Dynamic Interaction Between Vehicle and Infrastructure Experiment ...

Eidgenössische Materialprüfungs- und Forschungsanstalt Laboratoire fédéral d'essai des matériaux et de recherche Laboratorio federale di prova dei materiali e di ricerca Institut federal da controlla da material e da retschertgas Swiss Federal Laboratories for Materials Testing and Research

EMPA Ueberlandstrasse 129 CH-8600 Duebendorf Tel. +41-1-823 55 11 Fax +41-1-823 44 55

OECD IR 6 DIVINE Project Dynamic Interaction Between Vehicle and Infrastructure Experiment

Element 6, Bridge Research FINAL REPORT

By

Reto Cantieni and Walter Krebs, EMPA Duebendorf, Switzerland and Rob Heywood, QUT Brisbane, Australia

EMPA Test Report No. 153'031

Duebendorf, June 2000

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia, Summary

2 /169 8. June 2010

Report prepared at: EMPA, SWISS FEDERAL LABORATORIES FOR MATERIALS TESTING AND RESEARCH UEBERLANDSTRASSE 129 CH-8600 DUEBENDORF SWITZERLAND and QUT, QUEENSLAND UNIVERSITY OF TECHNOLOGY FACULTY OF BUILT ENVIRONMENT AND ENGINEERING 2 GEORGE STREET GPO BOX 2434 BRISBANE Q 4001 AUSTRALIA

COPYRIGHT © 2000 BY EMPA AND QUT ISBN 3-905594-05-6 Copies of this report are available, priced CHF 30.-- or USD 20.-- each, from: EMPA Department Structural Engineering Ueberlandstrasse 129 CH-8600 Duebendorf Switzerland Voice: +41 1 823 55 11 Fax: +41 1 823 44 55 e-mail: [email protected] www.empa.ch

PRINTED BY EMPA


3 /169 8. June 2010

CONTENTS SUMMARY ..............................................................................................................................18 S.1 General ........................................................................................................................18 S.2 The Problem to be Solved ............................................................................................18 S.3 Test Parameters...........................................................................................................19 S.3.1 Bridges, Pavements..............................................................................................19 S.3.2 Vehicles ................................................................................................................20 S.4 Test Procedures and Data Processing.........................................................................21 S.5 Dynamic Bridge Response ...........................................................................................21 S.5.1 The Results...........................................................................................................21 S.5.2 Discussion.............................................................................................................22 S.6 Conclusions..................................................................................................................23 1.

INTRODUCTION..............................................................................................................25 1.1 General ........................................................................................................................25 1.2 The Problem to be Solved ............................................................................................25 1.3 Previous Knowledge.....................................................................................................25 1.3.1 Experiment vs. Numerical Analysis .......................................................................25 1.3.2 Frequency Matching .............................................................................................26 1.3.3 Pavement Unevenness .........................................................................................28 1.3.4 Relationship "Natural Frequency" vs. "Maximum Span" of a Bridge .....................29 1.4 Strategy of the OECD DIVINE Bridge Research Project ..............................................31 1.5 Swiss Tests on Medium to Long-Span Bridges ............................................................33 1.6 Australian Tests on Short-Span Bridges ......................................................................34


4 /169 8. June 2010

1.7 Summary - Test Bridges and Vehicles .........................................................................35 2.

TEST PARAMETERS ......................................................................................................36 2.1 Bridges .........................................................................................................................36 2.1.1 Sort Bridge (CH) ...................................................................................................38 2.1.2 Deibüel Bridge (CH) ..............................................................................................39 2.1.3 Föss Bridge (CH) ..................................................................................................40 2.1.4 Lawsons Creek Bridge (AUS) ...............................................................................41 2.1.5 Coxs River Bridge (AUS) ......................................................................................42 2.1.6 Camerons Creek Bridge (AUS).............................................................................43 2.1.7 Cromarty Creek Bridge (AUS)...............................................................................44 2.2 Pavement Profiles ........................................................................................................45 2.2.1 Longitudinal Profiles of the Swiss Bridges ............................................................45 2.2.2 Longitudinal Profiles of the Australian Bridges......................................................45 2.2.3 Spectral Density of the Bridges Pavement Profile.................................................45 2.2.4 Unevenness Rating According to ISO/TC 108 ......................................................47 2.2.5 Axle-hop plank ......................................................................................................50 2.3 Vehicles........................................................................................................................51 2.3.1 The NRC Test Vehicle Used for the Swiss Bridge Tests.......................................51 2.3.2 The Test Vehicles Used in Australia .....................................................................53

3.

INSTRUMENTATION, DATA ACQUISITION ..................................................................57 3.1 Instrumentation of Swiss Bridges .................................................................................57 3.1.1 Sort Bridge ............................................................................................................59 3.1.2 Deibüel Bridge ......................................................................................................59


5 /169 8. June 2010

3.1.3 Föss Bridge...........................................................................................................59 3.2 Instrumentation of Australian Bridges...........................................................................59 3.2.1 Lawsons Creek Bridge ..........................................................................................60 3.2.2 Coxs River Bridge .................................................................................................60 3.2.3 Camerons Creek Bridge........................................................................................61 3.2.4 Cromarty Creek Bridge .........................................................................................61 3.3 Vehicle Instrumentation................................................................................................62 3.3.1 The NRC Vehicle ..................................................................................................62 3.3.2 The Boral Vehicle..................................................................................................64 3.3.3 The SA vehicle......................................................................................................64 3.4 Data Acquisition ...........................................................................................................64 3.4.1 Tests in Switzerland..............................................................................................64 3.4.2 Tests in Australia ..................................................................................................66 4.

TEST PROCEDURES AND PROGRAM .........................................................................68 4.1 Tests in Switzerland .....................................................................................................68 4.2 Tests in Australia..........................................................................................................69

5.

DATA PROCESSING ......................................................................................................71 5.1 Time Domain Analysis Methods as Used in Switzerland..............................................71 5.1.1 Bridge Response Data Processing Methods.........................................................71 5.1.2 Wheel Load Processing Methods .........................................................................71 5.2 Time Domain Analysis Methods as Used in Australia ..................................................72 5.2.1 Bridge Response Data Processing Methods.........................................................72 5.2.2 Wheel Load Processing Methods .........................................................................72


6 /169 8. June 2010

5.3 Frequency Domain Analysis Methods ..........................................................................73 5.3.1 General .................................................................................................................73 5.3.2 Wheel Load Spectra as Determined in Switzerland ..............................................73 5.3.3 Vehicle/Bridge Interaction as Investigated in Switzerland .....................................74 5.3.4 Frequency Domain Analysis Methods as Used in Australia ..................................74 6.

TEST RESULTS ..............................................................................................................75 6.1 Time Domain Analysis: Bridge Response Signals........................................................75 6.1.1 Sort Bridge ............................................................................................................75 6.1.2 Deibüel Bridge ......................................................................................................76 6.1.3 Föss Bridge...........................................................................................................78 6.1.4 Lawsons Creek Bridge ..........................................................................................80 6.1.5 Coxs River Bridge .................................................................................................82 6.1.6 Camerons Creek Bridge........................................................................................85 6.1.7 Cromarty Creek Bridge .........................................................................................90 6.2 Time Domain Analysis: Dynamic Wheel Load Signals .................................................93 6.2.1 The NRC Vehicle; Typical Wheel Load Time Signals ...........................................93 6.2.2 The NRC Vehicle; Dynamic Load Coefficients......................................................95 6.2.4 The Boral Vehicle; Typical Wheel Load Time Signals.........................................100 6.2.5 The Boral Vehicle; Dynamic Load Coefficients ...................................................105 6.3 Frequency Domain Analysis of Dynamic Wheel Load Signals ...................................109 6.3.1 The NRC Vehicle; Frequency Spectra ................................................................109 6.3.2 The NRC Vehicle; Mode Shapes ........................................................................114 6.3.3 The Boral Vehicle; Frequency Spectra ...............................................................118


7 /169 8. June 2010

6.4 Vehicle/Bridge Interaction: Interpretation of the Wheel Load Signals.........................119 6.4.1 Basics on Vehicle/Bridge Interaction: Medium to Long-Span Bridges ................119 6.4.2 Basics on Vehicle/Bridge Interaction: Short-Span Bridges..................................120 6.4.3 Swiss Tests.........................................................................................................121 6.4.4 Australian Tests ..................................................................................................126 6.5 Vehicle/Bridge Interaction: Combining Bridge Response and Wheel Load Spectra...126 6.5.1 Tests in Switzerland: General .............................................................................126 6.5.2 Deibüel Bridge, Steel-Suspended NRC Test vehicle ..........................................127 6.5.3 Deibüel Bridge, Air-Suspended NRC Test Vehicle .............................................127 6.5.4 Sort Bridge, Steel-Suspended NRC Test Vehicle ...............................................127 6.5.5 Sort Bridge, Air-Suspended NRC Test Vehicle ...................................................128 6.5.6 Tests in Australia ................................................................................................133 7.

DISCUSSION.................................................................................................................134 7.1 Dynamic Bridge Response .........................................................................................134 7.1.1 Sort Bridge (CH) .................................................................................................135 7.1.2 Deibüel Bridge (CH) ............................................................................................135 7.1.3 Föss Bridge (CH) ................................................................................................135 7.1.4 Lawsons Creek Bridge (AUS) .............................................................................136 7.1.5 Coxs River Bridge (AUS) ....................................................................................136 7.1.6 Camerons Creek Bridge (AUS)...........................................................................137 7.1.7 Cromarty Creek Bridge (AUS).............................................................................143 7.2 DIVINE Testing versus Swiss Tests ...........................................................................145 7.3 Dynamic Wheel Loads (DLC's) ..................................................................................148


8 /169 8. June 2010

7.4 Wheel Load Spectra as a Function of Vehicle Speed ................................................149 7.5 Frequency Matching...................................................................................................149 7.5.1 Medium to Long-Span Bridges in Switzerland ....................................................150 7.5.2 Short-Span Bridges in Australia ..........................................................................150 7.6 Vehicle/Bridge Interaction...........................................................................................152 7.6.1 Medium to Long-Span Bridges............................................................................152 7.6.2 Short-Span Bridges: Bridge Damping .................................................................152 7.7 Limiting the Dynamics of Wheel Loads and Bridge Response ...................................155 8.

CONCLUSIONS AND RECOMMENDATIONS..............................................................158 8.1 Basics, Basic Physics, Parameters ............................................................................158 8.2 The Basic Physics ......................................................................................................159 8.2.1 Dynamic Wheel Loads ........................................................................................159 8.2.2 Dynamic Bridge Response..................................................................................160 8.3 The Parameters..........................................................................................................161 8.3.1 Bridge Parameters ..............................................................................................161 8.3.2 Pavement Parameters ........................................................................................162 8.3.3 Vehicle Parameters.............................................................................................162 8.4 Conclusions................................................................................................................163 8.5 Recommendations .....................................................................................................164

9.

ACKNOWLEDGMENTS ................................................................................................167 9.1 Swiss Research..........................................................................................................167 9.2 Australian Research ...................................................................................................167

10. LIST OF ANNEXES .......................................................................................................169


9 /169 8. June 2010

LIST OF FIGURES

Fig. 1

Bridge deflection time history with the definitions of the dynamic increment φ [%], the fundamental frequency f [Hz] and associated damping decrement δ (or, damping coefficient ζ in percent of critical: ζ = δ/2π), (from [Cantieni, 1983])........................................................................................................27

Fig. 2

Dynamic increments as a function of the bridge fundamental frequency; single steel-suspended two-axle trucks (from [Cantieni, 1983]). .........................................27

Fig. 3

Dynamic increments as a function of the bridge fundamental frequency for passages with a single, steel-suspended two-axle vehicle crossing a 300 x 50 mm plank (from [Cantieni, 1983]). .............................................................................29

Fig. 4

Relationship between maximum span L and fundamental frequency f of beamtype highway bridges (from [Cantieni, 1983]). ...........................................................30

Fig. 5

Relationship between maximum span L and fundamental frequency f for shortspan Australian bridges (RC: reinforced concrete; PC: prestressed concrete; cont: continuous).......................................................................................................30

Fig. 6

Histogram of maximum span of roughly 200 highway bridges in Switzerland (from [Cantieni, 1983]). .............................................................................................32

Fig. 7

Histogram of maximum span of highway bridges in Australia (derived from [NAASRA, 1985]). .....................................................................................................33

Fig. 8

Amplification of the Single-Degree-of-Freedom bridge model to linear scales (Note: Bridges are arranged in order of frequency)...................................................37

Fig. 9

Photograph of the Sort Bridge. To the left the Ticino River, to the right the N2 motorway, both surpassed by the bridge (EMPA Photo No. 101'104/9)....................38

Fig. 10

Photograph of the Deibüel Bridge. To the left the bridge tested, to the right the "twin" bridge carrying two-way traffic during the tests (EMPA Photo No. 101'069/36). ..............................................................................................................39

Fig. 11

Photograph of the Föss Bridge (exact designation: Föss 2). In the background the arch bridge Föss 3 on the Gotthard Pass Highway (EMPA Photo No. 101'049/17). ..............................................................................................................40


10 /169 8. June 2010

Fig. 12

Photograph of the Lawsons Creek Bridge.................................................................41

Fig. 13

Photograph of the Coxs River Bridge. .......................................................................42

Fig. 14

Photograph of the Camerons Creek Bridge. .............................................................43

Fig. 15

Photograph of the Cromarty Creek Bridge ................................................................44

Fig. 16

Classification of roads according to [ISO, 1995]. ......................................................47

Fig. 17

Cross-section through the Australian axle-hop plank (AHP). ....................................50

Fig. 18

The NRC test vehicle. ...............................................................................................51

Fig. 19

The NRC test vehicle during the tests on Sort Bridge (EMPA Photo No. 101'070/30). ..............................................................................................................52

Fig. 20

The BA test vehicle crossing the bridge over Lawsons Creek...................................53

Fig. 21

Scheme of the wire-supported displacement measurement technology as used in Switzerland. The axial force in the wire is produced by stretching of the spring. Measuring of the stretched spring length allows checking of the nominal force being close to the nominal value of 300 N. .........................................58

Fig. 22

Wire-supported displacement measurement technology fitted to the bridge over Camerons Creek. ..............................................................................................60

Fig. 23

Instrumentation for measuring dynamic wheel forces ...............................................62

Fig. 24

NRC test vehicle instrumentation: an "undercover" shear gage and an accelerometer located close to the wheel hub can be seen (EMPA Photo No. 101'155/5)...........................................................................................................63

Fig. 25

The data acquisition systems located in the NRC vehicle's driver cabin: to the left: the Canadian MegaDac system, to right (orange) the Swiss PCM system (EMPA Photo No. 101'176/19). .................................................................................66

Fig. 26

Sort Bridge mid-span deflection for the steel-suspended NRC vehicle, v = 29.9 km/h; φ = 2.01%. .........................................................................................75

Fig. 27

Sort Bridge mid-span deflection for the air-suspended NRC vehicle, v = 75.2 km/h; φ = 25.45%. .......................................................................................75

Fig. 28

Dynamic increments vs. speed for the Sort Bridge mid-span deflection. ..................76


11 /169 8. June 2010

Fig. 29

Deibüel Bridge mid-span deflection for the steel-suspended NRC vehicle, v = 50.6 km/h; φ = 20.84%. .......................................................................................76

Fig. 30

Deibüel Bridge mid-span deflection for the air-suspended NRC vehicle, v = 37.3 km/h; φ = 3.73%. .........................................................................................77

Fig. 31

Dynamic increments vs. speed for the Deibüel Bridge mid-span deflection. .............77

Fig. 32

Föss Bridge mid-span deflection for the steel-suspended NRC vehicle, v = 14.9 km/h; φ = 3.86%. .........................................................................................78

Fig. 33

Föss Bridge mid-span deflection for the air-suspended NRC vehicle, v = 27.9 km/h; φ = 3.11%. .........................................................................................78

Fig. 34

Dynamic increments vs. speed for the Föss Bridge mid-span deflection. .................79

Fig. 35

Lawsons Creek Bridge, mid-span deflections D1, D3, D5 and D7, BS vehicle, v = 60 km/h. ..............................................................................................................80

Fig. 36

Lawsons Creek Bridge, mid-span deflections D1, D3, D5 and D7, BA vehicle, v = 99 km/h. ..............................................................................................................80

Fig. 37

Dynamic increments vs. speed (positive: west-bound) for Lawsons Creek Bridge, deflection D5. Solid squares: steel suspension, open squares: air suspension. ...............................................................................................................81

Fig. 38

Dynamic increment vs. speed for Lawsons Creek Bridge with axle-hop plank, deflection D5. Solid squares: steel suspension, open squares: air suspension. .......81

Fig. 39

Coxs River Bridge, mid-span deflection D(2,3), BS vehicle, v = 96 km/h. .................82

Fig. 40

Coxs River Bridge mid-span deflection D(2,3) for the BA vehicle, v = 77 km/h.........83

Fig. 41

Dynamic increments vs. speed for Coxs River Bridge, (negative speed: southbound, positive speed: northbound), deflection D(2,3). Solid squares: steel leaf suspension, open squares: air suspension; dotted line: air suspension, no shock absorbers...............................................................................84

Fig. 42

Dynamic increment vs. speed for Coxs River Bridge with axle hop plank, deflection D(2,3), northbound. Solid squares: steel-suspended vehicle, open squares: air-suspended vehicle; dashed line: air suspension with unequal spacing......................................................................................................................84


12 /169 8. June 2010

Fig. 43

Camerons Creek Bridge mid-span deflection for the BS vehicle, v = 80 km/h, northbound. ...............................................................................................................86

Fig. 44

Camerons Creek Bridge mid-span deflection for the BA vehicle, v = 62 km/h, northbound. ...............................................................................................................86

Fig. 45

Camerons Creek Bridge mid-span deflection for the SA air-suspended vehicle, v = 63 km/h, northbound. ..........................................................................................87

Fig. 46

Dynamic increments vs. speed for Camerons Creek Bridge, (negative speed: southbound, positive speed: northbound). Solid symbols: steel suspension, open symbols: air suspension; Solid lines D(4,8), dashed lines D(1,8). ....................87

Fig. 47

Dynamic increments vs. speed for Camerons Creek Bridge, (negative speed: southbound, positive speed: northbound); SA air suspended vehicle; Solid lines D(4,8), dashed lines D(1,8)...............................................................................88

Fig. 48

Dynamic increment vs. speed for Camerons Creek Bridge, D(4,8) with axle hop plank (AHP), northbound. Solid symbols: BS vehicle, open symbols: BA vehicle. Note that the deflection D(1,8) is not influenced by the AHP in either direction. D(4,8) is not affected in the southbound direction. ....................................88

Fig. 49

Cromarty Creek Bridge mid-span deflection for the BS vehicle, v = 91 km/h, northbound. ...............................................................................................................90

Fig. 50

Cromarty Creek Bridge mid-span deflection for the BA vehicle, v = 92 km/h, northbound. ...............................................................................................................90

Fig. 51

Dynamic increments vs. speed for Cromarty Creek Bridge (positive speed: northbound - D(3,3), negative speed: southbound - D(3,4)), Solid symbols: steel suspension, open symbols: air suspension. .....................................................91

Fig. 52

Dynamic increment vs. speed for Cromarty Creek Bridge with axle hop plank (north-bound only). Solid symbols: steel-suspended vehicle, open symbols: air-suspended vehicle. ..............................................................................................91

Fig. 53

Dynamic wheel load time signals recorded for the passage of the steelsuspended NRC vehicle over Deibüel Bridge at v = 48.1 km/h. For the sake of clarity, the static wheel loads have been offset from their true values. The yaxis scaling is however correct for the dynamic wheel load parts. The true static wheel loads are indicated in Fig. 55.................................................................94


13 /169 8. June 2010

Fig. 54

Dynamic wheel load time signals recorded for the passage of the airsuspended NRC vehicle over Deibüel Bridge at v = 55.3 km/h. For the sake of clarity, the static wheel loads have been offset from their true values. The yaxis scaling is however correct for the dynamic wheel load parts. The true static wheel loads are indicated in Fig. 57.................................................................95

Fig. 55

DLC's for the steel-suspended NRC vehicle on the Deibüel Bridge Approach..........96

Fig. 56

DLC's for the steel-suspended NRC vehicle on the Deibüel Bridge. .........................97

Fig. 57

DLC's for the air-suspended NRC vehicle on the Deibüel Bridge Approach. ............98

Fig. 58

DLC's for the air-suspended NRC vehicle on the Deibüel Bridge..............................99

Fig. 59

Trailer tri-axle group dynamic wheel load time signals for the passage of the BA vehicle over Camerons Creek Bridge at v = 59 km/h over the axle-hop plank (AHP).............................................................................................................100

Fig. 60

Trailer tri-axle group dynamic wheel load time signals for the passage of the BS vehicle over Camerons Creek Bridge at v = 62 km/h over the axle-hop plank (AHP).............................................................................................................101

Fig. 61

Trailer tri-axle group dynamic wheel load time signals for the passage of the BS vehicle over Coxs River Bridge at v = 69 km/h northbound over the axlehop plank.................................................................................................................102

Fig. 62

Trailer tri-axle group dynamic wheel load time signals for the passage of the BA vehicle over Coxs River Bridge at v = 80 km/h southbound over the axlehop plank.................................................................................................................102

Fig. 63

Trailer tri-axle group dynamic wheel load time signals for the passage of the BA vehicle over Coxs River Bridge at v = 57 km/h northbound; no shock absorbers on the rear axle, no axle-hop plank. .......................................................104

Fig. 64

Coxs River Bridge; DLC, BS vehicle, northbound, no planks. .................................105

Fig. 65

Coxs River Bridge; DLC, BA vehicle, northbound, no planks. .................................105

Fig. 66

Coxs River Bridge; Dynamic load coefficient (DLC), BA vehicle, northbound; No shock absorbers on the rear axle, no planks. ....................................................106

Fig. 67

Camerons Creek Bridge; Dynamic load coefficient (DLC), BS vehicle, northbound, no planks.............................................................................................107


14 /169 8. June 2010

Fig. 68

Camerons Creek Bridge; Dynamic load coefficient (DLC), BA vehicle, northbound, no planks. Note that axles were spaced at 1.54 m during this test rather than the 1.23 m used in this report. ..............................................................107

Fig. 69

Cromarty Creek Bridge; Dynamic load coefficient (DLC), BS vehicle, northbound, no planks.............................................................................................108

Fig. 70

Cromarty Creek Bridge; Dynamic load coefficient (DLC), BA vehicle, northbound, no planks.............................................................................................108

Fig. 71

Example of a wheel load spectral density for the steel-suspended NRC vehicle on the Deibüel Bridge, wheel No. 14, v = 37.1 km/h. ..............................................109

Fig. 72

Example of a wheel load spectral density for the air-suspended NRC vehicle on the Deibüel Bridge, wheel No. 14, v = 42.2 km/h. ..............................................110

Fig. 73

Example of a wheel load spectral density for the steel-suspended NRC vehicle on the Deibüel Bridge, wheel No. 10, v = 57.3 km/h. ..............................................110

Fig. 74

Example of a wheel load spectral density for the air-suspended NRC vehicle on the Sort Bridge, wheel No. 06, v = 70.3 km/h.....................................................111

Fig. 75

Peak frequencies of wheel loads vs. speed for steel suspension on Deibüel Bridge, wheel No. 10. The respective highest peaks are connected through a solid line. .................................................................................................................112

Fig. 76

Peak frequencies of wheel loads vs. speed for air suspension on Sort Bridge for wheel No. 10. The respective highest peaks are connected through a solid line. .........................................................................................................................112

Fig. 77

Mode shapes of the NRC vehicle operational vibrations (the steer axle shown on the left-hand side, the trailing axles on the right-hand side of the diagrams). Δφ = phase angle in the complex motion of the driving and trailing axles................115

Fig. 78

Frequency of vibration modes vs. speed for steel suspension on the Deibüel Bridge......................................................................................................................116

Fig. 79

Frequency of vibration modes vs. speed for air suspension on the Deibüel Bridge......................................................................................................................116


15 /169 8. June 2010

Fig. 80

Camerons Creek Bridge; Power spectral densities for dynamic wheel load drivers side center wheel. Top: vehicle BA, v = 47 km/h, axle-hop plank. Bottom: vehicle BS, v = 67 km/h, axle-hop plank. ...................................................118

Fig. 81

"Real" vehicle/bridge interaction reduced to a Two-Degree-of-Freedom-Model......119

Fig. 82

The vehicle reduced to a forcing function acting on the bridge as an SDOFsystem.....................................................................................................................121

Fig. 83

Frequency vs. speed of the body bounce vibration (Mode shape Type 4) on the Dei-büel Bridge for steel suspension.................................................................122

Fig. 84

Histogram of the frequencies of the body bounce vibration on the Deibüel Bridge for steel suspension. ....................................................................................122

Fig. 85

Frequency vs. speed of the body bounce vibration (Mode shape Type 4) on the Dei-büel Bridge for air suspension. ...................................................................123

Fig. 86

Histogram of the frequencies of the body bounce vibration on the Deibüel Bridge for air suspension.........................................................................................123

Fig. 87

Histogram of the frequencies of the body bounce vibration on the Föss Bridge for steel suspension. ...............................................................................................124

Fig. 88

Histogram of the frequencies of the body bounce vibration on the Föss Bridge for air suspension....................................................................................................124

Fig. 89

Histogram of the wheel load frequencies of the steel-suspended NRC vehicle body bounce vibration on the Sort Bridge. ..............................................................125

Fig. 90

Histogram of the wheel load frequencies of the air-suspended NRC vehicle body bounce vibration on the Sort Bridge. ..............................................................125

Fig. 91

Passage of the steel-suspended NRC vehicle over the Deibüel Bridge at v = 32.1 km/h: (a) Time signals of the test vehicle's wheel 06 and the induced bridge response WG 21, (b) power spectral densities for the wheel load and bridge signals, (c) cross-power spectrum magnitude and (d) phase angle between the two signals. .........................................................................................129

Fig. 92

Passage of the air-suspended NRC vehicle over the Deibüel Bridge at v = 49.4 km/h: (a) Time signals of the test vehicle's wheel 06 and the induced bridge response WG 21, (b) power spectral densities for the wheel load and


16 /169 8. June 2010

bridge signals, (c) cross-power spectrum magnitude and (d) phase angle between the two signals. .........................................................................................130 Fig. 93

Passage of the steel-suspended NRC vehicle over the Sort Bridge at v = 70.0 km/h: (a) Time signals of the test vehicle's wheel 06 and the induced bridge response WG 32, (b) power spectral densities for the wheel load and bridge signals, (c) cross-power spectrum magnitude and (d) phase angle between the two signals. .........................................................................................131

Fig. 94

Passage of the air-suspended NRC vehicle over the Sort Bridge at v = 40.2 km/h: (a) Time signals of the test vehicle's wheel 06 and the induced bridge response WG 32, (b) power spectral densities for the wheel load and bridge signals, (c) cross-power spectrum magnitude and (d) phase angle between the two signals. .........................................................................................132

Fig. 95

Dynamic increments as a function of bridge fundamental frequency for twoaxle trucks (from [Cantieni, 1983]) incorporating OECD DIVINE test data: Solid squares: steel suspension; Open squares: air suspension. ....................................145

Fig. 96

Dynamic increments as a function of bridge fundamental frequency for passages with with a two-axle truck crossing a 300 x 50 mm plank (from [Cantieni, 1983]) incorporating OECD DIVINE test data for a 300 x 25 mm plank: Solid squares: steel suspension; Open squares: air suspension. .................146

Fig. 97

Maximum dynamic increments for the Deibüel Bridge deflection at the midpoint of the middle span and various steel suspended vehicles (from [Cantieni, 1987, 1988]). Description of the vehicle type coding: see previous page. .......................................................................................................................147

Fig. 98

Speed influences how the dynamic components of axles within a group combine...................................................................................................................151

Fig. 99

Amplification of the Single-Degree-of-Freedom bridge model to semilogarithmic scales (Note: Bridges are arranged in order of frequency). ..................153

Fig. 100 Camerons Creek Bridge deflection time signals measured at D(1,8) and D(4,8), Test Vehicle SA, no planks, v = 63 km/h, northbound.................................154 Fig. 101 Combinations of wavelengths in the road profile and vehicle speed leading to excitation of heavy vehicle suspensions..................................................................156


17 /169 8. June 2010

LIST OF TABLES

Table 1

Main parameters of the test bridges including the test track pavement quality. ......20

Table 2

Maximum dynamic increments φmax for vertical bridge deflection measured in a significant bridge location for the passages without axle-hop plank. ....................22

Table 3

Main parameters of the bridges tested in Switzerland and Australia.......................35

Table 4

Summary of the test bridges main parameters. ......................................................37

Table 5

Classification of road profiles according to Table C.2 given in [ISO, 1995].............48

Table 6

Pavement unevenness rating according to [ISO, 1995] (N/B = north bound, W/B = west bound)..................................................................................................49

Table 7

Test vehicle dimensions, weights (nominal) and suspensions - Australia. ..............55

Table 8

Low frequency suspension characteristics of Australian test vehicles [Sweatman, 1994]. ..................................................................................................56

Table 9

Comparison of dynamic response of Camerons Creek bridge to test vehicles. ......89

Table 10 Comparison of dynamic response of Cromarty Creek bridge due to test vehicles. ..................................................................................................................92 Table 11 Observed lengths in wheel load diagrams and distances between axles of the test vehicle. ...........................................................................................................113 Table 12 Passages without axle-hop plank: Maximum dynamic increments φmax measured for bridge vertical displacement in a significant location.......................134 Table 13 Passages with axle-hop plank: Maximum dynamic increments φmax measured for bridge vertical displacement in a significant location........................................134 Table 14 Qualitative relationship between bridge span and the amplification of dynamic wheel forces induced by short, medium and long wavelengths in the road profile for vehicles traveling in excess of 50 km/h. ................................................157

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia

18 / 169 8. June 2010

SUMMARY S.1

General

The OECD DIVINE Project was started in November 1992 and finished in November 1997. OECD is the Paris based "Organization for Economic Co-operation and Development", DIVINE stands for "Dynamic Interaction Between Vehicle and Infrastructure Experiment". DIVINE was divided into six research elements. Three of them were related to the experimental investigation of the behavior of flexible pavements, one to road simulators, one to the numerical simulation of a heavy vehicle driving down a road on a flexible pavement and one, Research Element 6, to the dynamic interaction between vehicles and bridges. The general layout of DIVINE as well as the condensed results and the technical and policy implications of all the research elements are given in the Final Technical Report [OECD, 1998] (see Annex I for the list of References). Here, the Final Report for Research Element 6 is presented. S.2

The Problem to be Solved

The goal of the DIVINE Project Element 6, "Bridge Research", was to experimentally establish the consequences, in terms of the dynamic response of bridges, of changing heavy commercial vehicle suspensions from the well known steel to modern air suspensions. To reach this goal, dynamic load tests using a single test vehicle have been performed in 1994/95 on seven highway bridges, three of them located in Switzerland and four in Australia. As seen from the bridge part of the highway transportation infrastructure the following changes in the vehicle suspension's parameters are important: • Dynamic wheel load frequencies: The dominating body bounce frequency is shifted from f ≈ 3 Hz for the steel-suspended to f ≈ 1.5...1.8 Hz for the air-suspended vehicles. Axle-hop vibrations, however, occur at similar frequencies f ≈ 8...20 Hz for both suspension types. • Damping: Steel suspension damping is characterized by the fact that the steel leaf springs are locked for most speed and pavement unevenness conditions. The vehicle then vibrates on the tires only and the corresponding damping is low. For high speeds and/or rough pavements the steel leafs unlock and high damping is provided through inter-leaf Coulomb friction. With air suspensions, damping is provided through shock absorbers. Here, the suspension damping capacity is firstly influenced by the more or less optimum arrangement of the shock absorbers between axle and chassis. Secondly, high damping is given only as long as the shock absorbers are in good condition.


19 / 169 8. June 2010

Depending on the bridge's fundamental natural frequency f and the related maximum span length L, the questions to be answered by the OECD DIVINE Bridge Research Project were: •

Is there any risk, that medium to long-span bridges (f ≈ 1.5...1.8 Hz, L ≈ 60...70 m), will be extremely susceptible to the body bounce vibrational actions of vehicles equipped with modern suspensions? If yes: What can the maximum dynamic increment be?

•

Is there any risk that short-span bridges (f ≈ 10...15 Hz, L ≈ 10 m), may react more strongly to the axle-hop vibrational actions of heavy commercial vehicles equipped with modern suspensions than to those of steel-suspended vehicles?

Due to several kinds of restrictions, the problem of multiple presence of several vehicles on a bridge could not be subject of this project. This means that direct application of the results reported here for the design of Dynamic Load Allowance Prescriptions of Bridge Loading Codes is not always possible. S.3

Test Parameters

S.3.1 Bridges, Pavements Considering the bridge population in the respective countries, the three bridges tested in Switzerland were of the medium to long-span type whereas short-span bridges were tested in Australia. However, comparison of the respective test series and their results was possible for one of the bridges tested in each country. The shortest Swiss and the longest Australian bridge exhibited a similar fundamental natural frequency, f ≈ 4.5...5 Hz (Table 1). Contrary to tests performed in Switzerland earlier it was not possible to influence the overall quality of the test track's longitudinal pavement profile. As a consequence, the Swiss test track quality could not be changed from the actual smooth to very smooth condition and the same applied for the Australian test tracks which ranged between smooth to quite rough. The road profiles for each of the test tracks were measured and classified based on their unevenness spectra in accordance to [ISO, 1995].

OECD DIVINE, Element 6, Bridge Research

20 / 169

Final Report, EMPA Switzerland, QUT Australia

8. June 2010

Bridge (Country)

Max. span [m]

Description

1st freq. [Hz]

Damping [%]

Pavement Condition

Sort (CH)

70.0

5-span continuous prestressed concrete single-cell box girder

1.62

1.0

A

Deibüel (CH)

41.0

3-span continuous prestressed concrete single-cell box girder

3.01

0.8

A

Föss (CH)

31.0

3-span continuous prestressed concrete two-cell box girder

4.44

1.6

B

Lawsons (AUS)

23.3

1-span prestressed concrete girder and reinforced concrete slab

5.1

1.0

A

Coxs (AUS)

11.0

4 simply supported steel girder and reinforced concrete slab spans

10.2

4.5

A

Camerons (AUS)

9.1

4 simply supported prestressed concrete plank spans

11.3

1.5

B…C

Cromarty (AUS)

9.0

3-span natural timber girders and timber deck planks

9.5

2.6

B…C

Table 1 Main parameters of the test bridges including the test track pavement quality. S.3.2 Vehicles To reach the goal defined above, the experiments included dynamic bridge load tests using vehicles equipped with steel and air suspensions respectively. The test vehicle used in Switzerland was provided by the Canadian National Research Council (NRC), Centre of Surface Transportation Technology, Ottawa. This NRC test vehicle was a five-axle tractor-semitrailer where the suspension could be changed from steel to air on four of the five axles (the steer axle was always steel-suspended). All wheels were instrumented to measure dynamic wheel loads.


21 / 169 8. June 2010

It was not possible to use the NRC vehicle for the Australian tests. The Australian test vehicles were provided by Boral Transport and Shell Australia. They were a six-axle over the rear axle tip truck and a six-axle petrol tanker (both tractor-semitrailers) respectively. With the Boral vehicle, the trailing axles were instrumented and the suspensions could be exchanged between a steel and an air suspension. To cope with the requirements as closely as possible two prime-movers (tractors) equipped with steel and air suspension respectively were used. The Shell Tanker was available in an air-suspended version only and without any axle instrumentation. This vehicle is mentioned in the report in one significant bridge loading test only (Camerons Creek Bridge). S.4

Test Procedures and Data Processing

The bridge response and the dynamic wheel loads (where possible) were measured and acquired simultaneously during the passage of the test vehicles at speeds from crawl up to v = 50...100 km/h depending on the geometry of the respective test track. Data processing included time as well as frequency domain analysis. Time domain analysis of the bridge response signals yielded the dynamic bridge response as a function of speed, so-called (φ,v)-curves. φ is the bridge response dynamic increment being derived from the maximum static bridge deflection Astat and the maximum dynamic bridge deflection Adyn as defined in the following formula (ref. Fig. 1, p. 27):

φ=

A dyn − A stat ⋅ 100 [%] A stat

Frequency domain analysis included determination of dynamic wheel load spectra and, in the case of the Swiss tests, of the vehicle operational mode shapes. Investigation of dynamic vehicle/bridge-interaction processes based on interpretation of crosspower-spectra calculated from simultaneously acquired bridge response and dynamic wheel load signals proved to be more difficult than anticipated. It was however possible to solve all technical problems associated with signal synchronisation and processing and some results can be presented in this report. S.5

Dynamic Bridge Response

S.5.1 The Results The maximum dynamic increment for bridge vertical deflection for each of the bridge/vehicle parameter configurations tested is given in Table 2 together with other significant parameters.


22 / 169


Bridge

1st freq. [Hz]

8. June 2010

Damping Pavement [%] Condition

Frequency Matching

φmax steel

φmax air

[%]

[%]

Sort

1.62

1.0

A

body-bounce, air

10

26

Deibüel

3.01

0.8

A

body-bounce, steel

21

5

Föss

4.44

1.6

B

none

15

12

Lawsons

5.1

1.0

A

none

6

3

Coxs

10.2

4.5

A

axle-hop

28

18

Camerons

11.3

1.5

B…C

axle-hop (Boral)

105

75

axle-hop (Shell)

---

137

axle-hop

109

50

Cromarty

9.5

2.6

B…C

Table 2 Maximum dynamic increments φmax for vertical bridge deflection measured in a significant bridge location for the passages without axle-hop plank.

S.5.2 Discussion In summary, the response of the bridges to the passage of the test vehicles was consistent with the laws of physics in that the largest dynamic responses occurred when either the body bounce or axle-hop frequency of the test vehicle corresponded to the natural frequencies, especially to the fundamental natural frequency in bending, of the bridge. This effect is called "frequency-matching" and it results in "quasi-resonance" phenomena. Also, the surface profile of the bridge and its approaches (up to 50 m) was fundamental to the response of the vehicle suspension and in turn the dynamic response of the bridge. Furthermore, for short-span bridges, the damping capacity of both the vehicle and bridge were found to be important parameters influencing the dynamic bridge response. Frequency-matching yielded maximum dynamic increments φ = 21...26% for both cases, Sort/air and Deibüel/steel. The dynamic increments for the "wrong" bridge/vehicle combination were significantly smaller (φ = 5...10%). As a consequence of the smooth pavement surfaces the largest dynamic responses are still significantly smaller than those e.g. observed in the case of the Deibüel bridge under rough pavement surface conditions [Cantieni, 1983]).


23 / 169 8. June 2010

Simulation studies are required to investigate the effect of a rougher surface for the Sort/air case. The results of the tests on the Föss and Lawsons Creek bridges were as expected: Little dynamic bridge response in the case where no frequency-matching is possible. Their natural frequencies being almost identical, frequency-matching in the axle-hop range could be expected for both the Coxs River and Camerons Creek bridges. The dynamic response of the Coxs River Bridge was however less than expected. The smooth pavement surface conditions prevented high dynamic wheel loads. In addition, the very high levels of damping evident in the bridge due to friction between the non-composite steel girders and concrete slab reduced the bridge response amplification and hence the effects of the dynamic components of the wheel loads. The maximum dynamic increments for the air and steel-suspended vehicles were of the same order of magnitude as for the Sort and Deibüel Bridges. In contrast to Coxs River Bridge, the rough pavement surface and the moderate level of damping evident in the Camerons Creek Bridge resulted in considerable amplification of the axle-hop effects of the dynamic wheel loads. One of the results was a very large dynamic response (φ ≈ 137%, approximately 4.5 times the value allowed for in e.g. the Australian bridge design code) for the passage of the Shell air-suspended vehicle. Short wavelength roughness and the correct speed (v ≈ 60 km/h) were required to initiate this response. The maximum dynamic response due to the passage of the Boral air-suspended vehicle was φ ≈ 75%. This difference is believed to be a consequence of the Boral air suspension fitted to the trailer having the better damping performance than the one of the Shell tanker. The maximum dynamic response due to the passage of the Boral steel-suspended vehicle was φ ≈ 105%. Cromarty Creek Bridge is a natural timber girder bridge which is relatively soft and therefore deflects considerably during the passage of vehicles. This deflection adds significantly to the critical pavement profile wavelengths for body bounce. The maximum dynamic increment induced by the steel-suspended Boral vehicle, φ ≈ 109%, is of the same order of magnitude as for Camerons Creek Bridge and well above "design" values for "normal" pavements. The airsuspended Boral vehicle induced approximately half the dynamic bridge response of its steelsuspended counterpart. S.6

Conclusions

As expected, the DIVINE experiment answered the questions formulated at the beginning of the project in part only. This is due to the fact that the cases of short span bridges with a smooth bridge and approach profile and of medium to long span bridges with a rough bridge


24 / 169 8. June 2010

and approach profile could not be covered with the DIVINE tests. It was also not possible to investigate the problem of multiple vehicle presence on the bridges. It was however found that: • For frequency-matching conditions at f ≈ 1.5...1.8 Hz: No excessive bridge vibrations are to be expected with air-suspended vehicles involved for smooth pavement conditions. • For frequency-matching conditions at f ≈ 10 Hz: Excessive bridge vibrations are to be expected with steel as well as air-suspended vehicles involved for average to rough pavement conditions. The situation is especially severe if the shock absorbers of airsuspended vehicles are ineffective. Questions remaining open are: • For frequency-matching conditions at f ≈ 1.5...1.8 Hz: What could the maximum dynamic increment be in the case of average to rough pavement conditions? • For frequency-matching conditions at f ≈ 10 Hz: What pavement quality and vehicle suspension damping performance is necessary to avoid excessive bridge vibrations? To answer these questions, further tests or analytical studies are necessary.


