 2 Dif ference Long. - Dir. Trans. - Dir. Long . - Dir. Trans. - Dir. Long . - Dir. Trans. - Dir. (%) (%) {%) 1 92.40 0.00 91 .914 0 - 0 . 5 2 % n/a 2 0.00 99 .98 0 99 .965 n/a - 0 . 0 2 % 3 7.56 0.00 8.056 0 6 .50% n/a 4 0.00 0.00 0 0 n/a n/a 5 0.00 0.00 0 0 n/a n/a 6 0.02 0.00 0.012 0 - 3 3 . 3 3 % n/a 7 0.00 0.00 0 0 n/a n/a 8 0.00 0.00 0 0 n/a n/a 9 0.00 0.00 0 0 n/a ' n/a 10 0.00 0.00 0 0 n/a n/a CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 114 CO UJ o 2 UJ ZD o UJ or 2 i CO CO UJ 2 u_ u. h-CO Q CL CD 2 QT iS CQ CO b z> CO UJ or >-a z> H CO 0 or f— UJ & 1 CM I O UJ _ l CQ Difference (%) Sp sP sP *sP Np Np Sp SP NP Sp 0 S * C J * 0 S * 0 S " 0 S * 0 S " O ^ * O * * 0 S * T - T - O N - C N C M O T - O O CM CO CO 1 "it CM N yr ~ • s r c N j - v r c o c o c o r — i - N - c D c\i O) o L O L O L O T - q> r - . T T co ^ Lo co d •sj-' d d CM CM t - T - CM CM CM CM Difference (%) sP Sp Sp Sp Sp Sp Sp Sp Np Sp o*** c r 0 S 0 S * 0 S * o*** 0 S * o** o** o"** T - C O C D - ^ - T - T - O T - O O T - T - T - ' K3 j (Hz) i s co o m cn co co co N C D o c o m N ^ c o r - S T - T -C M ^ cd d ^ ' d d csi C N T - T - CM CM CM CM Difference (%) -13% -20% -19% -4% M /O -1% -1% -0% -1% 0% 0% K2 K = 1/2 Ki (Hz) T - T - I O C 0 O ) O ) ^ C N C D N C O O r - C O C M O ) O U ) r - T -CM co co d co" d d CM" CM' T - T - CM CM CM CM , Difference (%) -22% -33% -31% -6% -1% -2% -0% -1% 0% 0% K1 K = 1/4 Ki (Hz) 0 ) C 0 0 ) C D M - O C M L 0 O S q L O c o ^ c M O O ^ - T - ^ -CM CM CM d CO d O CM CM T - T - CM CM CM CM Base Model (HzL. K N C O C O O C O C O L n c D N (O N CO O) CO r- O (JD r - - r CM CO CO d Nf' d d CM CM T ~ T - CM CM CM CM vr 12 K\ CM 1 T — T — CM II II 11 II •Ks •Ks •Ks •Ks K2-K3-K4-CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 115 T A B L E 5.13 - P A R A M E T R I C S T U D Y R E S U L T S : B E A R I N G S T I F F N E S S -P A R T I C I P A T I N G M A S S Mode Calibrated Model Bearing Stiffness Model K1 Difference Long - Dir Trans. - Dir Long - Dir \ Trans - Dir Long - Dir Trans - Dir m (%) (%) m (%) <%> 1 92.40 0.00 80.14 0 -13.27% n/a 2 0.00 99.98 0 99.994 n/a 0.01% 3 7.56 0.00 19.848 0 162.40% n/a 4 0.00 0.00 0 0.001 n/a n/a 5 0.00 0.00 0 0 n/a n/a 6 0.02 0.00 0.006 0 -66.67% n/a 7 0.00 0.00 0 0 n/a n/a 8 0.00 0.00 0 0 n/a n/a 9 0.00 0.00 0 0 n/a n/a 1D 0.00 0.00 0 0 n/a n/a Mode Calibrated Mode* Bearing Stiffness Model K2 Difference Long - Dir Trans - Dir Long. - Dir. Trans - Dir Long - Dir Trans - Dir m m m {%) m 1 92.40 0.00 88.11 0.00 -4.65% n/a 0.00 99.98 0.00 99.99 n/a 0.01% 3 7.56 0.00 11.88 0.00 57.02% n/a 4 0.00 0.00 0.00 0.00 n/a n/a 5 0.00 0.00 0.00 0.00 n/a n/a 6 0.02 0.00 0.01 0.00 -50.00% n/a 7 0.00 0.00 0.00 0.00 n/a n/a 8 0.00 0.00 0.00 0.00 n/a n/a 9 0.00 0.00 0.00 0.00 n/a n/a 1D 0.00 0.00 0.00 0.00 n/a n/a Mode Calibrated Model Bearing Stiffness Model K3 Difference Long -Dir Trans - Dir Long. - Dir. Trans - Dir Long - Dir Trans - Dir *%> m <*> . \ m {%> m 1 92.40 0.00 93.69 0.00 1.40% n/a 2 0.00 99.98 0.00 99.97 n/a -0.01% 3 7.56 0.00 6.25 0.00 -17.40% n/a 4 0.00 0.00 0.00 0.00 n/a n/a 5 0.00 0.00 0.00 0.01 n/a n/a 6 0.02 0.00 0.03 0.00 66.67% n/a 7 0.00 0.00 0.00 0.00 n/a n/a 8 0.00 0.00 0.00 0.00 n/a n/a 9 0.00 0.00 0.00 0.00 n/a n/a 10 0.00 0.00 0.00 0.00 n/a n/a Mode Cafibrated Moder Bearing Stiffness Model K4 Difference Long - Dir. Trans - Dir Long - Dir Trans - Dir Long - Dir Trans - Dir {%) _ {%) (%} {%) m <%) • 1 92.40 0.00 94.41 0.00 2.18% n/a 2 0.00 99.98 0.00 99.95 n/a -0.03% 3 7.56 0.00 5.50 0.00 -27.34% n/a 4 0.00 0.00 0.00 0.00 n/a n/a 5 0.00 0.00 0.00 0.01 n/a n/a 6 0.02 0.00 0.04 0.00 144.44% n/a 7 0.00 0.00 0.00 0.00 n/a n/a 8 0.00 0.00 0.00 0.00 n/a n/a 9 0.00 0.00 0.00 0.00 n/a n/a 10 0.00 0.00 0.00 0.00 n/a n/a CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 116 CORRELATION ANALYSIS AND PARAMETRIC STUDY OF THE TEST STRUCTURE 117 118 CHAPTER 6 STRUCTURAL MODELLING - BATTER PILED WHARF A structural model of the experimentally measured batter piled, Oakland Outer Harbour Wharf, was developed using the commercially available finite element analysis package SAP90 (Version 5.4). A combination of three dimensional beam elements, shell elements, and mass elements was employed to create the model. Modelling techniques and rules utilised in general practice were followed in determining the parameters for the model framing system and for the individual member properties. The preliminary model was used to obtain an estimate of the global dynamic characteristics of the wharf prior to analysing strong motion data obtained for the wharf. The details of the batter piled wharf are given below. An explanation of the existing structural system is followed by an account of the assumptions used to determine individual member section and material properties. 6.1 SITE CONDITIONS The profile of the soil supporting the BPW consists of deep intermixed layers of clay and sand (Norris et al 1991). The soil profile varies over the length of the 490 m long wharf, but an average soil profile is shown in Fig. 6.1. This average profile indicates that the piles pass through two thick sand layers separated by a thin 1 m thick clay layer. The first sand layer is STRUCTURAL MODELLING - BATTER PILED WHARF 119 approximately 6 m deep and consists of loose to medium dense sand. The second sand layer is made up of a 13 m deep medium dense to dense silty, clayey sand. These layers are loosely referred to as Bay Mud. These three soil layers overlay several layers of clay and sand which in turn rest on bedrock, located approximately 150 m below the ground surface. The layers of sand and gravel achieve shear wave velocities ranging from approximately 150 m/s at the surface to 525 m/s near the bottom layers. According to the A A S H T O code, these shear wave velocities represent soft soil at the surface and medium dense soil near the bottom layers. While most of the piles are founded in the first three layers of soil, the lower layers of clay and sand contribute to the amplification of bedrock seismic motions. The geology at the Oakland Harbour is shown in Fig. 6.2. 6.2 STRUCTURE DESCRIPTION The structure is a batter piled wharf located at the Outer Harbour in Oakland, California and was designed and constructed in the mid 1970's. Overall site photographs are shown in Fig. 6.3. Typical plan and elevation views of the structure are shown in Fig. 6.4 and Fig. 6.5. The main wharf apron is made up of ten independent sections separated structurally by construction joints. These independent sections have lengths between successive control joints varying from 36.5 m to 88 m, making a total length of approximately 490 m. The width of the wharf is approximately 19 m and the total structure, including all the independent sections, has a total weight of approximately 18500 tonnes (181,500 kN). The main support system consists of vertical and batter solid square precast prestressed concrete piles spaced at 3.8 m in the east-west direction and 1.8 m, 2.7 m or 3.6 m in the north-south STRUCTURAL MODELLING - BATTER PILED WHARF 120 direction. A sheet piled wall, which mainly serves to isolate the wharf from the backlands to the east, supports the extreme east side of the wharf. The piles are cast integrally with cast-in-place concrete pile caps which are generally 1.2 m wide by 1.2 m deep and slabs range in depth from 460 mm to 625 mm. Batter piles typically frame into a pile cap while vertical piles generally frame directly into the slab. The pile caps span in the north-south direction along the longitudinal axis of the wharf. The piles, pile caps and slabs form integral units that make up each of the individual wharf sections between successive isolation joints. The piles, both battered and vertical, range from a free standing length through water and air of 4.0 m for the sheet steel wall at the east side of the structure to a free standing length of 13.75 m for the concrete precast pile at the most westerly side of the structure. Penetration depths for each of the piles could not be determined from the information available. This type of structure arrangement is a common form of construction for many existing wharves. 6.3 M O D E L L I N G ASSUMPTIONS The computer model for the BPW structure was developed in much the same manner as the V P W was developed. Three dimensional beam elements were used to model the piles and pile caps, shell elements for the top deck and sheet steel wall, and mass elements for additional mass not accounted for in the beam and shell elements. Assumptions that were made for modelling purposes such as depth to fixity, and individual member section and material properties were calculated using standard techniques as discussed in Chapter 2. Assumptions and modelling of member properties used for the preliminary model of the wharf are discussed in this section. STRUCTURAL MODELLING - BATTER PILED WHARF 121 The framing system consisted of the following: piles pile caps deck The components of the framing system were connected as follows. The piles, represented by beam elements for the concrete piles and shell elements for the steel sheet piles, were framed into pile caps, represented by a separate grid of beam elements or the slab represented by shell elements. The deck diaphram spanned the pile caps. Connections between the piles and the pile caps were assumed to be full moment connections. A feature of SAP90, the element rigid offset, was used on the pile elements to model the distance from the model node points (located in the mid-point of intersecting members) to the actual interface between the piles and the pile caps. The element rigid offset is explained in more detail in the SAP90 users manual (SRAC 1990). The deck members were assumed to transfer all in-plane and out-of-plane forces. Below ground, the point of full fixity was assumed to be five pile diameters below the ground profile determined using ground line #2 discussed in Chapter 2. The model created for this analysis consists of a 66 m section of the wharf. This section is assumed to be totally isolated from the rest of the wharf through the isolation joints. Fig. 6.6 shows an isometric view of the BPW model. Details of the SAP90 model are presented in the following sections. STRUCTURAL MODELLING - BATTER PILED WHARF 122 6.3.1 PILE PROPERTIES Two different types of piles were used in the model. Support piles for most of the main wharf consists of square precast prestressed concrete piles. Support piles for the east end of the main wharf apron is provided by steel sheet piles. The steel sheet piles also act as a retaining wall for the soil at the backlands. Concrete Pile Properties The geometry of the square concrete piles is shown in Fig. 6.7. The width of the pile is 460 mm between the flat sides. Section properties for calculating the stiffness of the preliminary model were calculated according to the gross concrete section using the dimensions given in the figure. Additional stiffness due to the reinforcing and prestressing steel was ignored. Also ignored was any reduction in stiffness due to cracking of the section under high lateral loads. Information regarding the concrete compressive strength (fc') was not available for the structure. Therefore the value of fc' was estimated based on strengths commonly used for structures of this type during the time that the structure was constructed. The 28 day concrete compressive strength was estimated to be 40 MPa. Material properties were calculated using the relationships established in Chapter 2. Table 6.1 summarizes the section and material property values used to model the concrete piles. STRUCTURAL MODELLING - BATTER PILED WHARF 123 Steel Sheet Pile Properties The assumed geometry of the steel sheet pile is shown in Fig. 6.8. Section properties for the preliminary model were calculated using the given geometry. Loss of section due to corrosion was ignored. Standard material properties for steel were used to model this section. Table 6.1 shows the section and material property values used to model the steel piles. 