1.

INTRODUCTION

1.1

General

25 / 169 8. June 2010

The OECD DIVINE Project was started in November 1992 and finished in November 1997. OECD is the Paris based "Organization for Economic Co-operation and Development", DIVINE stands for "Dynamic Interaction Between Vehicle and Infrastructure Experiment". DIVINE was divided into six research elements. Three of them were related to the experimental investigation of the behavior of flexible pavements, one to road simulators, one to the numerical simulation of a heavy vehicle driving down a road on a flexible pavement and one, Research Element 6, to the dynamic interaction between vehicles and bridges. The general layout of DIVINE as well as the condensed results and the technical and policy implications of the research elements are given in the Final Technical Report [OECD, 1997] (see Annex I for the list of References). Here, the Final Report for Research Element 6 is presented. 1.2

The Problem to be Solved

The goal of the DIVINE Project Element 6, "Bridge Research", was to experimentally establish the consequences, in terms of the dynamic response of bridges, of changing heavy commercial vehicle suspensions from the well known steel to modern air suspensions. To reach this goal, dynamic load tests using a single test vehicle have been performed in 1994/95 on seven highway bridges, three of them located in Switzerland and four in Australia. Due to several kinds of restrictions, the problem of multiple presence of several vehicles on a bridge could not be incorporated in this project. Similarly, pavement roughness could not be chosen as a parameter to be varied in these tests. 1.3

Previous Knowledge

1.3.1 Experiment vs. Numerical Analysis All information presented below stems from experimental investigations. Experience shows that a reliable analytical modeling of the process "a heavy vehicle crosses a bridge with a given longitudinal pavement profile" is not an easy task. The main problem to be solved is proper modeling of the vehicle suspension system. Software packages considering the respective details are on the market for some time now. They are however very sensitive to the parameters used in the vehicle models and should therefore be validated against


26 / 169 8. June 2010

experiments. This was done in exceptional cases only, also due to a lack in reliable experimental data. Consequently, the results of analytical studies were not suited to form the basis of the present project. Simple mechanical models as e.g. also discussed later in this paper are well suited to investigate qualitative trends but they can definitely not be used to determine the bridge response signal with reasonable accuracy. In the following paragraphs the main parameters influencing the dynamic interaction between vehicles and bridges are discussed. Some readers may miss the so-called speed parameter α relating the time the vehicles needs to cross a bridge and the period of the bridge fundamental natural frequency. This and other non-dimensional parameters can be found in most of the analytical investigations performed by various authors. Experience however shows that these parameters or of negligible importance versus that of the parameters treated below [Cantieni, 1983, 1992]. The speed parameter, as an example, becomes important for typical highway bridges for very high vehicle speeds only. 1.3.2 Frequency Matching Besides static deformations, highway bridges also exhibit dynamic responses to the passage of heavy, commercial vehicles. This fact was established about 150 years ago when investigations into this matter were first performed in the context of severe accidents with railway bridges [Willis, 1851]. Research into the experimental investigation of the dynamic response of highway bridges has significantly intensified in the last 30 years with the development of modern measurement, data acquisition and data processing methods. Related developments between 1847 and 1990 are described in [Cantieni, 1983, 1992]. Research Element 6, "Bridge Research", of the OECD DIVINE project is based on these reports. The dynamic increment φ compares the maximum dynamic bridge deflection Adyn and the maximum static deflection Astat. The definition of the dynamic increment φ as shown in Fig. 1 is used throughout this report.


27 / 169


Fig. 1

8. June 2010

Bridge deflection time history with the definitions of the dynamic increment φ [%], the fundamental frequency f [Hz] and associated damping decrement δ (or, damping coefficient ζ in percent of critical: ζ = δ/2π), (from [Cantieni, 1983]).

Dynamic Increment [%]

The main result given in [Cantieni, 1983] is that bridges with a fundamental natural frequency f ≈ 2.5...4 Hz are more susceptible to the dynamic actions of single two-axle heavy commercial vehicles equipped with steel leaf suspensions than bridges with a fundamental frequency lying outside this range. This is due to the fact that the wheel loads of heavy vehicles show a predominant frequency content in this same range [Cantieni, 1992]. This effect is known as "frequency-matching" leading to "quasi-resonance" effects with the corresponding strong dynamic bridge response (Fig. 2).

100 80 60 40 20

14 Hz

0 0

1

2

3

4

5

6

7

8

9

10

Fundamental Frequency f [Hz] Fig. 2

Dynamic increments as a function of the bridge fundamental frequency; single steelsuspended two-axle trucks (from [Cantieni, 1983]).

Processing of a large number of measured dynamic wheel loads also revealed that the fact of the wheel load frequencies predominantly lying close to f ≈ 3 Hz is a consequence of the


28 / 169 8. June 2010

specific behavior of steel leaf suspensions: Under normal conditions concerning pavement unevenness, the leaves are locked and the vehicle vibrates on its tires only. They unlock only when the pavement is very rough and/or the vehicle speed is high [Cantieni, 1992]. Under such circumstances the predominant wheel load frequency of a steel leaf sprung vehicle can decrease to f ≈ 1.5 Hz. 1.3.3 Pavement Unevenness The importance of pavement unevenness on the dynamic bridge response was known before the OECD DIVINE project commenced. It is well known that increases in pavement and/or vehicle speed result in higher dynamic wheel load amplitudes. For example, the scatter of the dynamic increment in the frequency-matching region, f ≈ 3 Hz, (see Fig. 2) has been shown to be a consequence of variations in the pavement unevenness [Cantieni, 1983]. Further evidence was provided through the results of tests performed on medium and short-span bridges where the test vehicle was driven over a 300 mm wide and 50 mm thick plank lying on the pavement [Cantieni, 1983]. Here, the dynamic bridge response was found to be roughly three times larger than for passages on the normal pavement (Fig. 3). Compared with the results for "normal" pavement conditions (Fig. 2), two ranges of bridge natural frequency where maximum response occurred can be distiguished in Fig. 3: f ≈ 2.5 Hz for low speeds, v = 10...15 km/h, and f ≈ 8...10 Hz for higher speeds, v = 40...80 km/h. These two ranges reflect the fact that, at low speeds, the plank excites vehicle body bounce vibrations (now at f ≈ 2.5 Hz) whereas at high speeds, axle-hop vibrations (f ≈ 8...15 Hz) are excited significantly. For very high speeds, neither body bounce nor axle-hop vibrations are excited by a plank of the type mentioned. At high speeds, the only vehicle element influenced by the plank will be the tire.


29 / 169


8. June 2010


300

200

100

0 0

1

2

3

4

5

6

7

8

9

10

Fundamental Frequency f [Hz] Fig. 3

Dynamic increments as a function of the bridge fundamental frequency for passages with a single, steel-suspended two-axle vehicle crossing a 300 x 50 mm plank (from [Cantieni, 1983]).

1.3.4 Relationship "Natural Frequency" vs. "Maximum Span" of a Bridge Concerning the relationship between the fundamental natural frequency f of a bridge and its maximum span L, the results derived from a large number of dynamic bridge tests are given in Fig. 4. Additional data concerned with short-span bridges collected in Australia is presented in Fig. 5. These figures show, that • bridges with f ≈ 1.5 Hz have a maximum span of roughly L ≈ 70 m ("long span"), • those with f ≈ 3 Hz of L ≈ 40 m ("medium span"), and • those with f ≈ 8...10 Hz of L ≈ 10 m ("short-span").

"Very long span" bridges with L > 100 m and f < 1.0 Hz are not the subject of this report. Such bridges may show critical dynamic response to e.g. wind forces but no dynamic interaction of the global structure with heavy commercial vehicles is to be expected. Heavy vehicles may exhibit roll motions at such low frequencies but the corresponding dynamic wheel load components are too small to be of concern. Furthermore, the critical loads on such bridges are due to queues of stationary vehicles.


30 / 169


Fig. 4

8. June 2010

Relationship between maximum span L and fundamental frequency f of beam-type highway bridges (from [Cantieni, 1983]).

Frequency f [Hz]

20

RC Girders (cont) P/S Deck Unit RC & PC Girders Steel Girders Timber Girders f=80/L f=100/L f=120/L

15

10

5

0 5

Fig. 5

10

15 20 Maximum Span L [m]

25

Relationship between maximum span L and fundamental frequency f for short-span Australian bridges (RC: reinforced concrete; PC: prestressed concrete; cont: continuous).


1.4

31 / 169 8. June 2010

Strategy of the OECD DIVINE Bridge Research Project

As previously mentioned, the goal of the DIVINE Project Element 6, "Bridge Research", was to experimentally establish the consequences, in terms of the dynamic response of bridges to the passage of a single heavy vehicle, of changing the vehicle suspensions from the well known steel to modern air suspensions. The change from steel to air suspension means: • A decrease of the vehicle body bounce frequency from f ≈ 3 Hz to f ≈ 1.5...1.8 Hz, • That the wheel load frequency no longer depends upon speed and pavement unevenness, • That the mechanisms affecting damping change from Coulomb friction to viscous damping (shock absorbers).

Concerning the frequency range of axle-hop vibrations, f ≈ 8...20 Hz, no significant difference between standard steel and modern air suspended vehicles is to be expected. The frequency of axle-hop vibrations is a function of the vehicle unsprung mass, the body suspension spring and damper and the tire spring. Under axle-hop conditions, the steel suspension is unlocked and its stiffness and hence natural frequency is similar to that of an air suspension. Differences in behavior of the two suspension systems can become apparent because of the different damping mechanisms. Considering medium to long span bridges, the question to be answered by the OECD DIVINE (Dynamic Interaction Between Vehicles and Infrastructure Experiment) Bridge Research Project was: •

Is there any risk, that bridges with a fundamental natural frequency, f ≈ 1.5...1.8 Hz, i.e. with a maximum span length, L ≈ 60...70 m, will be extremely susceptible to the body bounce vibrational actions of vehicles equipped with modern suspensions? If yes: What can the maximum dynamic increment be?

The reason for asking this question can be taken from Fig. 2: The limiting value for the dynamic increment for a bridge with a natural frequency, f = 1.5 Hz, is φ = 20% only. Considering short-span bridges, the problem that required investigation was: •

How would short-span bridges with a fundamental natural frequency, f ≈10...15 Hz, i.e. with a maximum span length, L ≈ 10 m, react to the axle-hop vibrational actions of heavy air-suspended vehicles compared to those of steel-suspended vehicles?


32 / 169 8. June 2010

The reason for asking this question was that no sufficient information on the dynamic behavior of short-span bridges was available at all. It could however be expected that frequencymatching effects at axle-hop frequencies would be a critical issue for both suspension systems. This was of considerable interest in view of the large number of short-span bridges in service and the relative importance of live loads for such bridges. Switzerland was primarily interested in medium to long span bridges whereas Australia's interest was concentrated on short-span bridges. On average, Swiss bridges have mediumlength spans (Fig. 6) and may be taken as representative for European bridges whereas Australia’s bridge stock consists mainly of short-span bridges (Fig. 7). As a consequence, the OECD DIVINE Bridge Research Project was performed on medium to long-span bridges in Switzerland and on short-span bridges in Australia. Hence, the complete spectrum of highway bridges could be covered (with the exception, as mentioned before, of bridges with very long spans L > 100 m). Testing procedures were kept as similar as possible. As a result of the geographical boundary conditions, it was not possible to make use of the same test vehicle in both cases.

Fig. 6

Histogram of maximum span of roughly 200 highway bridges in Switzerland (from [Cantieni, 1983]).


33 / 169


8. June 2010

50%

40%

30%

20%

10%

0% 0 to 5

5 to 10

10 to 15

15 to 20

20 to 25

25 to 30

30 to 35

35 to 40

> 40

Maximum Span L [m]

Fig. 7

1.5

Histogram of maximum span of highway bridges in Australia (derived from [NAASRA, 1985]). Swiss Tests on Medium to Long-Span Bridges

Two Swiss bridges with known natural frequencies were selected in order to produce frequency-matching between their fundamental natural frequency and the wheel load frequencies associated with the steel and air-suspended vehicle body bounce vibrations respectively. These bridges included the Deibüel Bridge with a fundamental frequency f ≈ 3 Hz and the Sort Bridge with f ≈ 1.6 Hz. For comparison, the Föss Bridge with f ≈ 4.4 Hz where no frequencymatching effects were to be expected was also selected. For the tests in Switzerland, the National Research Center of Canada (NRC) test vehicle (see Paragraph 2.3.1 and Annex H) was used. This vehicle was also used in tests performed in Europe in 1994 in the context of other OECD DIVINE Research Elements. The NRC test vehicle was a five-axle tractor-semitrailer where the suspensions could be changed from steel to air (with the exception of the steer axle where the steel suspension could not be exchanged). This vehicle was instrumented to continuously measure the dynamic wheel loads on all wheels and to carefully maintain constant speed during a test passage. All of the bridges in Switzerland were tested without using an artificial bump on the test track pavement. One (down-hill) driving direction with the NRC test vehicle passing on the bridge centerline was always chosen for the tests. The relevant test parameters including the test track evenness rating according to [ISO, 1995] are given in Table 3 below.


1.6

34 / 169 8. June 2010

Australian Tests on Short-Span Bridges

The bridges chosen for the Australian tests had a fundamental natural frequency of f ≈ 10 Hz in a first phase (Coxs River, Camerons Creek, Cromarty Creek). Here, frequency-matching effects with the test vehicle's axle-hop vibrations could be expected. Later, the Lawsons Creek Bridge with a natural frequency of f ≈ 5 Hz was also chosen to form the link with the Swiss bridges, especially with the Föss Bridge with f ≈ 4.4 Hz. For the Australian tests, a Boral six-axle tractor-semitrailer (see Paragraph 2.3.2 and Annex H) was used. The trailing axles of this vehicle were instrumented and their suspensions could also be changed from steel to air. Two different prime-movers with steel and air suspensions were used in the respective tests. However, it was not possible to instrument the tractor axles. The Boral vehicle is abbreviated as BS and BA for the steel and air-suspended versions respectively. In addition to the usual tests making use of the Boral vehicle, tests were performed with using an air-suspended "Shell Tanker", shortly described as vehicle SA. Tests using the SA vehicle were conducted at the Camerons Creek Bridge only. Owing to their extraordinary significance, the results of these tests have been included in the report. The four bridges in Australia were tested with and without a 25 mm plank, called the axle hop plank (AHP), lying on the pavement over a pier. Contrary to the Swiss tests, both driving directions were used and the test vehicle used the normal (left-hand side) driving lanes except for Coxs River Bridge where the vehicle passed on the centerline only. At Lawsons Creek Bridge the test vehicle used the middle of three driving lanes which corresponded with the bridge centerline. The relevant test parameters including the test track evenness rating according to [ISO, 1995] are given in Table 3.


35 / 169


1.7

8. June 2010

Summary - Test Bridges and Vehicles

Bridge (Country)

max. 1st damping pavem. bumps vehicle, suspen- meas. driving ζ [ISO] no of sions axles direcspan freq. [%] axles tions [m] [Hz]

Sort (CH)

70.0

1.62

1.0

A

none

NRC, 5

steel/air

5

1

Deibüel (CH)

41.0

3.01

0.8

A

none

NRC, 5

steel/air

5

1

Föss (CH)

31.0

4.44

1.6

B

none

NRC, 5

steel/air

5

1

Lawsons (AUS)

23.3

5.1

1.0

A

axlehop

Boral, 6

steel/air

3

2

Coxs (AUS)

11.0

10.2

4.5

A

axlehop

Boral, 6

steel/air

3

2

Camerons (AUS)

9.1

11.3

1.5

B…C

axlehop

Boral, 6 Shell, 6

steel/air air

3 0

2 2

Cromarty (AUS)

9.0

9.5

2.6

B…C

axlehop

Boral, 6

steel/air

3

2

Table 3 Main parameters of the bridges tested in Switzerland and Australia. The main parameters of the bridges, pavements and vehicles used on the testing program conducted in Switzerland and Australia are summarized in Table 3. Details concerning natural frequencies and damping of the bridges as well as the pavement roughness are given in Paragraph 2.2 and the Annexes A to G. Details of the test vehicles are given in Paragraph 2.3 and in Annex H. In Chapter 6, the measured bridge and vehicle responses are described. With this background the implications of the trend towards softer air-suspensions is discussed in Chapter 7 and conclusions are drawn in Chapter 8.


36 / 169


2.

8. June 2010

TEST PARAMETERS

In this Chapter, the test bridges, the test tracks pavement longitudinal profile data and the test vehicles are shortly presented. The bridge and vehicle instrumentation is described in Chapter 3. Details concerning the bridge geometry, structural model, plan view, elevations, cross section, modal parameters, pavement profiles and instrumentation are presented in detail graphically for all bridges in the Annexes A to G. Similarly, all details concerning the test vehicles are given in Annex H. 2.1

Bridges

Table 4 summarizes the test bridges main geometrical and dynamic parameters. In the following paragraphs a short description of each bridge is given.

Bridge (Country) Max span [m]

Description

1st freq. Damping [Hz] [%]

Sort (CH)

70.0

5 span continuous prestressed concrete single-cell box girder

1.62

1.0

Deibüel (CH)

41.0

3 span continuous prestressed concrete single-cell box girder

3.01

0.8

Föss (CH)

31.0

3 span continuous prestressed concrete twin-cell box girder

4.44

1.6

Lawsons (AUS)

23.3

1 span prestressed concrete girder and reinforced concrete slab

5.1

1.0

Coxs (AUS)

11.0

4 simply supported steel girder and reinforced concrete slab spans

10.2

4.5

Camerons (AUS)

9.1

4 simply supported prestressed concrete plank spans

11.3

1.5

Cromarty (AUS)

9.0

3 span natural timber girders and timber deck planks

9.5

2.6


37 / 169


8. June 2010

Table 4 Summary of the test bridges main parameters.

60 50 40 30 20

Camerons (AUS)

70

Cromarty (AUS) Coxs (AUS)

80

Sort (CH)

|Amplification Factor|

90

Foss (CH) Lawsons (AUS)

100

Diebuel (CH)

The frequency and damping of the bridges are compared graphically in Fig. 8 where the amplification factor is presented for a single degree of freedom dynamic system with the same frequency and damping as the bridge. As can be seen, the bridges selected represented the range of bridge frequencies where frequency-matching with air suspensions (f ≈ 1.5...1.8 Hz), steel suspensions (f ≈ 3 Hz), and axle-hop (f ≈ 8...20 Hz) was expected. In addition, two bridges were selected from outside this range.

10 0 0

Fig. 8

1

2

3

4

5

6

7 8 9 10 11 12 13 14 15 16 17 18 19 20 Forcing frequency [Hz]

Amplification of the Single-Degree-of-Freedom bridge model to linear scales (Note: Bridges are arranged in order of frequency).


38 / 169 8. June 2010

2.1.1 Sort Bridge (CH)

Fig. 9

Photograph of the Sort Bridge. To the left the Ticino River, to the right the N2 motorway, both surpassed by the bridge (EMPA Photo No. 101'104/9).

The Sort Bridge carries the Cantonal Highway over the National Highway N2 Basel - Chiasso and the Ticino River near Airolo, a village south of the Gotthard Pass. With a total length L = 258.8 m, the five spans measure roughly l = 36, 58, 70, 58 and 36 m. The depth of the one-cell box girder varies between 2 m and 2.8 m. The 11 m wide bridge deck provides for two traffic lanes and one walkway. The radius of curvature of the superstructure is R = 900 m. The superstructure, a prestressed concrete continuous beam, is connected to the substructure with roller bearings at the abutments and piers. The horizontal fixpoint is located at one of the piers. The four piers have a circular section with a 1.6 m diameter. The dynamic natural properties of the bridge were determined by an Ambient Vibration Test performed in September 1994 [Felber, 1995a]. The frequency and damping of the first vertical bending mode were determined to f = 1.62 Hz and ζ = 0.95% which corresponds to the values found in 1977 [Egger, 1977].


39 / 169 8. June 2010

2.1.2 Deibüel Bridge (CH)

Fig. 10

Photograph of the Deibüel Bridge. To the left the bridge tested, to the right the "twin" bridge carrying two-way traffic during the tests (EMPA Photo No. 101'069/36).

The Deibüel Bridge, located on the National Highway N4a between Zürich and Lucerne, is a straight structure with an overall length L = 110.3 m continuous over three spans l ≈ 37, 41 and 32 m. Its prestressed concrete one-cell box-girder is 1.80 m deep and 11.75 m wide and provides for two traffic lanes plus an emergency lane in the Zürich-Lucerne direction. The traffic in the Lucerne-Zürich direction uses a parallel bridge located some meters distant from the structure investigated. The cross section of the two rectangular piers is 1 m by 3 m. One of the piers is clamped into the box-girder. The connection between the other pier and the girder is designed as a concrete hinge whereas the abutments are horizontally free (PTFE pot bearings). This makes the bridge behave as a frame. The dynamic natural properties of the bridge were determined by an Ambient Vibration Test performed in March 1995 [Felber, 1995b]. The frequency and damping of the first vertical bending mode were determined to f = 3.01 Hz and ζ = 0.8% which corresponds to the values found in 1975 [Egger, 1975] and 1978 [Cantieni, 1992].


40 / 169 8. June 2010

2.1.3 Föss Bridge (CH)

Fig. 11

Photograph of the Föss Bridge (exact designation: Föss 2). In the background the arch bridge Föss 3 on the Gotthard Pass Highway (EMPA Photo No. 101'049/17).

The Föss Bridge is part of the Gotthard Pass Highway and crosses the Föss Creek on the southern side of the mountains at a height of 1’500 m above sea level. It has a total length L = 79 m and three spans l = 24, 31 and 24 m. The cross-section of this prestressed concrete structure is a two-cell box-girder of 1.59 m depth and 17.0 m width and allows for two through traffic lanes plus a merging lane for exit to Motto Bartola from both directions. The interior supports consist of two rectangular columns with a 1 m by 2 m cross section each. The superstructure is curved with a constant radius R = 200 m. Considering the degrees of freedom allowed by the steel bearings located at the pier heads and abutments, the static system of the superstructure can be understood as a girder, continuous over three spans with the fix point at one of the abutments. The dynamic natural properties of the bridge were determined by an Ambient Vibration Test performed in April 1995 [Felber, 1995c]. The frequency and damping of the first vertical bending mode were determined to f = 4.44 Hz and ζ = 1.64% which corresponds to the values found in 1976 [Cantieni, 1976].


41 / 169 8. June 2010

2.1.4 Lawsons Creek Bridge (AUS)

Fig. 12

Photograph of the Lawsons Creek Bridge.

The bridge over Lawsons Creek is a single span concrete structure situated 26 km west of Lithgow on the Great Western Highway, New South Wales, Australia. The bridge has a composite reinforced and prestressed concrete superstructure supported on reinforced concrete abutments. The superstructure consists of 9 precast pre-tensioned concrete "I"-girders acting compositely with a cast-in-situ reinforced concrete deck. The "I"-girders are 1’150 mm deep, 450 mm wide and they are spaced at 1’350 mm centers. The overall length of the single span bridge is L = 24.0 m with 23.3 m between supports. The girders are supported at both abutments on elastomeric bearings. The width between curbs is 11.7 m allowing for three traffic lanes. The third lane is an overtaking lane necessitated by the 3.4% grade. The bridge is straight with no skew. At abutment A the bridge is provided with an expansion joint, abutment B is fixed. The dynamic properties of the bridge were determined from the free vibrations of the bridge after the test vehicle had left the bridge. The frequency and damping of the first vertical bending mode were determined to be f = 5.1 Hz and ζ = 1.0%.


42 / 169 8. June 2010

2.1.5 Coxs River Bridge (AUS)

Fig. 13

Photograph of the Coxs River Bridge.

The four-span, two-lane bridge over the Coxs River is located in the township of Wallerawang, west of Lithgow, New South Wales, Australia. This bridge was constructed in 1945 using 670 mm steel "I"-beams (24″ x 7.5″ x 95 lb/ft) supporting a 180 mm thick reinforced concrete deck. There is no shear connection between the deck slab and the steel girders. Relative movement between the girders and the deck slab was observed. The bridge is supported on reinforced concrete piers and abutments and has a total length L = 46.1 m with four simply supported spans of l = 11.65 m, 11.45 m, 11.45 m and 11.55 m. All spans are fixed at one end and have an expansion joint at the other end. The expansion bearings were constructed from steel and brass strips. Movement in these bearings was not evident. The width between curbs is 6.5 m with a 1.2 m footpath which has been added to the western side of the bridge in recent years. This footpath has limited influence on the stiffness or the mass of the bridge. Since large loads are often carried over the bridge to the nearby power plant, the bridge has been strengthened by props positioned at mid-span. During the tests, the props were lowered to allow an investigation of the original structure. Using typical displacement time signals, the dynamic properties of the bridge were determined from the free vibration of the bridge after the passage of test vehicles. The frequency and damping of the first vertical bending mode were determined to be f = 10.2 Hz and ζ = 4.5%.


43 / 169 8. June 2010

This high level of damping was attributed to the friction between the deck slab and the steel girders. 2.1.6 Camerons Creek Bridge (AUS)

Fig. 14

Photograph of the Camerons Creek Bridge.

The bridge over Camerons Creek is located on Bucketts Way, a secondary highway north of Newcastle, New South Wales, Australia. The bridge consists of a composite prestressed and reinforced concrete superstructure supported on reinforced concrete piers and abutments. The bridge has an overall length of L = 36.6 m and the width between curbs is 8.53 m. The twolane bridge consists of four simply-supported spans of l = 9.14 m between pier centerlines. The bearings consist of strips of elastomeric rubber. The superstructure is fabricated from 600 mm wide prestressed precast concrete planks tied together with a reinforced concrete topping slab. The bottom of the planks are bound together by grouted transverse reinforcement. The reinforced concrete topping slab acts compositely with the prestressed concrete planks and provides lateral load distribution. The reinforced concrete substructure is supported on spread footings. This bridge is typical of hundreds of bridges constructed in NSW and is similar to prestressed concrete plank bridges throughout Australia. Using typical displacement time signals, the dynamic properties of the bridge were determined from the free vibration of the bridge after the passage of test vehicles. The frequency and damping of the first vertical bending mode were determined to be f = 11.3 Hz and ζ = 1.5%.


44 / 169 8. June 2010

2.1.7 Cromarty Creek Bridge (AUS)

Fig. 15

Photograph of the Cromarty Creek Bridge

This timber girder bridge over Cromarty Creek is located 2 km north of the Camerons Creek Bridge on Bucketts Way, a secondary highway north of Newcastle, New South Wales, Australia. The bridge is a three span timber girder structure L = 24.4 m long and 6.70 m wide. The ends of the natural round timber girders are made semi-continuous by vertical bolts into corbels. The three spans L = 7.62 m, 9.14 m and 7.62 m are supported on timber piers and abutments. Drawings of the bridge show three 9.14 m spans. As the bridge has undergone major rehabilitation, it was concluded that the end spans had been shortened to 7.62 m during this process. The bridge is on a skew of 25 degrees. The deck of the bridge consists of transverse timber planking and longitudinal running boards. Approximately ten thousand of these bridges remain in service and present a major management challenge for Australian authorities. Using typical displacement time signals, the dynamic natural properties of the bridge were determined from the free vibration of the bridge after the passage of test vehicles. The frequency and damping of the first vertical bending mode were determined to be f ≈ 9.5 Hz and ζ = 2.6%.


2.2

45 / 169 8. June 2010

Pavement Profiles

2.2.1 Longitudinal Profiles of the Swiss Bridges Measurement of the test track profiles was performed by the Swedish Company Laser RST (Road Survey Technology, Solna and Linköping, Sweden). Their mobile system consists of two laser systems and two accelerometers, one for the right wheel track and the other for the left wheel track. These devices are fixed to a beam which in turn was fixed to the trailer coupling of an otherwise standard EMPA car. (For a photograph see Annex B.) The laser systems measure the displacement between the car body and the pavement using the contactless reflective methodology. These systems consist of a laser emitter and a camera which observes the location of the laser beam spot on the pavement. The relative displacement between chassis and pavement surface is hence established. Accelerometers measure the car’s chassis acceleration. These acceleration signals are integrated twice by an on-board computer and provide a measure for the absolute chassis displacement. The on-board computer calculates the pavement longitudinal profile from the difference between the measured absolute and relative displacements. The measurement car traveled at speeds of v = 35 km/h and v = 70 km/h. The resolution of the Laser RST device in the spatial domain is 0.05 m. The profile of the pavement was measured in the scheduled two wheel paths of the test vehicle. The profiles of the wheel tracks are shown in the Annexes A, B and C (refer Figures A- 8 to A- 10, B- 10 to B- 12 and C- 13 to C- 15). 2.2.2 Longitudinal Profiles of the Australian Bridges Profiles of the Australian bridges were measured by the Roads and Traffic Authority of New South Wales using a laser based profilometer developed by ARRB Transport Research. This device has the capability to produce outputs of the profile and grade against distance and of the international roughness index (IRI). The outputs can be stored as ASCII files. Profile measurement accuracy is ±1.0 mm per 50 meters of smooth pavement. The system uses a methodology similar to that described for the Swiss tests. Spatial resolution of the profile data was 0.06 m. Sample profiles of the wheel tracks are shown in the Annexes D, E, F and G (refer Figures D- 3, E- 3, F- 3 and G- 3). 2.2.3 Spectral Density of the Bridges Pavement Profile For the Swiss pavements, from the pavement profiles measured by RST Sweden, the spatial domain pavement unevenness spectra were calculated by the Swedish Road and Traffic


46 / 169 8. June 2010

Research Institute, VTI, Linköping, Sweden. They are also given in the Annexes (refer Figures A- 11, A- 12, B- 13, B- 14, C- 15 and C- 16). Macros developed in the Matlab environment using procedures described in the ISO 8608 [ISO, 1991] were used to calculate power spectral densities for the Australian pavements (refer Figures D- 4, E- 4, F- 4 and G- 4).


47 / 169 8. June 2010

2.2.4 Unevenness Rating According to ISO/TC 108 From the spatial domain spectra the pavement unevenness could be rated according to ISO/TC 108 [ISO, 1995].

Fig. 16

Classification of roads according to [ISO, 1995].


48 / 169


8. June 2010

The classification of roads as shown in Fig. 16 needs some explanation. First of all, this figure, logarithmically scaled on both axes, assumes the longitudinal pavement profile power spectral density (PSD), Gd, to being a linear function of two possible physical x-axis values, Spatial Frequency, n [cycles/m], and Angular Spatial Frequency, Ω [rad/m]. This is based on experience with a large number of measured profiles. The same applies to the fact that the PSD's slope, the waviness w, is on the average constant, w = 2, independent on the overall unevenness rating which is defined by the position of the PSD-curve in the plotted figure. To classify the pavement profile, it is hence sufficient to determine the intersection of the PSD-line with a reference value on the x-axis, be it n0 or Ω0. These reference values are: • n0 = 0.1 [cycles/m] • Ω0 = 1.0 [rad/m] = 1.0 [1/m]

Furthermore, the relationship between the PSD's based on n and Ω respectively is: • Gd(n) = 16 • Gd(Ω) [m3]

The third x-axis value offering an easy physical interpretation of the PSD's, the wavelength, λ [m], equivalent to 1/n, is not feasible to calculate the PSD Gd. Road classification based on Gd(n) according to Table C.2 given in [ISO, 1995] is as follows:

Road Class

Degree of Roughness Gd(n0) [10-6 m3]

A

--

16

32

B

32

64

128

C

128

256

512

D

512

1'024

2'048

E

2'048

4'096

8'192

F

8'192

16'384

32'768

Table 5 Classification of road profiles according to Table C.2 given in [ISO, 1995].


49 / 169


8. June 2010

However, problems arose with meaningful rating of the pavement unevenness. Determination of the unevenness power spectral density and subsequent classification according to [ISO, 1995] is the standard method when heavy commercial vehicles are involved. However, this method is reliable for tracks of at least 1'000 m length only if wavelengths of up to λ = 100 m have to be considered. In none of the present cases this requirement was fulfilled. Whereas the method may be applied without too much concern for the longer Swiss bridges, this is most probably not the case for the short Swiss and all Australian bridges. To interpret the results from tests performed on a short span bridge, the actual road profile has to be taken into account. For the test bridge's pavements the results of the road classification according to [ISO, 1995] surfaces is given in Table 6. These classifications are presented despite the fact that the length of the road profiles are often less than the 1'000 m required to reliably discern wavelengths up to λ = 100 m.

Bridge Sort

Deibüel

Föss

Lawsons

Coxs River

Camerons

Cromarty

Wheel Path

Gd(n0) [10-6 m3]

Waviness w

Classification

right

8.5

2.0

A

left

7.0

2.1

A

right

10

2.2

A

left

17

2.1

A

right

40

2.6

B

left

40

2.6

B

outer W/B

6.6

2.1

A

inner W/B

9.0

2.3

A

inner N/B

7.1

2.0

A

outer N/B

7.2

2.2

A

outer N/B

160

2.5

C

inner N/B

120

2.4

B...C

outer N/B

120

2.5

B...C

inner N/B

120

2.5

B...C

Table 6 Pavement unevenness rating according to [ISO, 1995] (N/B = north bound, W/B = west bound).


50 / 169


8. June 2010

2.2.5 Axle-hop plank For the tests conducted in Australia, vehicle axle-hop vibrations were specifically excited by driving the test vehicles across a 300 mm x 25 mm plywood plank fixed to the road (Fig. 17). This bump is referred to as the axle-hop plank (AHP) in this report. For the sake of completeness it has to be mentioned that the plank usually used in Switzerland, but not referred to in this report, is also 300 mm wide but has a thickness of 50 mm which is twice the one of the Australian axle-hop plank.

25

300 mm

Fig. 17

Cross-section through the Australian axle-hop plank (AHP).


51 / 169


2.3

8. June 2010

Vehicles

2.3.1 The NRC Test Vehicle Used for the Swiss Bridge Tests The NRC test vehicle is shown in Fig. 18 and Fig. 19. Details of the suspensions and tires used are given in Annex H. The vehicle instrumentation is described in detail in Paragraph 3.3.1 and Annex H. The NRC test vehicle was a five-axle tractor-semitrailer with an overall length of 17.8 m, a height of 3.66 m and a width of 2.62 m and 3.05 m with and without mirrors respectively. The trailer axle spacing was 1.27 m for the air-suspended (see Fig. 18) and 1.37 m for the steel-suspended vehicle. All other measures given in Fig. 18 were the same for both suspensions. The weight of the empty vehicle was 150 kN. The trailer carried a tank divided into four compartments. Two of them were filled with a total of 30’000 l of water during the tests. The vehicle gross weight during the tests was 450 kN. The static wheel loads were exactly determined with using a mobile flat-bed scale provided by the Canton Zürich Police Department. Because the vehicle was too long, too wide and too heavy, special permits were required to operate the vehicle in Switzerland.

NRC-Test-Vehicle 2.62 m 3.05 m

Water

empty

empty

Water

16'000 l

6'500 l

11'500 l

14'000 l 3.66 m

3.24 m

1.52 m

9.24 m

1.27m

15.27 m 17.80 m Axle load :

Fig. 18

50 - 60 kN

200 kN

The NRC test vehicle.

200 kN


Fig. 19

52 / 169 8. June 2010

The NRC test vehicle during the tests on Sort Bridge (EMPA Photo No. 101'070/30).

There were two suspension systems available with the NRC vehicle: a)

A steel suspension with the steer axle equipped with a multi-leaf steel spring, the tractor axles with a walking beam and the trailing axles with a four-spring-suspension.

b)

An air suspension, with the tractor and trailer tandem axles being equipped with independent air bag suspensions plus shock absorbers for each axle but with the steer axle suspension remaining the same multi-leaf steel spring as mentioned above.

From previous tests performed at NRC the fundamental body bounce frequencies were known to be f ≈ 1.6 Hz and f ≈ 3 Hz for the air and steel-suspended vehicle respectively. The vehicle as well as on-board electronics were provided together with the driver and the technicians operating the electronics by the National Research Council Canada, Center for Surface Transportation Technology, Ottawa, Canada. It was driven from Ottawa to Halifax, shipped from there to Rotterdam and then again driven to Switzerland. After having been used for the Swiss bridge tests the vehicle continued its journey to France (DIVINE Research Element 5) and Finland (Research Element 2) before being shipped back to Halifax and Ottawa.


53 / 169 8. June 2010

For the bridge tests in Switzerland the vehicle arrived with the steel suspensions mounted. The air suspensions were shipped by air to Switzerland. The first part of the tests in Switzerland being finished the suspension system was changed from steel to air at AMP Hinwil, a facility of the Swiss Army. For the rest of the vehicle’s journey through Europe the air suspension remained on the vehicle and the steel suspension was sent directly from Switzerland back to Ottawa via air freight. 2.3.2 The Test Vehicles Used in Australia

Fig. 20

The BA test vehicle crossing the bridge over Lawsons Creek.

Six-axle articulated over the rear axle tip trucks were used in the Australian test program. They were nominally loaded to their 425 kN legal limit (steer = 60 kN, tandem = 165 kN, tri-axle group = 200 kN). Tests were conducted with both steel and air suspensions in order to compare the two types of suspension. A summary description of the prime-movers and the suspensions is presented in Table 7. The details of the test vehicles can be found in Annex H. The vehicle instrumentation is described in detail in Paragraphs 3.3.2 and 3.3.3. The Australian test vehicles and drivers were supplied by Boral Transport. The instrumented trailer suspensions were supplied by a co-operative venture that transcended international boundaries. The instrumented axles were donated by BPW Germany and air freighted to Australia. The BPW air-suspensions were donated by Transpec Australia with York Australia providing the 8-leaf steel suspensions. These suspensions were fitted into a removable frame that was consistent with Boral Transport’s fleet by Hamalex Australia.


54 / 169 8. June 2010

Boral, with assistance from Hamalex and Heggies Bulkhaul provided the facilities and the expertise to interchange the air and steel trailer suspensions during the testing. The prime-mover suspension was interchanged by swapping an air-suspended prime-mover with a steelsuspended model manufactured by the same company and with similar specifications. This required the removal of this vehicle from service and the replacement of its low profile tires so that the same tires were used for all tests. Mercedes Benz Australia helped make this possible. The BA and BS vehicles (see codes given in Table 7) were the basic test vehicles. These vehicles were instrumented to measure the dynamic wheel force for each of the trailer axles. The presence of the other vehicles provided an opportunity to extend the research and the database. Of particular interest is the air-suspended petrol or SA vehicle (Camerons Creek Bridge tests only). At the Coxs River Bridge tests the spacing between the axles of the tri-axle group of the BA vehicle was made unequal. This configuration is presented as "BA unequal".


55 / 169


Prime-Mover

Freightliner, air suspension

8. June 2010

Trailer

Gross Laden Mass (kN)

Vehicle Code

Over the rear axle tri-axle tipper, BPW air suspension

425

BA

60 kN

165 kN

4.07

200 kN

1.30

4.65

1.23

1.23

12.48 m

Over the rear axle tri-axle tipper, York 8 leaf steel suspension

Freightliner, Hendrickson walking beam, steel suspension 60 kN

165 kN

3.71

425

BS

200 kN

1.34

4.65

1.23

1.23

12.16 m


450

Petrol tanker, BPW air suspension with shock absorbers fitted diagonally

55 kN

165 kN

3.66

SA

230 kN

1.32

6.56

1.4

1.4

14.34 m



60 kN

165 kN

4.07

425

BA unequal

200 kN

1.30

4.03

1.85

1.23

12.48 m

Table 7 Test vehicle dimensions, weights (nominal) and suspensions - Australia.


56 / 169


8. June 2010

The Australian National Road Transport Commission commissioned a study to compare the low frequency dynamic response of a range of truck and trailer suspensions ([Sweatman, 1994]). The trailer suspensions used as part of the BA and BS vehicles were tested in this way. In addition, prime-movers fitted with similar suspensions to the BA and BS vehicles were tested. The suspensions were characterized in terms of natural frequency and damping. Testing involved slowly driving the vehicle off 80 mm blocks and recording the dynamic response of the vehicle in accordance with the European Community (EC) drop test which defines a "roadfriendly" suspension as one with a natural frequency of less than f = 2.0 Hz and damping coefficient greater than ζ = 20%. Test results are presented in Table 8.

Vehicle

BA Series, Freightliner and

Prime-mover

Trailer

Natural frequency [Hz]

Damping ratio [ %]


Damping ratio [ %]

1.45

20

1.4

35

2.75

6

3.2

10

BPW air suspensions BS Series, Hendrickson Walking beam and York 8 leaf steel suspensions

Table 8 Low frequency suspension characteristics of Australian test vehicles [Sweatman, 1994]. The air-suspended vehicle BA satisfied the requirements for "road-friendly" suspensions. The mechanical suspensions of vehicle BS did not. The contrast between the damping levels of the prime-mover and the trailer fitted with air suspensions is important. The 35% damping achieved by the trailer suspension was much higher than damping levels in the prime-mover (20%). At 20% damping, the prime-mover borders on “non-road-friendly”. Of all the suspensions, the trailer suspension exhibited the highest degree of damping. The steel suspensions used in the test program were at the low-damping/highfrequency end of the scale. Thus, the tested suspensions represent both ends of the spectrum of suspensions in good working condition. Test vehicles did not include air-suspended vehicles with poorly functioning dampers although for one test (Coxs River) the shock absorbers were removed from the rear axle of the BA vehicle.


3.