6.3.2 PILE C A P PROPERTIES The geometry of a typical pile cap is shown in Fig. 6.9. Section properties for the preliminary model were calculated according to the gross concrete section using the dimensions shown. Again, material information was not known, and therefore the material properties were calculated assuming compressive concrete strength of 30 MPa. Table 6.1 shows the section and material property values used to model the pile caps. 6.3.3 D E C K PROPERTIES Shell elements were used to model the deck slab. As noted previously the actual deck thickness varies from 460 mm to 625 mm. For modelling purposes the deck slabs were all assumed to be 540 mm in depth. Section properties for the slabs were calculated according to the gross concrete section using the assumed depth of 540 mm. Material properties for the deck elements were calculated assuming STRUCTURAL MODELLING - BATTER PILED WHARF 124 concrete with a compressive strength of 30 MPa. Table 6.1 shows the section and material properties used to model the deck diaphram. 6.3.4 D Y N A M I C A N A L Y S I S A dynamic analysis was performed using the above framing system, and section and material properties. A total of five eigenvalues (natural frequencies) and eigenvectors (mode shapes) were determined in the analysis. Natural frequencies, mode shapes and participating mass were calculated for each mode. Visual inspection of the natural frequencies and mode shapes showed that the first three modes of vibration are rigid body modes. Natural frequencies of the three modes range from 1.7 Hz to 2.1 Hz. The first analytical mode is a longitudinal mode at 1.7 Hz. The second mode is a transverse mode with a natural frequency of 1.9 Hz. The third mode is a torsional mode with a natural frequency of 2.1 Hz. Modes four, five and six were predominantly in-plane deck modes, while the remaining four modes were coupled in-plane deck and out-of-plane (vertical) deck modes. Natural frequency results of the first ten modes of vibration are shown in Table 6.2 and plots of the mode shapes are shown in Fig. 6.10. The participating mass results are presented in Table 6.3. The results showed that only the first three modes of vibration contribute significantly to the vibrational response of the structure. Participating masses are virtually zero for the in-plane vibrational modes after the third mode. Participating masses for the out-of-plane vibrational modes are significant for the higher modes, but vertical vibrations are generally ignored in seismic analyses, and thus were ignored here. STRUCTURAL MODELLING - BATTER PILED WHARF 125 CO UJ UJ o_ O cc CL _ l < cc UJ Q z < z o h-o UJ CO li UJ Q O < o < < UJ CL UJ _ i CO < jCO:; I'T3 & iro: ro: f<3> CO CO CO <D .2 CL S CO - I T 3 (On TJ CO ^ 40 CO CD O O O o o o o LO IO co co o o o o LO o CM O O O LO co I co 0 m CO:; i l l 01 S a . c: O: JO: ,C»: POj o o o o o o o o O CM CM CO CM fji CN m W m CO CM CO ro ro io c o LO CM _ £ co ro ro vt-mm* ::£§: :*»:::: :« J > |UJ CD M g co ro ro o T cr c CO co o CO r— ca ca o c c CM d ro ro a) o "5) _^ (D CD CD O E a ro i _ x: a ro CD CL CO Q Q STRUCTURAL MODELLING - BATTER PILED WHARF 126 co UJ o z UJ ZD o UJ or Li-ZD g 2 UJ h-I-to or u. JJ UJ Q O s < 2 < or < z UJ or D_ • CM CD UJ _ J CQ .< Natural Frequency (Hz) CD S O N CO LO Oi O CM CM CM CM CO a> xy o (D N CO O) ° Natural Frequency (Hz) I 3.24 4.32 9.38 21.85 24 08 WW •OW:« CM CO LO CO or o r-'2 CO CO < 0 2 1 o l -or 2 UJ h-I-co or u_ UJ Q O < o < 2 < or < 2 UJ or Q_ CO CD LU _ l CQ Vertical - Sum (%) _ _ C N - s J - C N C O C D i n C M d r-; T-" T-" c\i CD r-: Transv. - sum i (%) I m S S N c o o i o i r o o i o ) Q Cfi Q O) O) O) O) O) Ol O) c n o ) 0 ) 0 0 ) C 9 0 ) 0 0 ) Long - sum (%) 1 CNJCNJCDCOCDCDCDCDCDCD r - T - 0 ) 0 ) O C B 0 ) 0 ) C 8 r o x f ' t o i o J O J c n c n c n c B O ) Vertical - Dir 0 0.01 41.61 n 09 31.18 2.11 0.03 11.59 1.17 Transv. - Dir CO £ o ~ - o o o p o o o o Long - Dir CO _ o C\| „ -^J-• C D ^ O O O O O O O T _ e-i 0 0 Mode T - r M ( T * m ( D S c o c » ° STRUCTURAL MODELLING - BATTER PILED WHARF -| 27 18' 4' 43 ' S a n d C lay S a n d 80' C lay 20' C lay 65' C lay 25' 45' C lay C lay 85' S a n d 15' C lay 95" S a n d FIG. 6.1 - A V E R A G E S O I L P R O F I L E A T O A K L A N D O U T E R H A R B O U R W H A R F STRUCTURAL MODELLING - BATTER PILED WHARF 128 STRUCTURAL MODELLING - BATTER PILED WHARF -| 2 9 J J B I J M jnoqjBH JQjno pue|>|BO P M G I A uoflBAaig - 9 9 6\j J J B L I M J n o q j B H p u B p j B O i o M 9 | A U B | d - - B y N V l d TVOIdAl iNior NouoricJisNoo 01* 11 10 o "IVNIlAlcBl ^ / r d c J V H M 1 t Z'£Q[ 2SL Q'S9 2* LSI 0"9fr 3cJiN30ld3 Ol LU>I 96 jyvHM a a n i d y a u v a - o N i m a a o w i v y n i o n y i s STRUCTURAL MODELLING - BATTER PILED WHARF -| 32 STRUCTURAL MODELLING - BATTER PILED WHARF 133 STRUCTURAL MODELLING - BATTER PILED WHARF 134 STRUCTURAL MODELLING - BATTER PILED WHARF 135 136 CHAPTER 7 SEISMIC BEHAVIOUR OF A BATTER PILED WHARF SYSTEM The dynamic behaviour due to strong ground motion of a BPW at the Oakland Outer Harbour wharf is discussed in this chapter. This structure was instrumented with strong motion accelerometers by the California Strong Motion Instrumentation Program (CSMIP) following its construction in 1975/1976. The first set of strong motion measurements recorded from a well instrumented BPW was obtained from this structure during the 1989 Loma Prieta earthquake. The following text describes the structure, the measurement system, the strong motion data, and the data analysis. The extent of the data analysis included the following: examination of the strong motion records in the time domain to determine peak motions and relative differences between different signals; animation of the structure using the real time displacement response histories obtained for the wharf during the strong motion event; analysis of the strong motion records in the frequency domain to extract natural frequencies of a portion of the structure; and analysis of the strong motion records using response spectrum analysis to determine trends in the response of the structure under the seismic event. SEISMIC BEHAVIOUR - BATTER PILED WHARF 137 As noted in the previous chapter, a SAP90 model of a portion of the structure was created to supplement the above analyses. 7.1 M E A S U R E M E N T EQUIPMENT A N D EQUIPMENT L A Y O U T The measurement system used to record the strong motion data consists of twelve Kinemetrics FBA-1 force balanced accelerometers with a range of ±1.0 g. The FBA-1 accelerometer is the predecessor of the FBA-11 accelerometers used in the ambient vibration study described in Chapter 4. The accelerometer are connected to a CR-1 central recording system by fixed cables. The cables run through dedicated conduits that were cast directly into the structure during construction. The structure sensors are located in telephone utility pits which are situated in a collector trench at the west side of the wharf. The collector trench is in turn located directly adjacent to the west crane rail. The location of the strong motion sensors is shown in Fig. 7.1. Six are located at selected positions on the structure. The remaining six are mounted as tri-axial sensors. One is located in the free field near the south end of the structure while the other is located in the free field near the middle of the structure. The two structure sensors are oriented in the longitudinal direction and the four remaining sensors are oriented to detect transverse motions. Sensors for measuring vertical motions were not included in the array. Sensor numbers 6 and 7 are located at the south end of the structure and measure motion in the transverse and longitudinal directions respectively. These sensors were placed between the extreme south end of the wharf and the first separation joint. Sensor Numbers 4 and 5 are SEISMIC BEHAVIOUR - BATTER PILED WHARF -\ 33 located on the north and south ends of the 88 m long middle section of the wharf and measure motion in the transverse and longitudinal directions respectively. Sensor Numbers 8 and 9 are located at the extreme north end of the structure and measure motion in the transverse and longitudinal directions respectively. Seven other sections of the wharf separated by movement joints were not instrumented. The two free field stations each have a triaxial sensor setup. Its transverse and longitudinal sensors are oriented along the same axes as the structure's transverse and longitudinal sensors. The first free field setup is situated in the terminal area, 11 m east of the east side of the wharf and 75 m north of the south end of the wharf. The second free field setup is also situated in the terminal area, 20 m east of the east side of the wharf and 307 m north of the south end of the wharf. The centre line of the east crane rail is located in between the two free field stations, 3 m east of free field station #1 and 7 m west of free field station #2. 7.2 SEISMIC D A T A The seismic data for the O O H W is explained in the following sections. The background regarding the Loma Prieta earthquake is followed by a discussion on the origin of the data, a discussion of how the raw data were processed and by a discussion on the form of the data. 7.2.1 L O M A PRIETA E A R T H Q U A K E - BACKGROUND The Loma Prieta earthquake occurred in the Santa Cruz mountains near the San Andreas fault. Its epicentre was located approximately 16 km east of Santa Cruz and 33 km southwest of San SEISMIC BEHAVIOUR - BATTER PILED WHARF -| 39 Jose. The United States Geological Survey located the hypocentre at 37.037N by 121.883W, and at a depth of 18 km. The location of the epicentre of the earthquake, along with the location of the major faults in the vicinity and the location of strong motion instrumentation sites set out by CSMIP, is shown in Fig. 7.2. The earthquake occurred on October 17, 1989 at approximately 5:04 PM (PDT); its Richter Magnitude was 7.0. Peak horizontal ground accelerations near the epicentre measured as high as 0.64 g. The furthest station from the epicentre triggered by the earthquake was located 175 km north of the epicentre; it recorded a peak horizontal ground acceleration of 0.05 g. Modified Mercalli intensities (MMI) as high as level VIII were measured near the epicentre of the earthquake (Stover et al). In the vicinity of the OOHW, the MMI reached level VII. The Cypress Viaduct, two kilometres to the east of the OOHW, collapsed along a significant section of its length and to the north west a span of the Bay bridge fell off its bearings. The Seventh Street Terminal, which is directly adjacent to the OOHW, sustained substantial damage. At the Seventh Street Terminal, liquefaction of the soil caused slope failures and differential settlement between fill locations and piled locations. Inertial forces caused extensive damage to concrete batter piles. The location of the OOHW, the Cypress Viaduct, the Bay bridge and the Seventh Street pier were shown previously in Fig. 6.3. 7.2.2 ORIGIN OF DATA The O O H W seismic data from the 1989 Loma Prieta earthquake consisting of digitized and processed acceleration records from the twelve sensors located on the wharf and in the free field were obtained from CSMIP (1989). These raw acceleration records are presented in Appendix SEISMIC BEHAVIOUR - BATTER PILED WHARF <| 4Q C. Figure 7.2 shows the relative locations of the wharf site and the epicentre of the earthquake. As can be seen in this figure, the epicentre of the earthquake lies on a bearing of approximately N 170°E with respect to the wharf site. The CR-1 recorder was triggered approximately 35.0 s after the initial rupture of the earthquake. The sensors were positioned with the positive direction of the sensors on bearings of N 35°E and N 125°E for the longitudinal and transverse sensors respectively. According to the sensor layout and the position of the wharf relative to the epicentre of the earthquake, the radial propagation of ground waves moved through the site at a skew of approximately 45° to the long axis of the wharf. The insert in Fig. 7.2 shows approximately how the ground waves are thought to have moved past the wharf from the southwest to the northeast. The following section dicusses the manner in which the raw acceleration records obtained from CSMIP were processed. The raw records are attached in Appendix C. 7.2.3 PROCESSED STRONG MOTION D A T A The processed data set consists of records for corrected accelerations and calculated velocities and displacements. The correction of raw data was performed using a modified and improved version of the correction procedure used in the Caltech standard data processing project (Hudson 1979). In the correction procedure, the uncorrected data is subjected to a transducer correction. Following the transducer correction linear trends are removed from the data and then high and low pass filters are applied to produce the final corrected accelerogram. The data is then integrated and double integrated to produce corrected velocity and displacement records. The SEISMIC BEHAVIOUR - BATTER PILED WHARF 141 digitized data set released for public use by CSMIP consists of 2000 data points per record discretized at 0.02 second intervals. For analysis purposes, it must be noted that, due to the nature of the filters that were used in the data correction process, the data are only reliable for frequencies in the range of 0.2 Hz to 23.0 Hz. Additionally, the data are questionable for frequencies in the range of 0.1 Hz to 0.2 Hz and 23.0 Hz to 25.0 Hz. The data are not valid for frequencies in the range of 0.0 Hz to 0.1 Hz and 25.0 Hz and above. 7.2.3.1 C O R R E C T E D A C C E L E R A T I O N R E C O R D S Corrected acceleration records obtained from the 12 sensors are shown in Fig. 7.3. The peak structure accelerations were 0.45 g in the longitudinal direction and 0.32 g in the transverse direction. The peak free field accelerations were 0.29 g in the longitudinal direction and 0.28 g in the transverse direction. One free field sensor taking measurements in the vertical direction showed accelerations of up to 0.07 g, while another free field station, sensor #2, taking vertical readings was reported to have malfunctioned. 7.2.3.2 C O R R E C T E D V E L O C I T Y R E C O R D S Corrected velocity records obtained from the 12 sensors are shown in Fig. 7.4. The peak structure velocities were 52.6 cm/s in the longitudinal direction and 41.4 cm/s in the transverse direction. The peak velocities in the free field were 40.9 cm/s in the longitudinal direction and 41.6 cm/s in the transverse direction. The maximum measured velocity in the vertical direction was 10.4 cm/s. SEISMIC BEHAVIOUR - BATTER PILED WHARF 142 7.2.3.3 CORRECTED DISPLACEMENT RECORDS Corrected displacement records shown in Fig. 7.5 exhibit a maximum structure displacement of 11.2 cm in the longitudinal direction and 8.2 cm in the transverse direction. The maximum displacements in the free field for each of the principal directions were 9.4 cm and 8.7 cm in the longitudinal and transverse directions, respectively. The maximum measured displacement in the vertical direction was 1.6 cm. 7.2.4 DISCUSSION OF D A T A The CSMIP records for the longitudinal, transverse and vertical directions are presented in the following sections. 7.2.4.1 LONGITUDINAL DIRECTION The longitudinal acceleration, velocity and displacement records are discussed in detail in the following sections. Acceleration Data A large difference in the peak accelerations between the signals measured by the two structure longitudinal transducers was observed. The signal from the south end of the wharf showed a peak acceleration of 0.28 g (channel 7) while the signal from the north end of the wharf showed SEISMIC BEHAVIOUR - BATTER PILED WHARF 143 a peak acceleration of 0.45 g (channel 8). The two acceleration time histories are plotted on the same set of axes in Fig. 7.6. Note that the general shape of the two accelerographs are the same. Similar peaks and valleys can be found on both of the records except that a slight time shift is present. The magnitude of the time shift varies along the records, but can reach as high as 0.2 seconds in the strong motion portion of the record. The signals from the longitudinal transducers in the free field also showed a large difference in peak acceleration levels. The signal from the sensor at the south end of the wharf (channel 3) had a peak acceleration of 0.22 g while the signal from the sensor to the north (channel 12) had a peak acceleration of 0.29 g. The two time histories are graphed together in Fig. 7.7. The general shape of the two accelerographs are again very similar except for the time shift. The peak time shift between the free field signals is larger than the peak time shift between the structure signals. The free field and structure signals at each end of the wharf are plotted on the same set of axes in Fig. 7.8 and Fig. 7.9. The acceleration records from the south end of the structure (channels 3 and 7) are almost identical except that the structure sensor recorded higher peak levels. At the north end of the structure, the free field longitudinal sensor is located quite a distance from the structure sensor, while the transverse sensors are quite close to each other. The acceleration records were again similar, but not as closely matched as the records for the south sensors. The above findings indicate that the longitudinal accelerations recorded on the structure were highly influenced by the input accelerations from the free field. This occurrence was likely due to the fact that the structure is very stiff. Although the structure is very stiff, the higher S E I S M I C B E H A V I O U R - B A T T E R P I L E D W H A R F "| 44 acceleration levels measured on the wharf, as opposed to the free field, indicate that there was some amplification of the input signal. Velocity Data As was shown in Fig. 7.4, the two structure longitudinal sensors measured peak velocities of 52.6 cm/s and 37.1 cm/s for the north and south sensors respectively, while the free field longitudinal sensors measured peak velocities of 40.9 cm/s and 35.4 cm/s for the north and south sensors respectively. With the exception that the peak velocities are different and there is a slight time shift between the two records, the velocity time histories from both of the structure sensors are very similar. This is also true for the two free field sensors. The velocity time histories are also similar when the structure records are compared with the free field records at each end of the wharf. At the south end of the wharf, the velocity time histories of the structure and free field sensors are virtually identical with slightly higher velocities occurring on the structure. At the north end of the wharf the structure sensor had a significantly higher peak velocity than the free field sensor. The increased dissimilarity between the measurements in the free field and on the structure at the north end of the wharf can be attributed to the large distance between the free field sensor and the structure sensor. Displacement Data As was shown in Fig. 7.5, the two structure longitudinal sensors measured peak displacements of 11.2 cm and 9.9 cm for the north and south sensors respectively. The free field longitudinal SEISMIC BEHAVIOUR - BATTER PILED WHARF <| 45 sensors measured peak displacements of 9.4 cm and 8.8 cm for the north and south sensors respectively. The displacement time histories from both of the structure sensors are very similar, including the magnitude of the motion. This is also two for the two free field sensors. Amplification of the signal from the free field to the structure is apparent as it was for both the acceleration and velocity signals. 