INSTRUMENTATION, DATA ACQUISITION

3.1

Instrumentation of Swiss Bridges

57 / 169 8. June 2010

Detailed arrangements of the instrumentation used during testing are presented for each bridge in the appropriate Annex. The following general discussion applies to all of the bridges tested in Switzerland. The bridges were instrumented to determine their vertical dynamic displacement at several points. Their horizontal displacement in the longitudinal direction was also measured at one abutment. An additional channel was used to register the information on instant vehicle speed provided by several contact thresholds distributed along the test track. These rubber hose contact thresholds were positioned at the start of the test track, the start of the bridge and at the bridge end. The number of measurement channels was limited by the capacity of the 2.45GHz radio telemetry link used to send the bridge response signals to the test vehicle (eight channels). Vertical dynamic displacements were measured using the spring and wire technique. This system was adopted in preference to other methods such as the double integration of accelerations because of the confidence in the accuracy of the results established over many years of conducting dynamic tests. For the vertical displacement measurements, inductive displacement transducers (HottingerBaldwin HBM W20) with a measurement range of ±20 mm were installed inside the box girder. In all cases, the wire-supported measurement technique was applied. This technique uses a wire stretched between the bridge superstructure and the terrain underneath as a measurement basis. A force of nominally 0.3 kN is applied to the wire through mounting of a spring at one wire end. The wire force is controlled through measurement of the initial spring length. For the sake of simplicity Fig. 21 shows this technique schematically when the transducer is located close to the terrain underneath the bridge. With the transducer mounted inside a box beam, the spring is fixed to the underside of the bridge deck slab and the transducer coil to the upper side of the bottom slab using a steel bar screwed to this slab. In any case the transducer iron core (freely moving inside the coil) is fixed to the invar wire. With the transducer mounted close to the terrain, the wire reflects the superstructure movement (the iron core duplicating this movement) whereas with the transducer mounted inside the box cell the wire does not move and the bridge movement is duplicated through the movement of the coil. A systematic influence of the force variation in the wire due to the bridge movement, i.e. to the associated spring length (force) variation on the


58 / 169 8. June 2010

displacement measured has to be considered. This amounts to roughly +3% per 10 m wire length. The correction factor is always greater than 1.0. For the horizontal longitudinal displacement measurements, a transducer HBM W50 with a measurement range of ±50 mm was installed between the end of the bridge superstructure and an abutment wall.

Fig. 21

Scheme of the wire-supported displacement measurement technology as used in Switzerland. The axial force in the wire is produced by stretching of the spring. Measuring of the stretched spring length allows checking of the nominal force being close to the nominal value of 300 N.

For the Swiss tests, the abbreviation used to identify a displacement measurement point is WGxy. "WG" is the German short form for "displacement transducer", "x" identifies the bridge span number and "y" the measurement point location in the bridge cross section. The "test track" is the piece of highway (including the bridge under investigation) where the data acquisition system on-board of the test vehicle was active.


59 / 169 8. June 2010

3.1.1 Sort Bridge Vertical displacement was measured at one point in the mid-span region of all of the five spans. Because the terrain underneath the bridge was unusually difficult (a four-lane highway and a river respectively), positioning of the wires was possible at one point per span only. In some cases, the measurement points had to be slightly shifted from the theoretical mid-span point. The horizontal transducer was installed at the Bellinzona side abutment. The test track started 150 m from the beginning of the bridge and ended 70 m from its end. Refer to Annex A for further details. 3.1.2 Deibüel Bridge Vertical displacement was measured at two points located in the middle of each of the three spans. The horizontal transducer was positioned at the Zug side abutment. The test track started 120 m before the beginning of the bridge and ended 65 m from its end. Refer to Annex B for further details. 3.1.3 Föss Bridge Vertical displacement was measured at three points of the central span and at one point of both side spans. The horizontal transducer was located at the uphill St. Gotthard side abutment. The test track started 90 m from the beginning of the bridge and ended 100 m from its end. Refer to Annex C for further details. 3.2

Instrumentation of Australian Bridges

The following applies to all bridges tested in Australia. The bridges were instrumented to determine their vertical displacement at four points. Strain gauges and accelerometers supplemented the deflection data. In addition, axle detectors were used to record the time when each axle entered or exited the bridge. The speed and direction of the test vehicles was deduced from this data. A diagonal detector was also used to determine the lateral position of the test vehicles. For the vertical displacement measurements, inductive displacement transducers (HottingerBaldwin HBM W10 and W20) with a measurement range of ±10 mm and ±20 mm were installed on the underside of the bridges. As with the Swiss tests, the wire-supported measurement technique was applied.


Fig. 22

60 / 169 8. June 2010

Wire-supported displacement measurement technology fitted to the bridge over Camerons Creek.

For the Australian tests, the abbreviation used to identify a displacement measurement point is D(x,y). "D" means deflection measurement point, "x" identifies the span number and "y" the deck plank or girder number. Spans are numbered from Abutment A and girders/deck planks from the left. D(3,2), for example, refers to the third span and the second timber beam from the left for the Cromarty Creek bridge. 3.2.1 Lawsons Creek Bridge The displacement transducers were fitted at midspan of this single span bridge. They were arranged laterally to provide further information about the distribution of the static and dynamic loads. The axle hop plank (AHP) influenced the response for westbound runs only. Refer to Annex D for further details. 3.2.2 Coxs River Bridge Three of the four displacement transducers were concentrated in span 2 immediately after the axle-hop plank (AHP) for vehicles traveling in the northerly direction towards Lithgow. Higher speeds were possible in this direction of travel. The proximity of an intersection in the opposite direction reduced speeds for vehicles traveling to the south. The fourth displacement


61 / 169 8. June 2010

transducer monitored an end span in order to evaluate the frequency and damping of the bridge under free vibration conditions. In addition, a series of strain gauges were monitored. Refer to Annex E for further details. 3.2.3 Camerons Creek Bridge The first and last spans of the bridge over Camerons Creek were instrumented with displacement transducers at midspan in order to record the maximum effects of any roughness of the approach stretches and to contrast this with vehicles exiting the bridge having traveled over the relatively smooth bridge pavement. Three of the transducers were concentrated in span 4 immediately after the 300 mm x 25 mm axle-hop plank (AHP). Again, instrumentation of the end spans allowed for the evaluation of the natural frequency of the bridge. Refer to Annex F for further details. 3.2.4 Cromarty Creek Bridge The central four of six girders in span 3 of the bridge were monitored. A single span was chosen because of the inherent variability in natural timber beams and to provide the opportunity to compare the effects of the traffic entering the bridge and exiting the bridge. The AHP influenced the measured response for northbound traffic only. Refer to Annex G for further details.


62 / 169


3.3

8. June 2010

Vehicle Instrumentation

The dynamic wheel forces applied by the trucks to the bridges were measured utilising the instrumented axles fitted to the vehicles. In all cases, the dynamic wheel force was measured using a combination of strain gauge rosettes and accelerometers. The strain gauges were arranged to measure the principal strains induced by the shear in the stub axle and hence the shear force in the stub axle. This shear force then needs to be adjusted by the inertia effects of the mass between the shear transducer and the point of application of the dynamic wheel force. This accelerometer is utilised to estimate these inertial effects. This methodology is summarised Fig. 23. For more detailed information refer to the OECD IR2 report which contains a detailed discussion of the various methods of measuring dynamic wheel forces.

Strain gauges to measure strains due to shear

Outboard mass mo

FP = FV + moxa

where:

FV

FP

Dynamic wheel force applied to the pavement

FV

Dynamic shear force measured in axle stub

mo

The mass outboard of the strain gauges

a

The acceleration of the outboard mass (m).

mo x a Accelerometer & signal conditioning

FP

Fig. 23

Instrumentation for measuring dynamic wheel forces

3.3.1 The NRC Vehicle To continuously measure the dynamic wheel loads during the passages over the bridges, each wheel of the tractor and trailer was instrumented with an accelerometer and a shear force strain gauge located close to the wheel hubs (Fig. 23 and Fig. 24). Transformation of these strain and acceleration data to wheel load data is described in detail in [LeBlanc, 1992]. For the case of a malfunction of the shear gauges, additional bending moment strain gauges were fitted to the axles.


Fig. 24

63 / 169 8. June 2010

NRC test vehicle instrumentation: an "undercover" shear gage and an accelerometer located close to the wheel hub can be seen (EMPA Photo No. 101'155/5).

To continuously control the speed of the vehicle during the runs and to precisely determine the distance traveled by the vehicle, the measurement system consisted of three elements. Firstly, reflective tape strips were glued to the pavement. At the beginning and end of the test track four strips were positioned 190 mm apart from each other. These triggered the on-board data acquisition system described below. On the test track itself, the distance between each strip was 30 m. This means that the total test track length was a multiple of 30 m. Secondly, a reflective tape sensor fixed to the vehicle front bumper was activated each time the sensor crossed a reflective strip. This Datron Sensor measured the vehicle speed via time and distance between reflective tapes. An Odometer System measured the precise position of the truck on the test track and controlled the start and stop of the on-board signal data acquisition system by detecting the four white tape strips located at the start and end points of the test track. In case of rainy weather conditions the Odometer System also allowed manual triggering. In addition to its trigger function it yielded a stabilized digital speed signal for the Speed Indicator System located in the driver’s cabin. This was connected to a color-bar-graph display which indicated the speed


64 / 169 8. June 2010

deviation from a pre-selected target speed thus helping the driver to keep the vehicle speed as constant as possible. 3.3.2 The Boral Vehicle Each wheel of the trailing tri-axle groups of the BA and BS vehicles was instrumented with strain gauges measuring the principal strains due to shear in the stub axle and accelerometers to measure the acceleration of the outboard mass. The shear gauges were fitted by BPW in Germany and tested to demonstrate their linearity and insensitivity to torsion and bending influences. The Qflex QA700 accelerometers and the strain gauge conditioning were incorporated into a weatherproof box mounted under the axle adjacent to each wheel. This system was designed and manufactured by ARRB Transport Research. Speed was controlled by the drivers using the vehicle's speedometers and engine speed displays. Axle detectors placed at each abutment of the bridge were used to determine the average speed of each passage of a test vehicle. The variability of vehicle speed while on the bridge was relatively small given that the maximum length of bridge was 46.1 m. The vehicle based data acquisition system was triggered by an infra-red emitter/receiver mounted on the rear of the prime-mover in combination with reflective tape fixed to the pavement. 3.3.3 The SA vehicle No wheel load measuring devices were fitted to this vehicle.

3.4

Data Acquisition

3.4.1 Tests in Switzerland For the tests in Switzerland, two data acquisition systems were used. The central data acquisition unit was an NRC MegaDac system located on-board the NRC vehicle. A total of 24 channels (wheel loads and vehicle speed/position control) were recorded directly with a sampling rate s = 500 Hz. A second, independent PCM-data acquisition system was located in the EMPA measurement van positioned either close to a bridge abutment or underneith the bridge. The bridge response signals were acquired, digitized (sampling rate s = 370 Hz) and stored on magnetic tape.


65 / 169 8. June 2010

With the EMPA PCM (Pulse Code Modulation) data acquisition system the signals of the transducers passed first through amplifiers with amplification factors ranging through 0.1 / 0.2 / 0.5 / 1 / 2 / 5. After being passed through anti-aliasing filters the signals were digitized with a 12 bit resolution in the PCM modulator and subsequently recorded on a Stellavox magnetic tape recorder. In parallel, a maximum of seven bridge response and one contact threshold signals were sent via a 2.45-GHz radio telemetry link to the NRC test vehicle. A telemetry receiver and PCM D/A-converter were located on the NRC vehicle (Fig. 25). From this converter, the backtransformed analog signals were fed to the NRC MegaDac system and stored simultaneously with the vehicle signals (s = 500 Hz). The MegaDac was controlled by an on-board NRC computer using an Optim's TCS software package. After each bridge test, all data recorded with the MegDac was dumped to this on-board NRC computer and after each day of testing the data was also copied to an EMPA computer. The NRC on-board computer was also used for on-line quality control of the dynamic wheel load signals received. This on-line control soon revealed that the choice of the power supply for the EMPA equipment on-board the NRC vehicle was not optimal. The NRC equipment was run with a 110 V/60 Hz power generator fixed to the vehicle tractor. However, the EMPA equipment used the vehicle battery and a device to convert the 24 V DC current provided by this battery to the necessary 220 V/50 Hz AC current. This chopper obviously severely distorted the signals from the wheel load measurement instruments. After the power supply for the EMPA on-board equipment was changed from the chopper to an extra 220 V/50 Hz power generator placed on-board the NRC vehicle, the wheel load signals were no longer distorted. Unfortunately, the attempts undertaken to simultaneously acquire dynamic wheel load and bridge deflection signals were less successful than those described in [Cantieni, 1992]. In the DIVINE tests, the same radio telemetry equipment was used for simultaneous data acquisition as in [Cantieni, 1992]. However, because the more vulnerable receiving part of the equipment was now installed on-board the vehicle instead of in the stationary measurement van it had to be noticed that this part quit service after approximately half of the tests had been performed. It was not possible to take corrective measures, which means replacement of some broken electronic parts of the equipment, before the end of the tests. In addition, due to the presence of a high voltage line crossing the bridge near the south abutment the telemetry link never worked satisfactorily for the Sort Bridge. Having kept the possibility of such problems arising in mind, the triggering systems of the NRC and EMPA signal acquisition systems were connected via exact measurement of the respective contact thresholds. With help of these two trigger systems it was possible to synchronize the vehicle wheel loads stored with the Canadian MegaDac system and the bridge


66 / 169 8. June 2010

response signals stored with the EMPA PCM-system respectively. The uncertainty of the synchronization was estimated to be better than Δt = 0.01 s. This allowed calculation of the crosspower spectra between wheel load and bridge response signal with adequate phase accuracy in the frequency region of interest. This means that processing of the data concerning vehicle/bridge interaction will be difficult for tests where the telemetry did not work properly. However, individual signals from the vehicle and bridge instrumentation respectively were not affected by these problems.

Fig. 25

The data acquisition systems located in the NRC vehicle's driver cabin: to the left: the Canadian MegaDac system, to right (orange) the Swiss PCM system (EMPA Photo No. 101'176/19).

3.4.2 Tests in Australia Data from the Australian tests was collected using two identical Blastronics BMX data acquisition systems. One was mounted in the vehicle and the second recorded data from the bridge. The systems were triggered simultaneously via road-mounted devices. The bridge system was triggered via the axle detectors placed at each abutment. The vehicle instrumentation system was triggered by contact between an infra-red emitter/receiver attached to the


67 / 169 8. June 2010

vehicle sensing reflectors mounted on the road. The reflectors were positioned relative to the axle detectors so that both the bridge and vehicle data acquisition systems triggered when the steer axle of the test vehicle crossed an abutment in either direction. The data acquisition system’s pre-trigger recording system was used to record the response of the bridge and the dynamic wheel forces prior to the test vehicle entering the bridge. The sampling duration was sufficient to allow the vehicle to cross the bridge plus a further allowance to measure the frequency and damping from the free vibration of the bridge. The bridge and vehicle responses were recorded at a sampling rate s = 200 Hz except for speeds v < 30 km/h which were sampled at s = 50 Hz. The signal from the axle detectors was recorded at s = 1’000 Hz. By post-processing the axle detector signals, the velocity, i.e. speed and direction, of the vehicle entering, crossing and exiting the bridge was determined. The lateral position of the vehicle was determined in a similar manner. At the southern abutments, a second axle detector angled at 45° was positioned at a known distance from the abutment detector. This allowed the lateral position of the vehicle to be calculated using the geometry of the axle detectors, the average speed of the vehicle and the time taken to cross the two detectors. For tests on the Camerons and Cromarty Creek bridges, the performance of the infra-red emitter/receiver and the reflective strip mounted on the pavement was somewhat less than desired. On some occasions, the system did not trigger and required manual triggering. This resulted in a loss of correlation between the bridge and vehicle-based data acquisition systems. This problem was overcome for the testing of Lawsons Creek and Coxs River bridges by arranging for the infra-red sensor to be mounted much closer to the road surface. The bridge and vehicle transducers were conditioned adjacent to the transducers. The signals passed through a 50 Hz anti-aliasing filter and a programmable gain (1, 2, 4, 8, 16, 32, 64, 128 or 256) was applied before the signals were converted to a digital form using a 12 bit analog to digital converter. The digital signal was then transformed into engineering units and recorded directly to computer hard-drive for further processing. The instrumentation and data acquisition systems were powered from 12 V sources available at the bridge and in the vehicle.


4.

TEST PROCEDURES AND PROGRAM

4.1

Tests in Switzerland

68 / 169 8. June 2010

The test procedures were simple inasmuch as the vehicle was driven at a constant speed over the test track. The dynamic wheel load signals and the bridge response signals were recorded on-board the NRC vehicle and in the EMPA measurement van. Normally, the first couple of runs were performed at crawl speed to allow for the establishment of the quasistatic bridge response. The vehicle speed was then increased in steps of Δv ≈ 2...3 km/h until the practically achievable speed was reached. The driving direction was always the same. In all cases, this was the downhill direction because the usually significant slope made it impossible to reach a reasonable vehicle speed when driving uphill. The driving axis was identical with the bridge centerline in all cases. To keep the lateral deviation from the theoretical wheel path as small as possible, this lane was always marked with red cones (see Annexes A to C for details). Preliminary tasks were performed in February and March 1994 in Ottawa and Switzerland respectively. These tasks included the adaptation of the NRC and EMPA electronic equipment (telemetry system, data transfer from the NRC to the EMPA computer) and the preparation of the bridge instrumentation (application of the wires). Upon arrival of the NRC measurement technicians at EMPA, the reflective strips were fixed to the pavements of the test bridges on April 13 and 14, 1994. The NRC test vehicle and the air suspensions arrived at EMPA on April 15. The vehicle’s electronic equipment was then thoroughly checked, the water tanks filled and the static wheel loads determined. The latter was performed with using flat bed scales provided by the Canton Zürich Police. The tests using the NRC vehicle equipped with steel springs were performed between April 18 and 29. The suspensions were changed from steel to air at the Swiss Army Facility in Hinwil between May 2 and May 6. The tests with the air-suspended NRC vehicle took place between May 16 and May 27. The NRC vehicle left for tests in France May 30, 1994. Determination of the longitudinal profiles of the test pavements was performed in the first week of May 1994 on all test bridges. The Ambient Vibration Tests to determine the test bridge's modal parameters were performed between September 1994 and April 1995. For tests on the Sort Bridge, the traffic was managed by the Canton Ticino Highway Department staff. The traffic on the bridge was stopped for single tests and allowed to pass between


69 / 169 8. June 2010

the tests. This task was demanding because waiting times reached 15 minutes in extreme cases. Motorists did not always appreciate the delay. For tests on the Deibüel Bridge, the southbound traffic had to detour through one of the two lanes usually reserved for the northbound traffic. Two-way traffic was organized on the two northbound lanes. This was achieved by the Canton Zug Highway Traffic Department. No traffic management measures had to be taken for tests on the Föss Bridge because the Gotthard Pass Highway had not yet been opened to traffic. It was, however, opened to traffic the day after tests with the NRC vehicle equipped with air springs were finished. 4.2

Tests in Australia

In the Australian tests, the test vehicles traversed the bridges in both directions beginning at crawl speeds and then incrementing in Δv = 10 km/h intervals to top speeds of v = 100 km/h. The increment was reduced to Δv = 5 km/h around those speeds where increased dynamic activity was observed/predicted. The dynamic response of the bridge was monitored and additional runs were made at intermediate speeds as required. These additional speeds were particularly focused around v = 50...60 km/h. At Cromarty Creek, Lawsons Creek and Coxs River, it was not possible to reach v = 100 km/h in both directions for reasons of road horizontal geometry, steep incline and the proximity of an intersection respectively. To assist in the transverse positioning of the vehicle, lines were marked corresponding to the drivers side steer wheel. The tests were repeated with an axle-hop plank (AHP) fitted to the bridge. This AHP was used to excite the axle-hop frequencies of the truck suspensions. A body bounce bump (BBB) was also fitted to the bridge over Camerons Creek. This BBB simulated the long wavelength variations in the longitudinal profile. The interpretation of the BBB data was complicated by the long wavelength components of the road profile over the bridge. Consequently this data was of limited value and is not presented in this report. Furthermore the BBB was not continued as a feature of the testing program. Traffic control was undertaken by staff of the Northern and Western Regions of the Roads and Traffic Authority (RTA) of NSW under the direction of RTA Technology. This required the preparation of safety plans, advertising, advance warning signs, temporary dates, traffic lights and traffic controllers. The bridge tests were conducted in pairs - Camerons and Cromarty Creeks between 29 and 30 April 1994 and Coxs River and Lawsons Creek between 28 and 31 March 1995. In both cases the testing program began with the BA test vehicle with the tests being conducted on


70 / 169 8. June 2010

the first and second bridge. The suspensions were then changed (BS) and the tests conducted again on the second and then the first bridge. This sequence was adopted as it was easier to change the bridge instrumentation than to change the suspensions. The axle hop planks were only on the road for the duration of each test. The suspensions were changed at Boral’s Tomago workshop with assistance from Hamalex Australia during the Camerons and Cromarty Creek testing program. For these tests the vehicles were loaded with crusher fines at Boral’s Seaham quarry. The weigh-station at the quarry was used to weigh the test vehicles. These results were confirmed using portable scales. During the Coxs River/Lawsons Creek testing program, Heggies Bulkhaul workshop at Wallerawang provided the facilities to change the suspensions and instrument the vehicles. Boral Transport assisted by Hamalex Australia and Heggies Bulkhaul provided the labour for these actuaries. The tests vehicles were loaded with coal by Heggies Bulkhead and weighed at a local weigh-station.


5.

DATA PROCESSING

5.1

Time Domain Analysis Methods as Used in Switzerland

71 / 169 8. June 2010

5.1.1 Bridge Response Data Processing Methods For on-line control of the signal quality a PCM demodulator was connected to the tape recorder and the signals were back-transformed into analog form. This happened in order to plot the bridge response and contact threshold signals actually stored on tape on a 12-channel paper-strip recorder (ref. Chapter 6.1 and Annexes A to C for examples). Signal amplification could thus be adapted as necessary. In addition, preliminary signal-processing was achieved by feeding the signal of a representative channel into a tracking voltmeter which allowed for determination of the dynamic increment for the respective signal. Hence, it was possible to adapt the test schedule through including additional runs in the region where high dynamic increments were determined. The bridge response frequency content was checked on-line with the help of a two-channel ONO SOKKI CF 920 frequency analyzer. For data processing in the laboratory, the signals were played back in digital form from the tape via the PCM demodulator and a de-multiplexer being connected to a computer. Subsequent time domain signal processing was performed with software packages developed at EMPA. Besides determining the dynamic increment φ as defined earlier, the software also allowed for exact evaluation of the effective vehicle speed through processing of the contact threshold signals. The final result of the time domain analysis were the dynamic increment versus vehicle speed graphs, φ(v)-curves, for every bridge response signal measured. 5.1.2 Wheel Load Processing Methods After each day of testing, the wheel load instrumentation signals recorded in the NRC MegaDac system (and if the telemetry link had been working properly, also the bridge response signals) were dumped to an EMPA Toshiba computer. The wheel load instrumentation signals were subsequently transformed from strain gauge and acceleration signals into wheel load signals in ASCII format. For this purpose, the data was processed by two programs. With the TCS-Plus program, the raw strain gauge and acceleration data was checked on its quality by means of graphical display on the computer screen and then transformed into DSP format. These DSP Files were then processed by a Fortran program (named EMPA32A, written by NRC) which transformed them into wheel load signals in ASCII format. Time domain analysis of the wheel load signals included determination of the dynamic load coefficients (DLC's) as defined in Paragraph 5.2.3 and discussed in Chapter 6.


5.2

72 / 169 8. June 2010

Time Domain Analysis Methods as Used in Australia

5.2.1 Bridge Response Data Processing Methods The bridge response data was monitored visually via the computer graphics of the data acquisition system. As an additional measure, on-line processing also provided peak-to-peak values which were monitored to determine the levels of dynamic response and thus identify any speed ranges which needed to be supplemented. The data was recorded directly into computer files in a binary format which was subsequently converted into the ASCII format. In the case of the bridge responses, any deflection offsets were removed, the deflection waveforms were printed and the peak responses were determined for each run. Software developed in the Matlab language was used for both the time domain and the frequency domain analyses. The results were summarized in terms of deflection, speed, direction, vehicle suspension, and bumps. From this data the dynamic increment versus speed analyses were undertaken. Both directions of travel were presented on a single graph by assigning speeds in one direction (e.g. northbound) as positive and in the opposite direction negative. 5.2.2 Wheel Load Processing Methods The total dynamic force of the tri-axle group and the dynamic force for individual axles of the tri-axle group was determined using software developed by QUT. Parameters such as amplitude, minimum, maximum, standard deviation and mean value of the dynamic force as well as the dynamic load coefficient were calculated for all truck configurations and various velocities. The dynamic load coefficient (DLC) is the wheel load coefficient of variation and hence a simple measure of the magnitude of the dynamic variation of axle load for a specific combination of road roughness and speed. It has been determined in accordance with the following definition: DLC =

where

σ F

σ

= the standard deviation of the dynamic variation of the axle load and

F

= the mean value of the dynamic axle load.


5.3

73 / 169 8. June 2010

Frequency Domain Analysis Methods

5.3.1 General

Analyses in the frequency domain were performed using the Fast Fourier Transform (FFT) algorithm. Main general assumptions when performing FFT are that the time signal to be transfomed is stationary and ergodic. This assumption is reasonable for the analysis of dynamic wheel loads unless the pavement profile is very far from being a random-type function. Otherwise, the usual techniques concerning anti-aliasing and leakage errors were applied. When it comes to transformation of dynamic bridge response signals, FFT should theoretically not be applied because this response is of basically transient nature. However, application of FFT provides useful information on the bridge response signal frequency content. Similarly, calculation of crosspower-spectra derived from wheel load and bridge response spectra to investigate the dynamic vehicle/bridge interaction is from a scientific standpoint not a very sound procedure. It is however the only method available up to now. The resolution of frequency domain analyses is directly dependent on the length of the time signals to be transformed: the longer the signals, the better the frequency resolution achievable. Unsatisfactory resolution, i.e. Δf > 0.1 Hz, resulted in the cases of short bridges and moderate vehicle speeds and of long bridges and high vehicle speeds. 5.3.2 Wheel Load Spectra as Determined in Switzerland

Analysis in the frequency domain was performed with using software packages developed by EDI, Experimental Dynamics Investigations, Vancouver, Canada [Felber, 1993], [EDI, 1995]. The ASCII files containing the scaled wheel load signals (NRC) were split into time signals for a single wheel per file. Separate power spectra were calculated for both the segment before the bridge and the run over the bridge. With the help of the EDI software it was possible to calculate the power spectra of the wheel loads in an efficient manner. Moreover, with this software (having primarily been written for the processing of data resulting from Ambient Vibration Tests on civil engineering structures) it is possible to animate the corresponding operational mode shapes of the vehicle. With the help of this visual animation it was then possible to identify these mode shapes. The analysis was performed in the following steps: •

Calculating the actual wheel loads from the raw data with suitable calibration factors.

•

Splitting the data into separate data sets for every wheel.


74 / 169 8. June 2010

•

Defining the lengths of the data for the two parts of a run where a) the vehicle was on the approach and b) the vehicle was on the bridge.

•

Calculating the frequency spectra for the two cases a) and b) for every wheel.

•

Calculating the cross power spectra for the two cases a) and b) for all pairs of wheels.

•

Identifying the peaks and the corresponding frequencies in the frequency spectra for every wheel.

•

Plotting the frequencies of the identified peaks in the frequency spectra against velocity for the two cases a) and b) for every wheel.

•

Animating the vehicle's operational mode shape for all frequencies identified in the frequency spectra and identifying the corresponding shape.

•

Plotting the mode shapes and the corresponding frequencies against velocity for the two cases a) and b).

5.3.3 Vehicle/Bridge Interaction as Investigated in Switzerland

To analyze the dynamic interaction between vehicle and bridge the simultaneously acquired wheel load and bridge response signals were again processed with using the EDI software packages. From the frequency spectra of these signals the cross-power spectrum was determined. This provided the necessary magnitude and, of especial importance, the phase information related to a given pair of such signals which allowed interpretation of the vehicle/bridge interaction process. 5.3.4 Frequency Domain Analysis Methods as Used in Australia

The Blastronics data acquisition system software enabled calculation of the measured time signals power spectral densities, thus enabling the vehicle's dynamic wheel load and the bridge response frequencies to be monitored during testing. Power spectral densities (PSD) of the total dynamic force of the tri-axle group and for individual axles of the tri-axle group were produced for selected runs and truck configurations. The vehicle response was recorded while the vehicle was on the bridge with a short pre-trigger time. The signal corresponding to the vehicle on the approach is too short to be analyzed independently in the frequency domain. The PSD of the tri-axle group and independent axle dynamic forces were hence produced for the vehicle on the bridge only.


75 / 169


6.

TEST RESULTS

6.1

Time Domain Analysis: Bridge Response Signals

8. June 2010

Here, some selected representative results are presented only. These include typical deflection time signals and dynamic increment versus vehicle speed graphs, φ(v)-curves, for typical measurement points as well as a short description of the facts. The abbreviations used concerning the test parameters are described in Chapter 2. Chapter 7 contains a detailed discussion. Additional information concerning the time domain analysis can be found in the relevant annexes. 6.1.1 Sort Bridge a)

Typical deflection time signals

WG 32

3 mm

Fig. 26

2s

Sort Bridge mid-span deflection for the steel-suspended NRC vehicle, v = 29.9 km/h; φ = 2.01%.

WG 32

5 mm Fig. 27

1s

Sort Bridge mid-span deflection for the air-suspended NRC vehicle, v = 75.2 km/h; φ = 25.45%.


76 / 169


b)

8. June 2010

Dynamic increments 30


SORT Bridge WG 32 25

Static deflection: 5.68 mm 20

15

STEEL AIR

10 5 0 0

10

20

30

40

50

60

70

80

Vehicle Speed [km/h]

Fig. 28

Dynamic increments vs. speed for the Sort Bridge mid-span deflection.

For mid-span of the Sort Bridge largest span the dynamic increments are significantly higher for the air than for the steel-suspended vehicle for speeds v > 40 km/h (Fig. 28). The maximum values are φ = 26% and φ = 10% for steel and air suspension respectively. For lower speeds, the dynamic increments are very similar and with φ ≤ 5% at a rather low level. Additional information is given in Annex A. 6.1.2 Deibüel Bridge a)


WG 21 2.5 mm

Fig. 29

2s

Deibüel Bridge mid-span deflection for the steel-suspended NRC vehicle, v = 50.6 km/h; φ = 20.84%.


77 / 169


8. June 2010

WG 21 2s

2.5 mm Fig. 30

Deibüel Bridge mid-span v = 37.3 km/h; φ = 3.73%.

b)

deflection

for

the

air-suspended

NRC

vehicle,



DEIBÜEL Bridge WG 22 25

Static deflection: 2.96 mm 20 15

STEEL AIR

10 5 0 0

10

20

30

40

50

60

70

80


Fig. 31

Dynamic increments vs. speed for the Deibüel Bridge mid-span deflection.

For mid-span of the Deibüel Bridge largest span the dynamic increments are significantly higher for the steel than for the air-suspended vehicle for almost all speeds (Fig. 31). The maximum values are φ = 21% and φ = 5% for steel and air suspension respectively. Additional information is presented in Annex B.


78 / 169


8. June 2010

6.1.3 Föss Bridge a)


WG 22 5s

2 mm

Fig. 32

Föss Bridge mid-span v = 14.9 km/h; φ = 3.86%.

deflection

for

the

steel-suspended

NRC

vehicle,

WG 22 1 mm

Fig. 33

2s

Föss Bridge mid-span deflection for the air-suspended NRC vehicle, v = 27.9 km/h; φ = 3.11%.


79 / 169


b)

8. June 2010



FÖSS Bridge WG 22 20

Static deflection: 1.44 mm 15 STEEL AIR

10

5

0 0

10

20

30

40

50

60

Vehicle Speed [km/h] Fig. 34

Dynamic increments vs. speed for the Föss Bridge mid-span deflection.

For mid-span of the Föss Bridge largest span, the dynamic increments reach smaller values than for the respective points of the Deibüel and Sort bridges. The results for the steel and airsuspended vehicle are of comparable size with often slightly larger values for the steelsuspended vehicle (Fig. 34). The maximum values are φ = 15% and φ = 12% for the steel and the air suspension respectively. Additional information is presented in Annex C.


80 / 169


8. June 2010

6.1.4 Lawsons Creek Bridge Typical deflection time signals

Deflection [mm]

a)

0.5 0.0 -0.5 -1.0 -1.5 -2.0 -2.5 -3.0 -3.5

D1

D3, D7

D5

0

1

2

3

4

5

6

Time [s] Lawsons Creek Bridge, mid-span deflections D1, D3, D5 and D7, BS vehicle, v = 60 km/h.

Deflection [mm]

Fig. 35

0.5 0.0 -0.5 -1.0 -1.5 -2.0 -2.5 -3.0 -3.5

D1 D3, D7

D5

0

1

2

3

4

Time [s] Fig. 36

Lawsons Creek Bridge, mid-span deflections D1, D3, D5 and D7, BA vehicle, v = 99 km/h.

The dynamic component of the deflection time signals for the bridge over Lawsons Creek was relatively small. Fig. 35 and Fig. 36 illustrate this along with the fact that the bridge exhibits low damping (ζ = 1%). The deflection D5 is located under the centerline of the single-span bridge and is the basis of the φ(v)-curves that follow.


81 / 169


b)

8. June 2010

Dynamic increments

Fig. 37 shows φ(v)-curves for the steel (BS) and air (BA) suspended vehicles crossing the Lawsons Creek Bridge without axle hop plank and in Fig. 38 for the same cases with the axle hop plank (AHP) in place over abutment A. 20

Lawsons Creek Dynamic Increment [%]

15 10 5 0 -5 -100

Fig. 37

-50

0 Vehicle Speed [km/h]

50

100

Dynamic increments vs. speed (positive: west-bound) for Lawsons Creek Bridge, deflection D5. Solid squares: steel suspension, open squares: air suspension. 20

Lawsons Creek [Axle hop plank]


15 10 5 0 -5 0

Fig. 38

20

40 60 Vehicle Speed [km/h]

80

100

Dynamic increment vs. speed for Lawsons Creek Bridge with axle-hop plank, deflection D5. Solid squares: steel suspension, open squares: air suspension.

For runs without the axle hop plank, the steel-suspended as well as the air-suspended vehicle induced maximum dynamic increments were φ < 5%. When the axle-hop plank was fitted to the bridge, the maximum recorded dynamic increment was φ = 15% for the steel-suspended vehicle. The speed corresponding to maximum dynamic


82 / 169


8. June 2010

increment is approximately v = 70 km/h. The dynamic increment tended to increase with speed up to v = 70 km/h but decreased for higher speeds. Similar observations are described in [Cantieni, 1993]. The air-suspended vehicle induced dynamic increments φ < 5% for all runs with the axle-hop plank. Additional information is presented in Annex D. 6.1.5 Coxs River Bridge

Two additional tests involving the air-suspended vehicle were conducted at Coxs River. The shock absorbers were removed from the rear axle of the trailer tri-axle group and the distance between axles in the tri-axle group were made unequal. These tests were designed to help understand the influence of suspension damping and axle spacing on the response of a bridge where dynamic coupling at axle-hop frequencies was anticipated. Unfortunately, due to the high bridge damping the anticipated level of dynamic coupling did not occur, thus diminishing the effectiveness of the tests. a)

Typical deflection time signals 1

Deflection [mm]

0 -1 -2 -3 D(2,3)

-4 -5 -6 -7 0

Fig. 39

1

2 Time [s]

3

4

Coxs River Bridge, mid-span deflection D(2,3), BS vehicle, v = 96 km/h.


83 / 169


8. June 2010

1

Deflection [mm]

0 -1 -2 -3 D(2,3)

-4 -5 -6 0

Fig. 40

1

2 Time [s]

3

4

Coxs River Bridge mid-span deflection D(2,3) for the BA vehicle, v = 77 km/h.

It can be seen from Fig. 39 and Fig. 40 that the dynamic component of the deflections for the bridge over Coxs River was small and decayed rapidly, thus indicating a high level of bridge damping. Despite the bridge being nominally simply supported for each of the three spans there is also some indication of continuity in the response. Observations made while proof load testing the same bridge confirmed that the bridge behaved non-compositely. The concrete slab deck slipped relative to the steel girders during the passage of the test truck thus damping the bridge vibrations. b)

Dynamic increments

The φ(v)-curves presented in the following figures are valid for the mid-span of span 2 on the bridge centerline, location D(2,3), where the largest deflections were recorded.


84 / 169


8. June 2010

40

Coxs River


35 30 25 20 15 10 5 0 -5 -100

Fig. 41

-50


50

100

Dynamic increments vs. speed for Coxs River Bridge, (negative speed: southbound, positive speed: northbound), deflection D(2,3). Solid squares: steel leaf suspension, open squares: air suspension; dotted line: air suspension, no shock absorbers.


40

Coxs River [Axle hop plank]

30 20 10 0 -10 0

Fig. 42

20


80

100

Dynamic increment vs. speed for Coxs River Bridge with axle hop plank, deflection D(2,3), northbound. Solid squares: steel-suspended vehicle, open squares: air-suspended vehicle; dashed line: air suspension with unequal spacing.

For the runs without the AHP, the maximum dynamic increment is φ = 28% for the BS and less than φ = 20% for the BA vehicle. The effect of removing the shock absorbers from the BA vehicle trailer’s rear axle was small, for speeds up to v = 60 km/h for this highly damped bridge.


85 / 169 8. June 2010

When the AHP was used to excite the suspensions, maximum dynamic increments of φ = 37% for the steel and approximately φ = 30% for the air-suspended vehicle were recorded. The peak dynamic increment occurred at speeds between v = 60 km/h and v = 80 km/h. Making the axle spacing between the air-suspended trailer’s axles unequal reduced the dynamic increment φ. For the limited number of test speeds, the maximum dynamic increment with unequal spacing and the AHP in place was φ = 18% compared with φ = 27% for the same suspension with the trailer’s axles equally spaced. The dynamic increments tended to increase with speed for both air and steel-suspended vehicles. However, the relationship between the dynamic increments and vehicle speed is different for each suspension. Additional information is presented in Annex E. 6.1.6 Camerons Creek Bridge

In the case of this bridge, information is not only presented for the "standard" BS steel-suspended and the BA air-suspended vehicles but in addition for the SA air-suspended vehicle. If nothing else is mentioned, as usual, "steel-suspended" refers to the BS vehicle and "air-suspended" to the BA vehicle. Fig. 43 to Fig. 45 present the time domain responses of the first and last span of Camerons Creek bridge for northbound traffic. The dynamic response of the first span is much larger than that of the fourth span. The approach to span 1 exhibits short wavelength roughness not evident on the bridge. The dynamic response of D(1,8) for vehicles traveling in the southbound direction is reasonably consistent with D(4,8) for northbound traffic.


86 / 169


a)

8. June 2010

Typical deflection time signals 0.5

Deflection [mm]

0.0 -0.5 D (4,8)

-1.0 D (1,8)

-1.5 -2.0 -2.5 -1

Fig. 43

0

1 Tim e [s]

2

3

4

Camerons Creek Bridge mid-span deflection for the BS vehicle, v = 80 km/h, northbound. 0.5

Deflection [mm]

0.0 -0.5 D (4,8)

-1.0 D (1,8)

-1.5 -2.0 -1

0

1

2

3

4

Tim e [s]

Fig. 44

Camerons Creek Bridge mid-span deflection for the BA vehicle, v = 62 km/h, northbound.


87 / 169


8. June 2010

0.5

Deflection [mm]

0.0 -0.5

D(4,8)

-1.0 D(1,8)

-1.5 -2.0 -2.5 -3.0 0

Fig. 45

b)

1

2

Time [s]

3

4

5

Camerons Creek Bridge mid-span deflection for the SA air-suspended vehicle, v = 63 km/h, northbound. Dynamic increments

Dynamic increment vs. speed graphs, φ(v)-curves, are presented for the mid-span deflections in the first D(1,8) and fourth (D(4,8) spans respectively in Fig. 46 for BS and BA and in Fig. 47 for the SA vehicle. A comparison with Fig. 48 illustrates the sensitivity to the addition of the axle-hop plank for the BS and BA vehicles.

150 Dynamic Increment [%]

125

Camerons Creek

100 75 50 25 0 -25 -100

Fig. 46

-50


50

100

Dynamic increments vs. speed for Camerons Creek Bridge, (negative speed: southbound, positive speed: northbound). Solid symbols: steel suspension, open symbols: air suspension; Solid lines D(4,8), dashed lines D(1,8).


88 / 169


8. June 2010

150

Camerons Creek


125 100 75 50 25 0 -25 -100

Fig. 47

-50


50

100

Dynamic increments vs. speed for Camerons Creek Bridge, (negative speed: southbound, positive speed: northbound); SA air suspended vehicle; Solid lines D(4,8), dashed lines D(1,8). 120

Camerons Creek [Axle hop plank]


100 80 60 40 20 0 -20 0

Fig. 48

20


80

100

Dynamic increment vs. speed for Camerons Creek Bridge, D(4,8) with axle hop plank (AHP), northbound. Solid symbols: BS vehicle, open symbols: BA vehicle. Note that the deflection D(1,8) is not influenced by the AHP in either direction. D(4,8) is not affected in the southbound direction.

The dynamic response to various suspension and vehicle configurations is summarized in Table 9.


89 / 169


Vehicle

BS

8. June 2010

BA

SA

Dynamic Increment (φ )

≤ 110% ≤ 50% for speeds less than 40 km/h

≤ 75% typically less than 25%


Critical speed range

70 to 100 km/h

50 to 70 km/h

50 to 80 km/h

Axle-hop plank

confuses the behavior/response. φ increases to 60%

substantially increases φ to 115% and broadens the range of speed sensitivity to 40 to 80 km/h

was not tested

Influence of road profile

The largest φ was induced in D(1,8) by vehicles traveling in both directions. Addition of the AHP confused the behavior. It therefore appears that the BS vehicle was more sensitive to the long wave lengths in the profile.

In the southbound direction the φ is less than 30%. This contrasts with the northbound direction where the maximum φ occurred immediately after the repair or the AHP. In these cases speed was critical. The BA vehicle was more sensitive to short wave lengths in the profile, especially those that induced axle-hop.

Below 40 km/h there was very little reaction to any profile. At 60 km/h, the vehicle was sensitive to the roughness of the northbound approach. In southbound directions, the φ increased in comparison with the BA vehicle. Defects inducing axle-hop are the most significant.

Table 9

Comparison of dynamic response of Camerons Creek bridge to test vehicles.

Additional information is presented in Annex F.


90 / 169


8. June 2010

6.1.7 Cromarty Creek Bridge a)

Typical deflection time signals 2

Deflection [mm]

0 -2 -4 -6 D(3,3)

-8

-10 -12 -14 0.0

Fig. 49

0.5

1.0 Time [s]

1.5

2.0

Cromarty Creek Bridge mid-span deflection for the BS vehicle, v = 91 km/h, northbound. 2

Deflection [mm]

0 -2 -4 -6

D(3,3)

-8

-10 -12 -14 0.0

Fig. 50

b)

0.5

1.0 Time [s]

1.5

2.0

Cromarty Creek Bridge mid-span deflection for the BA vehicle, v = 92 km/h, northbound. Dynamic increment

When the dynamic increment was analyzed, consideration was given to the fact that maximum deflections occur at D(3,3) for the northbound runs and at D(3,4) for the southbound runs, i.e. the deflections that are under the lane in which the truck was traveling (refer Figure G- 2, Annex G). The φ’s have not been plotted for D(3,4) for northbound runs or for D(3,3) for southbound runs as the dynamic component adds to a relatively small static deflection leading to misleading values for the dynamic increment [Cantieni, 1983].


91 / 169


8. June 2010

125

Cromarty Creek


100 75 50 25 0 -25

-100

Fig. 51

-50


50

100

Dynamic increments vs. speed for Cromarty Creek Bridge (positive speed: northbound - D(3,3), negative speed: southbound - D(3,4)), Solid symbols: steel suspension, open symbols: air suspension. 125

Cromarty Creek [Axle hop plank]


100 75 50 25 0 -25 0

Fig. 52

20


80

100

Dynamic increment vs. speed for Cromarty Creek Bridge with axle hop plank (northbound only). Solid symbols: steel-suspended vehicle, open symbols: air-suspended vehicle.

The bridge response to various suspension and vehicle configurations is summarized in Table 10.


92 / 169


Vehicle Dynamic Increment (φ)

BS


Critical speed 60 to 90 km/h range Axle-hop plank

8. June 2010

marginally increases the bridge response for speeds greater than 40 km/h

BA

≤ 50% ≤ 25% for speeds less than 60 km/h 60 to 90 km/h

marginally increases the bridge response for speeds between 40 km/h and 60 km/h

Table 10 Comparison of dynamic response of Cromarty Creek bridge due to test vehicles. There is a general trend for the dynamic increment to increase with speed. This is more marked for the BS vehicle than the BA vehicle. The pattern for each girder seems to be repeated with some girders experiencing a higher proportion of dynamic load than others. This is believed to be a consequence of variability inherent in the stiffness of the natural timber girders and the fact that the contact between the top of some girders and the underside of the bridge deck planks is limited. The maximum dynamic increment registered for the BS vehicle was φ = 109% at v = 91 km/h and less than 50% for the BA vehicle. From the limited amount of data collected with the AHP in place, the changes in dynamic increment are small. As an example, the maximum dynamic increment for the BS vehicle was φ = 111% at v = 67 km/h with the axle-hop plank in position. Additional information is presented in Annex G.


6.2

93 / 169 8. June 2010

Time Domain Analysis: Dynamic Wheel Load Signals

The dynamic wheel load signals measured by the NRC test vehicle in Switzerland and the Boral test vehicle's trailer in Australia provide an opportunity to understand the loads applied to bridges and to help explain the sensitivity to subtle changes in speed, road roughness, suspension characteristics and the like. Here, some selected representative results are presented only. Measured time domain signals are presented along with the corresponding dynamic load coefficients. The abbreviations used concerning the test parameters are described in Chapter 2. 6.2.1 The NRC Vehicle; Typical Wheel Load Time Signals

Fig. 53 and Fig. 54 show two typical wheel load time signals for steel and air suspension respectively. The wheel numbering starts with No’s. 02 and 04 for the steer axle and continues with No’s. 06 to 12 for the tractor drive axles and No’s. 14 to 20 for the trailer axles. The vehicle's axles are however numbered from 1 (steer axle) to 5 (last trailing axle). Signals are presented for the vehicle's lefthand side wheels only.


94 / 169


8. June 2010

"Approach"

"Bridge"

100 90 80

[kN]

70 60

W_02

50

W_06

40

W_10

30

W_14 20

W_18 10 0 0

2

4

6

8

10

12

14

16

18

20

Time [s] Fig. 53

Dynamic wheel load time signals recorded for the passage of the steel-suspended NRC vehicle over Deibüel Bridge at v = 48.1 km/h. For the sake of clarity, the static wheel loads have been offset from their true values. The y-axis scaling is however correct for the dynamic wheel load parts. The true static wheel loads are indicated in Fig. 55.


95 / 169


8. June 2010

"Bridge"

"Approach" 100 90 80

[kN]

70 60

W_02

50

W_06

40

W_10

30

W_14 20

W_18 10 0 0

2

4

6

8

10

12

14

16

Time [s] Fig. 54

Dynamic wheel load time signals recorded for the passage of the air-suspended NRC vehicle over Deibüel Bridge at v = 55.3 km/h. For the sake of clarity, the static wheel loads have been offset from their true values. The y-axis scaling is however correct for the dynamic wheel load parts. The true static wheel loads are indicated in Fig. 57.

Although the time axis' scaling is not identical in the two figures it becomes quite clear without any sophisticated frequency domain analysis that the dominant wheel load frequency is not the same in the two cases: It is significantly higher for the steel than for the air-suspended vehicle. 6.2.2 The NRC Vehicle; Dynamic Load Coefficients

The dynamic load coefficient (DLC) is a simple measure of the magnitude of the dynamic variation of axle load [Sweatman, 1983]. The DLC is the coefficient of variation of the dynamic axle load and has been determined in accordance with the following definition: DLC =

σ F

where σ is the standard deviation of the dynamic variation of the axle load and F is the mean value of the dynamic axle load which usually corresponds to the static axle load.


96 / 169


8. June 2010

70

Mean Wheel Load W 02

[kN]

60

W 06

50 W 10

40

W 14 W 18

30 0

20

40

60

80

10%

DLC

W 02

8%

W 06

6% W 10

4% W 14

2%

W 18

0% 0

10%

20

40

60

80

60

80

Mean DLC (Axle 2 - 4)

8% 6% 4% 2% 0% 0

20

40

Velocity [km/h] Fig. 55

DLC's for the steel-suspended NRC vehicle on the Deibüel Bridge Approach.


97 / 169


8. June 2010

70


60 [kN]

W 06

50

W 10 W 14

40

W 18

30 0

20

40

60

80

10%

DLC

W 02

8%

W 06

6% W 10

4%

W 14

2%

W 18

0% 0

20

10%

40

60

80

60

80


8% 6% 4% 2% 0% 0

20

40

Velocity [km/h]

Fig. 56

DLC's for the steel-suspended NRC vehicle on the Deibüel Bridge.


98 / 169


8. June 2010

70


60 [kN]

W 06

50

W 10 W 14

40

W 18

30 0

20

40

60

80

10%

DLC

W 02

8%

W 06

6% W 10

4%

W 14

2%

W 18

0% 0

10%

20

40

60

80

60

80


8% 6% 4% 2% 0% 0

20

40


DLC's for the air-suspended NRC vehicle on the Deibüel Bridge Approach.


99 / 169


8. June 2010

70


60 [kN]

W 06

50

W 10 W 14

40

W 18

30 0

20

40

60

80

10%

DLC

W 02

8%

W 06

6% W 10

4%

W 14

2%

W 18

0% 0

10%

20

40

60

80

60

80


8% 6% 4% 2% 0% 0

20

40


DLC's for the air-suspended NRC vehicle on the Deibüel Bridge.


100 / 169


8. June 2010

6.2.4 The Boral Vehicle; Typical Wheel Load Time Signals a)

Camerons Creek

Fig. 59 and Fig. 60 present sample dynamic wheel loads applied by the air and steel-suspended vehicle's trailer respectively as they crossed the bridge over Camerons Creek. The dynamic loads applied by the front, central and rear axles of the tri-axle group are presented along with the total dynamic load applied by the six wheels of the tri-axle group. Span 1

Repair

350

300

Span 2

Span 3

Span 4

Total dynamic force for the tri-axle group

Axle force (kN)

250

200

150 Dynamic force for each axle 100

50

0 0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Time (s)

Fig. 59

Trailer tri-axle group dynamic wheel load time signals for the passage of the BA vehicle over Camerons Creek Bridge at v = 59 km/h over the axle-hop plank (AHP).


101 / 169


Span 1

Repair

350

8. June 2010

Span 2

Span 3

Span 4

Total dynamic force for the tri-axle group 300

Axle force (kN)

250

200

150 Dynamic force for each axle 100

50

0 0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Time (s)

Fig. 60

Trailer tri-axle group dynamic wheel load time signals for the passage of the BS vehicle over Camerons Creek Bridge at v = 62 km/h over the axle-hop plank (AHP).

Fig. 59 and Fig. 60 demonstrate that the peak dynamic wheel loads are generally larger for the steel-suspended test vehicle and that this applies to both the individual axles and the tri-axle group. The body bounce modes of vibration for the steel suspension exhibit higher frequencies and larger amplitudes than for the air suspension for these particular waveforms (refer Fig. 80). This is consistent with observations at other speeds for soft, highly damped air suspensions versus conventional suspensions. b)

Coxs River

The testing at Camerons Creek had identified the importance of dynamic coupling between the axle-hop frequencies and the bridge. It had further demonstrated the importance of road roughness and suspension damping. At Coxs River the opportunity was taken to undertake two additional experiments: • investigate unequal spacing in the tri-axle group, • investigate the influence of suspension damping. Dynamic axle load signals for each axle (front, center and rear) of the tri-axle group, together with the total dynamic force of the tri-axle group signal are presented in Fig. 61 and Fig. 62.


102 / 169


8. June 2010

Each time signal has been offset to facilitate comparison. The average force for each axle is 67 kN. For the tri-axle group, the average force is 200 kN.

Dynamic force (1 div = 50 kN)

Total Tridem Force

50 kN

BRIDGE Rear Axle

AH P

Centre Axle

Front Axle

0

1

2

3

4

5

6

7

8

Time (s)

Fig. 61

Trailer tri-axle group dynamic wheel load time signals for the passage of the BS vehicle over Coxs River Bridge at v = 69 km/h northbound over the axle-hop plank.


Total Tridem Force

50 kN

BRIDGE Rear Axle

AH P

Centre Axle

Front Axle

0

1

2

3

4

5

6

7

8

Time (s)

Fig. 62

Trailer tri-axle group dynamic wheel load time signals for the passage of the BA vehicle over Coxs River Bridge at v = 80 km/h southbound over the axle-hop plank.


103 / 169 8. June 2010

Again the steel-suspended vehicle exhibits greater variation of the axle force than the airsuspended vehicle. For the steel-suspended vehicle, the low frequency variations (body bounce) of the axle force clearly dominate the high frequency variations (axle-hop). For the air-suspended vehicle, the amplitude of the high frequency component of the axle force becomes comparable to the amplitude of the low frequency component. This was especially true after the suspensions were excited by the axle-hop plank (refer Fig. 62). Note again that the axle-hop frequencies are more evident in the tri-axle group dynamic wheel force for the air suspension than the steel suspension. The influence of shock absorbers

Damping in steel suspensions is generated by Coulomb friction between the steel leaves whereas, for air suspensions, damping is achieved through shock absorbers. The damping capacity of a suspension is essential for reducing the variation of the dynamic wheel force. Vehicles fitted with air suspensions and shock absorbers in poor condition are considered likely to result in unacceptably large dynamic responses in short-span bridges that exhibit some short wavelength roughness. The shock absorbers were removed from the rear axle of the tri-axle group of the air-suspended vehicle in order to investigate the consequences of worn dampers on the dynamic wheel forces. Major changes were observed in the dynamic axle force (Fig. 63). Firstly, the amplitude of the high-frequency component of the dynamic force of the rear axle increased by a factor of 2 compared with the front and center axle forces. Secondly, these high frequency responses continued for an extended period because of the reduced damping. Note that the large variation of the dynamic force of the rear axle has little influence on the response of the center and front axle, further illustrating the independent behavior of the axles.


104 / 169


8. June 2010


Total Tridem Force

BRIDGE

50 kN

Rear Axle

Centre Axle

Front Axle

0

1

2

3

4

5

6

7

8

Time (s)

Fig. 63

Trailer tri-axle group dynamic wheel load time signals for the passage of the BA vehicle over Coxs River Bridge at v = 57 km/h northbound; no shock absorbers on the rear axle, no axle-hop plank.

The removal of shock absorbers from one axle had a significant influence on the total dynamic force of the tri-axle group. As expected, the high frequency component (axle-hop) of the dynamic force was influenced by the absence of shock absorbers more than the low frequency component. Note that the pavement surface was categorized as an A class surface, the road over the bridge and approaches had no significant defects and no axle hop plank was used for this truck configuration. However, the suspensions were strongly excited especially by localized defects at abutments and the joints in the deck.


105 / 169


8. June 2010

6.2.5 The Boral Vehicle; Dynamic Load Coefficients

Dynamic Load Coefficient

Graphs of dynamic load coefficient versus speed of the vehicles are presented in Fig. 64 to Fig. 70. The DLC’s are presented for the front (FA), center (CA) and rear (RA) axles of the triaxle group and the total tri-axle group force (TRIDEM). It can be observed that the DLC increases with speed for both steel and air-suspended vehicles. The maximum DLC for the steel-suspended vehicle is approximately double the maximum DLC for the air-suspended one. Thus the steel-suspended vehicle induced a dynamic force whose amplitude was approximately double of that induced by its air-suspended counterpart. 0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 0

10

20

30

40

50

60

Speed (km/h) DLC - FA DLC - RA

90

100

90

100

DLC - CA TRIDEM

0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 0

10

20

30

40

50

Speed (km/h) DLC - FA DLC - RA

Fig. 65

80

Coxs River Bridge; DLC, BS vehicle, northbound, no planks. Dynamic Load Coefficient

Fig. 64

70

60

70

80

DLC - CA TRIDEM

Coxs River Bridge; DLC, BA vehicle, northbound, no planks.


106 / 169


8. June 2010

The responses of the steel and air-suspended vehicles also differ in that the DLC of all axles is similar for the BA vehicle whereas the DLC of individual axles are different for the BS truck. This behavior is especially marked for speeds above v = 60 km/h. In this case, the front axle exhibits the largest DLC indicating the largest variation of the dynamic force. For speeds up to v = 60 km/h for the BS vehicle, the DLC corresponding to the total dynamic force of the tri-axle group is less than the DLC of independent axles. For the BA vehicle, the DLC of the tri-axle group force is less than the DLC of individual axles for all runs.

Dynamic Load Coefficient

For the air-suspended vehicle configuration without shock absorbers on the rear axle, at speeds above v = 40 km/h, independent axles no longer have the same DLC. The rear axle (without shock absorbers) exhibits the largest DLC (refer Fig. 66). Note that the maximum speed for this truck configuration is v = 60 km/h compared with a maximum speed of v = 100 km/h for the basic configurations BA and BS. If the trend is followed, at v = 100 km/h, the rear axle would reach a DLC of 0.3. This is comparable with the DLC for the steel suspension. 0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

Trend DLC-RA

0

10

20

30

40

50

60

70

80

90

100

Speed (km/h) DLC - FA

Fig. 66

DLC - CA

DLC - RA

TRIDEM

Coxs River Bridge; Dynamic load coefficient (DLC), BA vehicle, northbound; No shock absorbers on the rear axle, no planks.


107 / 169


40%

8. June 2010

DLC

30% FA CA RA TRIDEM

20% 10% 0% 0

20

40

60

80

100

Speed [km/h] Fig. 67

Camerons Creek Bridge; Dynamic load coefficient (DLC), BS vehicle, northbound, no planks.

40%

DLC 30% FA CA RA TRIDEM

20% 10% 0% 0

20

40

60

80

100


Camerons Creek Bridge; Dynamic load coefficient (DLC), BA vehicle, northbound, no planks. Note that axles were spaced at 1.54 m during this test rather than the 1.23 m used in this report.


108 / 169


8. June 2010

40%


20% 10% 0% 0

20

40

60

80

100


Cromarty Creek Bridge; Dynamic load coefficient (DLC), BS vehicle, northbound, no planks.

40%


20% 10% 0% 0

20

40

60

80

100


Cromarty Creek Bridge; Dynamic load coefficient (DLC), BA vehicle, northbound, no planks.


109 / 169


6.3

8. June 2010

Frequency Domain Analysis of Dynamic Wheel Load Signals

6.3.1 The NRC Vehicle; Frequency Spectra

Fig. 71 and Fig. 72 show two typical wheel load frequency spectra for steel and air-suspended axles or, in the present case, wheels, at a moderate vehicle speed respectively. The wheel numbering starts with No’s. 02 and 04 for the steer axle and continues with No’s. 06 to 12 for the tractor drive axles and No’s. 14 to 20 for the trailer axles. There are significant peaks to be detected in these wheel load frequency spectra at f = 3.0 Hz in the case of steel suspension and at f = 1.5 Hz for air suspension. These peaks appear in almost all spectra for every wheel and all velocities. In addition there are peaks at lower frequencies for both suspension systems. In many cases, the frequency of these peaks increases with increasing vehicle speed.

1.6E+06 1.2E+06 8.0E+05 4.0E+05 0.0E+00 0.0

1.0

2.0

2.9

3.9

4.9

5.9

6.8

7.8

8.8

9.8

Frequency [Hz]

Fig. 71

Example of a wheel load spectral density for the steel-suspended NRC vehicle on the Deibüel Bridge, wheel No. 14, v = 37.1 km/h.


110 / 169


8. June 2010

2.0E+06 1.6E+06 1.2E+06 8.0E+05 4.0E+05 0.0E+00 0.0

1.0

2.0

2.9

3.9

4.9

5.9

6.8

7.8

8.8

9.8

Frequency [Hz]

Fig. 72

Example of a wheel load spectral density for the air-suspended NRC vehicle on the Deibüel Bridge, wheel No. 14, v = 42.2 km/h.

At higher speeds, additional peaks can be seen at higher frequencies of up to f = 10 Hz for steel (Fig. 73) and up to f = 20 Hz for air suspension (Fig. 74).

3.2E+06 2.4E+06 1.6E+06 8.0E+05 0.0E+00 0.0

2.0

3.9

5.9

7.8

9.8 11.7 13.7 15.6 17.6 19.5

Frequency [Hz]

Fig. 73

Example of a wheel load spectral density for the steel-suspended NRC vehicle on the Deibüel Bridge, wheel No. 10, v = 57.3 km/h.


111 / 169


8. June 2010

4.0E+05 3.0E+05 2.0E+05 1.0E+05 0.0E+00 0.0

2.0

3.9

5.9

7.8

9.8 11.7 13.7 15.6 17.6 19.5

Frequency [Hz]

Fig. 74

Example of a wheel load spectral density for the air-suspended NRC vehicle on the Sort Bridge, wheel No. 06, v = 70.3 km/h.

For further analysis, the frequency of the peaks identified in the wheel load spectra were plotted against speed. Fig. 75 and Fig. 76 show examples for the corresponding scatter plots for steel and air suspension. In both diagrams there are data depending on speed and data independent of speed. The frequency of the speed dependent part of the data increases linearly with speed. The corresponding scaling factor of the dominant line slope is L = 3.3 m. This line appears in all diagrams for every wheel and both suspension systems. This length corresponds well with the circumference of the wheels of the test vehicle. It is hence most probable that this part of the vibration is due to non-homogeneous characteristics in the vehicle wheels. In addition there exist lines with different slopes, corresponding to other lengths. Some of these lengths correspond well with the distances between different pairs of axles of the NRC test vehicle. The distances between different axles of the test vehicle and the characteristic lengths corresponding to the slopes of the observed lines in the wheel load scatter plots are listed in Table 11. Lines with slopes corresponding to distances between axles of the vehicle may be explained according to the so-called "wheel base filtering" effect: For spatial frequencies of the test track pavement profile corresponding to the distance between two axles of the vehicle, the wheels are loaded in phase when the vehicle drives along the test track. This effect leads to an amplification of the corresponding vibration. The remaining lines, not corresponding to distances between axles, may be caused by other peaks in the profiles of the test tracks. Frequencies that do not depend on speed are found at f = 2.8 Hz for steel and at f = 1.5 Hz for air suspension. In addition, non-speed-dependent peaks can be seen at very low frequencies


112 / 169 8. June 2010

(f < 1.5 Hz) and, for the steel suspension, at high frequencies (f > 10 Hz). In the case of air suspension high-frequency peaks (f = 9...18 Hz) are found to be speed-dependent.

Fig. 75

Peak frequencies of wheel loads vs. speed for steel suspension on Deibüel Bridge, wheel No. 10. The respective highest peaks are connected through a solid line.

Fig. 76

Peak frequencies of wheel loads vs. speed for air suspension on Sort Bridge for wheel No. 10. The respective highest peaks are connected through a solid line.


113 / 169


Observed lengths in wheel load diagrams [m]

8. June 2010

Distance between axles of the test vehicle [m]

Axles of test vehicle

22.2 15.37

1-5

14.00

1-4

12.13

2-5

10.6

10.76

3-5

10.6

10.61

2-4

9.4

9.24

3-4

4.8

4.76

1-3

3.3

3.28

1-2

1.6

1.52

2-3

1.4

1.37

4-5

14.1

11.1

7.8 5.9 5.3

2.2

1.1 0.9 Table 11 Observed lengths in wheel load diagrams and distances between axles of the test vehicle.


114 / 169 8. June 2010

6.3.2 The NRC Vehicle; Mode Shapes

Twelve different types of operational mode shapes could be identified for the NRC test vehicle. They are presented graphically in Fig. 77. The five axles of the test vehicle are shown schematically and the mode shapes are indicated as relative amplitudes of the ten wheels: Types 1, 2 and 3 characterize rolling of the vehicle around its longitudinal axis. In Type 1 all axles roll in phase. In Type 2 there is a phase shift between axles 2/3 and 4/5 of about 90° and in Type 3 the phase shift is 180°: the two axle pairs vibrate in opposite direction. Type 4 is the pure heave body bounce vibration with all axles moving up and down in phase. Type 5 is the pure pitch motion of the car body with all axles moving in a plane rotating around a transverse axis lying approximately in the middle of the vehicle length. The wheels of the respective axles move in phase. Type 6 is a "bending" vibration, the first and the last two axles moving up and down in phase, axles 2 and 3 moving out-of-phase with respect to the first and to the last axles. Type 7 is similar to Type 6, but with a phase shift of about 90° between the twin axles 2/3 and 4/5 respectively. This phase angle shows that the mode of vibration is not real but complex. Complex modes occur with highly damped systems. The resulting motion looks similar to the motion of a swimming dolphin. Types 8, 9 and 10 are axle-hop vibrations, with the wheels of the respective axles moving inphase. Types 11 and 12 are out-of-phase axle-hop vibrations of the wheels of the twin axles 2/3 and 4/5 respectively with the axis of rotation parallel to the longitudinal axis of the vehicle.

Most vehicle operational mode shapes could be characterized with the help of these twelve mode shapes and combinations of them.


115 / 169


8. June 2010

Mode Shapes 1

2

3

Δφ=0°

7

8 Δφ=90°

9 Δφ=180°

4

10

5

11

6

Fig. 77

Δφ=90°

Δφ=180°

12

Mode shapes of the NRC vehicle operational vibrations (the steer axle shown on the left-hand side, the trailing axles on the right-hand side of the diagrams). Δφ = phase angle in the complex motion of the driving and trailing axles.


116 / 169


8. June 2010

Fig. 78 and Fig. 79 show scatter plots of the wheel load peak frequencies and the associated mode shapes (symbol numbering see Fig. 77) versus speed for steel and air suspensions respectively.

1

2

3

4

5

6

7

8

9

10

11

12

6

Frequency [Hz]

5

4

3

2

1

0

0

Fig. 78

10

20

30

40

Velocity [km/h]

50

60

70

Frequency of vibration modes vs. speed for steel suspension on the Deibüel Bridge.

1

2

3

4

5

6

7

8

9

10

11

12

6

Frequency [Hz]

5

4

3

2

1

0

0

Fig. 79

10

20

30

40

Velocity [km/h]

50

60

70

Frequency of vibration modes vs. speed for air suspension on the Deibüel Bridge.


117 / 169 8. June 2010

In both plots a linear relationship between frequency and speed can be recognized for some of the data. Again, the slope of the dominant line corresponds to L = 3.3 m. The vibration modes change with increasing frequency. • At low frequencies the dominant vibration modes are the rolling types 1, 2 and 3. • At medium frequencies (f = 1.5 Hz in the case of air, f = 2.8 Hz in the case of steel suspension) the pure heave body bounce type 4 is dominant. Type 5, 6 and 7 are found at frequencies in the range f = 3...5 Hz for steel suspension. These mode shapes rarely occur in the case of air suspension. In the same frequency range vibration mode 8 canbe identified, only for air suspension. This mode is caused by the fact that the first axle of the vehicle couldn't be changed to air suspension. • For high frequencies, the information discussed here is taken from processing parts of the scatter plots shown in Fig. 75 and Fig. 76. At f = 10 Hz, mode shapes 9 and 10 (axle-hop) are found to be independent of speed for steel suspension and linearly dependent on speed in the case of air suspension. The slope of the lines corresponds to 1.6 m and 1.1 m. This length is nearly the same as the distance between the axle pairs 2/3 and 4/5 (see Table 11). The mode shapes 11 and 12 are exclusively found at highest frequencies (f = 15...20 Hz) for air suspension. They couldn't be seen in the case of steel suspension.


118 / 169


8. June 2010

6.3.3 The Boral Vehicle; Frequency Spectra

200

1.5 Hz

Magnitude

150 100 50 0 0

25

3.3 Hz

300 Magnitude

5 10 15 20 Frequency of force function (Hz)

12.3 Hz

200

100

0 0

Fig. 80

5 10 15 20 Frequency of force function (Hz)

25

Camerons Creek Bridge; Power spectral densities for dynamic wheel load - drivers side center wheel. Top: vehicle BA, v = 47 km/h, axle-hop plank. Bottom: vehicle BS, v = 67 km/h, axle-hop plank.

The power spectral densities shown in Fig. 80 highlight the body bounce (f = 1.5 Hz and f = 3.3 Hz) and axle-hop modes (f > 8 Hz). Both of these modes can be observed in the time signals given in Fig. 59 and Fig. 60. The amplitudes of the axle-hop components of the dynamic wheel loads are much smaller than the amplitudes of the dynamic wheel forces associated with the body bounce. This is true for individual axles and for the tri-axle group as a whole.


119 / 169


6.4

8. June 2010

Vehicle/Bridge Interaction: Interpretation of the Wheel Load Signals

6.4.1 Basics on Vehicle/Bridge Interaction: Medium to Long-Span Bridges

Vehicle/bridge-interaction is a quite complex process being mainly dependent on the physical properties of the two elements of the system: the vehicle and the bridge. However, the qualitative effects of this interaction can be demonstrated using a very simple mechanical model: Both, the vehicle and bridge are modeled as Single-Degree-of-Freedom-(SDOF)systems. Considering the vehicle oscillating in its fundamental body bounce mode, the corresponding mass, stiffness and damping parameters can be determined. The bridge is modeled while oscillating in its fundamental natural mode which also defines the modal mass, stiffness and damping. One of the main simplifications associated with this model relates to the fact that the vehicle is modeled as being stationary whilst it actually moves along the bridge. It is hence neglected that vehicle/bridge interaction is a transient process of often only a short duration. However, the model is quite appropriate for considering the situation when the vehicle is close to the bridge's mid-span. Real interaction occurs when the two SDOF-systems temporarily couple into a Two-Degreeof-Freedom-system (Fig. 81). This is only possible if the natural frequencies of the two SDOFsystems involved are close to each other or identical. In this case, the two natural frequencies inherent in the coupled system are different from the (identical) natural frequencies of the two SDOF-systems involved [Cantieni, 1992]. In other words, there will be a frequency shift in both, the bridge response and the wheel load spectra when interaction occurs.

Mveh kveh

cveh Mbr

kbr

Fig. 81

cbr

"Real" vehicle/bridge interaction reduced to a Two-Degree-of-Freedom-Model.


120 / 169 8. June 2010

If the frequencies involved are practically the same the natural frequencies of the coupled system vehicle/bridge depend firstly on the mass ratio and secondly on the damping coefficients of the two systems. Concerning the mass ratio, significant system coupling has been observed in the case where the modal bridge mass was approximately 20 times larger than the vehicle mass. In other words, "real" vehicle/bridge interaction occurs already if the vehicle mass is as small as 5% of the bridge modal mass. This is usually the case for medium to long-span bridges. The first and relatively easy step in trying to determine any "real" bridge/vehicle-interaction effect is hence analysis of either the bridge or dynamic wheel load signals separately and to look for possible frequency shifts. 6.4.2 Basics on Vehicle/Bridge Interaction: Short-Span Bridges

Vehicle/bridge interaction is a quite different problem here. The vehicle is usually longer than the bridge span, the vehicle/bridge mass ratio is much smaller than for the bridges discussed above and, last but not least, the interaction involves vehicle axle-hop instead of body bounce vibrations. The fact that interaction happens at the vehicle axle-hop frequencies changes the interaction process completely. With interaction in the body bounce frequency range the vehicle's gross mass is an active partner in the game because it moves relative to the bridge mass and the modeling SDOF + SDOF = 2DOF is in accordance to the laws of physics. Axle-hop means: The vehicle body does not move and the axle is vibrating between the vehicle body and the "rigid" surface. Here, the vehicle sprung mass is a passive member of the game and the above mentioned equation is valid with considering the vehicle unsprung mass only. It has not yet been possible to experimentally prove the existence of "real" vehicle/bridge interaction for this case. It seems however much more reasonable to change the modeling from "system coupling" SDOF + 2DOF = 2DOF to the model of a system (the bridge) being forced to vibrate by external forces (the dynamic wheel loads) without taking the vehicle masses involved into account. In this case, basic physics tells us that the dynamic response of the system forced to vibrate depends on its damping capacity only if the forcing frequency is equal to the natural frequency of the forced system. This is illustrated in the next paragraph.


121 / 169


8. June 2010

F(t) Mbr kbr

Fig. 82

cbr

The vehicle reduced to a forcing function acting on the bridge as an SDOF-system.

6.4.3 Swiss Tests

In the context of analyzing the measured dynamic wheel load signals, attempts were undertaken to check for frequency shifts when comparing the wheel loads measured while the vehicle was on the approach stretch with the ones measured while the vehicle crossed the bridge. As the Swiss tests were centered on medium to long-span bridges the chances to unveil "real" vehicle/bridge processes were quite real. As a matter of fact, close analysis of the measured wheel load signals revealed that a real vehicle/bridge interaction effect can be observed, when the steel-suspended NRC test vehicle (f = 2.8 Hz) crosses the Deibüel bridge (f = 3.03 Hz). Fig. 83 shows the frequency versus speed plot of the pure heave body bounce vibration (Type 4, Fig. 77) for the vehicle equipped with steel suspension on the Deibüel Bridge. Data for the two sectors a) "vehicle on the approach stretch" and b) "vehicle on the bridge" are plotted separately in this figure. The histogram of the corresponding frequencies is given in Fig. 84. The average frequency of vibration is reduced by Δf = 0.13 Hz when the vehicle is on the bridge.


122 / 169


8. June 2010

Before Bridge

On the Bridge

Frequency [Hz]

4 3.5 3 2.5 2 1.5 1 0

20

40

60

80

Velocity [km/h]

Fig. 83

Frequency vs. speed of the body bounce vibration (Mode shape Type 4) on the Deibüel Bridge for steel suspension. Before Bridge

On the Bridge

16 12 8 4 0 2.0 2.1 2.2 2.3 2.4 2.6 2.7 2.8 2.9 3.1 3.2 3.3

Frequency [Hz] Fig. 84

Histogram of the frequencies of the body bounce vibration on the Deibüel Bridge for steel suspension.

Fig. 85 and Fig. 86 show analogous data for the Deibüel Bridge and the vehicle equipped with air suspension. In this case the body bounce frequency of the vehicle is much lower than the lowest frequency of the bridge. No frequency shift can be seen when the vehicle crosses the bridge. This fits quite well with the expectations because vehicle/bridge interaction cannot be expected if the fundamental natural bridge frequency is not close to the dominating wheel load frequency.


123 / 169


8. June 2010

Before Bridge

On the Bridge

Frequency [Hz]

3 2.5 2 1.5 1 0.5 0 0

20

40

60

80

Velocity [km/h]

Fig. 85

Frequency vs. speed of the body bounce vibration (Mode shape Type 4) on the Deibüel Bridge for air suspension. Before Bridge

On the Bridge

20 16 12 8 4 0 0.9 1.0 1.1 1.2 1.3 1.5 1.6 1.7 1.8 2.0 2.1 2.2


Histogram of the frequencies of the body bounce vibration on the Deibüel Bridge for air suspension.

No effect is seen for the Föss bridge with neither of the two suspension systems. For both suspension systems the frequency of the bridge is significantly higher than the body bounce frequency of the vehicle. Fig. 88 and Fig. 89 show the histograms of the corresponding frequencies of the vehicle for the two sectors of the test track - before the bridge and on the


124 / 169


8. June 2010

bridge - and the two suspension systems. Again, no vehicle/bridge interaction is to be expected. Before Bridge

On the Bridge

16 12 8 4 0 2.0 2.1 2.2 2.3 2.4 2.6 2.7 2.8 2.9 3.1 3.2 3.3


Histogram of the frequencies of the body bounce vibration on the Föss Bridge for steel suspension. Before Bridge

On the Bridge

8 6 4 2 0 0.9 1.0 1.1 1.2 1.3 1.5 1.6 1.7 1.8 2.0 2.1 2.2


Histogram of the frequencies of the body bounce vibration on the Föss Bridge for air suspension.

For the Sort Bridge the situation is not very clear. Altough the first bridge natural frequency, f1 = 1.62 Hz, corresponds with the fundamental frequency of the air-suspended test vehicle, there are higher bridge natural frequencies, especially f3 = 2.98 Hz, which are in the range of the fundamental frequency of the steel-suspended vehicle. The data show a slight frequency shift in the wheel load spectra when the vehicle crosses the bridge compared to the frequency


125 / 169


8. June 2010

when the vehicle is on the approach stretch for both suspension systems (Fig. 89, Fig. 90). The frequency resolution Δf = 0.122 Hz, given by the length of the data set in the time domain, the sampling rate and the maximum number of data used by the software, is however too poor to reliably separate these peaks. Further investigations will be necessary to confirm that this shift really originates from a vehicle/bridge interaction. Before Bridge

On the Bridge

24 20 16 12 8 4 0 2.0 2.1 2.2 2.3 2.4 2.6 2.7 2.8 2.9 3.1 3.2 3.3


Histogram of the wheel load frequencies of the steel-suspended NRC vehicle body bounce vibration on the Sort Bridge. Before Bridge

On the Bridge

24 20 16 12 8 4 0 0.9 1.0 1.1 1.2 1.3 1.5 1.6 1.7 1.8 2.0 2.1 2.2


Histogram of the wheel load frequencies of the air-suspended NRC vehicle body bounce vibration on the Sort Bridge.


126 / 169 8. June 2010

6.4.4 Australian Tests

As may have become clear from the preceding remarks, the bridge/vehicle-interaction problem reduces to a "forced-vibration"-problem in the case of short-span bridges. No attempts have hence be undertaken to distinguish between wheel load spectra for the cases of the vehicle traveling on the approach stretches and on the bridge itself. 6.5

Vehicle/Bridge Interaction: Combining Bridge Response and Wheel Load Spectra

6.5.1 Tests in Switzerland: General

To reliably interpret the vehicle/bridge interaction phenomenon it is necessary to process dynamic wheel load and bridge response signals having been acquired simultaneously. As a result of the technical problems described in Paragraph 3.4.1, a very limited amount of simultaneously acquired vehicle and bridge signals is available now. Data processing can be performed by calculating cross-power-spectra from the respective time signals (refer Paragraph 5.3.1). Whereas the resulting amplitude relationships are of qualitative interest only the phase information is crucial inasmuch it allows to determine whether or not the motions of vehicle and bridge were in phase. In order to get the correct phase information the wheel load signal had to be inverted to take into account, that a high (positive) wheel load caused a negative deflection of the bridge. To ensure comparability of the results the analysis was performed for a fixed time interval of 10 seconds (exact: 10.24 s = 1'024 data points sampled at s = 100 Hz) centered at the maximum deflection of the bridge. This constant interaction length resulted in a constant frequency resolution of Δf = 1/T = 0.098 Hz for every run analyzed. On the other hand this fixed time interval corresponded to a different track length when the vehicle ran over a bridge with different speed. For this fixed time interval the power spectra of the wheel load and bridge response signals and the cross-power spectra between these signals were calculated and the corresponding phase angle determined. To reduce the number of possible combinations between the ten separately measured wheel load signals and the six bridge response signals for the Deibüel Bridge and five response signals for the bridges Sort and Föss the analysis was restricted to the cross-power spectra between the wheel load signals No. 06 (left wheel, first tractor drive axle, axle No. 2) and No. 14 (left wheel, first trailing axle, axle No. 4) and the signal of the displacement transducer fixed at the mid-point of the largest span of the bridges concerned.


127 / 169 8. June 2010

6.5.2 Deibüel Bridge, Steel-Suspended NRC Test vehicle

In the case of a Two-Degree-of-Freedom-System model of the coupled system vehicle and bridge (Fig. 81) with the uncoupled natural frequencies f = 2.8 Hz (vehicle) and f = 3.03 Hz (bridge) the natural frequencies of the coupled (undamped) system are f1 = 2.49 Hz and f2 = 3.41 Hz. Fig. 91 shows a) the wheel load signal of the test vehicle wheel No. 06 and the induced bridge response WG 21, b) the calculated power spectral densities for these signals, c) the cross-power spectrum magnitude, and d) and the phase lag between the two signals. In the diagram (b) the power spectral density of the bridge response is multiplied by a factor of 10. The vehicle speed was v = 32.1 km/h. For this run a significant peak in the power spectra and in the cross-power spectra can clearly be identified at f = 2.54 Hz. This frequency is significantly lower than the uncoupled natural frequencies of both vehicle and bridge. The phase angle at f = 2.54 Hz is α = 21 degrees, indicating that the motions of the vehicle and the bridge are in phase. At f = 3 Hz the phase lag jumps to α ≈ 150 degrees, indicating an out-of-phase vibration at higher frequencies. However, in the power spectra no significant peak is observed in the region of the second, out-of-phase natural frequency of the coupled system, f2 = 3.41 Hz. This is however no surprise, because the same was true for similar tests performed earlier [Cantieni, 1992]. 6.5.3 Deibüel Bridge, Air-Suspended NRC Test Vehicle

The natural frequency of the air-suspended vehicle, f = 1.6 Hz, is much lower than the fundamental natural frequency of the Deibüel bridge. As can be seen in Fig. 92, the dynamic bridge response is comparatively small and the same hence applies to the reliability of the analysis. Only very small peaks can be observed in the power spectrum of the bridge response (multiplied by a factor 100) at f = 1.37 Hz, 2.93 Hz and 4.10 Hz. There is no clear correlation between the power spectra of the wheel load signal with a peak at f = 1.76 Hz and the spectra of the bridge response. The cross-power spectra and the phase angle do not contribute much to clarify the interaction process between vehicle and bridge. Such a process is however not expected under the given circumstances. 6.5.4 Sort Bridge, Steel-Suspended NRC Test Vehicle

In this configuration, the natural frequency of the steel-suspended vehicle is higher than the fundamental natural frequency of the bridge. Again, the dynamic bridge response is small. A significant dynamic response is observed at high test vehicle velocities exceeding v = 50 km/h only. In Fig. 93 the power spectrum of the bridge response signal shows a peak at f = 1.56 Hz and a second small peak at f = 3.03 Hz. While there is no corresponding peak in the power spectrum of the wheel load in the range of the bridge fundamental frequency in vertical


128 / 169 8. June 2010

bending, f2 = 1.56 Hz, there is a large peak to be observed at f = 3.03 Hz. This frequency corresponds well with the frequency of the third bridge bending mode, f7 = 2.98 Hz. The phase angle is α = 164 degrees. Animation of the bridge vibration clearly showed that it is actually the third bridge bending mode which is responsible for this peak. Hence again: No coupling between the SDOF-systems of vehicle and bridge occurs if the dominant frequencies are not at least similar. 6.5.5 Sort Bridge, Air-Suspended NRC Test Vehicle

In the case of the air-suspended vehicle on the Sort Bridge the fundamental frequencies of the two sub-systems are similar, hence a frequency coupling can be expected. The natural frequencies of the coupled system can be calculated to be f1 = 1.494 Hz (in-phase motion of the two masses involved) and f2 = 1.735 Hz (out-of-phase motion) if the natural frequencies of the vehicle and bridge are f = 1.6 Hz and f = 1.62 Hz respectively. The power spectra in Fig. 94 show two peaks at f = 1.27 Hz and f = 1.76 Hz for the wheel load signal and at f = 1.37 Hz and f = 1.66 Hz for the bridge response signal. The phase lags are 39 and 148 degrees respectively, indicating an in-phase vibration of the coupled vehicle/bridge system in the frequency range f = 1.3 Hz and an out-of-phase vibration in the range f = 1.7 Hz. This behavior agrees very well with the 2DOF model of the coupled system. This is quite remarkable because up to now it has been possible to identify the in-phase vibration [Cantie-ni, 1992] but not the out-of-phase vibration of a coupled vehicle/bridge system. Fig. 94 gives the results of processing the signals measured for a vehicle speed v = 40.2 km/h. The characteristic features of the 2DOF-model coupling could however not be identified for higher vehicle speeds. Here, additional data processing with using longer time windows would be necessary.