7.2.4.2 TRANSVERSE DIRECTION The transverse acceleration, velocity and displacement records are discussed in detail in the following sections. Acceleration Data As was shown in Fig. 7.3, the four structure transducers in the transverse direction show very similar peak acceleration readings. The four signals have peak acceleration values ranging from 0.28 g to 0.32 g. The acceleration records for the structure sensors are plotted together in several combinations in Fig. 7.10 to Fig. 7.15. As can be seen, all of the acceleration records are quite similar, although the comparison between the signal from the sensors near the middle of the wharf (sensors 4 or 5) and each of the signals from the sensors at the ends of the wharf (sensors 6 or 9) is better than the direct comparison between the signals from the sensors at the ends of the wharf. This is expected since there is a greater distance between sensors in the latter case and thus there is a significant possibility of geological differences between the two sensor sites causing different vibration characteristics. As in the longitudinal signals, there is a time shift between similar peaks in each of the records. The magnitude of the time shift again S E I S M I C B E H A V I O U R - B A T T E R P I L E D W H A R F 146 varies throughout the record with the maximum time shift reaching 0.4 s. The signals from the transverse sensors in the free field also showed very evenly matched peak accelerations. The sensor at the south end of the wharf (channel 1) recorded a peak acceleration of 0.28 g while the sensor to the north (channel 10) recorded a peak acceleration of 0.27 g. The two time histories are graphed together in Fig. 7.16. The north and south free field readings are graphed along with readings from the nearest structure sensors in Fig. 7.17 and Fig. 7.18. The acceleration records at the south end of the structure are almost identical except that the structure sensor recorded slightly higher peak levels. At the north end of the structure, the acceleration records were again similar, but not as closely matched as the records for the south sensors. The acceleration records show that the transverse structure accelerations, like the longitudinal accelerations, are highly influenced by the input accelerations, but amplification of the input signal on the structure is again apparent. Velocity Data As was shown in Fig. 7.4, the four structure transducers in the transverse direction show similar peak velocity readings with peak velocities ranging from 34.4 cm/s for the most southerly sensor to 41.4 cm/s for the most northerly sensor. All the velocity records show similar trends but the comparison between the signals from the middle sensors (sensors 4 or 5) and each of the signals from the two end sensors (sensors 6 or 9) is better than the direct comparison between the two end sensors. This is expected since there is a greater distance between sensors in the latter SEISMIC BEHAVIOUR - B A T T E R PILED W H A R F 1 4 7 case and thus there is a greater possibility of geological differences between the two sensor sites causing different vibration characteristics. As was the case for the longitudinal signals, there is a time shift between corresponding peaks in each of the records. The signals from the transverse sensors in the free field also showed very evenly matched peak velocities. The sensor at the north end of the wharf recorded a peak velocity of 41.6 cm/s while the sensor to the north recorded a peak velocity of 37.6 cm/s. Comparisons between the signals from the free field and structure sensors at approximately the same location along the length of the wharf indicate that the velocity records at the south end of the structure are virtually identical. The exception is that the free field sensor recorded slightly higher peak levels. At the north end of the structure, the velocity records were again similar, but not as closely matched as the records for the south sensors. The results indicate that the transverse structure velocities, like the longitudinal velocities, are highly influenced by the input velocities, but unlike the longitudinal velocities, de-amplification of the peak input signal is apparent on the structure time history. Displacement Data As was shown in Fig. 7.5, the four structure transducers in the transverse direction show similar displacement readings with peak displacements ranging from 6.7 cm for the most southerly sensor to 8.2 cm for the most northerly sensor. The transverse sensors in the free field measured peak displacements of 8.7 cm for the south sensor and 8.0 cm for the north sensor. The displacement time histories show the same trends as were found in the acceleration and SEISMIC BEHAVIOUR - BATTER PILED WHARF 148 velocity time histories. The displacement time histories, like for the transverse sensors, show a de-amplification of the peak input signal. 7.2.4.3 V E R T I C A L DIRECTION No vertical transducers were placed on the structure and therefore no information was available for the structure motions. Of the two sensors placed vertically in the free field, one transducer measured a peak acceleration of 0.07 g and the other sensor was thought to have malfunctioned. The velocity and displacement data for the functioning transducer show a peak velocity of 10.4 cm/s and a peak displacement of 1.7 cm (see Fig. 7.3 to Fig. 7.5). The uncorrected acceleration time history record for the malfunctioning sensor (shown in Appendix C) shows what seems to be a fairly low amplitude, high frequency signal superimposed on a higher amplitude, lower frequency signal. Unconfirmed reports indicated that a container crane was located very close to free field station #1 during the earthquake. As noted in Section 7.1, the centre line of the crane rail is only 3 m away from the free field station. Because of the possible proximity of the crane during the earthquake, it is possible that the measured signal is a result of free field vibrations superimposed on a signal produced by the rocking of the container crane. This theory could not be substantiated because digitized data were not made available for this sensor by CSMIP. SEISMIC BEHAVIOUR - BATTER PILED WHARF 149 7.3 DATA ANALYSIS The data was analysed to determine the behaviour of the separate sections of the wharf between the isolation joints. Of special interest was the relative movement of the different sections. A spectral analysis and a response spectrum analysis was carried out to determine natural frequencies of the input motions and the structure. The spectral analysis results were compared with the results reported in a previous study (Norris and Siddharthan 1991). The following sections introduce the computer programs used to carry out the data analysis and summarize how the packages were used to obtain results. 7.3.1 COMPUTER PROGRAMS Relative movement analysis between separate sections of the wharf was carried out using the computer program Bacchus. Results of a spectral analysis using ULTRA was also used to assist in the relative movement analysis. Natural frequency determination analysis was carried out using the response spectrum analysis computer program S P E C E Q . Again ULTRA was used to assist in the frequency domain analysis. Bacchus is a time domain 3-D animation program developed at UBC by Vincent Latendresse and Dr. Carlos Ventura (Latendresse and Ventura 1994). This program accepts input for the geometry of the structure, the location of the sensors on the structure, the direction of the sensors, and the time history records for the sensors and animates the structure according to the measured data. This program is extremely useful for viewing the real time movements of the structure during a strong motion event. SEISMIC BEHAVIOUR - BATTER PILED WHARF 150 S P E C E Q is a program originally developed at C A L T E C H and modified by Dr. Carlos Ventura at UBC (Nigam and Jennings 1968, Ventura 1992). This program was used to carry out a response spectrum analysis of the recorded accelerations. Acceleration time history records and a user defined file of analysis parameters are the only inputs required for S P E C E Q . The program ULTRA is described in Chapter 4. 7.3.2 RELATIVE MOTION OF WHARF SECTIONS The displacement time histories taken directly from the data files were input to Bacchus for animation. It immediately became obvious from the visual examination of the results, that there was significant differential motion between the different portions of the structure. Exaggerated plots of the displacement of the entire wharf at different times during the earthquake are shown in Fig. 7.19. These plots show the relative movement of the separate structure sections. Little relative motion between the sensors was detected during the initial stages of the earthquake, but significant differential activity was apparent in the strong motion portion of the record. Further proof that there is differential motion in the separate sections of the wharf can be found in the time history records. Varying peak accelerations in the structure indicate that all the sensors are not moving in the same manner. This is especially apparent in the longitudinal structure sensors. The two structure sensors measured peak accelerations of 0.28 g and 0.45 g. SEISMIC BEHAVIOUR - BATTER PILED WHARF -| 51 The coherence between the four transverse sensor signals was calculated. Fig. 7.20 to Fig. 7.22 show the coherence between the signals from sensors 4 and 5, 4 and 6, and 4 and 9. The coherence between the signals from sensors 4 and 5, on the same section of wharf, is much better than the coherence between the signals from sensors on different sections of the wharf. As explained in chapter 2, a coherence near unity indicates that one record is more linearly related to the other than if the coherence is far from unity. In the case of the strong motion data, the results indicate that the response of the sensors on different sections of the structure are not as closely related as the response of sensors on the same section of the wharf. This is further evidence that the separate sections of the wharf moved somewhat independently. 7.3.3 D Y N A M I C CHARACTERISTICS A N A L Y S I S Frequency domain analysis using ULTRA and response spectrum analysis using S P E C E Q resulted in similar conclusions. Each set of analyses are explained further in the following sections. 