129 / 169


8. June 2010

Wheel load (inv) Bridge response 7

9

11

13

15

17

Time [s]

(a) 150000

2.54

PSD Wheel PSD Bridge

PSD

100000

50000

0 1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

3.0

3.5

4.0

4.5

5.0

3.0

3.5

4.0

4.5

5.0

Frequency [Hz]

(b) 30000

2.54 XSD

20000 10000 0 1.0

(c)

1.5

2.0

2.5

Phase Angle

180

90 21 0

(d) Fig. 91

1.0

1.5

2.0

2.5

Passage of the steel-suspended NRC vehicle over the Deibüel Bridge at v = 32.1 km/h: (a) Time signals of the test vehicle's wheel 06 and the induced bridge response WG 21, (b) power spectral densities for the wheel load and bridge signals, (c) cross-power spectrum magnitude and (d) phase angle between the two signals.


130 / 169


8. June 2010


7

9

11

13

15

Time [s]

(a) 20000 1.76


PSD

15000 1.37

10000

4.10

2.93 5000

0 1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

4.5

5.0

4.5

5.0

Frequency [Hz]

(b) 1000 1.37 XSD

1.76 500 4.20 0 1.0

(c)

1.5

2.0

2.5

3.0

3.5

4.0

Phase Angle

180 99

76

90 33 0

(d) Fig. 92

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Passage of the air-suspended NRC vehicle over the Deibüel Bridge at v = 49.4 km/h: (a) Time signals of the test vehicle's wheel 06 and the induced bridge response WG 21, (b) power spectral densities for the wheel load and bridge signals, (c) crosspower spectrum magnitude and (d) phase angle between the two signals.


131 / 169


8. June 2010


7

9

11

13

15

Time [s]

(a) 25000

3.03


PSD

20000 15000

1.56

3.32

2.15

1.95

10000 5000 0 1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

3.5

4.0

4.5

5.0

4.0

4.5

5.0

Frequency [Hz]

(b) 1500 1.56

3.03

XSD

1000 500 0 1.0

(c)

Phase Angle

180

1.5

2.0

2.5

163

3.0

164

90

0

(d) Fig. 93

1.0

1.5

2.0

2.5

3.0

3.5

Passage of the steel-suspended NRC vehicle over the Sort Bridge at v = 70.0 km/h: (a) Time signals of the test vehicle's wheel 06 and the induced bridge response WG 32, (b) power spectral densities for the wheel load and bridge signals, (c) crosspower spectrum magnitude and (d) phase angle between the two signals.


132 / 169


8. June 2010


13

15

17

19

21

Time [s]

(a) 3000

PSD Wheel

1.27

PSD Bridge 1.37 1.66

PSD

2000

1.76 3.42 1000

2.93

0 1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

Frequency [Hz]

(b) 750

1.27 1.76

XSD

500 250 0 1.0

(c)

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

2.5

3.0

3.5

4.0

4.5

5.0

Phase Angle

180 148 90 67 39 0

(d) Fig. 94

1.0

1.5

2.0

Passage of the air-suspended NRC vehicle over the Sort Bridge at v = 40.2 km/h: (a) Time signals of the test vehicle's wheel 06 and the induced bridge response WG 32, (b) power spectral densities for the wheel load and bridge signals, (c) cross-power spectrum magnitude and (d) phase angle between the two signals.


133 / 169 8. June 2010

6.5.6 Tests in Australia

Here, the limits for electronically triggered simultaneous data acquisition of bridge and vehicle signals were clear from the beginning. No radio-telemetry link could be installed. However, similar attempts to "mechanically" trigger both data acquisition systems were undertaken. Unlike the Swiss tests, corresponding data processing could not be achieved in due time for the production of this report. It is also noted that the time the instrumented tri-axle group is on the instrumented span is small: e.g. Δt = 0.5 s at v = 100 km/h for the test vehicle crossing Camerons Creek Bridge. This limits the resolution associated with cross-power-spectra analyses. For the limited number of cross-power-spectra’s calculated for Camerons Creek Bridge there was no discernible change in the frequencies from the fundamental frequency of the bridge.


134 / 169


7.

DISCUSSION

7.1

Dynamic Bridge Response

8. June 2010

The maximum dynamic increment for bridge vertical deflection for each of the bridge/vehicle parameter configurations tested is given in Table 12 and Table 13 for the cases "without" and "with axle-hop plank" respectively together with other significant parameters. Bridge

1st freq. [Hz]

Damping Pavement Frequency Matching φmax steel [%] [%] Condition

φmax air [%]

Sort

1.62

1.0

A

body-bounce, air

10

26

Deibüel

3.01

0.8

A

body-bounce, steel

21

5

Föss

4.44

1.6

B

none

15

12

Lawsons

5.1

1.0

A

none

6

3

Coxs

10.2

4.5

A

axle-hop

28

18

Camerons

11.3

1.5

B…C

axle-hop (Boral)

105

75

axle-hop (Shell)

---

137

axle-hop

109

50

Cromarty

9.5

2.6

B…C

Table 12 Passages without axle-hop plank: Maximum dynamic increments φmax measured for bridge vertical displacement in a significant location.

Bridge

1st freq. [Hz]

Damping Pavement Frequency Matching φmax steel [%] Condition [%]

φmax air [%]

Lawsons

5.1

1.0

A

none

13

1

Coxs

10.2

4.5

A

axle-hop

37

27

Camerons

11.3

1.5

B…C

axle-hop (Boral)

60

115

Cromarty

9.5

2.6

B…C

axle-hop

111

40

Table 13 Passages with axle-hop plank: Maximum dynamic increments φmax measured for bridge vertical displacement in a significant location.


135 / 169 8. June 2010

The response of each of the bridges tested, considering vertical deflections measured for significant bridge locations only, is discussed below: 7.1.1 Sort Bridge (CH)

The frequencies of the air-suspended vehicle body-bounce mode and the fundamental natural bridge mode were similar (f ≈ 1.6 Hz). The air-suspended vehicle induced maximum dynamic increments (φ = 25%) significantly larger than those induced by the steel suspension (φ = 10%). However, for a vehicle speed of v < 40 km/h, the dynamic increments are very small (φ < 5%) and no influence of the vehicle suspension system on the bridge response can be detected. As a consequence of the smooth pavement surface the largest responses are smaller than those given in the next paragraph for a rough pavement and frequency matching at f ≈ 3 Hz. Simulation studies are required to investigate the effect of a rougher surface in the case of frequency matching at f ≈ 1.6 Hz. 7.1.2 Deibüel Bridge (CH)

The frequencies of the steel-suspended vehicle body-bounce mode and the fundamental natural bridge mode were similar (f ≈ 3 Hz). The largest dynamic bridge response (φ = 21%) was induced by the steel-suspended vehicle with only very limited dynamic effects induced by the air-suspended vehicle (φ ≤ 5%). This is true for almost all vehicle speeds. Comparison with maximum values achievable for rough but still acceptable pavement surface conditions can be based on earlier tests performed on the Deibüel Bridge ([Cantieni, 1987, 1988], Paragraph 7.2, DIVINE testing versus Swiss Tests, Fig. 97). These tests yielded that dynamic increments φ ≈ 80% for rigid trucks and φ ≈ 50% for articulated vehicles (both steel-suspended) are reached for a medium pavement quality with unevenness parameters close to the limits as prescribed in the Swiss Code [SIA, 1991]). 7.1.3 Föss Bridge (CH)

The bridge fundamental natural frequency, f = 4.44 Hz, was higher than the body-bounce frequencies of both the steel and air-suspended vehicles but lower than their axle-hop frequencies. As expected, no frequency-matching and no quasi-resonance phenomena were hence to be observed and the maximum dynamic response was similar and small (φ < 15%) for both suspensions. The dynamic increments tended to slightly increase with vehicle speed. This reflects the fact that the vehicle dynamic wheel loads were just transmitted and neither amplified nor attenuated. As with the Sort Bridge, the situation for rough pavement conditions is not known.


136 / 169 8. June 2010

7.1.4 Lawsons Creek Bridge (AUS)

As with Föss, no frequency-matching occurred and the maximum dynamic response was consequently very small (φ ≤ 6%). This could be expected given the bridge fundamental natural frequency, f = 5.1 Hz, lying between the vehicle body-bounce and axle-hop frequencies. The maximum dynamic increment of the bridge due to the air-suspended vehicle is very similar to that for the steel-suspended vehicle (φ = 3%, φ = 6%). This most probably reflects the fact of the dynamic wheel loads being of the respective magnitude. The introduction of the axle-hop plank (300 mm x 25 mm) saw larger dynamic wheel loads and bridge responses for the steel-suspended (φ < 13%) than for the air-suspended vehicle (φ < 1%). The addition of the AHP at the abutment only had limited effects on the bridge dynamic response. This is as expected because the bridge attenuates the effects of dynamic wheel loads in the axle-hop frequency range (Fig. 99) and the vehicle suspension damps the axle-hop vibrations before the vehicle has moved onto the load sensitive portion of the bridge. 7.1.5 Coxs River Bridge (AUS)

The maximum dynamic increments for the Coxs River Bridge were moderate and larger for the steel-suspended than for the air-suspended vehicle (BS: φ = 28%, BA: φ = 18%). This was surprising as this bridge has a natural frequency, f = 10.2 Hz, that was considered likely to amplify the vehicle axle-hop vibrations similarly to the Camerons Creek Bridge (refer next Paragraph) which was tested just prior to the Coxs River Bridge. There are two factors which may be responsible for the observed differences in dynamic bridge response behavior: • The damping coefficient is higher by a factor of 4.5 for Coxs River Bridge compared to Camerons Creek Bridge. The concrete slab and the steel girders of Coxs River Bridge act non-compositely. This was confirmed by the observation that the slab slipped over the flange during the passage of test vehicles. As a consequence, the steel girder and the concrete slab acted as two leaves of a leaf suspension with considerable inter-leaf Coulomb friction. As a result, the bridge was highly damped which limited its dynamic response to the components of the dynamic wheel forces with frequencies similar to the bridge. • The pavement profile of Coxs River Bridge shows significantly less axle-hop vibration inducing short-wave unevenness than the Camerons Creek Bridge profile (see the profiles given in Annexes E and F respectively).


137 / 169 8. June 2010

Unequal axle spacing of the air suspended tri-axle group increased the dynamic increments significantly (φ = 30%) whereas the removal of the shock absorbers on the rear axle of the air suspended tri-axle group resulted in little change. The addition of the axle-hop plank did not increase the dynamic increments significantly. This increase was stronger for the air-suspended than for the steel-suspended vehicle, but still, the response to the steel-suspended was larger than to the air-suspended vehicle (BS: φ = 37%, BA: φ = 27%). It was concluded that due to the high damping, the bridge was not significantly amplifying the effects of axle-hop vibrations but simply transmitting the body-bounce variations in wheel load. Consequently the response to the air-suspended vehicle was smaller than to the steelsuspended vehicle. 7.1.6 Camerons Creek Bridge (AUS) a)

General

As with Coxs River Bridge the natural frequency of Camerons Creek Bridge, f = 9.1 Hz, is in the vehicle axle-hop vibration's range. In contrast to Coxs River Bridge, the only average levels of damping evident in Camerons Creek Bridge allowed considerable amplification of the axle-hop components of the dynamic wheel loads. Deflection measurement points D(1,8) and D(4,8) were located on the bridge centerline in the middle of bridge spans 1 and 4, the southern and northern bridge end spans respectively (the two middle spans were not instrumented, refer Annex F). Given a smooth road one would expect their response to be similar for both northbound and southbound traffic. This was not the case. The differences in behavior between span 1 and 4 can be attributed to features in the road profile and the introduction of the axle-hop plank. The road profile included significant axle-hop vibration inducing short-wave unevenness in the southern approach in form of a cold-mix repair (refer Annex F). The reinforced concrete surface of the bridge itself was quite smooth and thus free of axle-hop inducing features. As a result, the dynamic bridge response of span 1 (i.e. the closest to the southern abutment) was much larger for the northbound driving direction compared with the southbound driving direction where the truck had traversed three smooth spans before crossing span 1. Further insights into the influence of pavement profile were made possible through the fitting of a 300 x 25 mm axle-hop plank over pier 3 and thus immediately before span 4 for northbound traffic. When the test vehicles traveled in the northbound direction significant axle-hop was


138 / 169 8. June 2010

induced immediately prior to span 1 by the cold-mix repair and in span 4 when the axle-hop plank was in place. When the test vehicles traveled in the southbound direction neither the cold-mix repair nor the axle-hop plank had any significant influence on the instrumented spans (i.e. 1 & 4). The response of these spans is discussed with and without the AHP below. b)

Without axle-hop plank

For the BA air-suspended vehicle, the dynamic increment induced in bridge span 1, D(1,8), is similar for the two traveling directions except for a narrow range of speeds for vehicles traveling north. Generally, the dynamic increment is less than φ ≈ 25%. However for speeds around v = 60 km/h for the test vehicle traveling north, the dynamic increment jumps to almost φ = 75%. This is related to the bridge response to the axle-hop vibrations induced by the coldmix repair located immediately before bridge span 1, the first to be crossed for northbound traffic. By the time the vehicle has reached bridge span 4 most of the axle-hop has been damped out and D(4,8) registers again a dynamic increment φ ≈ 25%. The reverse did not occur for the southbound traffic as the northern approach to bridge span 4 (the first to be crossed now) does not exhibit the pavement unevenness features required to induce axle-hop despite a relatively large depression (i.e. long wavelength roughness). The dynamic increment for D(4,8) is never larger than φ ≈ 5% for southbound traffic. For the SA air-suspended vehicle, similar effects with even more pronounced results are to be observed. For northbound traffic, the dynamic increments are smaller than φ ≈ 25% with the exception of the sharp peak at v ≈ 60 km/h where the largest dynamic increment, φ ≈ 137%, was induced in D(1,8) for the northbound direction. It has to be mentioned, that in this case, the bridge deflection cycles went from positive to negative a number of times thus involving approximately 10 fatigue cycles while the vehicle was still on the bridge. Again, for bridge span 4, D(4,8), the dynamic increment was not larger than φ ≈ 25%. For southbound traveling, the dynamic increments were similar for the bridge spans 4 and 1, φ ≈ 60...70%. For bridge span 4, D(4,8), this corresponds to the maximum dynamic increment induced by the three vehicle configurations traveling in either direction. For the BS steel-suspended vehicle, similar effects as for the BA vehicle, but now at higher φlevels are to be observed. For northbound traffic, the difference between the "standard" dynamic increment, now φ < 50%, and the cold-mix repair plank induced peak in the dynamic increment for D(1,8), φ ≈ 105%, was similar to the respective result for the BA vehicle, but the critical speed was now v ≈ 80 km/h. For southbound traveling, the maximum dynamic increments were φ ≈ 30% for bridge span 4, D(4,8), and φ ≈ 70% for bridge span 1.


139 / 169


8. June 2010

An intermediate summary yields that BA is the least aggressive vehicle in all cases but that for rough approach, SA produces higher dynamic increments than BS, whereas for smooth approach, the contrary is true: • northbound, first span to be crossed (after the cold-mix repair), D(1,8): No. 1: SA vehicle (φ ≈ 137%),

No. 2: BS vehicle (φ ≈ 105%),

No. 3 BA vehicle (φ ≈ 75%)

• southbound, last span to be crossed (after three smooth spans), D(1,8): No. 1: BS vehicle (φ ≈ 70%),

No. 2: SA vehicle (φ ≈ 60%),


• southbound, first span to be crossed (no cold-mix repair), D(4,8): No. 1: SA vehicle (φ ≈ 60%),

No. 2: BS vehicle (φ ≈ 35%),


• northbound, last span to be crossed (after three smooth spans), D(4,8): No. 1: BS vehicle (φ ≈ 50%),

No. 2: SA vehicle (φ ≈ 25%),


It can be concluded that the SA vehicle is more sensitive to short-wave pavement unevenness than the other two vehicles. Generally, it has to be mentioned that the experiments performed on the Camerons Creek Bridge pavement surface as actually present on site yielded dynamic increments which are close to the limit given in the Swiss Code [SIA, 1991] for the "Single Vehicle Load Model", φ ≈ 80%, for southbound traffic and clearly above this limit for northbound traveling of two of the three test vehicles (φmax ≈ 137%, φmax ≈ 105%). Under the AUSTROADS Bridge Design Code [AUSTROADS, 1992] dynamic load allowance of 25%, two vehicles would induce the same peak deflection as one vehicle with a 100% dynamic increment. It is clear that the dynamic increment often substantially exceeds the Dynamic Load Allowance for this short-span bridge under a single, legally-loaded vehicle.

A reason for the high dynamic axle-hop response, particularly of the SA vehicle, to the coldmix repair bump could be the following: At critical speed, each of the tri-axle group axles vibrated at the natural frequency of the bridge and in phase with each other. As a result, dynamic coupling between the vehicle axle-hop vibrations and the Camerons Creek Bridge was strong. Such an in-phase axle-hop vibration of all axles of a multiple-axle aggregate is not possible for steel-suspended aggregates because the load-equalizing devices preclude inphase axle-hop vibrations to be generated by random or specific features. This is discussed further in the following section.


140 / 169 8. June 2010

An explanation is then required for the obvious differences in the effects of the two air-suspended vehicles BA and SA. These vehicles were similar in that the prime-movers were the same and both trailers were fitted with air suspensions made by the same manufacturer. They differed in the following areas: 1) a tip truck (BA) versus a tanker (SA), 2) gross weight 425 kN versus 452 kN, 3) tri-axle spacing of 1.23 m versus 1.4 m, 4) 4.65 m versus 6.56 m spacing between the tandem and tri-axle groups and 5) vertical dampers mounted behind the axle versus inclined dampers mounted in front of the axle. It is suggested that the less geometrically efficient location of the dampers in the SA vehicle is the major reason for the increased dynamic bridge response to this vehicle. Reference to the bridge response time histories illustrates that the BA and SA vehicles induced much larger and more stress reversals compared with the BS vehicle where the short wave-length roughness induced axle-hop and the vehicles traveled at critical speeds. Thus the combination of short wavelength roughness and air suspensions is potentially very damaging from a fatigue standpoint. This is especially significant if the damping on the air suspensions is ineffective. This observation requires further investigation. c)

With axle-hop plank

The "standard" axle-hop plank as used in the other Australian bridge tests was fitted to both lanes of the road on the pier between bridge spans 3 and 4. Here, the BS and BA vehicles were involved only. The influence of the AHP is only significant for D(4,8) for vehicles traveling north. Consequently the comments that follow refer only to D(4,8) and northbound traffic. For the BA air-suspended vehicle the introduction of the axle-hop plank increases the maximum dynamic increment from φ ≈ 25% to φ ≈ 115% which is larger than the φ ≈ 75% associated with the cold mix repair bump and D(1,8). The AHP also results in an increased range of speeds (v = 40 km/h to v = 70 km/h) where large dynamic increments are evident for the BA vehicle. For the BS steel-suspended vehicle the introduction of the axle-hop plank increases the maximum dynamic increment from φ ≈ 40% to φ ≈ 60% which is less than the φ ≈ 110% associated with the cold mix repair and D(1,8). Thus, for the steel suspension, the introduction of the AHP was less significant than the cold-mix repair. The reverse was true for the BA suspension. It is of interest to note that for D(1,8) and southbound traffic the effect of the BS vehicle (φ ≈ 70%) was substantially larger than that of the BA vehicle (φ ≈ 30%). It is suggested that this reflects the ability of the BA suspension to smooth out the long wavelengths compared with the BS vehicle. When short wavelengths dominate, the BA vehicle results in significant increases in bridge response. When both short and long wavelengths are present then it is more difficult to predict which suspension will dominate.


141 / 169 8. June 2010

Preliminarily summarizing again, the effect of the AHP was far more significant for the airsuspended BA vehicle compared with the steel-suspended BS vehicle. For northbound traffic, BA induced a dynamic increment in the bridge span adjacent to the plank about twice as large as BS did (D(4,8), φ ≈ 115% versus φ ≈ 60%). The AHP also increased the range of speeds over which large dynamic responses (both magnitude and number of fatigue cycles) were evident for the BA vehicle. d)

Dynamic Wheel Loads

The dynamic wheel load power spectral densities (refer Paragraph 6.3.3, c, Fig. 80) highlight the body-bounce (f = 1.5 Hz and f = 3.3 Hz) and axle-hop modes (8 < f < 15 Hz). Both the axle-hop and the body-bounce modes can be observed in the dynamic wheel load time signals presented in Paragraph 6.2.4, a. The amplitudes of the axle-hop components of the dynamic wheel loads are much smaller than the amplitudes of the dynamic wheel forces associated with body-bounce. This is true for individual axles and for the tri-axle group as a whole. For short-span bridges with natural frequencies corresponding to those of axle-hop, the effects of the axle-hop components of the dynamic wheel forces are amplified whereas the bodybounce effects are essentially transmitted without amplification. Thus the total response of the bridge is the sum of the direct effects of the body-bounce forces plus the amplified effects of the axle-hop forces. From the bridge standpoint, the load applied to the bridge by the axle-group is more important than the loads applied by individual axles. Thus, it is how the axle-hop components in each dynamic wheel force add that is important. Inspection of the wheel load signals (Paragraph 6.2.4, a) shows that the axle-hop components in the total dynamic wheel force for the tri-axle group is largest for the BA vehicle whereas the body-bounce effects are largest for the BS vehicle. The mechanisms of load sharing between axles in a group are evidently significant, especially at axle-hop frequencies. Air-bag suspensions are load sharing through pressure equalization. However, under high frequency dynamic loads this equalization does not occur due to the lack of time for the air to travel between the air bags. Consequently each air-suspended axle acts independently of the others for dynamic loads. Air suspensions do not share the high frequency components of the axle loads. Steel leaf suspensions include a series of rockers which provide the load sharing when the vehicle is stationary but also transmit dynamic forces between axles (refer Annex H for typical details of steel suspensions). Thus when the first axle in a group strikes a short bump, a portion of the shock is transmitted to the other axles in a manner which is likely to be out of phase


142 / 169 8. June 2010

with the first axle. As the axle group continues over the bump, this 'cross-talk' continues, thus diminishing the opportunity for each axle to vibrate independently or in phase. A similar argument applies for the steel walking beam suspensions such as the one fitted to the BS prime mover. These differences are further exacerbated by the speed of the vehicle. At a critical speed, the time between each axle striking an axle-hop exciter is equal to the natural period of vibration of each axle. When this occurs, each axle in an air-suspended group vibrates in phase and the effects add for the total group. With the steel suspensions used in these tests, the 'cross-talk' diminished this effect. Thus, one would expect speed to be a critical issue for air suspensions and that the critical speed (vcrit.ah) would be approximately the axle-hop frequency (fah) times the axle spacing (s): vcrit.ah

≈ s • fah

In the case of an axle-hop natural frequency of f = 12 Hz and a spacing of 1.30 m (drive tandem), the critical speed would be vcrit.ah = 12 • 1.30 • 3.6 = 56 km/h. This is similar to the 60 km/h peaks evident in the φ(v)-graphs for the BA and SA vehicles. The peaks associated with the steel suspensions are for higher speeds. e)

Summary

Interpretation of the tests performed on Camerons Creek Bridge is a quite difficult task. This is a consequence of the pavement surface conditions at the site. One the one hand, the overall quality of the test track pavement surface was quite poor. On the other hand, the surface of the bridge deck itself was quite smooth except for some long wavelengths associated with the cambering of the bridge deck. Above all, there was a cold-mix repair containing short wavelength roughness on the northbound approach to the bridge just prior to bridge span 1. Tests performed with a 300 x 25 mm axle-hop plank over the pier between bridge spans 3 and 4 enables the effect of a known profile feature to be evaluated. It is evident that the dynamic response of short-span bridges such as the bridge over Camerons Creek is a complex interaction between the suspension type, the short and long wavelength components of the road profile, vehicle speed, bridge frequency and damping. Despite the complexity, several things have become clear: • Under the pavement conditions valid for Camerons Creek Bridge, for all vehicle configurations tested bridge dynamic increments occur which are definitely larger than any Bridge Design Code would account for.


143 / 169 8. June 2010

• Under such pavement conditions, air-suspended axles can produce effects due to axle-hop which are higher than those induced by steel-suspended axles. However, body-bounce vibrations induced effects are stronger for steel than for air-suspended vehicles. • Under such pavement conditions, the vehicle suspension damping capacity is of crucial importance for the magnitude of the dynamic bridge response. • There is evidence that air-suspended vehicles with the shock absorbers mounted in a geometrically inefficient position are inducing higher dynamic increments than steel-suspended vehicles do. However, the same may not be true if the shock absorbers of air suspensions are mounted in an efficient position. • Lacking of load-sharing between axles of individually suspended multiple-axle aggregates yields the possibility of all of the axles vibrating in-phase and hence producing significant vehicle/bridge-interaction at axle-hop frequencies. Here, the load-sharing devices commonly used for steel-suspended multiple axle aggregates can be very beneficial because such an interaction is not possible and the bridge response remains moderate. 7.1.7 Cromarty Creek Bridge (AUS)

The deflection time signals corresponding to the peak dynamic response of Cromarty Creek to the passage of the BA and BS test vehicles are presented in Paragraph 6.1.7, a. The strong high frequency components of the dynamic bridge response evident at Camerons Creek can not be observed at Cromarty Creek. Despite this, the dynamic increments registered at Cromarty Creek were large for the BS vehicle with φmax ≈ 109%. For the BA vehicle φmax ≈ 50% or half of that for the BS vehicle. The dynamic load allowance (DLA) specified in the AUSTROADS Bridge Design Code [AUSTROADS, 1992] for bridges of this natural frequency is φ = 25%. With values commonly exceeding φ = 50% and a maximum dynamic increment of φ = 109%, it is clear that the φ often substantially exceeds the DLA by up to 350% for this short-span bridge under a single legally loaded vehicle. The addition of the axle-hop plank at Cromarty Creek resulted in small increases in φ for both vehicles over the limited speed range that was recorded. Again, this is in contrast with the Camerons Creek bridge where large increases were observed for the BA vehicle in particular. It could be suggested from the fundamental natural frequency of the Cromarty Creek Bridge, f = 9.5 Hz, that significant dynamic bridge response forced by vehicle axle-hop vibrations would occur. Despite having similar dynamic increments to Camerons Creek it is suggested that the behavior of Cromarty Creek is different.


144 / 169 8. June 2010

The 30° skew and the inconsistent transfer of load from the deck planks to the girders both contribute to the dynamic wheel forces being received by the girders in a "confused" manner. Under these conditions there is less likelihood that the effects of the axle-hop components of the dynamic wheel forces in the axle group will add. In addition, the bridge over Cromarty Creek is relatively highly damped (ζ = 2.6%), thus reducing the magnitude of amplification possible. Consequently the axle-hop components are less important for the bridge over Cromarty Creek. This is consistent with the relatively small effects of fitting the AHP to the bridge. However it does not explain the rather large dynamic increments recorded in this bridge. The dynamic wheel forces recorded for this bridge show dynamic load coefficients for the BS vehicle to be approximately twice those for the BA vehicle - a similar ratio is evident in the dynamic increments. Also the DLC’s for the BS increase with speed as the dynamic increments do. The maximum value of the DLC for the tri-axle group is ≈ 20% which would correspond to a peak dynamic tri-axle group force approximately 1.5 times the stationary force. This would explain approximately half of the maximum dynamic increment recorded. Cromarty Creek Bridge is a relatively soft natural timber girder bridge and therefore deflected considerably (10...12 mm compared with 1...2 mm at Camerons Creek) during the passage of test vehicles. At v = 90 km/h (25 m/s) the critical wavelengths required to induce body-bounce effects in the BA and BS vehicles are approximately 8 m and 17 m respectively. Thus the deflection adds significantly to the critical wavelengths for body-bounce, especially for the BS vehicle where the critical wavelength is very similar to spans (7.6 m and 9.1 m). It is believed that the softness of the Cromarty Creek Bridge is significant. This additional "roughness" excites the body-bounce modes of the steel suspensions more than those of the air suspended vehicle. These effects are transmitted directly to the bridge without significant amplification or attenuation. The result is that the bridge experiences significantly larger forces and hence dynamic increments for the steel suspended vehicle. This is believed to be the major reason for the observation that the BS vehicle induces dynamic increments which are approximately twice those of the BA vehicle. In summary, the response of the Cromarty Creek bridge was surprising. Its skew, the high level of damping, its softness, and inconsistent / loose nature of the timber deck and its connections to the girders all combine to reduce the amplification of axle-hop effects. The result is that both the body-bounce and axle-hop behavior of the vehicles are important.


7.2

145 / 169 8. June 2010

DIVINE Testing versus Swiss Tests

The maximum dynamic increment φmax for each DIVINE bridge/vehicle-configuration is compared with the results of similar Swiss testing involving single two-axle rigid trucks fitted with steel leaf suspensions in Fig. 95. The following comments are made: • In the body-bounce frequency range (f = 1.5...4 Hz), the dynamic increments are largest for the vehicle (suspension) that has a natural body-bounce frequency matching the fundamental natural frequency of the bridge. • In the case of frequency matching at f ≈ 1.5 Hz, the maximum dynamic increment measured coincides with the solid line given in Fig. 95 indicating the upper limit of dynamic increments observed with earlier tests. • In the axle-hop frequency range (f = 8...15 Hz), dynamic increments lying well above the "Swiss test limit" have been observed with the exception of the results for a highly damped bridge. • In between the body-bounce and axle-hop frequency ranges (f = 4...8 Hz), the dynamic bridge response is relatively small and the type of truck suspension is less important.


• Very limited testing has been previously undertaken on bridges with axle-hop frequencies.

150

100 50

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Fundamental Frequency f [Hz]

Fig. 95

Dynamic increments as a function of bridge fundamental frequency for two-axle trucks (from [Cantieni, 1983]) incorporating OECD DIVINE test data: Solid squares: steel suspension; Open squares: air suspension.

The tests conducted with a 300 x 25 mm plank during the OECD Divine tests on short-span bridges in Australia are compared with Swiss tests using a 300 x 50 mm plank in Fig. 96. The


146 / 169 8. June 2010


combination of a larger plank and a two-axle rigid truck is more severe than the smaller plank used in association with six-axle articulated vehicles. It is suggested that the vibrations excited in a two-axle rigid vehicle crossing a severe plank are worse than the more complex responses of articulated vehicles fitted with tandem and tri-axle groups. 300

200

100

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Fundamental Frequency f [Hz]

Fig. 96

Dynamic increments as a function of bridge fundamental frequency for passages with with a two-axle truck crossing a 300 x 50 mm plank (from [Cantieni, 1983]) incorporating OECD DIVINE test data for a 300 x 25 mm plank: Solid squares: steel suspension; Open squares: air suspension.

Further comparison of the DIVINE test results with those of Swiss tests performed earlier is possible. The results given in the two preceding figures encompass a large number of bridges tested but they do not consider the test track pavement surface quality and the test vehicles were not at all similar to those used in the DIVINE tests. However, there are results available from tests having been performed on the Deibüel Bridge with known pavement longitudinal profile and using 14 different vehicles in 1978 (Fig. 97). The pavement quality according to [ISO, 1995] was C. The vehicles included two-axle rigid trucks (Nos. 11, 12, 13, 14) three- and four-axle rigid trucks (Nos. 20, 21, 24, 25), four- and five axle tractor-trailers (Nos. 42, 52) and three- to five-axle tractor-semitrailers (Nos. 71, 74, 76). Furthermore, tests with multiple vehicle presence were conducted (2LH, 2MH, 2MN, 4MH, 4MN). The three character coding for the latter tests has the following meaning: • First character: Number of vehicles involved, • Second character: L = 160 kN two-axle truck, M = 280 kN four-axle truck, • Third character: H = all vehicles in line, N = vehicles side by side, in the case of four vehicles, pairs of two L and M vehicles in line respectively.


Fig. 97

147 / 169 8. June 2010

Maximum dynamic increments for the Deibüel Bridge deflection at the mid-point of the middle span and various steel suspended vehicles (from [Cantieni, 1987, 1988]). Description of the vehicle type coding: see previous page.

The results presented in Fig. 97 show that for body-bounce frequency-matching conditions and a quite rough pavement profile • rigid trucks with three and four axles are on the average the dynamically most aggressive vehicles, • two-axle trucks are less aggressive, • combined and articulated vehicles are less aggressive than other types of single vehicles, • multiple presence of vehicles reduces the dynamic increment to about 25...50% of what can be achieved with a single vehicle. The NRC test vehicle is closest to the vehicle types 74 (2S2) and 76 (2S3). The maximum dynamic increments reached for the types 74 and 76 were φ ≈ 48% and φ ≈ 27% respectively. The parameters having been found to significantly influence the bridge dynamic response under the passage of a heavy commercial vehicle are discussed in the following paragraphs in more detail.


7.3

148 / 169 8. June 2010

Dynamic Wheel Loads (DLC's)

The dynamic load coefficient (DLC) increased with speed and road roughness for both air and steel suspensions. This is consistent with the literature. For the smooth pavements in Switzerland the difference in DLC’s for the steel and air suspensions were not large (DLC < 10%). The DLC’s were slightly higher for the steel suspended vehicle than for the air-suspended. For the rough pavement conditions of the Australian test tracks, the steel-suspended vehicle applied the largest dynamic wheel forces. These were associated with the truck body-bounce modes rather than the axle-hop ones. The dynamic component for the steel suspension was approximately twice that of the air suspensions. The DLC varies between axles of a tri-axle group. The DLC’s were larger for the front axle of the tri-axle group. This was more evident for the steel-suspended vehicle. The sum of the axle loads that make up each axle group is the important parameter for determining the bridge response. The DLC’s for the tri-axle groups were approximately two-thirds of the DLC’s for individual axles for both air and steel suspensions. The DLC’s for the steel tri-axle group were approximately twice those of the air-suspended tri-axle group. The removal of the shock absorbers on the rear axle of the air-suspended tri-axle group caused uncontrolled axle-hop and induced significant changes in the dynamic wheel forces. The DLC for the air-suspended axle without shock absorbers (≈ 20%) was similar to the steel suspension however the variability was at the axle-hop frequencies rather than the bodybounce frequencies. Vehicles fitted with air suspensions and ineffective shock absorbers are considered likely to result in unacceptably large dynamic responses in lowly damped short-span bridges that exhibit some short wavelength roughness. Ineffective damping results from worn dampers remaining in service or as a consequence of design. Shock absorbers positioned close to the pivot point of the trailing arm and/or at an inclination are geometrically less efficient than those placed as far away from the pivot and perpendicular to the trailing arm. The dampers fitted to the tri-axle group of the air suspended test vehicle (BA) were highly effective. This was not the case e.g. for the SA vehicle. Larger dynamic responses than those induced in short-span bridges by the BA test vehicle can be expected to be induced by air suspended vehicles with less effective damping.


7.4

149 / 169 8. June 2010

Wheel Load Spectra as a Function of Vehicle Speed

The frequency of the peaks identified in the wheel load spectra and the associated mode shapes of the NRC vehicle are plotted against speed in Paragraph 6.2.1, c. These results show that there are vehicle vibrational modes which depend on speed and others which do not. The frequency of the speed dependent modes increases linearly with speed. The corresponding scaling factor of the dominant line slope is L = 3.3 m. This line appears in all diagrams for every wheel and both suspension systems. This length corresponds well with the circumference of the wheels of the test vehicle. It is hence most probable that this part of the vibration is due to non-homogeneous characteristics in the vehicle wheels. In addition there exist lines with other slopes, corresponding to other lengths. Some of these lengths correspond well with the distances between different pairs of axles of the NRC test vehicle. Lines with slopes corresponding to distances between axles of the vehicle may be explained according to the so-called "wheel base filtering" effect: For spatial frequencies of the test track pavement profile corresponding to the distance between two axles of the vehicle, the wheels are loaded in phase when the vehicle drives along the test track. This effect leads to an amplification of the corresponding vibration. The remaining lines, not corresponding to distances between axles, may be caused by other peaks in the profiles of the test tracks. Frequencies that do not depend on speed are found at f = 1.5 Hz for air and at f = 2.8 Hz for steel suspension. In addition, non-speed-dependent peaks can be seen at very low frequencies (f < 1.5 Hz) and, for the steel suspension, at high frequencies (f > 10 Hz). In the case of air suspension high-frequency peaks (f = 9...18 Hz) are found to be speed-dependent. In summary, the relationship between the wheel load spectra and the vehicle speed is of quite complex nature. 7.5

Frequency Matching

This is obviously the basic parameter. Without frequency matching, no significant dynamic bridge response and hence no quasi-resonance effects are possible (see the exception mentioned below considering "abnormal" pavement unevenness conditions). Furthermore, the dynamic wheel load magnitude is important. Considering frequency matching in the vehicle body-bounce frequency region the wheel load magnitude depends mainly on the pavement longitudinal profile. Considering frequency matching in the vehicle axle-hop frequency region, additional parameters are affecting the dynamic wheel loads: Inter-axle load equalizing devices, efficiency of the shock absorbers damping capacity for air suspensions.


150 / 169 8. June 2010

However, frequency matching alone does not necessarily suffice to result in excessive bridge vibrations. Large dynamic bridge response may be due to large dynamic wheel loads excited by very rough pavement conditions even without the "help" of frequency matching effects. 7.5.1 Medium to Long-Span Bridges in Switzerland

In case the bridge natural frequency and the dominant wheel load frequency matched, higher dynamic increments were observed than under non-frequency-matching conditions. This applies for the Deibüel Bridge and the steel-suspended vehicle and for the Sort Bridge and the air-suspended vehicle respectively. In the case of the Föss Bridge, where the bridge fundamental frequency lies outside of the wheel load frequency for both suspensions, the dynamic increments are relatively small. The maximum values for the dynamic increments under frequency-matching conditions, φ = 21...26%, are however comparatively small (see Table 12). This is due to the fact that the pavements were smooth to very smooth. Whereas the case of rough pavements is well covered by other experiments for steel-suspended vehicles, this does not hold for air-suspended vehicles. Here, additional research is necessary. As it is not easy to change bridge pavements this problem will be investigated using analytical methods. The bridge and vehicle models available for this purpose can be calibrated based on the experimental results described here. 7.5.2 Short-Span Bridges in Australia

It has been shown that forcing of short-span bridge vibrations through vehicle axle-hop vibrations does occur and that this is most likely when a short-wave pavement unevenness is crossed by the vehicle at critical speed. Critical speed is dependent on the pavement unevenness wavelength and the ratio vehicle axle-hop frequency / bridge natural frequency. It has been observed that this forcing may or may not be much stronger for air rather than steel suspensions. The test results show that besides pavement unevenness, there are two factors influencing the dynamic wheel load magnitude. (a) dynamic inter-axle load equalization and (b) shock absorbers damping efficiency for air-suspensions. Considering (a): Air-suspended axles seem to operate independently as far as dynamic wheel loads are concerned whereas steel-suspended axles do not. Considering (b) the effectiveness of dampers is crucial in controlling axle-hop in air-suspensions. For dampers to be effective they need to be maintained in good condition and be designed to ensure adequate damping. The effects of the factor (a) mentioned can be illustrated with the help of a short arithmetic investigation. Consider the case of air suspensions fitted to an axle group crossing a sharp defect that excites axle-hop. Interaction between axles in the group does not exist. Thus, the way that the components of each wheel add is sensitive to speed and axle spacing. Fig. 98 il-


151 / 169


8. June 2010

Dynamic axle group loads (t)

lustrates the total dynamic loads applied by single, tandem and tri-axle groups for five speeds between v = 50 and v = 100 km/h. It has been assumed that each axle is excited at time t = 0, axles are spaced at 1.23 m and that the 10% dynamic component of each axle vibrates at only one frequency. In this case, the frequency chosen corresponded to the axle-hop range with f = 11.3 Hz, the fundamental natural frequency of the bridge over Camerons Creek. At v = 50 km/h, each axle is excited one cycle apart and the dynamic component of each axle adds for both the tandem and the tri-axle. As speed increases, the dynamic components begin to move out of phase and at some speeds cancel each other, e.g. at v = 75 km/h for the triaxle group. 25 Triaxle group 20 Tandem 15 Single axles 10 5

Speed (km/h) 62.5 75 87.5

50

100

0 0

0.2

0.4

0.6

0.8

1

Time (s)

Fig. 98

Speed influences how the dynamic components of axles within a group combine.