7.3.3.1 FREQUENCY D O M A I N A N A L Y S I S Because the separate sections of the structure seemed to move independently, spectral analysis was restricted to the 88 m long middle section of the wharf. This was the only location where two sensors measuring in the same direction were available. Only data from sensors 4 and 5 on the middle section of the wharf and sensor 10 adjacent to the middle section of the wharf were used for detailed analysis. For reference purposes, the power spectral density (PSD) calculated using ULTRA for each of the signals are included in Appendix D. SEISMIC BEHAVIOUR - BATTER PILED WHARF 152 During detailed analysis, it became apparent that many factors were causing difficulties in analysing the structure when using the same techniques as was utilized for the ambient vibration analysis. These factors fell into three basic categories: structure related factors instrumentation related factors and input vibration related factors The following are some of the factors which caused analysis difficulties: multiple input points for the structure vibrations due to the multiple supported structure; the bias of the input motions on the motion of the structure; the physical distance between the various sensors causing a time lag in which the input ground accelerations reached the different sensors; the small number of sensors used to instrument the structure; and the possible shifting of the dynamic characteristics of the structure over the course of the earthquake. The first third and fifth factors can be related to the structure, the fourth factor is related to the instrumentation system and the second factor is related to the input vibrations. The influence of each of the factors is explained below. One of the most powerful tools for determining natural frequencies of a structure under strong motion input is the frequency response function. An accurate and complete frequency response SEISMIC BEHAVIOUR - BATTER PILED WHARF 153 function relies on the assumption of a linear relationship between uncorrelated multiple inputs and the output of a system. The wharf structure sensors were subjected to inputs from each pile location with the multiple inputs caused by the same seismic event. This indicated that all the inputs were correlated and thus the frequency response function became incomplete and thus unreliable for final analysis. For comparison purposes, strong motion analysis of a building is much simpler because the system is basically a single input/single output problem. The input excitation can be measured using an accelerometer at the foundation location and is a known quantity. This record can then be used in a frequency response analysis with output records from transducers placed at other locations in the building. This type of analysis has been done successfully using ULTRA on several occasions (Ventura 1993, Ventura and Stienke 1993). Because a frequency response analysis could not be relied upon to produce the natural frequencies of the structure, an analysis of the Fourier spectra was also conducted. Fourier spectra were calculated using the power spectral density function. The input vibrations for the structure were caused by the ground waves of the seismic event. It was shown earlier that due to the stiffness of the structure, motions of the structure parallel the free field motions. Because of this, spectral analysis of the records from the structure shows a high bias towards the spectral characteristics of the input motions. Since the seismic motions contain a high level of energy, the spectral characteristics of the input motions can overshadow some of the spectral characteristics of the structure. This makes anaylsis using the Fourier spectra difficult because some structural spectral peaks fall into the background. Many high peaks in the structure PSD are associated with the input spectrum, therefore care must be taken to ensure that peaks in the SEISMIC BEHAVIOUR - BATTER PILED WHARF 154 Fourier spectra are structure peaks and not input spectrum peaks. The PSD for one of the structure sensor signals and the adjacent free field sensor signal is graphed on the same set of axes in Fig. 7.23. The PSD for the signals from the longitudinal sensors at the south end of the structure are plotted in Fig. 7.24 for comparison purposes. Note the number of spectral peaks in the structure spectrum which coincide with peaks in the input spectrum. With the input motion being so prevalent in the structure signals, it becomes important to know when the input reaches the various points in the structure. As explained earlier, the ground waves move from the south end of the wharf to the north end and there is a time shift when the same input ground waves reach different locations in the structure. Acceleration time histories of sensor numbers 6 and 9 were graphed previously in Fig. 7.15. As can be seen, the time at which the input motion seems to reach the two sensors is quite different, arid in fact the time difference changes during the course of the seismic event. The change in the time lag for the ground waves between different sensors can likely be attributed to the difference in wave speed for ground waves of different frequencies. As the natural frequency of the ground vibrations changed during the course of the event, the time difference changed. The time difference for which the ground waves arrived at the different sensors was as high as 0.4 s between the transverse sensors at the extreme north and south ends of the structure. The time difference for which the input signal reaches the different sensors causes an initial phase difference between the data from the different sensors. This makes it difficult to use phase relationships between two sensors to assist in natural frequency identification. In ideal situations, natural frequencies correspond to situations where two structure signals are exactly in-phase or 180° out of phase, or where a structure signal and a free field signal are 90° out of phase. SEISMIC BEHAVIOUR - BATTER PILED WHARF -| 55 The limited number of sensors used to instrument the structure made it impossible to make anything more than very general conclusions about the vibration characteristics of the structure. While spectral information could be obtained from individual sensors, mode shapes could not be used to verify natural frequencies. With the exception of the middle section of the wharf, no individual section of the structure had more than one sensor reading in the same direction. Previous experience from analysing ambient vibration data showed that the interpretation of vibration data for natural frequencies becomes increasingly easier with an increasing numbers of measurement locations. The shift in the dynamic characteristics of the structure over the course of the seismic event, although not measured, is a recognized phenomenon. When a concrete member cracks during a seismic event, it can undergo a significant change in its stiffness and thus undergo a complete change in its dynamic characteristics. Regardless of the difficulties encountered in analysing the structure, evidence was gathered to determine the input motion characteristics and several potential transverse natural frequencies of the structure. Natural Frequency Analysis and Results Analysis was performed by examining the power spectra, the coherence between the signals from the sensors, the frequency response function, and by conducting a preliminary finite element analysis. \ SEISMIC BEHAVIOUR - BATTER PILED WHARF 156 The power spectra were analysed to determine the potential natural frequencies of the input ground motions. The power spectra for the signals recorded in the free field are plotted individually in Fig. 7.25 to Fig. 7.28 and together in Fig. 7.29 and Fig. 7.30. It can be noted that the significant input motion in the longitudinal direction is spread over a broad range starting at 0.5 Hz and ending at 1.8 Hz. The significant motion in the transverse direction is located in two frequency ranges 0.6 Hz to 0.8 Hz and 1.2 Hz to 1.8 Hz. Particularly high peaks were located at 0.60 Hz, 0.75 Hz, 1.05 Hz, 1.30 Hz, and 1.45 Hz in the PSD for the longitudinal signals and at 0.70 Hz, 1.3 Hz, 1.4 Hz, 1.55 Hz and 1.70 Hz in the transverse signals. The PSD for the records from a structure sensor and a free field sensor were plotted previously in Fig. 7.23. Note that the PSD from the structure sensor signal generally follows the PSD of that from the free field sensor, but there are several locations where the PSD's do not coincide. The non-coinciding spectral peaks were potential candidates for natural frequencies of the structure. Frequencies which showed promise for being natural frequencies of the structure in the 0 Hz to 4 Hz range were 0.83 Hz, 1.37 Hz, 2.39 Hz, 3.03 Hz and 3.23 Hz. These five frequencies were studied further to determine if any of them corresponded to natural frequencies of the structure. The coherence function between the signals from the two structure sensors and signals from the free field sensors was analysed for further evidence of which frequencies if any corresponded to natural frequencies. Fig. 7.31 and 7.32 show the coherence function plots for the records from sensors 4 and 10 and sensors 5 and 10 respectively. The coherence function for the signal from sensors 4 and 5 was shown previously in Fig. 7.20. The coherence is very good for almost all the frequencies. On the basis of coherence, only one frequency, 2.39 Hz, could be discarded as a potential natural frequency. SEISMIC BEHAVIOUR - BATTER PILED WHARF 1 5 7 The measurements from the free field sensor were then subtracted from the measurements for both structure sensors. The resulting data represents the acceleration of the structure relative to the ground motions. The A N P S D functions for these two data ranges are shown in Fig. 7.33. The ANPSD's for the structure acceleration signals show that out of the four remaining potential frequencies, only 1.37 Hz and 3.03 Hz remain strong. The frequency response function between a signal from a structure sensor and the signal from a free field sensor is plotted in Fig. 7.34. Interpretation of the frequency response plot shows that the structure has significant energy at 3.03 Hz in relation to the input motion, but not at 1.37 Hz. The first trial computer model explained in Chapter 6 was also used to make an estimate of the natural frequencies of the structure. The results of the analysis using the computer model were shown in Table 6.2. The numbers show that the first natural frequency is 3.25 Hz. This indicates that the observed peak at a natural frequency of 3.03 Hz is more likely to be the first natural frequency of the system. The final indication is that 3.03 Hz is the first natural period of the structural system. 7.3.3.2 RESPONSE SPECTRUM A N A L Y S I S A response spectrum shows the maximum response of a family of single degree of freedom systems covering a range of natural frequencies when subjected to a known set of input vibrations. SEISMIC BEHAVIOUR - BATTER PILED WHARF <| 53 Response spectrum analysis was performed on the strong motion data using the computer program S P E C E Q . Input for the analysis consisted of the corrected acceleration time histories taken directly from the CSMIP supplied data. Damping for the system was assumed to be 5%. Output from the analyses were graphed such that trends in the response could be observed. Graphs of the acceleration response spectra for each of the sensor signals at 5% damping is provided in Appendix E. Selected response spectra are plotted together on the same set of axes in this section for comparison purposes. Longitudinal Signals The acceleration response spectra calculated from accelerations measured at the two free field stations are graphed together in Fig. 7.35, while the acceleration response spectra for the free field and structure stations at the north end and for the free field and structure stations at the south end of the structure are graphed together in Fig. 7.36 and Fig. 7.37 respectively. Fig. 7.35 shows that the spectral values calculated using the signal from the free field station to the north are greater than the spectral values calculated using the signal from the free field station to the south at virtually the full frequency range of frequencies that was examined (0 Hz to 5 Hz). The response spectrum for the free field station to the south only has significantly greater response between the frequencies of 2.24 Hz to 2.46 Hz. Significant peak responses occur at approximately 1.1 Hz, 1.5 Hz, and 2.4 Hz for the south station, while significant peak responses occur at approximately 1.1 Hz, 1.5 Hz and 3.0 Hz for the north station. These peak response frequencies are potential natural frequencies of the free field soil column. SEISMIC BEHAVIOUR - BATTER PILED WHARF 159 Fig. 7.36 shows that the longitudinal sensors at the north end of the structure are quite a distance apart, therefore many of the differences in the responses may be due to the distance between the sensors. The figure indicates that the response spectra for the acceleration signals for both sensors have, similar trends. While the trends are similar, the spectral values for the structure signal are greater at all frequencies. The similar trends in the spectra indicate that the structure motions are dependent on the free field motions. The greater response in the structure spectra indicates that amplification of the free field motions can be detected at the structure. Fig. 7.37 showing the acceleration response spectra for the free field and structure signals at the south end of the structure exhibits the similarity between the two spectra. Unlike the response spectra for the signals from the north sensors, the magnitude of the spectral values calculated from both the free field signal and the structure signal are comparable. Fig. 7.37 indicates that amplification of the free field motions occurs at frequencies below 1.9 Hz, between 2.85 Hz and 3.15 Hz and above 3.4 Hz for the structure. Transverse Signals The response spectra for the acceleration signals from the two free field sensors are graphed together in Fig. 7.38. The graph for the transverse records show that no regular trend can be detected in the spectral values for the input signals except that high energy levels are present between the range of 1.2 Hz to 2.0 Hz. The response spectrum calculated from the signal to the south shows higher values between the frequency ranges of 1.4 Hz to 1.65 Hz, 3.3 Hz to 3.65 Hz and 4.2 Hz to 5.0 Hz than the response spectrum calculated from the signal to the north. In all other frequency ranges the response is the same or the north spectrum shows higher values. SEISMIC BEHAVIOUR - BATTER PILED WHARF 160 Significant peaks in the spectrum from the south signal are located at 1.55 Hz, 2.4 Hz and 4.8 Hz. Significant peaks in the spectrum from the north signal are located at 1.55 Hz, and 2.5 Hz. Fig. 7.39 shows the response spectra for the acceleration signals from the north transverse free field and structure sensors on the same graph. Because more sensors were placed on the structure in the transverse direction than in the longitudinal direction, the structure sensor is at approximately the same location along the wharf as the free field sensor. The possible discrepancies between the spectral values in the response spectrum for the signals measured in the longitudinal direction from the northern pair of sensors due to their distance apart should not have occurred with the north transverse sensors. The figure indicates that the spectral values calculated from the structure signal are higher than the spectral values calculated from the free field signal between 1.7 Hz and 2.3 Hz and from 2.9 Hz to 5.0 Hz. The latter frequency range shows the greatest difference between the two spectra. The larger spectral values for the structure signal in the relatively high frequency range indicates that the structure is fairly stiff in the transverse direction. This is consistent with the presence batter piles in the transverse direction. Significant spectral peaks for the structure occur at 1.6 Hz, and 3.15 Hz. The peak at 1.6 Hz corresponds with a peak in the input signal spectrum, but the peak at 3.15 Hz seems to stand alone. This indicates a possible structure natural frequency at 3.15 Hz which is generally consistent with the natural frequency found at 3.03 Hz in the previous section. Fig. 7.40 displays the response spectra for the acceleration signals measured at the two south transverse sensors on the same graph. The spectrum calculated from the structure signal is higher than the spectrum calculated from the free field signal between 2.9 Hz and 4.65 Hz. Significant spectral peaks for the structure occur at 1.6 Hz and 3.4 Hz. These findings are in SEISMIC BEHAVIOUR - BATTER PILED WHARF 161 general conformance with the findings from the north signals. Slight differences can be attributed to the fact the north and south sensors are on different sections of the wharf structure. 7.4 S U M M A R Y Strong motion records obtained for the O O H W were analysed to determine the behaviour of the wharf under a seismic event. It was found that the sections of the structure separated by isolation joints moved independently. Peak accelerations in the free field reached 0.29 g and 0.28 g in the longitudinal and transverse directions respectively. Peak accelerations on the structure were amplified and reached 0.45 g and 0.32 g in the longitudinal and transverse directions respectively. Peak velocities in the free field reached 40.9 cm/s and 41.6 cm/s in the longitudinal and transverse directions respectively. Amplification of the peak free field velocities were noted in the longitudinal direction and de-amplification of the free field velocities were noted in the transverse direction for the structure. The peak structure velocities reached 41.4 cm/s and 52.6 cm/s in the longitudinal and transverse directions respectively. Peak displacements in the free field reached 9.4 cm and 8.8 cm in the longitudinal and transverse directions respectively. Peak displacements on the structure showed amplification of the peak free field displacements in the longitudinal direction and de-amplifiation of the peak free field displacements in the transverse direction. Peak structure displacements reached 11.2 cm and 8.2 cm in the longitudinal and transverse directions respectively. Similar motions in adjacent sensors in the free field and on the structure indicate that the structure motions are in most part governed by the motions of the free field. SEISMIC BEHAVIOUR - BATTER PILED WHARF -| Q2 Frequency analysis showed that the significant energy of the seismic event was concentrated between the ranges of 0.6 Hz to 0.8 Hz and 1.2 Hz to 1.8 Hz in the transverse direction and between 0.5 Hz to 1.8 Hz in the longitudinal direction. A natural frequency of the structure was determined to be at 3.05 Hz. Response Spectrum analysis was performed on the data. Results showed that amplification of the spectral values indicating amplification of ground motions occurred mainly at frequencies above 2.9 Hz. This shows that the structure is excited mainly at higher frequencies and thus indicating that the structure is quite stiff. Significant spectral peaks from the spectra calculated using the structure signals which do not correspond with spectral peaks from the spectra calculated using the free field signals show a potential natural frequency of the structure at 3.2 Hz in the transverse direction in the middle portion of the structure. SEISMIC BEHAVIOUR - BATTER PILED WHARF -\ 54 FIG. 7.2 - L O C A T I O N O F K N O W N F A U L T S IN T H E VICINITY O F T H E L O M A P R I E T A E A R T H Q U A K E ( I N S E T S H O W S R A D I A L P R O P A G A T I O N O F G R O U N D M O V I N G P A S T T H E W H A R F S ITE) SEISMIC BEHAVIOUR - BATTER PILED WHARF «| 55 Channel 1 Max. Accel. = .28 g 0.45 0 -0.45 Channel 3 Max. Accel. = .22 g Channel 4 Max. Accel. = .28 g Channel 5 Max. Accel. = .31 g Channel 6 Max. Accel. = .32 g c o 2 8 0.45 0 -0.45 0.45 0 -0.45 Channel 7 Max. Accel. = .28 g Channel 8 Max. Accel. = .45 g 0.45 0 -0.45 Channel 9 Max. Accel. = .30 g 0.45 0 -0.45 Channel 10 Max. Accel. = .27 g 0.45 0 -0.45 Channel 11 Max. Accel. = .07 g 0.45 0 -0.45 Channel 12 Max. Accel. = .29 g F IG . 7.3 - C O R R E C T E D A C C E L E R A T I O N R E C O R D S F R O M S T R O N G M O T I O N S E N S O R S L O C A T E D @ O O H W D U R I N G T H E L O M A P R I E T A E A R T H Q U A K E SEISMIC BEHAVIOUR - BATTER PILED WHARF "| 66 60 0 -60 Channel 1 Max. Vel. = 37.6 cm/s 60 Channel 3 -fin ~ Max. Vel. = 35.4 cm/s 60 0 -60 Channel 4 Max. Vel. = 40.6 cm/s Channel 5 Max. Vel. = 40.2 cm/s Channel 6 7 8 Max. Vel. = 52.6 cm/s 60 0 -60 Channel! Max. Vel. = 41.5 cm/s 60 0 -60 Channel 10 Max. Vel. = 41.6 cm/s 60 Channel 11 -60 Max. Vel. = 10.4 cm/s 60 0 -60 Channel 12 Max. Vel. = 40.9 cm/s FIG. 7.4 - C O R R E C T E D V E L O C I T Y R E C O R D S F R O M S T R O N G M O T I O N S E N S O R S L O C A T E D @ O O H W D U R I N G T H E L O M A P R I E T A E A R T H Q U A K E SEISMIC BEHAVIOUR - BATTER PILED WHARF -| QJ Channel 1 Max. Displ. = 8.0 cm Channel 3 Max. Displ. = 8.8 cm 10 0 -10 Channel 4 Max. Displ. = 7.9 cm Channel 5 E .u. c CO E 8 m o. co Q 10 0 -10 Max. Displ. = 7.4 cm Channel 6 Max. Displ. = 6.7 cm Channel 7 Max. Displ. = 9.9 cm Channel 8 Max. Displ. = 11.2 cm 10 0 -10 Channel 9 Max. Displ. = 8.2 cm Channel 10 Max. Displ. = 8.7 cm 10 0 -10 Channel 11 Max. Displ. = 1.7 cm 10 0 -10 Channel 12 Max. Displ. = 9.4 cm F IG . 