The above assumptions are an over-simplification. In suspensions such as steel leaf and "walking beam" suspensions, dynamic interaction between axles reduces the tendency for the dynamic components to add at some speeds. At other speeds, dynamic interaction causes the dynamic components to cancel or reduce. It is suggested that for air suspensions, interaction between axles is minimal and the phenomenon illustrated in Fig. 98 is possible. The behavior is further complicated by a spectrum of frequencies existing simultaneously. However, it is the frequencies that closely approximate the natural frequencies of the bridge that are the most important. Further, dynamic components are related to speed and, at low speeds, the energy to drive the dynamic component is substantially reduced. Despite the crudeness of the assumptions, it is clear that the axle-hop dynamic components within a group tend to add more at the speeds which approximate the speeds associated with the highest dynamic response of the bridge over Camerons Creek, for example. Similar behavior occurs at body-bounce frequencies within a group and within entire vehicles.


7.6

152 / 169 8. June 2010

Vehicle/Bridge Interaction

The basics of vehicle/bridge interaction using very simple models are described in the Paragraphs 6.4.1 and 6.4.2 for medium to long-span and short-span bridges respectively. 7.6.1 Medium to Long-Span Bridges

"Real" vehicle/bridge system coupling was clearly identified for the case of the NRC test vehicle equipped with steel suspensions (f1 ≈ 2.9 Hz) crossing the Deibüel Bridge (f1 = 3.03 Hz). The average vehicle body-bounce frequency observed while the vehicle was on the bridge is smaller by Δf = 0.13 Hz compared to the frequency when the vehicle is on the approach stretch. Hence system coupling could experimentally be proven to occur while the vehicle was crossing the bridge. A smaller frequency shift, Δf ≈ 0.1 Hz, is observed, when the vehicle with air suspension crosses the Sort Bridge whose fundamental frequency is f1 = 1.6 Hz. Because the vehicle/bridge mass ratio is smaller for the Sort than for the Deibüel Bridge, the expected frequency shift is smaller too. For the case Sort Bridge / air-suspended vehicle it can be noted that is was possible, probably for the first time ever, to identify not only the in-phase motion of the coupled vehicle/bridge system but also the out-of-phase motion. Furthermore, a system coupling effect is observed for the steel-suspended vehicle and the Sort Bridge. In this case the natural frequency of the vehicle (f = 2.8 Hz) is higher than the fundamental frequency of the bridge. A possible explanation for this effect could be that frequency coupling with higher natural modes of the bridge (f7 = 2.98 Hz) might occur. However, in the case of the Sort bridge, it has to be considered that the frequency resolution, which is limited by the length of the test track on the bridge and vehicle speed, is too poor to detect this system coupling unambiguously by analyzing the wheel load data of the test vehicle by itself. For further investigations data of the bridge response will have to be considered in the analysis too. Knowledge of what bridge modes are actually excited by the vehicle crossing should allow identification of the degree of interaction between vehicle and bridge more clearly. 7.6.2 Short-Span Bridges: Bridge Damping

It has been shown in Paragraph 6.4.2 that for short-span bridges it is much more reasonable to change the modeling from "system coupling" SDOF + SDOF = 2DOF to the SDOF model of a system (the bridge) being forced to vibrate by external forces (the dynamic wheel loads) without taking the vehicle masses involved into account. The basic physics of SDOF systems tells us that the dynamic response is a function of the ratio of the natural circular frequency of


153 / 169


8. June 2010

the bridge ωn to the frequency of the forcing function ω and the damping of the bridge ζ. The classical expression for the magnitude of the amplification factor M is as follows:

M =

1 ⎡ ⎛ω⎞ ⎢1 − ⎜ ⎟ ⎢⎣ ⎝ ωn ⎠

2 2

2 ⎤ ⎛ ⎞ 2 ω ⎥ + 4ς ⎜ ⎟ ⎝ ωn ⎠ ⎥⎦

In the case of frequency matching, ω/ωn = 1, the above expression reduces to |M| = 1/2ζ. That is, basic physics tells us that the dynamic response of the forced system depends only on its damping capacity if the forcing frequency is equal to the natural frequency of the forced system.

Camerons (AUS)

Coxs (AUS)

Cromarty (AUS)

Lawsons (AUS)

Foss (CH)

100.0

Sort (CH)

|Amplification Factor|

1000.0

Diebuel (CH)

The magnitude of the amplification factor versus the frequency of the forcing function for each of the bridges tested is presented in Fig. 99. Superimposed on this diagram are the frequencies associated with the body-bounce and axle-hop components of the dynamic wheel forces. It becomes clear from this figure that effects of the wheel forces are transmitted (M ≈ 1), amplified (M > 1) or attenuated (M < 1) by the bridge.

Body-bounce Air Body-bounce Steel Axle-hop

10.0

1.0

0.1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Forcing frequency [Hz] Fig. 99

Amplification of the Single-Degree-of-Freedom bridge model to semi-logarithmic scales (Note: Bridges are arranged in order of frequency).


154 / 169


8. June 2010

As can be seen from Fig. 99 long span bridges amplify the effects associated with bodybounce and attenuate axle-hop effects. Conversely short-span bridges amplify the effects associated with axle-hop. In addition, short-span bridges transmit with limited amplification the effects associated with body-bounce. This makes the interpretation of the dynamic response of short-span bridges more complex. For example, the relatively soft Cromarty Creek bridge induces large dynamic wheel forces due to its deflected profile. These are transmitted directly to the structure and lead to dynamic increments which increase with speed. The high levels of bridge damping, skew and natural frequencies at the low end for coupling with axle-hop mean that there is only limited axle-hop frequencies present in the bridge response. The dynamic increment versus speed relationships are largely related to increasing body-bounce forces associated with body-bounce. In the case of Camerons Creek dynamic response is dominated by its natural frequency corresponding to axle-hop frequencies, its low damping and the existence of short wavelength roughness in the approaches to the bridge. The axle-hop vibrations initiated by the short wavelengths in the profile excited the bridge at its resonance frequency resulting in up to 10 large dynamic cycles during the passage of the air suspended vehicles. Without the short wavelength roughness, the axle-hop components are small and the dynamic response reduces dramatically. This is illustrated in Fig. 100 which compares the response of the first and fourth spans of Camerons Creek during the passage of an air suspended test vehicle. The large response of the first span can be attributed to the short wavelength roughness in the approach whereas the approach to the third span is the smooth bridge deck which has minimal short axle-hop initiating roughness. 0.5

Deflection [mm]

0.0 -0.5

D(4,8)

-1.0 D(1,8)

-1.5 -2.0 -2.5 -3.0 0

1

2

Time [s]

3

4

5

Fig. 100 Camerons Creek Bridge deflection time signals measured at D(1,8) and D(4,8), Test Vehicle SA, no planks, v = 63 km/h, northbound.


155 / 169 8. June 2010

The amplitude of the dynamic components of the wheel forces associated with axle-hop are amplified by short-span bridges. The magnitude of this amplification can be significantly influenced by the amount of damping in the bridge. For example, the Coxs River bridge response was relatively small because its frequency was at the lower limit for dynamic coupling with axle-hop frequencies and it has a high level of damping. The dynamic responses of Sort, Deibüel, Föss and Lawsons were all comparatively small. These bridges exhibit smooth profiles and thus relatively small dynamic components of the wheel forces. The dynamic increments are largely consistent with the single degree of freedom model. 7.7

Limiting the Dynamics of Wheel Loads and Bridge Response

Limiting the dynamic component of wheel forces is the most fundamental way to minimize the dynamic response of bridges. The results of the DIVINE project have shown that this is achieved by maintaining a smooth road profile and providing good suspensions to heavy vehicles. In terms of the road profile, the wavelengths that are important vary from bridge to bridge, from vehicle to vehicle and with speed. A simplified model based on a single axle moving over a sinusoidal profile is used to illustrate this concept. Again, basic physics tells us that when the profile excites the vehicle suspension at its natural frequency, larger dynamic wheel forces result. The frequency f that the suspension “sees” is a function of the vehicle speed v and the wavelength λ such that f = v/λ. This relationship is presented in Fig. 101.


156 / 169


8. June 2010

30 1.0 Hz

Wavelength [m]

25 Body-bounce: Air

20 15

2.0 Hz 2.5 Hz

10

Body-bounce: Steel

4.0 Hz 5

8 Hz 15 Hz

Axle-hop: Air & Steel

0 0

20

40 60 Speed [km/h]

80

100

Fig. 101 Combinations of wavelengths in the road profile and vehicle speed leading to excitation of heavy vehicle suspensions. After considering the body-bounce frequencies for air suspensions (1.0 to 2.0 Hz) and steel suspensions (2.5 to 4.0 Hz), and the axle-hop frequencies (8 to 15 Hz) the range of critical wavelengths can be observed as a function of suspension type and vehicle speed. The effects of multi-axle groups and the concepts of wheel-base filtering complicate this but the overall concept is clear: • The higher the speed, the longer the critical wavelengths • For vehicle speeds of 50 to 100 km/h • short wavelengths (λ < 3 m say) excite axle hop • short to medium wavelengths (3 < λ < 10 m say) will be amplified by steel suspensions at their body-bounce frequencies • medium to long wavelengths (10 < λ < 30 m say) will be amplified by air suspensions at their body-bounce frequencies If this is combined with the concepts of vehicle/bridge interaction then the following qualitative model of the influence of the various wavelengths in the profile on dynamic bridge response results (refer Table 14).


157 / 169


8. June 2010

Frequency

Span

Short wavelengths Short to medium (λ < 3 m and either air wavelengths or steel suspensions) (3 < λ < 10 m and steel suspensions)

Medium to long wavelengths (10 < λ < 30 m and air suspensions)

>15 Hz

80 m

Local effects only

Table 14 Qualitative relationship between bridge span and the amplification of dynamic wheel forces induced by short, medium and long wavelengths in the road profile for vehicles traveling in excess of 50 km/h. The following strategies are possible to limit the dynamic response of bridges: • Ensure the road profile is smooth and thus reduce the amplitude of the dynamic components of the wheel forces whose effects can be amplified by the bridge. The longer wavelengths in the profile that excite body-bounce are important for medium span bridges. Both long and short wavelength roughness (mis-aligned joints, pot-holes, rough repairs, abrupt transitions in overlays…) will prove detrimental to short-span bridges. • Increase the level of damping in the bridge (for short-span structures). This is unlikely to be practical. • In the case of new short-span bridges, construct skewed bridges and thus reduce the possibility that each wheel in a multi-axle group is vibrating in phase with each other and the bridge. • Encourage multiple axle groups in vehicles with soft highly damped suspensions. Being assured of the effectiveness of dampers over the life of the vehicle is a significant issue for this strategy.

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia, Conclusions

8.

CONCLUSIONS AND RECOMMENDATIONS

8.1

Basics, Basic Physics, Parameters

158 / 169 8. June 2010

The tests performed in the context of the DIVINE Project showed that to better understand the vehicle/bridge interaction problem it is necessary to distinguish between the basic problem, the basic physics and the related parameters. The basics are:

• Dynamic wheel loads (and vehicle mass) are acting upon a bridge, generating a dynamic bridge response. The basic physics are:

• Dynamic wheel loads are influenced by • pavement unevenness (road profile), • the vehicle suspension system, • vehicle speed. • The dynamic bridge response is influenced by • Frequency matching, i.e. dynamic wheel load frequency content vs. the bridge natural frequency(ies), • Bridge damping, • Wheel load magnitude. The parameters are:

• Pavement longitudinal profile, • Vehicle parameters, • vehicle geometry (vehicle type, number of axles, axle spacing(s)), • suspension system (stiffness and damping characteristics), • suspension mechanisms (load equalization, cross-talk), • vehicle speed, • vehicle gross weight and axle loads. • Bridge parameters • bridge geometry (total length, lengths of the individual spans…), • bridge mass and stiffness distribution, boundary conditions, • damping capacity.


8.2

159 / 169 8. June 2010

The Basic Physics

8.2.1 Dynamic Wheel Loads

The dynamic wheel loads of a heavy vehicle occur in the body-bounce and the axle-hop frequency ranges. Their amplitudes and frequencies are dependent upon the longitudinal pavement profile, the suspension characteristics and the vehicle speed. The dynamic wheel loads, as quantified by the dynamic load coefficient (DLC), increased with speed and road roughness for both air and steel suspensions. This is consistent with the literature. For the smooth pavements in Switzerland the difference in DLC’s for the steel and air suspensions were not large (DLC < 10%). The DLC’s were slightly higher for the steel-suspended vehicle than for the air-suspended vehicle. For the smooth to quite rough pavement conditions of the Australian test tracks, the steelsuspended vehicle applied the largest dynamic wheel forces. The dynamic component for the steel suspension was approximately twice that of the air suspensions. This is related to bodybounce vibrations. The DLC varies between axles of the tri-axle group. The DLC’s were larger for the front axle of the tri-axle group. This was more evident for the steel-suspended vehicle. The sum of the axle loads that make up each axle group is the important parameter for determining the dynamic response of short-span bridges. The DLC’s for the tri-axle groups were approximately twothirds of the DLC’s for individual axles for both air and steel suspensions. The DLC’s for the steel tri-axle group were approximately twice those of the air-suspended tri-axle group. The removal of the shock absorbers on the rear axle of the air-suspended tri-axle group caused uncontrolled axle-hop and induced significant changes in the dynamic wheel forces. The DLC for the air-suspended axle without shock absorbers was similar to the steel suspension however the variability was at the axle-hop frequencies rather than the body-bounce frequencies. Vehicles fitted with air suspensions and ineffective shock absorbers resulted in unacceptably large dynamic responses in a lowly damped short-span bridge that exhibited short wavelength roughness. Ineffective damping results from worn dampers remaining in service or as a consequence of design. Shock absorbers positioned close to the pivot point of the trailing arm and/or at an inclination are geometrically less efficient than those placed as far away from the pivot and perpendicular to the trailing arm. The dampers fitted to the tri-axle group of the air


160 / 169 8. June 2010

suspended test vehicle (BA) were highly effective. This is not the case with all air-suspended vehicles. Larger dynamic responses than those induced in short-span bridges by the BA test vehicle can be expected to be induced by air suspended vehicles with less effective damping. 8.2.2 Dynamic Bridge Response a)

General (Frequency Matching)

The response of the bridges to the passage of the test vehicles was consistent with the laws of physics in that the largest dynamic responses occurred when either the body-bounce or axlehop frequencies of the test vehicle corresponded to the natural frequencies, especially to the fundamental natural frequency in bending, of the bridge. This effect is described as frequency-matching. The system's vehicle and bridge may then couple and quasi-resonance effects occur. For medium to long-span bridges, this phenomenon can be described in a qualitative manner by a Two-Degree-of-Freedom-System model of the coupled system. Quantitative results cannot be derived from this simple model. A more sophisticated model should take into account at least the real road profile and the exact mode of vibration of the vehicle when it enters the interacting zone on the bridge. In addition, the case of higher bridge modes interacting with the vehicle should be covered too. In the case of short-span bridges the model of the bridge as a SDOF-system excited by a dynamic force is more appropriate. b)

Medium to Long-Span Bridges

Here, frequency-matching occurs at the vehicle body-bounce frequencies. These are not identical for steel and air suspensions. Consequently: Air-suspended vehicles will induce larger dynamic responses in bridges with natural frequencies in the vicinity of f = 1.5...1.8 Hz (maximum span L ≈ 60...70 m). Due to the smooth pavement conditions of the test tracks the increases observed were however insufficient to warrant any increases in allowance for dynamic effects made in bridge design. Steel-suspended vehicles will induce larger dynamic responses in bridges with natural frequencies f = 2...4 Hz (maximum span L ≈ 30...50 m).


c)

161 / 169 8. June 2010

Short to Medium-Span Bridges

Here, frequency-matching does not occur at either the vehicle body-bounce or axle-hop frequencies. Consequently: Bridges with natural frequencies f ≈ 4...8 Hz (maximum span L ≈ 15...30 m) do not respond significantly to the passage of heavy vehicles independent of their suspension system. d)

Short-Span Bridges

Here, both the vehicle body-bounce and axle-hop vibrations are important. Axle-hop vibrations only occur when the pavement longitudinal profile exhibits significant short-wavelength unevenness. Frequency-matching occurs at the vehicle axle-hop frequencies. The effects of the vehicle body-bounce vibrations are transmitted to the structure without amplification. Consequently: Short-span bridges with natural frequencies f = 8...15 Hz (maximum span L ≈ 8...12 m) may respond significantly to both, steel and air-suspended vehicles. This response is reduced in the case when the bridge damping capacity is high. The air-suspended vehicles induced smaller bridge response when axle-hop exciting features were not present in the road profile i.e. when only body-bounce vibrations were excited. When vehicle axle-hop was excited and the vehicles travelled at a critical speed (e.g. v ≈ 60 km/h) the air-suspended vehicles induced large dynamic responses which were repeated many times. This effect was less pronounced with steel-suspended vehicles. 8.3

The Parameters

8.3.1 Bridge Parameters

The parameters of interest in the bridge vehicle interaction problem are the bridge fundamental natural frequency and associated damping. These are determined by the bridge mass, stiffness and boundary conditions. Besides the natural frequencies which determines the frequency-matching effect, the level of bridge damping proved to be very significant for short-span bridges but not for medium to long span bridges. Medium to long span bridges show damping values ζ = 0.8...2% whereas for short-span bridges the range is ζ = 1...5%.


162 / 169 8. June 2010

8.3.2 Pavement Parameters

The surface profile of the bridge and its approaches proved to be fundamental to the response of the vehicle suspension and hence in turn the dynamic response of the bridge. In the case of a smooth bridge and approach profile, the influence of the vehicle suspension is insignificant. The pavements of the Swiss bridges and approaches proved to be very smooth to smooth whereas they were smooth to quite rough for the Australian bridges and approaches. There were bumps at the ends of the bridges in all cases. However, such bumps do not affect the dynamic response of medium to long span bridges whereas they prove to be very important for short-span bridges. There are two reasons for this: a) Such bumps induce high frequency axle-hop vibrations to which short-span bridges are susceptible only and, b) these bumpinduced vibrations damp off rather quickly if the suspension system damping is at the high level required. Specific suspension exciting features in the profile are important. Their location with respect to the bridge, their amplitude and wavelength as compared with the bridge and vehicle natural frequencies are the major factors that determine the magnitude and frequencies of the wheel loads and their damaging effects to the structure. When air-suspended vehicles travelled at critical speeds over axle-hop inducing features then large dynamic responses and multiple fatigue cycles were observed in a short-span bridge. These responses were up to 4.5 times the dynamic load allowance specified in bridge design codes. This phenomenon was less severe for steel-suspended vehicles. 8.3.3 Vehicle Parameters

As the unevenness of the profile increases, the importance of vehicle suspensions increases. Changing heavy vehicle suspensions from stiff and lowly-damped to soft and highly-damped will result in reduced dynamic wheel forces except when sharp discontinuities are encountered or when dampers are ineffective in their operation. The fact that there is very limited dynamic load sharing in air suspensions allows the axles in a group to vibrate in phase at axle-hop frequencies. "Cross-talk" between conventional steel suspensions limits their ability to vibrate in phase at axle-hop frequencies. This difference in behaviour was crucial for the strength of the dynamic response of short-span bridges to the crossing of air-suspended vehicles. It should be noted that for short-span bridges only multiple axle groups were tested and that the test vehicles were loaded at their legal loads. It is anticipated that single axle groups would


163 / 169 8. June 2010

produce large dynamic increments for such bridges over a wider range of speeds. Similarly multiple vehicle events and heavier vehicles affect the dynamic response of bridges. 8.4

Conclusions

As expected, the DIVINE experiment answered the questions formulated at the beginning of the project in part only. This is due to the fact that the cases of short-span bridges with a smooth bridge and approach profile and of medium to long span bridges with a rough bridge and approach profile could not be covered with the DIVINE tests. It was however found that: • For frequency-matching conditions at f ≈ 1.5...1.8 Hz: No excessive bridge vibrations are to be expected with air-suspended vehicles involved for smooth pavement conditions. • For frequency-matching conditions at f ≈ 10 Hz: Excessive bridge vibrations are to be expected with steel as well as air-suspended vehicles involved for average to rough pavement conditions. The situation is especially severe if the shock absorbers of airsuspended vehicles are ineffective. Questions remaining open are: • For frequency-matching conditions at f ≈ 1.5...1.8 Hz: What could the maximum dynamic increment be in the case of average to rough pavement conditions? • For frequency-matching conditions at f ≈ 10 Hz: What pavement quality and vehicle suspension damping performance is necessary to avoid excessive bridge vibrations? To answer these questions, further tests or analytical studies are necessary.


8.5

164 / 169 8. June 2010

Recommendations

Recommendation 1:

That guidelines be developed to facilitate the identification and ranking of bridge and approach profiles.

The maintenance of smooth approaches and profiles across bridges is a very important factor in reducing damage to bridges. Long wavelength roughness is important for all bridges in that it induces body-bounce forces which are transmitted with limited amplification in short-span bridges and are amplified by medium and long span bridges. Axle-hop inducing short wavelength roughness is less important on the main longitudinal elements of medium to long span bridges because no frequency-matching and consequently no quasi-resonance effects occur. However, the presence of short wavelength roughness such as pot holes and mis-aligned joints will result in axle-hop vibrations that are amplified by short-span bridges and short-span sub-elements of longer-span bridges. Bridges with a smooth road profile are largely insensitive to vehicle suspension. It is intuitively understood that a bumpy profile is not good for a bridge. However bridge design codes provide little or no advice to those designing or evaluating the strength of bridges or maintaining bridges in relation to the significance of the road profile. This research has shown that both road profile and bridge natural frequency and damping are important elements that can be controlled by those responsible for bridges. Road profile measurement is now routine. Signal processing technologies are being invoked to improve the quantitative summaries of the data. The measurement of dynamic wheel forces and bridge responses can also be achieved. Element 4 of the OECD DIVINE program has demonstrated that dynamic vehicle models can accurately reproduce dynamic wheel forces. Many researchers have also demonstrated that the vehicle/bridge interaction problem can be solved analytically. There is a need and the tools are in place to develop and validate bridge code models that include road profile and bridge natural frequency in the recommendations for dynamic load allowances. Recommendation 2:

That the influence of vehicle suspension on bridge response be investigated over a wide range of profile unevenness, vehicle configuration and level of overload.

The trend for heavier vehicles continues and the ageing infrastructure of bridges will be asked to carry increased loads. Thus the demand to a closer understanding of the vehicle/road profile/bridge system is increasing. The OECD DIVINE project investigation has been restricted to a limited number of vehicles and bridges. The research has however shown that the dynamic response of bridges is sensitive to subtle changes in the road profile, the type of bridge and the vehicle suspension. For example, the number of axles in a group, the


165 / 169 8. June 2010

effectiveness of dampers in air suspensions, and the position of short wavelengths in the road profile are likely to be important for short-span bridges. It is also known that the level of overload and the number of vehicles on a bridge both tend to decrease the dynamic effects in bridges. Recommendation 3:

That recommendations for bridge design codes be developed that express the dynamic load allowance as a function of road profile and bridge frequency.

Current codes provide a range of recommendations regarding the dynamic load allowance (DLA). These range from a constant value for the DLA to the DLA being a function of the bridge average span to the DLA being a function of the fundamental natural frequency of the bridge for frequencies up to 10 Hz (i.e. spans up to L ≈ 15 m). The road profile does not generally rate a mention despite being the fundamental parameter in determining the amount of dynamic effects in both short and medium span bridges. The Swiss code is an exception in that it defines an upper limit of roughness that is consistent with the bridge design code. The OECD DIVINE research has shown that both the profile and the bridge natural frequency are important in determining the dynamic response of both short and medium span bridges. Thus the current DLA versus frequency relationships should be extended to include shortspan bridges (i.e. with frequencies up to 15 Hz, say) and a range of road profiles. Recommendation 4:

That the effectiveness of the suspension damping of heavy vehicles in service be investigated.

There is evidence from the OECD DIVINE project that the damping of air suspensions is important in limiting axle-hop vibrations and consequently the dynamic response of short-span bridges in particular. Under some conditions the increase in fatigue damage can be large. However, the range of bridges that experience such damaging events is not well understood. There are many variables involved in determining the level of performance of dampers. These involve both the geometric arrangements of the damper as part of the suspension system and its damping performance. The technology to cost effectively determine if a damper is operating effectively whilst it is fitted to a heavy commercial vehicle is not yet available to the transport industry. It is hence quite possible that vehicles can operate with ineffective dampers, thus increasing the potential to damage bridges. The performance of dampers and the consequences on short-span bridges requires further investigation.


Recommendation 5:

166 / 169 8. June 2010

That short-span bridges be monitored for increased fatigue damage due to the introduction of soft suspensions.

One of the major findings of the OECD DIVINE research has been the large dynamic responses measured in short-span bridges. Large numbers of damaging fatigue cycles have been observed, especially where the road profile is uneven and the damping in soft suspensions is limited. This is potentially a significant danger for short-span bridges and short-span sub elements of longer-span bridges. This situation needs to be monitored through the measurement of the dynamic response of short-span bridges with a view to identifying any change in behaviour with the introduction of soft suspensions.


9.

ACKNOWLEDGMENTS

9.1

Swiss Research

167 / 169 8. June 2010

The authors wish to thank all persons having been involved in the planning, execution and processing of the tests reported here. Among many others, this includes the colleagues of NRC Canada, VTI Sweden, RST Sweden, TNO Holland, Rijkswaterstaat Holland, TRL United Kingdom, The Technical University of Prague, Czech Republic, AMP Hinwil, EMPA Dübendorf and, last but not least, the members of the OECD IR6 expert group and of the OECD RTR Secretariat, Paris. It was a special pleasure working together with the highly professional team of NRC in a quite complex project. These professionals included the most ingenious test driver, Ed Feser, the electronics specialists Ron Senn and Yves Séguin, the computer specialist and engineer in charge at the front, Pierre LeBlanc, the mostly back-stage operating chief-manager, Scotty MacLeod and, last but not least, John Woodrooffe, whose enthusiasm was essential for making the whole operation happen. Thank you very much, all of you! Thanks to Sara Barella and Vladimir Bily (Technical University of Prague) who carried out the time domain analysis of the bridge response signals. Thanks also to Andreas Felber (author of the software packages used) who initially supported Walter Krebs, being responsible for the frequency domain analysis of the wheel load signals and the analysis of the vehicle/bridge interaction processes. We are also indebted to the Highway Administrations of the Cantons Zug and Ticino for helping us to solve the traffic management problems free of any charge. For financial support thanks are due to the EMPA Research Council, the EMPA Board of Directors and the Swiss Federal Highway Administration representing the Swiss Delegation at OECD RTR. Special thanks to Prof. Eggimann, President of EMPA, for his interest and his support in this work. 9.2

Australian Research

This research is the outcome of the efforts of a group of people dedicated to improving the utility of Australia’s bridges and the productivity of the Australian transport system without jeopardizing safety and longevity. The research would not have been conducted without the ongoing support of the OECD DIVINE Australian Reference Group and its chairman Mr Geoff Youdale and Mr Mike Veysey.


168 / 169 8. June 2010

The research has benefited greatly from being a part of the multi-disciplinary OECD DIVINE (Dynamic Interaction between Vehicles and Infrastructure Experiment) International Scientific Group. The contribution of its Chairman - Dr Peter Sweatman of Road User Research and all the members of the group has been important in broadening the vision and application of the research. The groups knowledge of vehicles and suspensions have been of fundamental assistance. The leadership of Dr Reto Cantieni of Switzerland has been invaluable. The Australian research was funded independently of the OECD DIVINE budget. It comes as an in-kind contribution by Australian governments and industry to the DIVINE research project. The field research was funded by the Roads and Traffic Authority (RTA) of New South Wales and VicRoads with in-kind contributions from the National Road Transport Commission, Boral Transport (trucks), Transpec Australia (air suspension), BPW Germany (instru-mented axles), York Australia (steel suspensions), Hamelex Australia (suspension fabrication), Mercedes Benz (video and trucks), Heggies Bulkhaul (workshop), Blastronics (data acquisition and field support). The Australian Road Research Board provided signal conditioners for the instrumented axles and Blastronics provided data acquisition systems. Queensland Transport provided support for the final preparation of this report. Throughout, the Queensland University of Technology (QUT) has provided the environment, the technical and equipment support necessary to conduct this research. Special thanks must be expressed to Professors Keith Wallace and Rod Troutbeck who encouraged the continued involvement in the OECD DIVINE research. The leadership and support of Mr Ray Wedgwood, Chief Bridge Engineer, Roads and Traffic Authority of New South Wales has been of fundamental importance. Mr Wije Ariyaratne, Mr Rod Oates and their RTA teams, Mr Geoff Boully and Mr Armando Giufre and their VicRoads team, Mr Paul Urquhart and Boral Transport staff helped the field work flow very smoothly. Difficult traffic control situations were professionally resolved and trucks suspensions changed with ease and co-operation. Sorin Moldoveanu (QUT) has participated in the program from its inception. He has undertaken much of the data analysis for both the bridge and vehicle response data. Arthur Powell provided expert advice regarding instrumentation ably assisted by the QUT team. Steve Hickey (Blastronics) went the extra mile to ensure quality data was collected and stored in their data acquisition systems. Their assistance is gratefully acknowledged. Mr Bob Pearson of the National Road Transport Commission, Mr Ray Wedgwood, Mr Warwick Pattinson of VicRoads and other members of the Austroads Project Team 3.E.51 have provided the environment and the assistance necessary to apply the knowledge to the bridge network.


169 / 169


10.

8. June 2010

LIST OF ANNEXES No. of Pages

Annex A

Sort Bridge

19

Annex B

Deibüel Bridge

21

Annex C

Föss Bridge

20

Annex D

Lawsons Creek Bridge

16

Annex E

Coxs River Bridge

10

Annex F

Camerons Creek Bridge

11

Annex G

Cromarty Creek Bridge

10

Annex H

Details of Test Vehicles

20

Annex I

References

5

Total No. of Pages Main Report

169

Total No. of Pages Main Report plus Annexes

302



OECD IR 6 DIVINE Project Element 6, Bridge Research

Annex A Sort Bridge (CH)


A.2 / 19

Final Report, EMPA Switzerland, QUT Australia, Annex A

8. June 2010

Contents A.1 DESCRIPTION OF THE SORT BRIDGE ............................................................................3 A.1.1 Location ........................................................................................................................3 A.1.2 Geometry, Material .......................................................................................................3 A.1.3 Piers, Bearings..............................................................................................................5 A.2 AMBIENT VIBRATION TEST OF THE SORT BRIDGE.......................................................6 A.2.1 General .........................................................................................................................6 A.2.2 Test Description ............................................................................................................6 A.2.2.1 Instrumentation.......................................................................................................6 A.2.2.2 Measurement Set-ups ............................................................................................6 A.2.2.3 Signal Acquisition ...................................................................................................7 A.2.2.4 Signal Processing...................................................................................................7 A.2.3 Results ..........................................................................................................................8 A.2.3.1 Natural Frequencies ...............................................................................................8 A.2.3.2 Vertical Bending Mode Shapes ............................................................................10 A.3 BRIDGE INSTRUMENTATION FOR THE DIVINE TESTS ...............................................11 A.4 PAVEMENT PROFILES FOR THE DIVINE TESTS ..........................................................12 A.5 RESULTS OF THE DIVINE TESTS ON THE SORT BRIDGE ..........................................14 A.5.1 Bridge Response Time Signals ...................................................................................14 A.5.2 Dynamic Increments ...................................................................................................17


A.3 / 19


A.1 A.1.1

8. June 2010

DESCRIPTION OF THE SORT BRIDGE Location

The Sort bridge is located near Ambri-Piotta in the Canton Tessin where it carries the two-lane Cantonal Road over the National Highway N2 Basel-Chiasso and the Ticino River. A photograph of the bridge is given in the main report Paragraph 2, Test Parameters, p. 33. A.1.2

Geometry, Material

Length along the bridge centerline

258.79 m

Number of spans Length of spans [m]

5 36.0, 58.42, 69.95, 58.42, 36.0

Longitudinal slope [%]

2.02

Width of carriageway [m]

11.0

Number of lanes

2 plus a pedestrian walkway on one side

Horizontal curvature R [m] Cross section type Box girder height in [m] Material Table A- 1

900 single cell box girder 2.0 ÷ 2.8, variable Prestressed, post-tensioned concrete

Main Geometrical Parameters of the Sort Bridge.

The mass and stiffness distribution of the Sort Bridge are quite complicated because the cross-sectional dimensions as well as the box girder height are not constant over the bridge length. Detailed information is presented in Fig. A- 1 and Fig. A- 2.


A.4 / 19 8. June 2010

258.79

Ticino 5.4

69.95

2.8

P3

P3

¯ 1.6

58.42

P4

P4

36.00

S2

S2 Airolo

Airolo S.Gottardo

S.Gottardo


11.0

R = 900 P1

14.3

P1

58.42

P2

P2

10

A = 272.807 A = 272.807 Bellinzona

Bellinzona

S1

S1

36.00

A = 272.807

Fig. A- 1 Sort Bridge longitudinal section and plan view (dimensions: m).

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia, Annex A

A.5 / 19 8. June 2010

Fig. A- 2 Sort Bridge mid-span cross section (dimensions: m, H = 2.0...2.8 m). A.1.3 Piers, Bearings All piers are circular, reinforced concrete columns with a diameter of 1.6 m. The individual piers have the following lengths: P1: 5.5 m, P2: 6.0 m, P3: 12.0 m, and P4: 11.5 m. The bridge is supported on roller bearings which permit longitudinal movement at all supports except at pier P2 where longitudinal restraint is provided. Pier P2 is equipped with a single fixed bearing, piers P1, P3 and P4 with a single roller bearing. The two roller bearings located at a 8.9 m distance at both abutments are the only points where torsional forces can be taken.


A.2 A.2.1

A.6 / 19 8. June 2010

AMBIENT VIBRATION TEST OF THE SORT BRIDGE General

The theory of Ambient Vibration Testing, AVT, is described in detail in [Felber, 1993]. Further details on the technology as used by EMPA and on the tests performed on the Sort Bridge can be taken from [EDI, 1995] and [Felber, 1995a]. For the tests performed on the Sort Bridge, ambient vibrations were measured at 125 points over a period of two days. Main sources of ambient excitation included traffic moving over the bridge, air pressure waves generated by the traffic moving on the N2 underneath the bridge and wind. From the acceleration time records a total of 19 modes of vibration were identified. The lowest transverse mode was f1 = 1.04 Hz and the first vertical mode was f2 = 1.62 Hz. The identified dynamic characteristics can now be used to calibrate a computer model of the bridge. A.2.2

Test Description

A.2.2.1 Instrumentation The bridge vibrations were measured using Brüel & Kjaer 8306 accelerometers with a sensitivity of 10 V/g and a resolution of 10-6 m/s2. The signals produced by the accelerometers were amplified using an EMPA made pre-amplifier. The same pre-amplifier provided also the 28 V DC-signal used to drive the accelerometers. The pre-amplified signals were then filtered and amplified using an eight channel DIFA/SCADAS front-end. Finally the signals were amplified and digitized using a 16 bit Keithley 575 Module with a AMM2 board. A.2.2.2 Measurement Set-ups The 125 locations where the vertical bridge vibrations were measured were spread approximately uniformly over the length of the bridge. For the short end spans the measurement points were placed at every quarter of the span and for the intermediate spans they were placed at every fifth of the span. Finally for the central span they were placed at every sixth of the span. In the transverse direction three measurements were taken; one at each side of the box on the bottom flange and the third measurement at the outer edge of the sidewalk. The location of the individual measurement points and their numbering can be seen from Fig. A- 3.


A.7 / 19 8. June 2010

Fig. A- 3 Location and Numbering of Measurement Points. The measurement campaign encompassed 34 measurement set-ups whereby the measurement points 64 and 44 (both vertical and transversal) were chosen as reference points. As a consequence, four accelerometers remained as instruments being roved over structure until the complete measurement point grid was covered. A.2.2.3 Signal Acquisition The technical parameters of the signal acquisition procedure were: • Sampling rate s = 20 Hz • Frequency range f = 0...10 Hz • Low Pass Filter: 12.5 Hz • Number of points per Channel: 32 k (8 segments of 4'096 points) In total 272 acceleration time histories were recorded at 125 measurement locations. This corresponds to 34 MB (in binary format). A.2.2.4 Signal Processing The recorded accelerations were analysed using the programs U2, V2 and P2 [EDI, 1995]. The program P2 was used to generate Averaged Normalized Power Spectral Densities, ANPSD's, for the identification of natural frequencies. The program U2 was used to compute the Modal Ratios from the recorded data. V2 was subsequently used to visualise and animate the deflected shapes corresponding to the computed Modal Ratios. These programs were used for preliminary data analysis on site and for further detailed data analysis at the office. To avoid the possibility of overlooking a natural frequency because the computed PSD might have been recorded on a location corresponding to a modal node, PSD's were computed for all recordings. First, all of these PSD's were normalised to give them equal weight, after which they were averaged. The resulting ANPSD computed from all vertical records was then used to identify all the vertical modes.


A.2.3

A.8 / 19 8. June 2010

Results

A.2.3.1 Natural Frequencies The natural frequencies of the structure were identified using the ANPSD functions described above. As an example, the ANPSD for all vertical signals is given in Fig. A- 4. (Other ANPSD's mentioned below are given in [Felber, 1995a].) Furthermore, for the natural frequencies' identification, ANPSD's for the added and subtracted vertical signals allowing easy differentiation between flexural and torsional modes were also used. With the technology available at EMPA, all time signals are stored on hard disk and hence can subsequently be arithmetically treated according to the requirements. Finally, the transverse bending modes were identified based on the ANPSD for the measured transverse signals. The ANPSD for all the vertical signals (Fig. A- 4) shows dominant peaks at f = 1.62, 2.45, 2.98 and 5.77 Hz. Minor peaks can be seen at f = 4.31, 4.51, 4.60, 7.24, 7.3 and 8.84 Hz. The ANPSD for all the transverse signals indicates that the dominant transverse modes are f = 1.04, 1.73, 2.55, 2.85, 2.98 Hz. In addition minor peaks which are not seen in the vertical ANPSD are indicated at f = 5.68, 6.13 and 8.08 Hz.

Fig. A- 4 ANPSD for the Sort Bridge vertical acceleration signals. All the identified natural frequencies are listed inTable A- 2. Being of primary interest in the present context, the following detailed discussion is restricted to the the vertical bending modes. A complete discussion can be found in [Felber, 1995a].


A.9 / 19


8. June 2010

Mode No.

Frequency [Hz]

Period [sec]

Description

1

1.04

0.97

1st Transverse

2

1.62

0.62

1st Vertical Bending

3

1.73

0.58

2nd Transverse

4

2.45

0.41

2nd Vertical Bending

5

2.55

0.39

3rd Transverse

6

2.85

0.36

1st Torsion

7

2.98

0.33

3rd Vertical Bending

8

4.04

0.25

4th Transverse

9

4.31

0.23

2nd Torsion

10

4.51

0.22

4th Vertical Bending

11

4.60

0.22


12

5.68

0.18

5th Transverse

13

5.77

0.17


14

6.13

0.16

3rd Torsion

15

7.24

0.14

4th Torsion

16

7.30

0.14


17

7.65

0.13


18

8.08

0.12

6th Transverse

19

8.84

0.11

5th Torsion

Table A- 2 Sort Bridge; Identified Natural Frequencies and Associated Mode Type.


A.10 / 19 8. June 2010

A.2.3.2 Vertical Bending Mode Shapes Elevation views of the eight identified vertical bending modes are shown in Fig. A- 5. All of these modes were clearly identified, however, for the higher modes the shapes are not as well defined because the measurement grid spacing was quite coarse for these modes. Of special interest is the fact that the frequency of Mode 2, f2 = 1.62 Hz, is close to the dynamic wheel load frequency of an air-suspended heavy vehicle. Furthermore, it has to be noted that the frequency of Mode 7, f7 = 2.98 Hz, is close to that of a steel-suspended vehicle. The shape of Mode 7 is quite simple which makes it easy to be excited by a passing vehicle. Some vehicle/bridge interaction in the 3 Hz-region can hence be expected.

Fig. A- 5 Elevation view of the Sort Bridge vertical modes shapes.


A.3

A.11 / 19 8. June 2010

BRIDGE INSTRUMENTATION FOR THE DIVINE TESTS

Fig. A- 6 Sort Bridge instrumentation: Five vertical displacement measurements. The lane used by the test vehicle (always in direction Airolo-Bellinzona) is also indicated.

Fig. A- 7 Sort Bridge instrumentation: Test track defined by white strips (Nos. 0...16) glued to the pavement surface at intervals of 30 m and at intervals of 0.19 m at the beginning and end of the track. The EMPA contact thresholds No. 1...3 are also indicated in the figure.


A.12 / 19


A.4

8. June 2010

PAVEMENT PROFILES FOR THE DIVINE TESTS

The methodology used to determine the pavements longitudinal profiles is described in Section 2.2.1 of the report’s main part. The profiles for the left and right wheel tracks are presented in the following figures. It can be seen from Fig. A- 10 that the differences in the profiles for the two wheel paths are generally small. From the longitudinal profiles, pavement unevenness spectra were calculated. These are also presented graphically in Fig. A- 11 and Fig. A- 12. The unevenness spectra allow a classification of the pavement evenness according to [ISO 1995]. Sort left 20

Profile [mm]

15 10 5 0 -5 -10 -15 -20 0

100

200

300

400

500

400

500

Distance [m] Fig. A- 8 Longitudinal profile in the lefthand wheel path.

Profile [mm]

Sort right 25 20 15 10 5 0 -5 -10 -15 -20 -25 0

100

200

300

Distance [m] Fig. A- 9 Longitudinal profile in the righthand wheel path.


A.13 / 19


8. June 2010

Sort 20

15

profile [mm]

10 sort mr1

5

sort ml1 sort mr2

0 0

50

100

150

200

250

300

350

400

srt ml2

-5

-10

-15 distance [m ]

Fig. A- 10 Longitudinal profiles in both wheelpaths.

Fig. A- 11 Spectrum of the left track pavement longitudinal profile; classification according to [ISO 1995].


A.14 / 19 8. June 2010

Fig. A- 12 Spectrum of the right track pavement longitudinal profile; classification according to [ISO 1995].

A.5 A.5.1

RESULTS OF THE DIVINE TESTS ON THE SORT BRIDGE Bridge Response Time Signals

For the purpose of illustration, Fig. A- 13 and Fig. A- 14 give a set of five vertical (one per bridge span) and the horizontal bridge dynamic deflection for a passage of the steel and airsuspended NRC vehicle respectively. The graphs, i.e. the vehicle speed, were chosen to allow a representative insight into the dynamic bridge behavior. It can be seen from Fig. A- 13 that the dynamic bridge response is close to nil for the steelsuspended vehicle in the large middle span (WG 32, L = 70 m). However, significant dynamic response with f ≈ 3 Hz can be observed in all other spans. This means that the Sort Bridge shows a relatively rare vehicle/bridge interaction with a higher than the fundamental bridge mode involved. Fig. A- 14 shows a quite significant dynamic interaction between the air-suspended vehicle and the bridge. As expected, the fundamental bridge bending mode, f = 1.62 Hz, is involved here.


A.15 / 19 8. June 2010

1s/div WG 12 1.5 mm

WG 22 3.0 mm

WG 32 3.0 mm

WG 41 2.5 mm

WG 52 1.5 mm

0.5 mm

H

Contact threshold's tickmarks

Fig. A- 13 Sort Bridge time signals for the crossing of the steel-suspended NRC vehicle at a speed v = 29.9 km/h.


A.16 / 19


8. June 2010

1s/div WG 12 1.5 mm

WG 22 3.0 mm

W G 32 5.0 mm

WG 41 2.5 mm

WG 52 1.5 mm

H 0.5 mm Contact threshold's tickmarks

Fig. A- 14 Sort Bridge time signals for the crossing of the air-suspended NRC vehicle at a speed v = 75.2 km/h.


A.17 / 19


A.5.2

8. June 2010

Dynamic Increments

The definition of the dynamic increment φ as used throughout this report can be found in Chapter 1, Introduction, p. 23. Here, the results of the DIVINE tests, i.e. the φ(v)-functions are given for the five vertical displacement measurement points as given in Fig. A- 6. Each of the graphically displayed functions allows direct comparison of the results for the tests performed with the NRC-vehicle (see Annex H) equipped with air and steel suspensions respectively.




20 STEEL AIR

15

10

5

0 0

20

40

60


Fig. A- 15 Dynamic Increments measured for point WG 12.

80


A.18 / 19


8. June 2010



Static deflection: 4.17 mm 25 20 STEEL AIR

15

10

5

0 0

20

40

60

80




Static deflection: 5.68 mm Dynamic Increment [%]

25

20 STEEL AIR

15

10

5

0 0

20

40

60



80


A.19 / 19


8. June 2010




15 STEEL AIR 10

5

0 0

20

40

60

80






15 STEEL AIR 10

5

0 0

20

40

60



80




Annex B Deibüel Bridge (CH)


B.2 / 21

Final Report, EMPA Switzerland, QUT Australia, Annex B

8. June 2010

Contents B.1 DESCRIPTION OF THE DEIBÜEL BRIDGE .......................................................................3 B.1.1 Location ........................................................................................................................3 B.1.2 Geometry, Material .......................................................................................................3 B.1.3 Piers, Bearings..............................................................................................................4 B.2 AMBIENT VIBRATION TEST OF THE DEIBÜEL BRIDGE .................................................5 B.2.1 General .........................................................................................................................5 B.2.2 Test Description ............................................................................................................5 B.2.2.1 Instrumentation.......................................................................................................5 B.2.2.2 Measurement Set-ups ............................................................................................5 B.2.2.3 Signal Acquisition ...................................................................................................6 B.2.2.4 Signal Processing...................................................................................................6 B.2.3 Results ..........................................................................................................................7 B.2.3.1 Natural Frequencies ...............................................................................................7 B.2.3.2 Mode Shapes .........................................................................................................8 B.2.4 Comparison with Previous Test Results........................................................................9 B.3 BRIDGE INSTRUMENTATION FOR THE DIVINE TESTS ...............................................11 B.4 PAVEMENT PROFILES FOR THE DIVINE TESTS ..........................................................12 B.5 RESULTS OF THE DIVINE TESTS ON THE DEIBÜEL BRIDGE .....................................15 B.5.1 Bridge Response Time Signals ...................................................................................15 B.5.2 Dynamic Increments ...................................................................................................18


B.3 / 21


B.1 B.1.1

8. June 2010

DESCRIPTION OF THE DEIBÜEL BRIDGE Location

The Deibüel Bridge is located near Baar (Canton Zug) where it carries two lanes of the N4a National Highway between Zürich and Lucerne over a small valley and river. The two other lanes in the Lucerne-Zürich direction are making use of a different bridge which is separated from the bridge under discussion by 3 m horizontally and 2 m vertically. This bridge has also five spans instead of the three-span structure under discussion. A photograph of the bridge is given in the main report Paragraph 2, Test Parameters, p. 34. B.1.2

Geometry, Material

Length [m] Number of spans Lenght of spans [m]

110.3 3 36.95, 41.0, 32.35


2.36


11.75

Horizontal curvature R [m] Cross section type Box girder height [m] Material

0 single cell box girder 1.8 Prestressed, posttensioned concrete

Fig. B- 1 Deibüel Bridge longitudinal section and plan view (dimensions: m).

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia, Annex B

B.4 / 21 8. June 2010

Fig. B- 2 Deibüel Bridge mid-span cross section (dimensions: m). B.1.3 Piers, Bearings The two piers have a rectangular cross section of 3.0 x 0.90 m. Their length is 3 m and 8 m respectively. At the abutments, two PTFE pot bearings allow free movement of the bridge in the horizontal longitudinal direction. The short pier is clamped into the superstructure whereas the connection between the longer pier and the superstructure is designed as a concrete hinge. Experience however showed that correct modeling of the bridge requires clamping in of this pier also. This makes the bridge behave as a frame.

Fig. B- 3 Deibüel Bridge theoretical static system and boundary conditions.


B.2 B.2.1

B.5 / 21 8. June 2010

AMBIENT VIBRATION TEST OF THE DEIBÜEL BRIDGE General

The theory of Ambient Vibration Testing, AVT, is described in detail in [Felber, 1993]. Further details on the technology as used by EMPA and on the tests performed on the Deibüel Bridge can be taken from [EDI, 1995] and [Felber, 1995b]. Ambient vibrations were measured at 65 points over a period of one day. Main sources of ambient excitation included traffic moving over the bridge and wind. From the acceleration time records a total of 11 modes of vibration were identified. The first vertical mode was f1 = 3.01 Hz and the lowest transverse mode was f2 = 3.79 Hz. The identified dynamic characteristics can now be used to calibrate a computer model of the bridge. The bridge was known from earlier tests to significantly interact with heavy commercial vehicles yielding shifts in the frequency of the fundamental mode of as much as Δf = 0.25 Hz. It was therefore of special interest to investigate the effect of this interaction on the results of an ambient vibration test. Experience showed, against the expectations of some experts involved that this interaction between bridge and vehicle vibrations doesn’t cause any problems in the identification of the primary bending modes. The amount of time where the vibration frequency of the bridge is shifted due to interaction effects seems to be small in comparison with the time where only cars or no vehicles at all are crossing the bridge. B.2.2

Test Description

B.2.2.1 Instrumentation The bridge vibrations were measured using Kinemetrics FBA-11 and FBA-23 accelerometers with a sensitivity of 10 V/g and a resolution of 10-6 m/s2. The signals produced by the accelerometers were amplified using an EMPA made pre-amplifier. These pre-amplified signals were then filtered and amplified using an eight channel DIFA/SCADAS front-end. Finally the signals were amplified and digitized using a 16 bit Keithley 575 Module with a AMM2 board. B.2.2.2 Measurement Set-ups The bridge carries two lanes of highway traffic and one emergency lane. The sensors were mounted to the railings on both sides of the bridge deck. Local highway and police authorities insisted that the emergency lane and the outer traffic lane were closed in order to reduce the risk of accidents during the measurements. This closure was in effect from 8 am to 4 pm. The upstream side of the bridge was instrumented with triaxial accelerometers and the downstream side was instrumented with uniaxial vertical accelerometers. The individual sensor


B.6 / 21 8. June 2010

locations are shown in Fig. B- 4. The solid squares indicate the location of the four reference sensors and the circles denote the locations covered by the roving instruments. A more detailed description of the individual measurement setups is given in [Felber, 1995b].

Fig. B- 4 Schematic plan of the sensor locations of the Deibüel bridge. B.2.2.3 Signal Acquisition The data was acquired using the following settings: • Sampling Rate s = 80 Hz • Low Pass Filter: 12.5 Hz • Attenuation: 42db or 36 dB (123.8 or 62.1 mm/s2/V) • Number of points per Channel: 32 k (8 segments of 4’096 points) In total 288 acceleration time histories were recorded at 65 measurement locations. This corresponds to 36 MB (in binary format). B.2.2.4 Signal Processing The recorded accelerations were analysed using the programs U2, V2 and P2 [EDI, 1995]. The program P2 was used to generate Averaged Normalized Power Spectral Densities, ANPSD's, for the identification of natural frequencies. The program U2 was used to compute the Modal Ratios from the recorded data. V2 was subsequently used to visualise and animate the deflected shapes corresponding to the computed Modal Ratios. These programs were used for preliminary data analysis on site and for further detailed data analysis at the office. To avoid the possibility of overlooking a natural frequency because the computed PSD might have been recorded on a location corresponding to a modal node, PSD's were computed for


B.7 / 21 8. June 2010

all recordings. First, all these PSD's were normalised to give them equal weight, after which they were averaged. The resulting ANPSD's computed from all vertical records was then used to identify all the vertical modes. B.2.3

Results

B.2.3.1 Natural Frequencies The natural frequencies of the structure were identified using the ANPSD functions described above. As an example, the ANPSD for all vertical signals is given in Fig. B- 5. (Other ANPSD's mentioned below are given in [Felber, 1995b].) Furthermore, for the natural frequencies' identification, ANPSD's for the added and subtracted vertical signals allowing easy differentiation between flexural and torsional modes were also used. With the technology available at EMPA, all time signals are stored on hard disk and hence can subsequently be arithmetically treated according to the requirements. Finally, the transverse bending modes were identified based on the ANPSD for the measured transverse signals. The ANPSD for all the vertical accelerations recorded at deck level shown in Fig. B- 5 indicates that the first three vertical modes have natural frequencies of f = 3.01, 4.24, and 5.43 Hz. In the region of f ≈ 6...9 Hz three very small peaks can be observed and finally two more dominant peaks appear at f = 10.02 Hz and 11.39 Hz.

Fig. B- 5 ANPSD of the Deibüel Bridge vertical acceleration signals.


B.8 / 21 8. June 2010

B.2.3.2 Mode Shapes A total of eleven modes could be identified from the ambient vibration data. These mode shapes are depicted in elevation and plan in Fig. B- 6.

Fig. B- 6

Elevation and plan view of the Deibüel Bridge mode shapes.


B.9 / 21 8. June 2010

Of primary interest in the present context are the three first bending modes, f1 = 3.01 Hz, f2 = 4.24 Hz and f3 = 5.43 Hz. This means that vehicle/bridge interaction can be expected for steel-suspended vehicles. A discussion of the other modes shown in Fig. B- 6 can be found in [Felber, 1995b]. B.2.4 Comparison with Previous Test Results The Deibüel Bridge has been the subject of several analytical and experimental dynamic investigations since the late 1970's [Cantieni, 1983]. In 1977 a dynamic computer model of the bridge was created based originally on the technical drawings available [Cantieni, 1983]. Then later in 1977 and 1978 forced vibration tests with a servohydraulic shaker were performed to determine the dynamic characteristics of the bridge with and without the concrete pavement [Cantieni, 1992]. In the context of the DIVINE project, an ambient vibration study was performed on the bridge in March 1995 firstly to investigate if the bridge fundamental natural frequency had changed in the meantime. In addition, the mode shapes and frequencies determined from the ambient vibration test could be compared with those obtained in 1977 and 1978. Finally, the more detailed mode shapes obtained from the ambient vibration study can now be used to calibrate analytical models which will be used to simulate vehicle/bridge interaction in order to predict dynamic amplification factors. The frequencies obtained from the ambient vibration test in 1995 and the results obtained from the forced vibration tests performed by Cantieni on the paved bridge in 1978 [Can-tieni, 1992] are listed in Table B- 1. These frequencies compare quite favourably. The ambient frequencies are generally slightly lower than those observed seventeen years earlier. This may be due to some deterioration such as normal micro-cracking. The results obtained from the tests in 1977 for the bridge without pavement (Table B- 1) do not warrant a close comparison.


B.10 / 21


Mode

Mode Description

8. June 2010

Frequency [Hz] Ambient 1995

Forced 1978 with pavement

Forced 1977 no pavement

1

1st vertical bending

3.01

3.03

3.07...3.11

2

1st transverse bending

3.79

3.72

not reported

3

2nd vertical bending

4.24

4.28

4.34...4.36

4

3rd vertical bending

5.43

5.49

5.53...5.60

5

2nd transverse bending

6.66

6.70

not reported

6

1st torsion

7.52

7.50

not reported

7

3rd transverse bending

8.57

8.70

8.2

8

4th vertical bending

10.02

10.0

9.79

9

5th vertical bending

11.39

not reported

10.2

10

combined bending & torsion

12.09

12.2

not reported

11

combined bending & torsion

13.89

14.1

13.5

Table B- 1

Comparison of Ambient and Forced Vibration Test Results for the Deibüel Bridge.


B.3

B.11 / 21 8. June 2010

BRIDGE INSTRUMENTATION FOR THE DIVINE TESTS

Fig. B- 7 Deibüel Bridge instrumentation: Six vertical displacement transducers. The lane used by the test vehicle (always in direction Sihlbrugg-Baar, i.e. from the right to the left) is also indicated.

Fig. B- 8

Deibüel Bridge instrumentation: Test track defined by white strips (Nos. 0...10) glued to the pavement surface at intervals of 30 m and at intervals of 0.19 m at the beginning and end of the test track (triggering of the data acquisition system onboard the test vehicle). The EMPA contact thresholds Nos. 1...3 are also indicated in the figure.


B.4

B.12 / 21 8. June 2010


The methodology used to determine the pavements longitudinal profiles is described in Section 2.2.1 of the report’s main part. Fig. B- 9 shows the laser profilometer as used to measure the pavement profiles. The profiles for left and right wheel tracks are presented in the following figures. It can be seen from Fig. B- 12 that the differences in the profiles for the two wheel paths are generally small. From the longitudinal profiles, pavement unevenness spectra were calculated. These are also presented graphically (Fig. B- 13, Fig. B- 14). The unevenness spectra allow a classification of the pavement evenness according to [ISO, 1995].

Fig. B- 9 The Laser RST portable profilometer fixed to an EMPA car, here on the Deibüel test track (EMPA Photo No. 101'176/11).


B.13 / 21


Deibüel left 15

Profile [mm]

10 5 0 -5 -10 -15

0

20 40 60 80 100 120 140 160 180 200 220 240 260 280

Distance [m] Fig. B- 10 Longitudinal profile in the lefthand wheel path.

Deibüel right

Profile [mm]

15 10 5 0 -5 -10 0

20 40 60 80 100 120 140 160 180 200 220 240 260 280

Distance [m] Fig. B- 11 Longitudinal profile in the righthand wheel path.

8. June 2010


B.14 / 21


8. June 2010

Deibüel 15

10

profile [mm]

5 deibüel_r

0 0

50

100

150

200

deibüel_l

-5

-10

-15 distance [m]

Fig. B- 12 Longitudinal profiles in both wheel paths.

Fig. B- 13 Spectrum of the left track pavement longitudinal profile; classification according to [ISO, 1995].


B.15 / 21 8. June 2010

Fig. B- 14 Spectrum of the right track pavement longitudinal profile; classification according to [ISO, 1995].

B.5 B.5.1

RESULTS OF THE DIVINE TESTS ON THE DEIBÜEL BRIDGE Bridge Response Time Signals

For the purpose of illustration, Fig. B- 15 and Fig. B- 16 give a set of the six vertical (two per bridge span) and the horizontal bridge dynamic deflection for a passage of the steel and airsuspended NRC vehicle respectively. The graphs, i.e. the vehicle speed, were chosen to allow a representative insight into the dynamic bridge behavior. It can be seen from Fig. B- 15 that the dynamic bridge response is quite significant for the steel-suspended vehicle in all bridge spans. This means that, as expected, the Deibüel Bridge interacts with the steel-suspended vehicle. Fig. B- 16 shows that a significant dynamic interaction between the air-suspended vehicle and the bridge, again as expected, does not exist.


B.16 / 21


8. June 2010

1s/div WG 11 1.5 mm

WG 12 1.5 mm

WG 21 2.5 mm

WG 22 2.5 mm

WG 31 2.5 mm

WG 32 2.5 mm

H 0.2 mm


Fig. B- 15 Deibüel Bridge time signals for the crossing of the steel-suspended NRC vehicle at a speed v = 50.6 km/h.


B.17 / 21


8. June 2010

1s/div WG 11 1.5 mm

WG 12 1.5 mm

WG 21 2.5 mm

WG 22 2.5 mm

WG 31 2.5 mm

WG 32 2.5 mm

H 0.2 mm


Fig. B- 16 Deibüel Bridge time signals for the crossing of the air-suspended NRC vehicle at a speed v = 37.3 km/h.


B.5.2

B.18 / 21 8. June 2010

Dynamic Increments

The definition of the dynamic increment φ as used throughout this report can be found in Chapter 1, Introduction, p. 23. Here, the results of the DIVINE tests, i.e. the φ(v)-functions are given for three of the six vertical displacement measurement points (one per bridge span, the results for their counterparts in the same span are similar) as given in Fig. B- 7. Each of the graphically displayed functions in Fig. B- 18 to Fig. B- 22 allows direct comparison of the results for the tests performed with the NRC-vehicle (see Annex H) equipped with air and steel suspensions respectively.


B.19 / 21


8. June 2010



25

20 STEEL

15

AIR

10

5

0 0

10

20

30

40

50

60

70

80


Fig. B- 17 Dynamic increments measured for point WG 11. DEIBÜEL Bridge WG 12 30


25

20 STEEL

15

AIR

10

5

0 0

10

20

30

40

50

60


Fig. B- 18 Dynamic increments measured for point WG 12.

70

80


B.20 / 21


8. June 2010



25

20 STEEL

15

AIR

10

5

0 0

10

20

30

40

50

60

70

80


Fig. B- 19 Dynamic Increments measured for point WG 21.



25

20 STEEL AIR

15

10

5

0 0

10

20

30

40

50

60


Fig. B- 20 Dynamic Increments measured for point WG 22.

70

80


B.21 / 21


8. June 2010



25

20 STEEL AIR

15

10

5

0 0

10

20

30

40

50

60

70

80


Fig. B- 21

Dynamic Increments measured for point WG 31. DEIBÜEL Bridge WG 32 30


25

20 STEEL AIR

15

10

5

0 0

10

20

30

40

50

60


Fig. B- 22

Dynamic Increments measured for point WG 32.

70

80




Annex C Föss Bridge (CH)


C.2 / 20

Final Report, EMPA Switzerland, QUT Australia, Annex C

May 30, 2000

Contents C.1 DESCRIPTION OF THE FÖSS BRIDGE ............................................................................3 C.1.1 Location ........................................................................................................................3 C.1.2 Geometry, Material .......................................................................................................3 C.1.3 Piers, Bearings .............................................................................................................5 C.2 AMBIENT VIBRATION TESTS OF THE FÖSS BRIDGE ....................................................6 C.2.1 General .........................................................................................................................6 C.2.2 Test Description ............................................................................................................6 C.2.2.1 Instrumentation ......................................................................................................6 C.2.2.2 Measurement Set-ups ............................................................................................6 C.2.2.3 Signal Acquisition ...................................................................................................7 C.2.2.4 Signal Processing...................................................................................................7 C.2.3 Results..........................................................................................................................8 C.2.3.1 Natural Frequencies ...............................................................................................8 C.2.3.2 Mode Shapes .........................................................................................................9 C.3 BRIDGE INSTRUMENTATION FOR THE DIVINE TESTS ...............................................11 C.4 PAVEMENT PROFILES FOR THE DIVINE TESTS ..........................................................12 C.5 RESULTS OF THE DIVINE TESTS ON THE FÖSS BRIDGE ..........................................15 C.5.1 Bridge Response Time Signals...................................................................................15 C.5.2 Dynamic Increments ...................................................................................................18


C.3 / 20


C.1 C.1.1

May 30, 2000

DESCRIPTION OF THE FÖSS BRIDGE Location

The Föss Bridge is part of the Gotthard Pass Highway and crosses the Föss Creek on the southern side of the mountains at a height of 1’500 m above sea level. The carriageway allows for two through traffic lanes plus a merging lane for exit to Motto Bartola from both directions. A photograph of the bridge is given in the main report Paragraph 2, Test Parameters, p.35. C.1.2

Geometry, Material

Length along the bridge centerline [m]

79.0

Number of spans

3

Lenght of spans [m]

24.0, 31.0, 24.0


7.27


17.0

Horizontal curvature R [m]

200

Cross section type

two-cell box girder

Box girder height [m]

≈ 1.60

Transversal slope [%]

6.2

Material

Prestressed, post-tensioned concrete

Table C- 1 Main geometrical parameters of the Föss Bridge. The box girder height is constant over the length of the structure but its width varies. Detailed information on the bridge geometry is given in Fig. C- 1 and Fig. C- 2.


C.4 / 20


May 30, 2000

S2

S.Gottardo

P2 P1 1.59

S1 Airolo

1.0

24.0

31.0

24.0

79.0

24.0

S2

6.35 6.35

16.0

R = 200

0.50

24.0

S1

P2

31.0

P1

S.Gotta rdo

0.50

Foss

Airolo

Fig. C- 1 Föss Bridge longitudinal section and plan view (dimensions: m).

2.15

1.00 0.30 0.35

tu h2 td

tw

6.35

6.35 Foss

Fig. C- 2 Föss Bridge mid-span cross section (dimensions: m).

2.15

1.88

tw

td

tw

3.60

0.53

1.00

0.50

0.35

1.00

tu

2.50

h1

0.30

0.20

0.72

2.15

0.35

2.10

8.00

0.20

8.00

0.35

0.50


C.5 / 20


C.1.3

May 30, 2000

Piers, Bearings

The interior supports consist of two rectangular columns with a 1 m by 2 m cross section each. In a given interior support section, the two piers are 8.4 m apart. On each pier (P1, P2), there is one roller bearing at the support. At the St.Gotthard end (S2), there are three roller bearings 4.95 m apart. At the other end (S1), there are three fixed supports.

P2

P1

S2 S1

P

S2.2

Foss

S1

24.0 31.0 24.0

10.0 8.0

Fig. C- 3 Föss Bridge boundary conditions and theoretical static system.

4.95

4.2

S2.2 P

4.95

4.2

P

4.2

4.95 Airolo

4.95

S1

P

4.2

S2.1 S1

S.Go ttardo

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia, Annex C

C.2 C.2.1

C.6 / 20 May 30, 2000

AMBIENT VIBRATION TESTS OF THE FÖSS BRIDGE General

The theory of Ambient Vibration Testing, AVT, is described in detail in [Felber, 1993]. Further details on the technology as used by EMPA on the tests performed on the Föss Bridge can be taken from [EDI, 1995] and [Felber, 1995c]. For the tests performed on the Föss Bridge, ambient vibrations were measured at 71 points over a period of three days. As the traffic on the bridge was very sparse during the measurements the main sources of ambient excitation were wind and microseismicity. From the acceleration time records a total of 5 modes of vibration could be identified. The lowest transverse mode was f1 = 4.03 Hz and the first vertical mode was f2 = 4.44 Hz. The identified dynamic characteristics can now be used to calibrate a computer model of the bridge. C.2.2

Test Description

C.2.2.1 Instrumentation The bridge vibrations were measured using eight Brüel & Kjaer 8306 accelerometers with a sensitivity of 10 V/g and a resolution of 10-6 m/s2. The signals produced by the accelerometers were amplified using an EMPA made pre-amplifier. The same pre-amplifier provided also the 28 V DC-signal used to drive the accelerometers. The pre-amplified signals were then filtered and amplified using an eight channel DIFA/SCADAS front-end. Finally the signals were amplified and digitized using a 16 bit Keithley 575 Module with an AMM2 board. C.2.2.2 Measurement Set-ups The 71 measurement points were spread approximately uniformly over the length and width of the structure (Fig. C- 4). No measurement points were located at the piers and abutments. Measurement direction was mainly vertical. Some measurements were however taken in both horizontal directions also. Measurement points Nos. 44 and 55 (both vertical) were chosen for the measurement phase centered on the vertical vibrations. For the subsequent phase centered on horizontal vibrations points Nos. 16, 36 and 51 (all horizontal transverse) were the reference points. The remaining acclerometers were roved over the structure. The measurement campaign encompassed 24 measurement set-ups.


C.7 / 20 May 30, 2000

Fig. C- 4 Location and Numbering of Measurement Points.

C.2.2.3 Signal Acquisition The data was acquired using the following settings: • Sampling Rate s = 100 Hz • Possible frequency range f = 0...50 Hz, frequency range used: f = 0...10 Hz • Low Pass Filter: 12.5 Hz • Attenuation: 42db or 36 dB (123.8 or 62.1 mm/s2/V) • Number of points per Channel: 32 k (8 segments of 4’096 points) In total 192 acceleration time histories were recorded at 71 measurement locations. This corresponds to 27 MB (in binary format). C.2.2.4 Signal Processing The recorded accelerations were analysed using the programs U2, V2 and P2 [EDI, 1995]. The program P2 was used to generate Averaged Normalized Power Spectral Densities, ANPSD's, for the identification of natural frequencies. The program U2 was used to compute the Modal Ratios from the recorded data. V2 was subsequently used to visualise and animate the deflected shapes corresponding to the computed Modal Ratios. These programs were used for preliminary data analysis on site and for further detailed data analysis at the office. To avoid the possibility of overlooking a natural frequency because the computed PSD might have been recorded on a location corresponding to a modal node, PSD's were computed for all recordings. First, all these PSD's were normalised to give them equal weight, after which they were averaged. The resulting ANPSD's computed from all vertical records was then used to identify all the vertical modes.


C.8 / 20


C.2.3

May 30, 2000

Results

C.2.3.1 Natural Frequencies The natural frequencies of the structure were identified using the ANPSD functions described above. As an example, the ANPSD for all vertical signals is given in Fig. C- 5. (Other ANPSD's mentioned below are given in [Felber, 1995c].) Furthermore, for the natural frequencies' identification, ANPSD's for the added and subtracted vertical signals allowing easy differentiation between flexural and torsional modes were also used. With the technology available at EMPA, all time signals are stored on hard disk and hence can subsequently be arithmetically treated according to the requirements. Finally, the transverse bending modes were identified based on the ANPSD for the measured transverse signals. The ANPSD for all the vertical signals (Fig. C- 5) shows dominant peaks at f ≈ 4.0, 4.44, and 7.9 Hz. A minor peak can be seen at f ≈ 4.76 Hz. Further information can be taken from the ANPSD'd for the added and subtracted vertical signals respectively as well as from those for the horizontal transverse and longitudinal signals ([Felber, 1995c]).

ANPSD Föss Vertical All 100

ANPSD

7.9

4.44

10

3.98

1

4.76

0.1

0.01 0

1

2

3

4

5

6

7

8

9

10

Frequenz [Hz]

Fig. C- 5 ANPSD for the Föss Bridge vertical acceleration signals. All the identified natural frequencies are listed in Table C- 2. Of special interest is the fact that the first vertical bending mode frequency, f2 = 4.44 Hz, lies outside of the dominant wheel load frequency range of vehicles equipped with suspensions of any type.


C.9 / 20


May 30, 2000

Mode No.

Frequency [Hz]

Description

1

4.03

First Horizontal Transverse

2

4.44

First Vertical Bendnig

3

6.45

Second Vertical plus Transverse Bending

4

7.10

Torsion

5

7.76

Second Horizontal Transverse

Table C- 2

Föss Bridge; Identified Natural Frequencies and associated Mode Type.

C.2.3.2 Mode Shapes The shapes of the five modes identified are presented graphically below.

Fig. C- 6 Föss Bridge Mode No. 1, f1 = 4.03 Hz.






C.10 / 20 May 30, 2000


C.11 / 20


C.3

May 30, 2000

BRIDGE INSTRUMENTATION FOR THE DIVINE TESTS 2 3

P1

31.0

24.0

S1

1 P2 S2

24.0

16.0

15.0

12.0

12.0

11.0 13.0

WG 21

WG 22 WG 23

WG 32

S.Go ttardo H

WG 12

Foss

Airolo

Cross section

2.00

2.00 0.90 1.10

WG 21

WG 12/22/32

0.15

WG 23 0.15

0.15 6.35

6.35 Foss

Fig. C- 11 Föss Bridge instrumentation: Five vertical displacement measurements. The lane used by the test vehicle (always in direction St. Gotthard-Airolo) is also indicated. S1

9

8

30.07 3 x 0.19

3 6

7

30.08

30.10

P1

5

30.10

P2

S2 2 3

4

30.12

30.08

1 2

30.8

1

30.05

0

30.05 4 x 0.19

Fig. C- 12 Föss Bridge test track defined by white strips (Nos. 0...16) glued to the pavement surface. The EMPA contact thresholds Nos. 1...3 are also indicated in the figure.


C.12 / 20


C.4


Föss left 15

Profile [mm]

10 5 0 -5 -10 -15 -20 0

20 40 60 80 100 120 140 160 180 200 220 240 260 280

Distance [m] Fig. C- 13 Longitudinal profile in the lefthand wheelpath.

Profile [mm]

Föss right 15 10 5 0 -5 -10 -15 -20 -25 -30 0

20 40 60 80 100 120 140 160 180 200 220 240 260 280

Distance [m] Fig. C- 14 Longitudinal profile in the righthand wheel path.

May 30, 2000


C.13 / 20


May 30, 2000

Foss 15

10

profile [mm]

5

0 0

20

40

60

80

100

120

140

160

180

200

Foss mr Foss ml

-5

-10

-15

-20 Distance [m]

Fig. C- 15 Longitudinal profiles in both wheel paths.

Fig. C- 16 Spectrum of the left track pavement longitudinal profile; classification according to [ISO, 1995].


C.14 / 20 May 30, 2000

Fig. C- 17 Spectrum of the right track pavement longitudinal profile; classification according to [ISO, 1995].


C.5 C.5.1

C.15 / 20 May 30, 2000

RESULTS OF THE DIVINE TESTS ON THE FÖSS BRIDGE Bridge Response Time Signals

For the purpose of illustration, Fig. C- 18 and Fig. C- 19 give a set of the five vertical (one per end-span, three in the middle span) and the horizontal bridge dynamic deflection for a passage of the steel and air-suspended NRC vehicle respectively. The graphs, i.e. the vehicle speed, were chosen to allow a representative insight into the dynamic bridge behavior. It can be seen from Fig. C- 18 that the dynamic bridge response is not very large for the steelsuspended vehicle in all bridge spans. It is larger in the side spans than in the middle span. This means that the Föss Bridge, as expected, shows practically no signs of dynamic vehicle/bridge interaction. Fig. C- 19 shows that the bridge response to the crossing of the air-suspended vehicle is close to nil in all bridge spans. The difference in dynamic bridge response for the steel and airsuspended vehicle may be due to the relative distance to the bridge fundamental mode, f = 4.44 Hz from the respective wheel load frequencies, f = 3 Hz for the steel and f = 1.6 Hz for the air-suspended vehicle.


C.16 / 20


May 30, 2000

1s/div WG 12 0.5 mm

WG 21 0.5 mm

WG 22 1.0 mm

WG 23 0.5 mm

WG 32 0.5 mm


Fig. C- 18 Föss Bridge time signals for the crossing of the steel-suspended NRC vehicle at a speed v = 14.9 km/h.


C.17 / 20


May 30, 2000

1s/div WG 12 0.5 mm

WG 21 0.5 mm

WG 22 0.5 mm

WG 23 0.5 mm

WG 32 0.5 mm


Fig. C- 19 Föss Bridge time signals for the crossing of the air-suspended NRC vehicle at a speed v = 27.9 km/h.


C.18 / 20


C.5.2

May 30, 2000

Dynamic Increments

The definition of the dynamic increment φ as used throughout this report can be found in Chapter 1, Introduction, p. 23. Here, the results of the DIVINE tests, i.e. the φ(v)-functions are given for one vertical displacement measurement point per bridge span as given in Fig. C- 11. Each of the graphically displayed functions (Fig. C- 20 to Fig. C- 24) allows direct comparison of the results for the tests performed with the NRC-vehicle (see Annex H) equipped with air and steel suspensions respectively.

FÖSS Bridge WG 12

25



15 STEEL AIR 10

5

0 0

10

20

30

40


Fig. C- 20 Dynamic Increments measured for point WG 12.

50

60


C.19 / 20


May 30, 2000




15 STEEL AIR 10

5

0 0

10

20

30

40

50

60

Vehicle Speed [km/h] Fig. C- 21 Dynamic Increments measured for point WG 21. FÖSS Bridge WG 22 25



15 STEEL AIR 10

5

0 0

10

20

30

40

Vehicle Speed [km/h] Fig. C- 22 Dynamic Increments measured for point WG 22.

50

60


C.20 / 20


May 30, 2000




15 STEEL AIR 10

5

0 0

10

20

30

40

50

60


Fig. C- 23 Dynamic Increments measured for point WG 23. FÖSS Bridge WG 32 25



15 STEEL AIR 10

5

0 0

10

20

30

40


Fig. C- 24 Dynamic Increments measured for point WG 32.

50

60




Annex D Lawsons Creek Bridge (AUS)


D.2 / 15

Final Report, EMPA Switzerland, QUT Australia, Annex D

May 30, 2000

Contents D.1 DESCRIPTION AND INSTRUMENTATION OF THE LAWSONS CREEK BRIDGE ...........3 D.2 PAVEMENT LONGITUDINAL PROFILE .............................................................................5 D.2.1 Longitudinal Profile .......................................................................................................5 D.2.2 Pavement Profile Classification ....................................................................................5 D.3 BRIDGE RESPONSE..........................................................................................................6 D.3.1 Typical Waveforms - Deflection versus Time................................................................6 D.3.2 Power Spectral Densities..............................................................................................8 D.4 DYNAMIC WHEEL LOADS .................................................................................................9 D.4.1 Dynamic Axle Force Waveforms...................................................................................9 D.4.1 Dynamic Axle Force Power Spectral Densities ...........................................................11 D.5 VEHICLE/BRIDGE INTERACTION: CROSS POWER SPECTRAL DENSITIES ..............14

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia, Annex D

D.1

D.3 / 15 May 30, 2000

DESCRIPTION AND INSTRUMENTATION OF THE LAWSONS CREEK BRIDGE

Fig. D- 1 Lawsons Creek Bridge, Lithgow, NSW.


D.4 / 15


May 30, 2000

cL 0.50

11.70

0.50

[email protected]

3%

D1 S1

0.61

0.18 3%

D3 S3

D5 S5

D7 S7

1.15

D9 S9

SECTION to Lithgow E/B

to Bathurst W/B

24.00

Abut A

Abut B

3.43 %

ELEVATION S9 axle detector ADA

Expansion joint

W/B

Axle Hop Bump (AHB)

S7

D7

S5

D5

S3

D3

S1

D1

Fixed

axle detector ADB

E/B

PLAN Fig. D- 2 Lawsons Creek Bridge, Lithgow, NSW; Geometry and Instrumentation.

cL


D.5 / 15


D.2 D.2.1

May 30, 2000

PAVEMENT LONGITUDINAL PROFILE Longitudinal Profile 0.015

Elevation [m]

0.01 0.005 0

-0.005 -0.01 Span 1

-0.015 290

300

310

320 330 Chainage [m]

340

350

360

Fig. D- 3 Longitudinal profile, East bound, Centerline, Passengers and drivers side. Pavement Profile Classification ity, Gd(n) [m^3/cycle]

D.2.2

0

10

100.0

10.0

Wavelength [m]

1.0

0.1

Lawsons Creek -2

10

H G

Fig. D- 4 Lawsons Creek - Unevenness Power Spectral Density of the Pavement Profile.


D.6 / 15


D.3 D.3.1

May 30, 2000

BRIDGE RESPONSE Typical Waveforms - Deflection versus Time 0.5 0.0 Deflection (mm)

-0.5 -1.0 -1.5 -2.0 -2.5 -3.0 -3.5 -4.0 0

5

10

15

20

25

Time (s)

Fig. D- 5 Lawsons Creek, BS, No Bumps, 7 km/h - Midspan Deflection D5. 0.5

Deflection (mm)

0.0 D1

-0.5 -1.0 -1.5

D7

-2.0 D3

-2.5 -3.0 -3.5

D5

-4.0 0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Time (s)

Fig. D- 6 Lawsons Creek, BS, No Bumps, 60 km/h - Midspan Deflections D1, D3, D5 and D7. 0.5 0.0 Deflection (mm)

-0.5 -1.0 -1.5 -2.0 -2.5 -3.0 -3.5 -4.0 0.5

1.0

1.5

2.0

2.5

3.0

Time (s)

Fig. D- 7 Lawsons Creek, BS, AHP, 72 km/h - Midspan Deflection D5.

3.5

4.0


D.7 / 15


May 30, 2000

0.5

Deflection (mm)

0.0 -0.5 -1.0 -1.5 -2.0 -2.5 -3.0 -3.5 -4.0 0

5

10

15

20

25

Time (s)

Fig. D- 8 Lawsons Creek, BA, No Bumps, 7 km/h - Midspan Deflection D5. 0.5

Deflection (mm)

0.0 D1

-0.5 -1.0

D7

-1.5 -2.0

D3

-2.5 -3.0 -3.5

D5

-4.0 0

0.5

1

1.5

2

2.5

3

3.5

Time (s)

Fig. D- 9 Lawsons Creek, BA, No Bumps, 99 km/h - Midspan Deflections D1, D3, D5 and D7. 0.5 0.0 Deflection (mm)

-0.5 -1.0 -1.5 -2.0 -2.5 -3.0 -3.5 -4.0 0

0.5

1

1.5

2

2.5

Time (s)

Fig. D- 10 Lawsons Creek, BA, AHP, 89 km/h - Midspan Deflection D5.

3

3.5


D.8 / 15


D.3.2

May 30, 2000

Power Spectral Densities 0.12 5.1 Hz

| Amplitude|

0.1 0.08 0.06

7.4 Hz

0.04 0.02 0

0

5

10

15

20

Frequency (Hz) Fig. D- 11 Power Spectral Density of the free vibration component of the bridge deflection D5, BA, 89 km/h, W/B, AHP.


D.9 / 15


D.4

DYNAMIC WHEEL LOADS Dynamic Axle Force Waveforms


D.4.1

May 30, 2000

Total Tridem Force

BRIDGE 100 kN

Rear Axle Centre Axle Front Axle

0

1

2

3

4

5

6

7

8

Time (s)


Fig. D- 12 Dynamic force waveform of BS, W/B, no bumps, 60 km/h. BRIDGE Total Tridem Force 100 kN AHB Rear Axle Centre Axle Front Axle

0

1

2

3

4

5

Time (s)

Fig. D- 13 Dynamic force waveform of BS, W/B, AHP, 72 km/h.

6

7

8


D.10 / 15



May 30, 2000

BRIDGE Total Tridem Force 100 kN Rear Axle Centre Axle Front Axle

0

1

2

3

4

5

6

7

8

Time (s)


Fig. D- 14 Dynamic force waveform of BA, W/B, no bumps, 99 km/h.

BRIDGE Total Tridem Force 100 kN AHB Rear Axle Centre Axle Front Axle

0

1

2

3

4

5

Time (s)

Fig. D- 15 Dynamic force waveform of BA, W/B, AHP, 89 km/h.

6

7

8


D.11 / 15


D.4.1

May 30, 2000

Dynamic Axle Force Power Spectral Densities 14000

3.1 Hz |Amplitude|

12000

10000

8000

2.0 Hz 6000

4000

4.1 Hz 2000

0 0

5

10

15

20

Frequency (Hz)

a. Power Spectral Density of the Front Axle Dynamic Force. 12000

3.1 Hz |Amplitude|

10000

8000

4.1 Hz

6000

4000

2000

0 0

5

10

15

20

Frequency (Hz)

b. Power Spectral Density of the Center Axle Dynamic Force. 15000

|Amplitude|

3.1 Hz 4.1 Hz 10000

5000

0 0

5

10

15

20

Frequency (Hz)

c. Power Spectral Density of the Rear Axle Dynamic Force. Fig. D- 16 Power Spectral Density of independent axles, BS, W/B, 72 km/h, AHP.