7.5 - C O R R E C T E D D I S P L A C E M E N T R E C O R D S F R O M S T R O N G M O T I O N S E N S O R S L O C A T E D @ O O H W D U R I N G T H E L O M A P R I E T A E A R T H Q U A K E SEISMIC BEHAVIOUR - BATTER PILED WHARF 168 BUnuatHHUMECBUn Is- co t q ^ c v i o c N i ^ t o ( T J o d d d d d rr I I I U -[6] uoijejaiaoov SEISMIC BEHAVIOUR - BATTER PILED WHARF 169 i i [6] uoiiBjapoov SEISMIC BEHAVIOUR - BATTER PILED WHARF 170 [6] uojiejeieoov SEISMIC BEHAVIOUR - BATTER PILED WHARF 171 a UJ UJ HI Q D 0 z O LL. —i a: • < co x UJ p g O CO g X UJ UJ X I-or O r2 co uj Q —i or pj o O UJ < or Q UJ uj or o UJ or x o h-o X r-O CO CD I S © X [6] U O I | B J 9 | 9 0 0 V SEISMIC BEHAVIOUR - BATTER PILED WHARF 172 in J Z J Z ! O I O "* Pill o est o o d o • i [6] UOIl82JO|900y SEISMIC BEHAVIOUR - BATTER PILED WHARF 173 C O J Z J Z o o [6] uoiiBje|8Dov SEISMIC BEHAVIOUR - BATTER PILED WHARF JZ O 1 O 1 CN O CM O d t PI uoijBjaieoov SEISMIC BEHAVIOUR - BATTER PILED WHARF •] 7 5 tn CD JZ. JZ a 0 •^ f CM O CM O O O O [6] uoipjapoov SEISMIC BEHAVIOUR - BATTER PILED WHARF 176 °S O CM [6] UOIJBJ3|900V SEISMIC BEHAVIOUR - BATTER PILED WHARF 177 SEISMIC BEHAVIOUR - BATTER PILED WHARF 178 [6] uoiejaiaoov [6] • j o i p j a i o o o y SEISMIC BEHAVIOUR - BATTER PILED WHARF 180 O UL is UL CO o o co .co. O CU CM <-Q Z < Q _l UJ UL u y LU X — 11 CO X UJ > CO z 1 o F-2 oo CD < X CO Q X o O UJ c Q X I -O GO UJ X X o h-00 X CO UJ £ 2 ° =• co X 00 UJ III 3 O X c o [6] U0!|BJ9|00aV BwnoHiavB ONiana s±N3i/eo\ndsia BIAIIX ivay nvoidAi -evi • NOI1V1S CT13ld 33yd mm^ma 3dVHS a s w a o d s a wmmmm 3dVHS a s w y o d s a N n aNaoan S9S'8l = au!l • • • S36'9l=8UJ!i 0 • £1 • m--m o n • • ^^^^^^^^^^^ i • • IQl d U v H M a a n i d a a i i v s - a n o i A V H a a OIIAISBS SEISMIC BEHAVIOUR - BATTER PILED WHARF <| Q2 SEISMIC BEHAVIOUR - BATTER PILED WHARF 183 80U9J9L|0Q SEISMIC BEHAVIOUR - BATTER PILED WHARF -| 84 80U8J8qOQ SEISMIC BEHAVIOUR - BATTER PILED WHARF 185 SUOIIHIAI A|isuea i B J j o s d s SEISMIC BEHAVIOUR - BATTER PILED WHARF 186 SEISMIC BEHAVIOUR - BATTER PILED WHARF -| 87 LL CD CM O LL suoi||!|Aj Al!SU8Q |Bjp9dS J8M0d SEISMIC BEHAVIOUR - BATTER PILED WHARF <| 88 O LL, O CO CO CM O SUOI||!|AJ Aj!su8Q iBjpeds j9M0d SEISMIC BEHAVIOUR - BATTER PILED WHARF 189 CM O 00 CD ^ CN O U-T — T — suo;n!i/\| SEISMIC B E H A V I O U R - BATTER PILED W H A R F -| g f j SUOI||!|A| SEISMIC BEHAVIOUR - BATTER PILED WHARF -| g-| A J I S I B Q iBjpads SEISMIC BEHAVIOUR - BATTER PILED WHARF -| g2 SEISMIC BEHAVIOUR - BATTER PILED WHARF 193 t 1 s , 1 I ' • o CN T- 00 CO TJ- CM O T-^  d o d o 80U9J9U0Q SEISMIC BEHAVIOUR - BATTER PILED WHARF 195 SEISMIC BEHAVIOUR - BATTER PILED WHARF 196 SEISMIC BEHAVIOUR - BATTER PILED WHARF esuodsey SEISMIC BEHAVIOUR - BATTER PILED WHARF -| QQ SEISMIC BEHAVIOUR - BATTER PILED WHARF 199 LO —I CO N I a c 3 cr Q: c. CM LL esuodsay o O 0 UJ UJ 01 • IT eg O Q _J UJ u L L Q ID fc CD Z O , i CD Z CL < Q io u. ©• < UJ u 0_ Q CO z UJ UJ CO I Z h-O Z) o_ o CO CO UJ . or !< I 0 SEISMIC BEHAVIOUR - BATTER PILED WHARF 200 SEISMIC BEHAVIOUR - BATTER PILED WHARF 203 CHAPTER 8 SUMMARY AND CONCLUSIONS The main purpose of the work conducted in this thesis was to attempt to assess the accuracy of different modelling assumptions used in determining the dynamic properties of open piled wharves. In doing so, it was hoped to determine which modelling parameters are most important in the modelling process. A secondary purpose of the research was to analyse strong motion data relating to an open piled wharf. 8.1 S U M M A R Y Two separate open piled wharf structures were analysed as part of this research. One, a local vertical piled wharf, was instrumented with portable accelerometers for ambient vibration measurements. The other, a batter piled wharf located in Oakland, California, was instrumented with permanent accelerometers and was subjected to strong motion vibrations from the 1989 Loma Prieta earthquake. The work performed for this thesis included the following: Literature research - Literature on the construction and design of open piled wharves was researched. Particular attention was given to standard practice for seismic analysis, SUMMARY AND CONCLUSIONS 204 especially the modelling of the structures for dynamic analyses. Additionally, literature was researched for information on measurement and analysis methods used in this thesis, such as background on ambient vibration measurement theory and practice, as well as correlation methods for matching analytical results to experimental results. Preliminary structure modelling - Preliminary computer models were made for each of the two research structures. Parameters used to model the structures were determined from standard values and relationships used in practice. Field testing - The vertical piled wharf was instrumented for ambient vibration and free vibration measurements. Prior to conducting field tests, several issues had to be resolved. Determining a suitable test structure and obtaining use of the structure from the owners was an initial necessity. Subsequently, available equipment systems had to be evaluated, and a final system chosen for use in the vibration measurements. Once the instrumentation system was chosen, a testing scheme was created and data collected. The data obtained in the field measurements were analysed to determine the dynamic properties of the structure. These results were compared to the preliminary analysis results. Correlation analysis and parametric study - A correlation analysis was performed between the test data obtained for the vertical piled wharf and the analytical model. A new analytical model with dynamic properties matching the measured properties was created. The new computer model was subsequently used in a parametric study. The purpose of the parametric study was to determine which modelling parameters were most important SUMMARY AND CONCLUSIONS 205 in the dynamic modelling of a piled wharf structure. Parameters studied were the pile length to fixity, structural member section properties, structure mass, and non-structural restraints. Analysis of strong motion data - Strong motion data was obtained from the California Strong Motion Instrumentation Program. This data was rigorously analysed for indications of differential motion between different sections of the structure that were separated by isolation joints. The data were also analysed to determine the dynamic characteristics of one of the sections of the wharf. The ambient vibration field testing and subsequent analysis resulted in the detection of seven natural frequencies of the site and the structure. Natural frequencies ranged from 0.3 Hz for the free field mode to 18.9 Hz for the third deck bending mode of the structure. A list of the measured natural frequencies is as follows: overall site mode 0.3 Hz first longitudinal mode 2.6 Hz first torsional mode 3.8 Hz first transverse mode 3.9 Hz first deck bending mode 7.9 Hz second deck bending mode 11.3 Hz third deck bending mode 18.9 Hz SUMMARY AND CONCLUSIONS 206 The order of the first rigid body torsional mode and the first rigid body transverse mode determined by the preliminary analytical model and by the experimental results were interchanged. Free vibration measurements were used to determine the damping of the structure. Results of the analysis showed that the damping for the third mode o f vibration is approximately 8% by the log decrement method o f damping determination. Damping for the other modes o f vibration could not be determined from the available data. Comparison o f the preliminary computer analysis to the results determined in the field testing showed that the frequency results did not match very well. Discrepancies in the two sets o f results ranged anywhere from 13% to 51%. Also, two of the modes were interchanged, as mentioned above, and some o f the analytical frequencies were not detected in the field measurements. A new correlated model used for a parametric study was developed by calibrating the model to results determined in the field testing. Analysis of the batter piled wharf subjected to strong motion vibrations showed that the individual sections of the wharf moved independently o f one another, although the wharf motions are highly influenced by the free field motions. Frequency analysis resulted in the determination o f one transverse natural frequency at 3.03 Hz for one o f the independent sections of the structure. This frequency was approximately verified by the results of the preliminary analytical model and a response spectrum analysis. The response spectrum analysis showed that amplification of the ground vibrations generally occurred SUMMARY AND CONCLUSIONS 207 above 2.9 Hz in both of the principal directions. Peaks in the structure response spectrum in the transverse direction show that a natural frequency of the middle portion of the wharf may occur at approximately 3.2 Hz. 8.2 CONCLUSIONS The objective of this project was to assess the accuracy of different modelling assumptions used in determining the dynamic properties of open piled wharves. It was found that the relative importance of the parameters examined in the study were as follows: restraint from non-structural members; pile length to fixity; pile and pile cap stiffness; and structure mass. The first two parameters can make a large difference in the dynamic properties of the structure, while the final two parameters are relatively insignificant when reasonable values are selected for each of the parameters. Based on the results of this study, it appears that the natural frequencies of vibration of open piled wharves under strong motion conditions can be estimated satisfactorily using current methods of determining the parametric values used in the analytical model. In contrast these methods are not appropriate for evaluation of the dynamic response of wharves at low levels of vibration. SUMMARY AND CONCLUSIONS 208 8.3 FURTHER RESEARCH Several items were not included in the scope of this study. In this connection, the following research and additional work is recommended: re-analyse the vertical piled wharf computer model using a different finite element program to incorporate gap/friction elements not available in SAP90; incorporate the effects of added mass and additional damping due to the presence of the surrounding water; conduct additional ambient vibration measurements on the vertical piled wharf using more measurement locations; conduct forced vibration tests on the vertical piled wharf; conduct linear and non-linear time history analyses of the vertical pile wharf; conduct an ambient vibration survey on the Oakland Outer Harbour Wharf using more measurement locations than were employed in the strong motion measurement program; and perform research on the lateral capacity of piles on slopes. 209 REFERENCES Abdel-Ghaffar A.M. , Scanlan R.H. (1985). Ambient Vibration Testing of the Golden Gate Bridge: I. Suspended Structure. Journal of Engineering Mechanics, ASCE, Vol.111, (pp.463-482). Bendat J .S. , Piersol A . G . , (1986). Random Data Analysis and Measurement Procedures. New York: John Wiley & Sons. Brownjohn J.M.W. (1988). Assessment of Structural Integrity by Dynamic Measurements. Doctoral Dissertation, University of Bristol, United Kingdom. Buckle I.G., Mayes R.L., Button M.R. (1986). Seismic Design and Retrofit Manual for Highway Bridges. Report to the United States Department of Transportation, Federal Highway Administration, Report No. FHWA-IP-86-6. California Department of Conservation, Division of Mines and Geology, Office of Strong Motion Studies, (1989). Strong Motion Ground Shaking from the Loma Prieta Earthquake of October 17,1989 and its relation to Near-Surface Geology in the Oakland Area. Report No. O S M S 89-07. Clough R.W., Penzien J . (1975). Dynamics of Structures. New York, New York: McGraw Hill. Collins M.P., Mitchell D. (1991). Prestressed Concrete Structures. Englewood Cliffs, New Jersey: Prentice Hall. Cooper S .E . (1991). Ductile Frames are Tough for Earthquakes. Civil Engineering, ASCE, Vol. 58, (pp. 61-63). Davisson M.T. (1970). Lateral Load Capacity of Piles. Highway Research Record No. 333 - Pile Foundations: High way Research Board. Diehl J .G . (1991). Ambient Vibration Survey: Application Theory and Analytical Techniques, Application Note No. 3. Pasadena, CA.: Kinemetrics. REFERENCES 210 Douglas B.M., Reid W.H. (1982). Dynamic Tests and System Identification of Bridges. Journal of the Structural Division ASCE, Vol. 108, (pp 2295-2312). Erickson B.P., Anderson D.G., Wittkop R.C. (1988). Designing to a Fault. Civil Engineering, ASCE, Vol. 58, (pp. 58 - 60). Ewins D.J. (1984). Modal Testing: Theory and Practice. New York, New York: John Wiley and Sons. Felber A .J . (1993). Development of a Hybrid Bridge Evaluation System. Doctoral Dissertaion, University of British Columbia, Canada. Fontinos G.C. , Serventi G. , Toda Y. (1988). Wharf Design and Construction Port of Oakland. Proc. Pod's 83, ASCE, (pp. 141-151). Gaythwaite J . (1990). Design of Marine Facilities for the Berthing, Mooring and Repair of Vessels. New York, New York: Van Nostrand Reinhold. Hudson D.E. (1964). Resonance Testing of Full-Scale Structures. Journal of the Engineering Mechanics Division, ASCE, Vol. 90, (pp. 1-19). Hudson D.E. (1977). Dynamic Tests of Full-Scale Structures. Journal of the Engineering Mechanics Division, ASCE, Vol. 103, (pp. 1141-1157). Hudson D.E. (1979). Reading and Interpreting Strong Motion Accelerograms. Earthquake Engineering Research Institute. Berkeley, California. Horyna, T. (1995). Dynamic Analysis of Bridges with a Laminated Wood Deck. Master's Thesis Dissertation, University of British Columbia, Canada. Kemp B.G., Ventura C .E . , Anderson D.L., Felber A .J . (1995). Ambient Vibration Measurement of Ruskin Dam for Seismic Assessment. Proc. Seventh Canadian Conference on Earthquake Engineering, Montreal Quebec, (pp. 641-648). REFERENCES 211 Latendresse V., Ventura, C . E . (1994). Bacchus Users Guide. Department of Civil Engineering, Department Software Collection, University of British Columbia, Canada. Luz E. (1987). Experimental Modal Analysis of Large Scale Structures. Proc. International Conference on Mechanical Dynamics, (pp. 257-262). Shenyang, China. McNutt S.R., Sydnor R.H. Ed. (1990). The Loma Prieta (Santa Cruz Mountains), California, Earthquake of 17 October 1989, California Department of Conservation, DMG, Special Publication 104 (pp. 29-46). Ngok M.T. (1982). The Determination of Structural Dynamic Properties of Three Buildings in Vancouver, BC from Ambient Vibration Surveys. Master's Thesis Dissertation, University of British Columbia, Canada. Nigbor R.L. (Undated). Golden Gate Bridge Ambient Vibration Survey, Application Note No. 26. Pasadena, CA.: Kinemetrics. Norris G. , Siddharthan R. (1991). Soil-Foundation-Structure Behaviour at the Oakland Outer Harbour Wharf. Center for Civil Engineering Earthquake Research, Report No. 91-2, University of Nevada, Reno. Quinn A.D. (1971). Design and Construction of Ports and Marine Structures. New York, New York: McGraw Hill. Okamoto S. (1984). Introduction to Earthquake Engineering. Tokyo, Japan: University of Tokyo Press. Paultre P., Proulx J . , Duron Z.H., Mai T.M., Im O. (1992). Dynamic Testing of Outardes 3 Gravity Dam. Proc. Tenth World Conference on Earthquake Engineering, Madrid, Spain, (pp 3571-3577). Schuster N.D. (1994). Dynamic Characteristics of a 30 Storey Building During Construction Detected From Ambient Vibration Measurements. Master's Thesis Dissertation, University of British Columbia, Canada. REFERENCES 212 Shakal A., Huang M., Reichle M., Ventura C , Cao T., Sherburne R., Savage M., Darragh R., and Petersen C. (1989). CSMIP Strong Motion Records from the Santa Cruz Mountains (Loma Prieta), California Earthquake of 17 October 1989, C D M G , Office of Strong Motion Studies, #OSMS 89-06. Shakal A., DeLisle M., Reichle M., and Darragh R. (1989). Strong Motion Ground Shaking from the Loma Prieta Earthquake of October 17,1989 and its relation to Near-Surface Geology in the Oakland Area, Calif. Dept. of Conservation, DMG, Special Publication 104, pp. 29-46. Shrivastava S.M., Hunt S .G . , (1989). Seismic Design of New Wharf for Squamish Terminals, British Columbia, Canada. Proc. Port's 89, ASCE, (pp. 654-663) The Society for Experimental Mechanics, Inc. (1993). Modal Analysis: Theory and Application Course. Torseth D.E., (1984). Port of Seattle Seismic Waterfront Design. Lifeline Earthquake Engineering Specialty Conference, ASCE, (147-153). Trifunac M.D. (1972). Comparison Between Ambient and Forced Vibrations Experiments. Earthquake Engineering and Structural Dynamics. Vol. 1, (pp. 133-150). Tsuchida H. (1980). Introduction to Earthquake Resistant Desgin for Port Structures. Earthquake Resistant Structure Laboratory, PHRI, Ministry of Transport, Japan. Tsuchida H. (1980). Lecture Notes on Earthquake Resistant Design of Trestle Type Piers, Sheetpile Bulkheads, and Cellular Bulkheads. Earthquake Resistant Structure Laboratory, PHRI, Ministry of Transport, Japan. Ventura C . E . (1992). Speceq User's Guide - Version Modified for PC . Department of Civil Engineering, Department Software Collection, University of British Columbia, Canada. Ventura C . E . (1993). MAC User's Guide. Department of Civil Engineering, Department Software Collection, University of British Columbia, Canada. REFERENCES 213 Ventura C .E . , Stienke, R (1993). Elastic Dynamic Analysis of a 49 Storey Instrumented Steel Structure Shaken By Loma Prieta Earthquake. Department of Civil Engineering Internal Report, University of British Columbia, Canada. Werner S.D., Hung S.J . (1984). Seismic Response of Port and Harbour Facilities. Lifeline Earthquake Engineering Specialty Conference, ASCE, (pp. 154-175). Wilson E.L. and Habibullah A., 1992. "SAP90 User's Manual". Berkeley, California. A P P E N D I X A S A M P L E S E T O F AMBIENT VIBRATION TIME HISTORIES APPENDIX A - SAMPLE SET OF AMBIENT VIBRATION TIME HISTORIES APPENDIX A - SAMPLE SET OF AMBIENT VIBRATION TIME HISTORIES 216 APPENDIX A - SAMPLE SET OF AMBIENT VIBRATION TIME HISTORIES 217 APPENDIX A - SAMPLE SET OF AMBIENT VIBRATION TIME HISTORIES 219 APPENDIX A - SAMPLE SET OF AMBIENT VIBRATION TIME HISTORIES 220 APPENDIX A - SAMPLE SET OF AMBIENT VIBRATION TIME HISTORIES 222 A P P E N D I X B S A M P L E S E T O F F R E E VIBRATION TIME HISTORIES SAMPLE SET OF FREE VIBRATION TIME HISTORIES 224 E ° « co co so N LO O IT) i n C o ® co LE > C +2 < T3 "5. £ < T3 0 S i CD > W" •M _ GOs-CD ° h-j> CO •M > O CO coal E io m O CO Oco + ci q b r»o o--6 ' +d — + q d coo o--d ' + d — + o d C O esq o r g d ' +d — + q d d ' 1 SAMPLE SET OF FREE VIBRATION TIME HISTORIES 225 1 o 05 cr -co ° ^ ~ » - ~ CO J- o.-IS O «-9 LO O i n LD c s c o 2 'co CO UL 3 > C © . c < "5 CD "D 3 "5. < CO. "D _CD C L £ S i CD CO •M CO. CD \—. +-* > co GLS| E q d CO + o + o qj, cn.- . - N d ' +d — + o d CN d ' +d — + q d C O 5TP op CMr- T - C O O ' +d — + o d d • SAMPLE SET OF FREE VIBRATION TIME HISTORIES 226 SAMPLE SET OF FREE VIBRATION TIME HISTORIES 227 0) 4 -£ ° o o >-CM o m i o "3 CO H P , in o in Sir O E < CO "O CD = all IO..E sl 0 CO _ C O -CD +-> > CO • O oo Oo q o" oo Oo OO Oo od dq _,- Jo CN) • -o 9 ; CN CO S A M P L E S E T O F F R E E V I B R A T I O N T I M E H I S T O R I E S 228 o o • E s q o '5 cr > a a tn N 3* * 9 LO O IT) C m c O £ CO ul o > c CD CO JZ < "co CD "5. £ < CO o 1_ c > CO • M _ <u; H J CO o CO E IO l a 00 + o 6 in.- .-in - ' + d — + o d 5<°. <o.s d ' + d — + o d cjT'0. <°.FJ d ' + d — + o d d ' 5 A P P E N D I X C O A K L A N D O U T E R H A R B O U R W H A R F U N C O R R E C T E D A C C E L E R A T I O N TIME HISTORIES OAKLAND OUTER HARBOUR WHARF UNCORRECTED ACCELERATION TIME HISTORIES s t r u c t u r a l Response - Area J 230 o a 3 o T3 C a o o 1 :vc to o p. C3 o t o CO •I 55 o •H w 4J U 7) o c: >4-i c f I: 4-1 o H * J CO U a> to i o to CM A P P E N D I X D O A K L A N D O U T E R H A R B O U R W H A R F P O W E R S P E C T R A L DENSITIES O A K L A N D O U T E R H A R B O U R W H A R F - P S D FUNCTIONS 232 O CN O CO CO NT CN O suo|||!i/\| A}!SU9Q |BJp9dS J8M0<-J O A K L A N D O U T E R H A R B O U R W H A R F - P S D F U N C T I O N S 233 CO CD suoi||!iAj AijsueQ |Bjpads Ja/v\od CN OAKLAND OUTER HARBOUR WHARF - PSD FUNCTIONS 234 SUOJIMIAJ A J J S U S Q iBjpeds J9M0d OAKLAND OUTER HARBOUR WHARF - PSD FUNCTIONS 235 A J I S U G Q |BJjoeds J9M0d OAKLAND OUTER HARBOUR WHARF - PSD FUNCTIONS 236 OAKLAND OUTER HARBOUR WHARF - PSD FUNCTIONS 237 I ,—I 1 1 1 o CO CO -3" CN O suojiniAi OAKLAND OUTER HARBOUR WHARF - PSD FUNCTIONS 238 OAKLAND OUTER HARBOUR WHARF - PSD FUNCTIONS 239 IT) m SUOJIHIAJ A J J S U S Q iBjpeds JaMOd OAKLAND OUTER HARBOUR WHARF - PSD FUNCTIONS 240 suojiniAl OAKLAND OUTER HARBOUR WHARF - PSD FUNCTIONS 241 OAKLAND OUTER HARBOUR WHARF - PSD FUNCTIONS 242 O suoj||!|/\| A P P E N D I X E O A K L A N D O U T E R H A R B O U R W H A R F A C C E L E R A T I O N R E S P O N S E S P E C T R A OAKLAND OUTER HARBOUR WHARF - ACCELERATION RESPONSE SPECTRA 244 L 1 1 1 1 1 O C N T - CO CO ^ CM T*~ o o o d asuodsay OAKLAND OUTER HARBOUR WHARF - ACCELERATION RESPONSE SPECTRA 246 I 1 I I I I Q CN T - 00 CO CM ^ • o o" o* d 9SU0dS9y OAKLAND OUTER HARBOUR WHARF - ACCELERATION RESPONSE SPECTRA 247 CN r- CO CO ^ CN o o o d esuodse^j OAKLAND OUTER HARBOUR WHARF - ACCELERATION RESPONSE SPECTRA 248 LO o o o o o o o CN O CO CD CN asuodsay OAKLAND OUTER HARBOUR WHARF - ACCELERATION RESPONSE SPECTRA 2 4 9 oo d co d d CN d esuodsey OAKLAND OUTER HARBOUR WHARF - ACCELERATION RESPONSE SPECTRA 250 I I I I I I I I to C D ^ C N T - O O C D ' ^ - C N T-1 i - : d o d o e s u o d s e y OAKLAND OUTER HARBOUR WHARF - ACCELERATION RESPONSE SPECTRA 2 5 1 I 1 1 1 1 1 O CM T - CO CO CM ^ • d o ci d e s u o d s e y OAKLAND OUTER HARBOUR WHARF - ACCELERATION RESPONSE SPECTRA OAKLAND OUTER HARBOUR WHARF - ACCELERATION RESPONSE SPECTRA 253 OAKLAND OUTER HARBOUR WHARF - ACCELERATION RESPONSE SPECTRA 254 e s u o d s e y 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0050354/manifest

Comment

Related Items