D.12 / 15


May 30, 2000

2000 1800 1600

|Amplitude|

1400 1200

2.4 Hz

1000 800 600 400 200 0 0

5

10

15

20

Frequency (Hz)

a. Power Spectral Density of the Front Axle Dynamic Force. 1200

2.4 Hz

|Amplitude|

1000

800

600

400

200

0 0

5

10

15

20

Frequency (Hz)

b. Power Spectral Density of the Center Axle Dynamic Force. 2500

|Amplitude|

2000

1500

1000

1.8 Hz 500

2.4 Hz

0 0

5

10

15

20

Frequency (Hz) c. Power Spectral Density of the Rear Axle Dynamic Force. Fig. D- 17 Power Spectral Density of independent axles, BA, W/B, 89 km/h, AHP.


D.13 / 15


May 30, 2000

x 10 4 12

3.1 Hz |Amplitude|

10

8

4.1 Hz

6

4

2

0 0

5

10

15

20

Frequency (Hz)

Fig. D- 18 Power Spectral Density of the dynamic tridem force, BS, W/B, 72 km/h, AHP. 8000

2.4 Hz

|Amplitude|

7000 6000 5000 4000 3000 2000 1000 0 0

5

10

15

20

Frequency (Hz)

Fig. D- 19 Power Spectral Density of the dynamic tridem force, BA, W/B, 89 km/h, AHP.


D.14 / 15


D.5

May 30, 2000

VEHICLE/BRIDGE INTERACTION: CROSS POWER SPECTRAL DENSITIES 300

|Amplitude|

250

200

150

100

3.2 Hz

50

5.1 Hz 0 0

5

10

15

20

Frequency (Hz)

Fig. D- 20 Cross power spectra of deflection D5 and the dynamic tridem force of BS, W/B, 60 km/h, no bumps. 700

0.8 Hz 600

|Amplitude|

500

400

300

200

100

3.2 Hz 0 0

5

10

15

20

Frequency (Hz)

Fig. D- 21 Cross power spectra of deflection D5 and the dynamic tridem force of BA, W/B, 99 km/h, no bumps. 7

5.1 Hz 6

|Amplitude|

5

4

3

2

1

0 0

5

10

15

20

Frequency (Hz)

Fig. D- 22 Cross power spectra of deflection D5 and the dynamic tridem force of BS, W/B, 60 km/h, no bumps. Signals high pass filtered at 5 Hz.


D.15 / 15


May 30, 2000

1.2

5.0 Hz 1

|Amplitude|

0.8

3.3 Hz

0.6

9.0 Hz 6.6 Hz

0.4

0.2

0 0

5

10

15

20

Frequency (Hz)

Fig. D- 23 Cross power spectra of deflection D5 and the dynamic tridem force of BA, W/B, 99 km/h, no bumps. Signals high pass filtered at 5 Hz.




Annex E Coxs River Bridge (AUS)


E.2 / 10

Final Report, EMPA Switzerland, QUT Australia, Annex E

May 30, 2000

Contents E.1 DESCRIPTION AND INSTRUMENTATION OF THE COXS RIVER BRIDGE ....................3 E.2 PAVEMENT LONGITUDINAL PROFILE .............................................................................5 E.2.1 Longitudinal Profile ........................................................................................................5 E.2.2 Pavement Profile Classification .....................................................................................5 E.3 BRIDGE RESPONSE ..........................................................................................................6 E.3.1 Typical Waveforms - Deflection versus Time.................................................................6 E.3.2 Typical Waveforms - Strain versus Time .......................................................................9 E.3.3 Power Spectral Density ...............................................................................................10

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia, Annex E

E.1

E.3 / 10 May 30, 2000

DESCRIPTION AND INSTRUMENTATION OF THE COXS RIVER BRIDGE

Fig. E- 1 Coxs River Bridge, Lithgow, NSW, Australia (BS test vehicle)

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia, Annex E

Fig. E- 2 Coxs River - Geometry and Instrumentation layout.

E.4 / 10 May 30, 2000


E.5 / 10


E.2

May 30, 2000

PAVEMENT LONGITUDINAL PROFILE

E.2.1 Longitudinal Profile 0.025 0.02

Elevation [m]

0.015 0.01

0.005 0

-0.005 -0.01 -0.015 -0.02

Span 1

Span 2

Span 3 Span 4

-0.025 300

320

340 360 Chainage [m]

380

400

Fig. E- 3 Coxs River - Longitudinal profile, North bound, Centreline, Passengers Side.

y, Gd(n) [m^3/cycle]

E.2.2 Pavement Profile Classification 0

10

100.0

10.0

Wavelength [m]

1.0

0.1

Coxs River -2

10

H G

Fig. E- 4 Coxs River - Unevenness Power Spectral Density of the Pavement Profile.


E.6 / 10


E.3

May 30, 2000

BRIDGE RESPONSE

E.3.1 Typical Waveforms - Deflection versus Time

1 0

Deflection (mm)

-1 -2 -3 -4 -5 -6 -7 -8 0

1

2

3

4

5

Time (s)

Fig. E- 5 Coxs River, BS, No Bumps, 96 km/h - Midspan Deflection D(2-3).

1 0

Deflection (mm)

-1 -2 -3 -4 -5 -6 -7 -8 0

1

2

3

4

Time (s)

Fig. E- 6 Coxs River, BS, AHP, 74 km/h - Midspan Deflection D(2-3).

5


E.7 / 10


May 30, 2000

1 0 0

1

2

3

4

5

Deflection (mm)

-1 -2 -3 -4 -5 -6 -7 -8

Time (s)

Fig. E- 7 Coxs River, BA, No Bumps, 96 km/h - Midspan Deflection D(2-3).

1 0

Deflection (mm)

-1 -2 -3 -4 -5 -6 -7 -8 0

1

2

3

4

Time (s)

Fig. E- 8 Coxs River, BA, AHP, 68 km/h - Midspan Deflection D(2-3).

5


E.8 / 10


May 30, 2000

1 0

Deflection (mm)

-1 -2 -3 -4 -5 -6 -7 -8 0

1

2

3

4

5

Time (s)

Fig. E- 9 Coxs River, BA (Unequal Spacing), AHP, 68 km/h - Midspan Deflection D(2-3).

1 0

Deflection (mm)

-1 -2 -3 -4 -5 -6 -7 -8 0

1

2

3

4

5

Time (s)

Fig. E- 10 Coxs River, BA, [No Shock Absorbers on the Rear Axle], No Bumps, 68 km/h Midspan Deflection D(2-3).


E.9 / 10


May 30, 2000

E.3.2 Typical Waveforms - Strain versus Time

50

BS 42.5 1.23 No Bumps

25 0

Strain (microstrains)

-25 -50 -75

BS 42.5 1.23 AHB

-100 -125 -150 -175 -200 -225 -250 0.5

1

1.5

2

2.5

3

Time (s)

Fig. E- 11 Coxs River, BS, No Bumps, 96 km/h and BS, AHP, 74 km/h - Midspan Strain S(2-3).

50

BA 42.5 1.23 No Bumps

Strain (microstrains)

0

-50

BA 42.5 1.23 No Shock Absorbers

BA 42.5 1.85-1.23 Unequal Spacing

-100

-150

-200

-250

-300 1.00

1.50

2.00

2.50

3.00

3.50

Time (s)

Fig. E- 12 Coxs River, BA, No Bumps, 77 km/h, BA, No shock absorbers on the rear axle, No Bumps, 61 km/h and BA (Unequal spacing), AHP, 57 km/h - Midspan Strain S(2-3).


E.10 / 10


May 30, 2000

E.3.3 Power Spectral Density

10.2 Hz

0.03

13.3 Hz

0.025 0.02 0.015 0.01 0.005 0

0

5

10 Frequency (Hz)

15

Fig. E- 13 PSD of the free vibration, BA, 58 km/h, N/B, No bumps.

20




Annex F Camerons Creek Bridge (AUS)


F.2 / 11

Final Report, EMPA Switzerland, QUT Australia, Annex F

May 30, 2000

Contents F.1 DESCRIPTION AND INSTRUMENTATION OF THE CAMERONS CREEK BRIDGE.........3 F.2 PAVEMENT LONGITUDINAL PROFILE .............................................................................5 F.2.1 Longitudinal Profile........................................................................................................5 F.2.2 Pavement Profile Classification .....................................................................................5 F.3 BRIDGE RESPONSE ..........................................................................................................6 F.3.1 Typical Waveforms - Deflection versus Position of Vehicle...........................................6 F.3.2 Typical Waveforms - Deflection versus Time ................................................................7 F.3.3 Dynamic Increments......................................................................................................9 F.4 RESULTS OF STATIC TESTS ..........................................................................................11

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia, Annex F

F.1

F.3 / 11 May 30, 2000

DESCRIPTION AND INSTRUMENTATION OF THE CAMERONS CREEK BRIDGE

Fig. F- 1

Bridge over Camerons Creek, Bucketts Way, New South Wales, Australia.

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia, Annex F

Fig. F- 2

F.4 / 11 May 30, 2000

Bridge over Camerons Creek, Bucketts Way, 32 km North of Raymond Terrace, New South Wales, Australia.


F.5 / 11


F.2 F.2.1

May 30, 2000

PAVEMENT LONGITUDINAL PROFILE Longitudinal Profile 0.04 0.03

Elevation [m]

0.02 0.01 0

-0.01 -0.02 -0.03 Span 1

Span 2

Span 4

Span 3

-0.04 0

10

20

30 40 Chainage [m]

50

60

70

Fig. F- 3 Camerons Creek Bridge longitudinal profile - northbound, passengers and drivers side wheel line. Pavement Profile Classification sity, Gd(n) [m^3/cycle]

F.2.2

Fig. F- 4

0

10

100.0

10.0

Wavelength [m]

1.0

0.1

Camerons Creek -2

10

H G -4

10

Power Spectral Density (PSD) of road profile


F.6 / 11


F.3

BRIDGE RESPONSE Typical Waveforms - Deflection versus Position of Vehicle

Deflection (mm)

F.3.1

May 30, 2000

1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 -2.5

Crawl 60 km/h 100 km/h

0

5

10

15

20

25

30

35

40

Distance to steer axle (m)

Deflection (mm)

Fig. F- 5 Deflection versus position of vehicle for various speeds: Camerons Creek Bridge, D(4-8), Truck BA, Axle hop plank.

1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 -2.5

Crawl 62 km/h 103 km/h

0

5

10

15

20

25

30

35

40


Fig. F- 6 Deflection versus position of vehicle for various speeds: Camerons Creek Bridge, D(4-8), Truck BS, Axle hop plank.


F.7 / 11


F.3.2

May 30, 2000

Typical Waveforms - Deflection versus Time

0.5

Deflection [mm]

0.0 -0.5 D(4,8)

-1.0 D(1,8)

-1.5 -2.0 -2.5 -1

Fig. F- 7

0

1 Tim e [s]

2

3

4

Deflections versus time waveforms, Truck BS, no planks, 80 km/h, northbound Camerons Creek Bridge.

0.5

Deflection [mm]

0.0 -0.5 D(4,8)

-1.0 D(1,8)

-1.5 -2.0 -1

0

1

2

3

4

Tim e [s]

Fig. F- 8

Deflections versus time waveforms, Truck BA, no planks, 62 km/h, northbound Camerons Creek Bridge.


F.8 / 11


May 30, 2000

0.5

Deflection [mm]

0.0 -0.5

D(4,8)

-1.0 D(1,8)

-1.5 -2.0 -2.5 -3.0 0

Fig. F- 9

1

2

Time [s]

3

4

5

Deflections versus time waveforms, Truck SA, no planks, 63 km/h, northbound Camerons Creek Bridge.


F.9 / 11


F.3.3

May 30, 2000

Dynamic Increments

Dynamic Increment

150% 125% 100%

D(1,8)

75%

D(4,3) D(4,8)

50%

D(4,14)

25% 0% -120 -100

-80

-60

-40

-20

0

20

40

60

80

100

120

Velocity (km/h)

(a) air suspension - BA

Dynamic Increment

150% 125% 100%

D(1,8)

75%

D(4,3) D(4,8)

50%

D(4,14)

25% 0% -120 -100 -80

-60

-40

-20

0

20

40

60

80

100 120

Velocity (km/h)

(b) steel suspension - BS

Dynamic Increment

150% 125% 100%

D(1,8)

75%

D(4,3) D(4,8)

50%

D(4,14)

25% 0% -120 -100 -80

-60

-40

-20

0

20

40

60

80

100

120

Velocity (km/h)

(c) air suspension - SA Fig. F- 10

Dynamic increment versus speed - Camerons Creek, no planks.


F.10 / 11


May 30, 2000

Dynamic Increment

150% 125% 100%

D(1,8)

75%

D(4,3)

50%

D(4,8)

25%

D(4,14)

0% -120 -100 -80

-60

-40

-20

0

20

40

60

80

100

120

Velocity (km/h)

(a) air suspension - BA AHP

Dynamic Increment

150% 125% 100%

D(1,8)

75%

D(4,3)

50%

D(4,8) D(4,14)

25% 0% -120 -100 -80

-60

-40

-20

0

20

40

60

80

100

120

Velocity (km/h)

(b) steel suspension - BS AHP Fig. F- 11

Dynamic increment versus speed - Camerons Creek, Axle hop plank in place.


F.11 / 11


F.4

May 30, 2000

RESULTS OF STATIC TESTS

L

N/B truck

D(4,3)

C

S/B truck

D(4,14)

D(4,8)

0

Deflection (mm)

0.2 0.4 0.6 0.8 1 1.2 1.4 Model N/B

Fig. F- 12

Model S/B

Test N/B

Test S/B

Camerons Creek - Comparison of theoretical and experimental deflections (BA stationary).




Annex G Cromarty Creek Bridge (AUS)


G.2 / 10

Final Report, EMPA Switzerland, QUT Australia, Annex G

May 30, 2000

Contents G.1 DESCRIPTION AND INSTRUMENTATION OF THE CROMARTY CREEK BRIDGE ........3 G.2 PAVEMENT LONGITUDINAL PROFILE.............................................................................5 G.2.1 Longitudinal Profile .......................................................................................................5 G.2.2 Pavement Profile Classification ....................................................................................5 G.3 BRIDGE RESPONSE..........................................................................................................6 G.3.1 Typical Waveforms - Deflection versus Position of Vehicle ..........................................6 G.3.2 Typical Waveforms - Deflection versus Time................................................................7 G.3.3 Dynamic Increments .....................................................................................................8 G.4 DYNAMIC WHEEL LOADS...............................................................................................10

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia, Annex G

G.1

G.3 / 10 May 30, 2000

DESCRIPTION AND INSTRUMENTATION OF THE CROMARTY CREEK BRIDGE

Fig. G- 1 Bridge over Cromarty Creek, Bucketts Way, 34 km North of Raymond Terrace, New South Wales, Australia.


G.4 / 10


May 30, 2000

9.14 m

Fig. G- 2 Cromarty Creek - Geometry and instrumentation layout


G.5 / 10


G.2 G.2.1

May 30, 2000

PAVEMENT LONGITUDINAL PROFILE Longitudinal Profile 0.04 0.03 Elevation [m]

0.02 0.01 0

-0.01 -0.02 Span 1

Span 2

Span 3

-0.03 0

10

20

30 40 Chainage [m]

50

60

70

Fig. G- 3 Cromarty Creek Bridge longitudinal profile - northbound. Pavement Profile Classification ensity, Gd(n) [m^3/cycle]

G.2.2

0

10

100.0

10.0

Wavelength [m]

1.0

0.1

Cromarty Creek -2

10

-4

10

Fig. G- 4 Power Spectral Density (PSD) of road profile

H G F


G.6 / 10


G.3

BRIDGE RESPONSE Typical Waveforms - Deflection versus Position of Vehicle

Deflection (mm)

G.3.1

May 30, 2000

2 0 -2 -4 -6 -8 -10 -12

crawl 45km/h 92 km/h

-14 -16 15

20

25

30

35

40

45

50


(a) D(3-3), Truck BA, No bumps 2

Deflection (mm)

0 -2 -4 crawl

-6 -8

48 km/h

-10 -12

91 km/h

-14 -16 15

20

25

30

35

40

45

50


(b) D(3-3), Truck BS, No bumps Fig. G- 5 Deflection versus position of vehicle for various speeds - Cromarty Creek


G.7 / 10


G.3.2

May 30, 2000

Typical Waveforms - Deflection versus Time 2 0 Deflection (mm)

-2 -4 -6

45 km/h

-8 -10

Dynamic increment 40%

-12 -14 -16 15

20

25

30

35

40

45

50


(a) D(3-3), Truck BA, with AHP, 45 km/h 2 0 Deflection (mm)

-2 -4 -6

67 km/h

-8 -10

Dynamic increment 111%

-12 -14 -16 20

25

30

35

40

45


(b) D(3-3), Truck BS, with AHP, 67 km/h Fig. G- 6 Deflection versus time waveforms - Cromarty Creek.

50

55


G.8 / 10


G.3.3

May 30, 2000

Dynamic Increments

Dynamic Increment

150% 125% D (3,2) D (3,3) D (3,4) D (3,5)

100% 75% 50% 25% 0% -120 -100 -80

-60

-40

-20

0

20

40

60

80

100 120

V elocity (km /h )

(a) air suspensions BA

Dynamic Increment

150% 125% D(3,2) D(3,3) D(3,4) D(3,5)

100% 75% 50% 25% 0% -120 -100 -80

-60

-40

-20

0

20

40

60

80

100 120

Velocity (km/h)

(b) steel suspensions BS Fig. G- 7 Dynamic increment versus velocity - Cromarty Creek, six-axle articulated vehicle, no bumps.


G.9 / 10


May 30, 2000

Dynamic Increment

150% 125% D(3,2) D(3,3) D(3,4) D(3,5)

100% 75% 50% 25% 0% -120 -100 -80 -60 -40 -20

0

20

40

60

80 100 120

Velocity (km/h)

(a) air suspension - BA, AHP

Dynamic Increment

150% 125% 100% D(3,2) D(3,3)

75% 50% 25% 0% -120 -100 -80 -60 -40 -20

0

20

40

60

80 100 120

Velocity (km/h)

(b) steel suspension - BS, AHP Fig. G- 8 Dynamic increment versus speed - Cromarty Creek, Axle hop plank.


G.10 / 10


DYNAMIC WHEEL LOADS

Wheel force (kN)

G.4

May 30, 2000

Repair Span 1 Span 2 Span 3

80 70 60 50 40 30 20 10 0 -10 -20 0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

4.5

5

Time (s)

(a) air suspension, 64 km/h over the axle hop bump (AHP), BA.

Wheel force (kN)

Repair

Span 1 Span 2 Span 3

80 70 60 50 40 30 20 10 0 -10 -20 0

0.5

1

1.5

2

2.5

3

3.5

4

Time (s)

(b) steel suspension, 67 km/h over the axle hop plank (AHP), BS. Fig. G- 9 Dynamic wheel forces - Cromarty Creek, drivers side center wheel of the trailer triaxle group.




Annex H Details of Test Vehicles


H.2 / 20

Final Report, EMPA Switzerland, QUT Australia, Annex H

May 30, 2000

Contents H.1 THE NRC TEST VEHICLE USED FOR THE SWISS BRIDGE TESTS...............................3 H.1.1 Steel Suspensions ........................................................................................................5 H.1.1.1 The Tractor Steer Axle Steel Suspension ..............................................................5 H.1.1.2 The Tractor Drive Axles Walking Beam Suspension ..............................................5 H.1.1.2 The Trailer Four-Spring Suspension ......................................................................7 H.1.2 Air Suspensions ............................................................................................................8 H.1.2.1 The Tractor Drive Axles Air Suspension.................................................................8 H.1.2.2 The Trailer Air Suspension ...................................................................................10 H.1.3 Tire Type.....................................................................................................................11 H.1.4 Static Wheel Loads .....................................................................................................12 H.2 TEST VEHICLES USED FOR THE AUSTRALIAN BRIDGE TESTS ................................13 H.2.1 General .......................................................................................................................13 H.2.2 Steel Suspensions ......................................................................................................16 H.2.2.1 The prime-mover walking beam suspension (BS vehicle)....................................16 H.2.2.2 The trailer steel suspension (BS vehicle) .............................................................17 H.2.3 Air Suspensions ..........................................................................................................18 H.2.3.1 The trailer air suspension (BA vehicle).................................................................18 H.2.3.2 The trailer air suspension (SA vehicle).................................................................19 H.2.4 Tires............................................................................................................................19 H.2.5 Instrumentation ...........................................................................................................19


H.3 / 20


H.1

May 30, 2000

THE NRC TEST VEHICLE USED FOR THE SWISS BRIDGE TESTS

The test vehicle is shown in Fig. H- 1 and Fig. H- 2, details of the suspensions used are given in Table H- 1. The NRC test vehicle is a five-axle tractor-semitrailer with an overall length of 17.8 m, a height of 3.66 m and a width of 2.62 m and 3.05 m (with and without mirrors respectively). The trailer axle spacing was 1.27 m for the air-suspended (see Fig. H- 1) and 1.37 m for the steel-suspeded vehicle. All other measures given in Fig. H- 1 were the same for both suspensions. The weight of the empty vehicle was 150 kN. The trailer carried a tank divided into four compartments. Two were filled with a total of 30’000 l of water during the tests. The vehicle gross weight during the tests was 450 kN. Because the vehicle was too long, too wide and too heavy, special permits were required to operate the vehicle in Switzerland.

NRC-Test-Vehicle 2.62 m 3.05 m

Water

empty

empty

Water

16'000 l

6'500 l

11'500 l

14'000 l 3.66 m

3.24 m

1.52 m

9.24 m

1.27m

15.27 m 17.80 m Axle load :

50 - 60 kN

Fig. H- 1

200 kN

The NRC test vehicle.

200 kN


H.4 / 20


May 30, 2000

Fig. H- 2 The NRC test vehicle on the Sort Bridge (EMPA Photo No. 101'070/30).

Axle Nr. Axle Type

Steel Suspension

Air Suspension

1

Steer

Steel leaf

2

Tractor Drive

Hendrickson Walking Beam

Neway ARD 244-6

3

Tractor Drive

Hendrickson Walking Beam

Neway ARD 244-6

4

Trailing

Four Spring Suspension

Neway AR95-14

5

Trailing

Four Spring Suspension

Neway AR95-14

Table H- 1

Suspension systems of the NRC test vehicle used for the tests in Switzerland.

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia, Annex H

H.5 / 20 May 30, 2000

H.1.1 Steel Suspensions H.1.1.1 The Tractor Steer Axle Steel Suspension As no further information is available on this suspension a photograph must suffice.

Fig. H- 3 Photograph of the NRC test vehicle steer axle suspension (EMPA Photo No. 101'175/24). H.1.1.2 The Tractor Drive Axles Walking Beam Suspension The walking beam suspension is illustrated in Fig. H- 4 and Fig. H- 5. The axles are fixed to a rigid beam which pivots at its center or balance point thus facilitating ideal static load sharing. Two spring elements are used, one for each side of the chassis. These elements are located between the frame rail and the walking beam. Specifications of the walking beam suspension are as follows: Manufacturer:

Hendrickson Mfg. (Canada) Ltd.

Model:

RTE 440 (extended leaf tandem)

Combined Axle Rated Capacity:

200 kN


Axle spacing:

1.52 m

Outer tire track width:

2.44 m

Spring elements:

steel leaf spring, 2 stage, Part Number 45322

H.6 / 20 May 30, 2000

Fig. H- 4

The NRC vehicle tractor drive axles walking beam suspension.

Fig. H- 5

Photograph of the walking beam suspension (EMPA Photo No. 101'069/10).


H.7 / 20 May 30, 2000

H.1.1.2 The Trailer Four-Spring Suspension The four spring suspension is illustrated in Fig. H- 6 and Fig. H- 7. Both the position of the axles on the trailer chassis and the axle spacing are variable within certain limits. For the tests, 1.27 m was chosen for the the axle spacing. Specifications of the four-spring suspension are as follows: Manufacturer:

Reyco Canada Inc. (a subsidiary of Reyco Industries Inc.)

Model:

2113-FAB-222-WB-14-C-50-3564


200 KN

Axle spacing

1.27 m


2.59 m

Spring elements:

multi leaf spring - T-3564

Fig. H- 6

The NRC vehicle trailing axles four spring suspension.

Fig. H- 7

The four spring suspension (EMPA Photo No. 101'049/5).


H.8 / 20


H.1.2

May 30, 2000

Air Suspensions

The air-suspended axles are mechanically independent of each other. Air supply to the air bags is regulated by two time-delayed, height-sensing valves, one on each side of the vehicle. The air bags on each side of the vehicle are plumbed in parallel thus achieving load equalization yet maintaining quasi-static vehicle roll resistance. Mechanical roll stiffness of the suspensions is achieved by the use of trailing members semi-rigidly fastened to the axle (trailer suspension) or a torsion tube (tractor suspension) to form an anti roll bar. This transfers roll moments of the vehicle to vertical forces at the wheels. Specifications of the air suspensions are as follows. H.1.2.1 The Tractor Drive Axles Air Suspension Manufacturer:

Neway (A division of Lear Siegler Inc.)

Model:

Tandem axle drive suspension ARD 244-6, Serial # C904157 EM


200 kN

Tandem axle spacing:

1.37 m


2.59 m

Spring element:

Air bag part number 905-57-031

Fig. H- 8 The NRC vehicle tractor drive axles air suspension system.


Fig. H- 9

The tractor air suspension (EMPA Photo No. 101'175/36).

H.9 / 20 May 30, 2000


H.10 / 20


May 30, 2000

H.1.2.2 The Trailer Air Suspension Manufacturer:

Neway (A division of Lear Siegler Inc.)

Model:

Tandem axle trailer suspension ARD 95-14, Serial # Lead - C9926210LK


227.3 kN

Tandem axle spacing tested:

1.27 m

Spring element:

Air bag part number 905-57-020

Fig. H- 10 The NRC vehicle trailing axles air suspension system.

Fig. H- 11 The trailing axles air suspension (EMPA Photo No. 101'175/32).


H.11 / 20 May 30, 2000

H.1.3 Tire Type The same tires were used with both suspension systems. The steer axle was fitted with normal single tires whereas the tractor drive and the trailing axles were equipped with twin tires. All tires were inflated to about 100 psi at the beginning of the testing day and, in contrary to the experiences of earlier tests, this pressure remained practically constant throughout testing. Due to excessive braking and the consequent heating of brakes and tires, tire pressure rose significantly during earlier tests. The tire types are summarized below.

Axle No.

Tire type

1

Uniroyal 14/80 R20

2 and 3

Michelin 12R 22.5

4 and 5

Michelin 11R 22.5

Table H- 2

Tire types fitted to NRC test vehicle


H.12 / 20


May 30, 2000

H.1.4 Static Wheel Loads Before beginning of the tests, the static wheel loads were measured for vehicles fitted with both steel and air suspension. The loads given in Table H- 3 were determined using a mobile weigh bridge provided by the Police Department of the Canton Zürich.

Static wheel loads [kN] Steel suspension

Air suspension

Axle No.

driver side (lefthand)

passenger side

driver side

passenger side

1

30.7

28.1

33.4

25.7

2

54.3

52.6

53.8

50.8

3

56.3

48.6

53.7

47.2

4

53.2

50.7

54.5

46.8

5

54.4

49.8

57.1

49.8

total

248.9

229.8

252.5

220.3

Total Table H- 3

478.7

472.8

Static wheel loads of the NRC vehicle used for the tests in Switzerland.


H.2

H.13 / 20 May 30, 2000

TEST VEHICLES USED FOR THE AUSTRALIAN BRIDGE TESTS

H.2.1 General Details of the three test vehicles used in the Australian tests are presented below. The BA and BS (Fig. H- 12, Fig. H- 13) test vehicles shared the same tip trailer. The SA (Fig. H- 14) test vehicle was only used during the Camerons Creek Testing. The dimensions, suspensions and axle loads are summarized in Table H- 4. All of these vehicles are 6 axle tractor semi-trailers.

Fig. H- 12 The BA test vehicle, Lawsons Creek.


H.14 / 20 May 30, 2000

Fig. H- 13 The BS test vehicle, Lawsons Creek. Note axle hop plank behind rear axle.

Fig. H- 14 The SA test vehicle crossing Camerons Creek


H.15 / 20


Prime-Mover


May 30, 2000

Trailer

Gross Laden Mass (kN)

Vehicle Code


425

BA

60 kN

165 kN

200 kN

4.65

1.23

Over the rear axle tri-axle tipper, York 8 leaf steel suspension

425

4.07

1.30

1.23

12.48 m

Freightliner, Hendrickson walking beam, steel suspension

60 kN

165 kN

3.71

BS

200 kN

1.34

4.65

1.23

1.23

12.16 m


Petrol tanker, BPW air suspension with shock absorbers fitted diagonally

55 kN

450

165 kN

3.66

SA

230 kN

1.32

6.56

1.4

1.4

14.34 m


Over the rear axle tri-axle tipper, BPW air suspension 60 kN

165 kN

4.07

425

BA unequal

200 kN

1.30

4.03

1.85

1.23

12.48 m

Table H- 4

Test vehicle dimensions, weights (nominal) & suspensions - Australia.


H.16 / 20


Vehicle

BA Series, Freightliner and

May 30, 2000

Prime-mover

Trailer


Damping ratio [%]


Damping ratio [%]

1.45

20

1.4

35

2.75

6

3.2

10

BPW air suspensions BS Series, Hendrickson Walking beam and York 8 leaf steel suspensions

Table H- 5

H.2.2

Low frequency suspension characteristics of Australian test vehicles [Sweatman, 1994]

Steel Suspensions

H.2.2.1 The prime-mover walking beam suspension (BS vehicle)

Fig. H- 15 Hendrickson walking beam suspension fitted to the prime-mover of the BS test vehicle


H.2.2.2 The trailer steel suspension (BS vehicle)

Fig. H- 16

York 8 leaf steel suspension fitted to trailer of BS vehicle.

H.17 / 20 May 30, 2000


H.2.3

H.18 / 20 May 30, 2000

Air Suspensions

H.2.3.1 The trailer air suspension (BA vehicle)

Fig. H- 17 BPW air suspensions, technical drawing and picture of one axle removed from test vehicle.


H.19 / 20 May 30, 2000

H.2.3.2 The trailer air suspension (SA vehicle)

Fig. H- 18 Air suspension fitted to trailer on SA test vehicle.

H.2.4

Tires

The tires fitted to each axle of the Australian test vehicles were 11R 22.5. The tire pressures were 700 kPa. H.2.5

Instrumentation

The tri-axle groups on the BA and BS test vehicles were instrumented in order to measure the dynamic wheel forces as the test vehicles crossed the bridge. Strain gauge rosettes were fitted to the ends of each axle by BPW Germany. These rosettes were orientated and wired so as to be sensitive to shear but insensitive to bending and torsion effects. In addition, the axle acceleration was also recorded. The axle shear force and the acceleration of the mass outboard of the shear gauges were then combined analytically to estimate the dynamic wheel force applied to the pavement (refer Fig. H- 19).


H.20 / 20


Strain gauges to measure strains due to shear

Outboard mass mo

May 30, 2000

FP = FV + moxa where:

FV

FP

mo x a

Dynamic wheel force applied to the pavement

FV Accelerometer & signal conditioning

Dynamic shear force measured in axle stub

mo

FP

The mass outboard of the strain gauges

a

Fig. H- 19

(1)

The acceleration of the outboard mass (m).

Instrumentation for measuring dynamic wheel forces




Annex I References

OECD DIVINE, Element 6, Bridge Research Final Report, EMPA Switzerland, QUT Australia, Annex I

I.2 / 5 May 30, 2000

[1]

AUSTROADS, Bridge Design Code, AUSTROADS (1992).

[2]

Cantieni, R., Ladner, M., Ponte sulla Föss 2 - Belastungsversuche vom 29. August 1975. EMPA-Untersuchungsbericht Nr. 34'110/3 (1976)

[3]

Cantieni, R., Dynamic Load Tests on Highway Bridges in Switzerland - 60 Years Experience of EMPA. EMPA Report No. 211, Swiss Federal Laboratories for Materials Testing and Research, Dübendorf, Switzerland (1983)

[4]

Cantieni, R., Stossempfindlichkeit von Strassenbrücken unter rollendem Verkehr, Dynamische Versuche an der Bergspurbrücke Deibüel. Forschungsberichte Nr. 169 A und 169B, Eidg. Verkehrs- und Energiewirtschaftsdepartement, Bundesamt für Strassenbau, Bern (1988)

[5]

Cantieni, R., Dynamic Behavior of Highway Bridges Under the Passage of Heavy Vehicles. EMPA Report No. 220, Swiss Federal Laboratories for Materials Testing and Research, Dübendorf, Switzerland (1992)

[6]

Cantieni, R., Dynamic Testing of a Two-Lane Highway Bridge. Proc. Int. Conf. "Traffic Effects on Structures and Environment", Štrbské Pleso, Czechslovakia, December 1-3 (1987)

[7]

EDI Experimental Dynamic Investigations Ltd., U2, V2 & P2 Manual Version 2.1, Vancouver, B.C., Canada, April (1995)

[8]

Egger, G., Ladner, M., Bergspurbrücke Deibüel - Statische und dynamische Belastungsversuche vom 7. Oktober 1975. EMPA-Untersuchungsbericht Nr. 34'779/1 (1975)

[9]

Egger, G., Ladner, M., Ponte Cavalcavia Sort - Belastungsversuche vom 15. April 1977. EMPA Untersuchungsbericht Nr. 37'629 (1977)

[10]

Felber, A.J., Development of the Hybrid Bridge Evaluation System. Ph.D. Dissertation, Department of Civil Engineering, University of British Columbia, Vancouver, B.C., Canada (1993)

[11]

Felber, A., Cantieni, R., Ambient Vibration Survey of the Sort Bridge, EMPA Test Report No 153’031/1, April (1995a)

[12]

Felber, A., Cantieni, R., Ambient Vibration Survey of the Deibüel Bridge, EMPA Test Report No 153’031/2, April (1995b)


I.3 / 5 May 30, 2000

[13]

Felber, A., Cantieni, R., de Smet, C.A.M., Ambient Vibration Survey of the Föss Bridge, EMPA Test Report No 153’031/3, April (1995c)

[14]

Felber, A.J., Cantieni, R., Introduction of a New Ambient Vibration Testing System, Description of the System and of Seven Bridge Tests, EMPA R&D Report No. 156'521 (1996)

[15]

International Organization for Standardization, Draft International Standard ISO DIS 8608, Mechanical vibration - Road surface profiles - Reporting of measured data (1991)

[16]

International Organization for Standardization, International Standard ISO 8608:1995(E), Mechanical vibration - Road surface profiles - Reporting of measured data (1995)

[17]

LeBlanc, P.A., Woodrooffe, J.H., Papagiannakis, A.T., A Comparison of the Accuracy of Two Types of Instrumentation for Measuring Vertical Wheel Load. Proc. 3rd Int. Symposium on Heavy Vehicles Weigths and Dimensions, Cambridge, UK, 28 June - 2 July 1992, pp. 86-94, Thomas Telford, London (1992)

[18]

Dynamic Interaction Between Vehicles and Infrastructure Experiment (DIVINE), Technical Report. Report 71048, DSTI/DOT/RTR/IR6(98)1/FINAL, OECD/OCDE, Paris (1998)

[19]

NAASRA, National Association of State Road Authorities, ‘Review of Road Vehicle Limits’, Sydney, Australia (1986).

[20]

SIA, Actions on Structures. SIA Code 160(1989), Swiss Society of Engineers and Architects, Zürich (1991)

[21]

Sweatman, P.F., Ranking of the road friendliness of heavy vehicle suspensions: Low frequency dynamics. National Road Transportation Commission technical working paper No 13 (1994).

[22]

Willis, R., An Essay on the Effects Produced by Causing Weights to Travel Over Elastic Bars. Addendum to: Barlow, P., A Treatise on the Strength of Timber, Cast and Malleable Iron. John Weale, London (1851)


I.4 / 5 May 30, 2000

References given below are not referred to in this report [23]

Aurell, J., The Influence of Different Suspension Parameters on Dynamic Wheel Loads. IMechE Conference on Road Wear: The Interaction Between Vehicle Suspensions and the Road, London (1991).

[24]

Cantieni, R., Barella, S., Swiss Testing of Medium Span Bridges. Proceedings of Dynamic Loading of Heavy Vehicles and Road Wear Mid-Term Seminar, OECD DIVINE Project, AUSTROADS, Sydney (2&3 Feb 1995), Paper 18.

[25]

Chan, T.H.T., O’Connor, C., Vehicle Model for Highway Bridge Impact. Journal of Structural Engineering 116, No, 7 (1990)

[26]

Green, M.F., Cebon, D., Cole, D.J., Effects of Vehicle Suspension Design on the Dynamics of Highway Bridges. (in press) Journal of Structural Engineering, ASCE.

[27]

Heywood, R.J., Truck suspensions and the dynamic response of the short-span bridges over Camerons and Cromarty Creek, NSW. Physical Infrastructure Center Research Report 94-25 (1994).

[28]

Heywood, R.J., Yarriambiack Creek Bridge - Dynamic Response and Truck Suspensions. Queensland University of Technology, Physical Infrastructure Centre Research Report 94-26, p.35, (1994).

[29]

Heywood, R.J., Dynamic Response of Douglas Fir Stress Laminated Timber Girder Bridge-Yabba Creek No. 4. Queensland University of Technology, Physical Infrastructure Centre Research Report 94-24, p.35, (1994).

[30]

Heywood, R.J., Wedgwood, R., Ariyaratne, W., Truck suspensions and the dynamic response of the short span bridges over Camerons and Cromarty Creek, NSW. (Draft Report), Queensland University of Technology, Physical Infrastructure Centre Research Report (1994).

[31]

Heywood, R.J., ‘Road-Friendly’ Suspensions and Short Span Bridges. ARRB Research Report ARR 260, Dynamic Interaction of Vehicles and the Infrastructure, Proceedings of DIVINE Special Session: 17th ARRB Conference (ed K.G. Sharp), pp. 39-65, (1995).

[32]

Heywood, R.J., Short-Span Bridge Friendly Suspensions - Research Element 6 of the OECD DIVINE Project. (in press) 4th International Symposium on Heavy Vehicle Weights and Dimensions, Ann Arbor, (1995).


I.5 / 5 May 30, 2000

[33]

Heywood, R.J., Are ‘Road-Friendly’ Suspensions ‘Bridge-Friendly’? - The OECD DIVINE Project. (in press) 4th International Bridge Engineering Conference, Transport Research Board, San Francisco (1995).

[34]

Heywood, R.J., “DIVINE Element 6 - Dynamic Bridge Testing - Short Span Bridges.” Proceedings of Dynamic Loading of Heavy Vehicles and Road Wear Mid-Term Seminar, OECD DIVINE Project, AUSTROADS, Sydney, Paper 19, (1995).

[35]

Ministry of Transportation and Communication, Ontario Highway Bridge Design Code Commentary, TRR 950/1, (1984).

[36]

Nowak, A.S., Bridge Load Models. Austroads conference: Bridges - Part of the Transport System (ed R.J. Heywood), pp. 503-524, (1991).

[37]

O’Connor, C., Pritchard, R.W., Impact Studies on Small Composite Girder Bridge. ASCE Journal of Structural Engineering 111, No 3, pp 641-653, (1985).

[38]

O’Connor, C. & Pritchard, R.W., Impact study on small composite timber bridge. Journal of Struct. Eng. 111, (1985).

[39]

Prem, H., A Laser-based highway-speed road profile measuring system. The dynamic of vehicles on roads and tracks, International Association for Vehicle System Dynamics (1988).

[40]

Sweatman, P.F., A study of dynamic wheel forces in axle group suspensions of heavy vehicles. ARRB Special Report 27 (1983).

[41]

Sweatman, P.F., A Study of Dynamic Wheel Forces in Axle Groups Suspensions of Heavy Vehicles. Australian Road Research Board Special Report 27 (1983).

[42]

Wright, D.T., Green, R., Highway Bridge Vibrations, Part II, Ontario Test Programme. Report 5, Queen’s University, Kingston, Ontario, (1962).

Dynamic Interaction Between Vehicle and Infrastructure Experiment ...

Dynamic Interaction Between Vehicle and Infrastructure Experiment ...

Suggest Documents

dynamic interaction between a lumped mass vehicle ... - Science Direct

Dynamic Interaction Between Microtubules and the Nucleus

Dynamic interaction between retinal and ... - Semantic Scholar

Vehicle-to-Infrastructure and Vehicle-to-Vehicle ... - Semantic Scholar

A Study on the Dynamic Interaction between Three-Axle Vehicle and ...

Dynamic magnetostatic interaction between amorphous ferromagnetic

vehicle-bridge interaction element for dynamic analysis

Dynamic responses of vehicle ballasted track interaction system for ...

Dynamic Interaction Between a Two-Phase Submarine Landslide and

Interaction between SPH Fluids and Dynamic ... - Semantic Scholar

Dynamic interaction between fetal adversity and a genetic score ...

Fluid dynamic interaction between train and noise ... | ScienceDirect

The Interaction Between Dynamic Marketing Capability and Service

Dynamic Interaction between Machine, Tool, and Substrate in Bobbin ...

Interaction between Fiscal and Monetary Policy in a Dynamic ...

connected vehicle/infrastructure university ... - VTechWorks

Electric Vehicle Policies, Fleet, and Infrastructure ...

Electrified Vehicle Technology Trends, Infrastructure Implications, and ...

OLEV Electric Vehicle Infrastructure Strategy

Modeling Behavioral Experiment Interaction and ... - GMRV

Interaction between exopolysaccharide and

Interaction of theory and experiment: examples

Identification in dynamic networks: identifiability and experiment

Experiment and Simulation of Dynamic Voltage