Luận văn: Thí nghiệm tải thực địa & mô hình cầu giàn thép, T. D. Khanh

長岡技術科学大学 The Nagaoka University of Technology The Graduate School of Engineering Faculty of Civil and Environmental Engineering Field Load Experiments and Computer Modeling of a Steel Truss Bridge for Assessment A Thesis in Civil Engineering by TRAN DUY KHANH Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Science Nagaoka, Niigata July 2015 i The thesis of TRAN DUY KHANH was reviewed and approved by the following: Takeshi MIYASHITA Associate Professor of Civil Engineering Eiji IWASAKI Professor of Civil Engineering Takumi SHIMOMURA Professor of Civil Engineering ii An abstract of Field load experiments and computer modeling of a steel truss bridge for assessment In Japan, it can be surely predicted that in 2030s, more than 50 percent of existing bridges, which were primarily constructed during the post World War II, a rapid economic growth between 1955 and 1975, is assumed more than 50 years of age. The research investigated an in-service, multi-span old steel through truss highway bridge constructed and opened to the traffic in 1937. The bridge, also known as a city symbol, which is on the National Highway No.351, spans the Shinano River, in Nagaoka City, Japan. The bridge was constructed with materials of steel for frames and concrete for the deck, consists of 13 spans composing of 2 cantilever anchored spans, and 11 suspension spans and cantilever central anchored spans alternatively connected via hinges at upper level and pins at lower level with a total length of 850 meters.

To ensure the normal service of the aging bridge and meet the transportation demand of citizens, a set of full-scale field load tests including static, dynamic load and short-term monitoring experiments was performed on the bridge to identify the current health status, and characterize the response to practical load conditions. The responses of the excited bridge was recorded continuously and simultaneously at 200 Hz- sampling frequency through an array of 100 strain gages installed on 25 identical and key members, and each of 26 displacement sensors and accelerometers instrumented at each span center point. The exciters used in these tests were three-axle 20-ton dump trucks for controlled tests and normal daily traffic for the short-term monitoring measurement. A three-dimensional finite element model of the bridge superstructure was constructed with geometry and structural arrangement extracted from recovered drawings.

The model, simulated the in-situ tests, was validated manually through a comprehensive comparison of analytical results and measurement data in terms of internal forces, vertical displacements. The axial force extracted from the upper chords and diagonal chords instrumented in the static loading tests was compared the computational axial iii forces from the static load analysis. Similarly, the maximum vertical deflections at each span middle point were also an objective for model validation when making a comparison between test and analytical results. The comprehension of the transverse direction behavior of the structure through the load distribution to the truss characterized by the dynamic load test utilizing a single known-weight truck controlled to travel on the different traffic lanes.

Bending natural frequencies based on power spectral density extracted from the measured acceleration data via Fast Fourier Transform and the peak picking method was compared to computational analysis generated frequencies for the model validation process. The boundary condition of the current bridge was qualified with changes in bearing support types and various stiffness level of Gerber hinges in three directions of vertical, horizontal and rotational. Based on the controlled dynamic loading tests and short-term monitoring with daily traffics, dynamic loading allowance were observed by using a digital signal filter tool called low-pass Butterworth with a cut-off frequency and filtering order. The maximum static response were filtered and compared to the corresponding practical dynamic response to obtain a dynamic loading allowance.

The correlation between the impact factor and measuring types of strain and vertical displacement were examined. The dynamic loading factor magnitude is strongly correlated with instrumented member maximum static strain with a general decrease trend in the static strain increase. In contrast, the dynamic loading allowance shows a quite stable trend in the maximum deflection increase. The calibrated model of the bridge is used to perform a load rating analysis based on Manual for Bridge Evaluation published 2013 and described by AASHTO.

The load rating factors of those instrumented truss members with the values greater than one implies the bridge is safe for the applied loading. iv Table of Contents List of Figures .ix List of Tables. xiv Chapter 1 Introduction. Objectives of the research.

Scopes of research. 5 Chapter 2 Bridge description, field loading tests and monitoring. Field loading tests and monitoring. Background and purpose.

Static loading test. Dynamic loading test. Short-term monitoring. Measurement processing for developing a finite element model.

Thermal axial stress. 33 Chapter 3 Finite element model validation. Finite element (FE) model construction. Parameter study: Boundary condition study.

Thermal induced axial stress. Lateral distribution stiffness. 62 Chapter 4 Dynamic loading allowance. Definition of dynamic loading allowance.

Specification prescribed dynamic load allowance. Constant dynamic load allowance. Span length varying dynamic load allowance. Frequency varying dynamic load allowance.

Dynamic field testing of bridges. Other dynamic load allowance studies. Determine the low-pass cut-off frequency. Determine the filtering order.

Filtering by Matlab. Controlled dynamic load test data. Short-term monitoring data. Impact factor versus strain measurement.

Impact factor versus displacement measurement. 81 Chapter 5 Bridge Assessment by AASHTO Code. Load rating procedure. FE model analysis.

Load rating results. 90 Chapter 6 Conclusions and recommendations. 139 viii List of Figures Figure 1-1 Composition of bridge’s age to the total number of bridge in Japan 2 Figure 1-2 Total number of highway bridges and construction year in Japan. 3 Figure 2-1 Location of the target study bridge.

7 Figure 2-2 A view of Chyosei Bridge-Niigata, Japan. 8 Figure 2-3 Cross-section of the bridge at mid-span and support. 9 Figure 2-4 Elevation of the Chyosei Bridge with condition of hinges and bearings as designed. 9 Figure 2-5 Typical cross-sections.

11 Figure 2-6 Upper (a) and lower (b) bracing systems. 12 Figure 2-7 Location of sensors in the field test. 17 Figure 2-8 Instrumentation in the field tests: a) Strain gages in Upper chord; b) Strain gages in Diagonal member; c) Displacement transducer in Lower chord; d) Accelerometer in Vertical member; e) Accelerometer over wheel guard of the pedestrian bridge. 19 Figure 2-9 Strain gage location in instrumented members: a) Foil strain gages in Upper chord Layout; b) Foil strain gages in Diagonal member Layout; c) Detail of Upper chord Layout; d) Detail of Diagonal member Layout.

20 Figure 2-10 Test truck configuration. 21 Figure 2-11 Cross-section of 2 trucks utilized for loading tests. 21 Figure 2-12 Load configurations used for the first load test. 22 Figure 2-13 Load configuration used for the second load test.

22 Figure 2-14 Dimensions of the trucks used for the load tests. 23 ix Figure 2-15 Patterns for static load tests. 24 Figure 2-16 Location of test trucks in dynamic loading test. 26 Figure 2-17 Typical strain gauges’ location and internal force notation.

28 Figure 2-18 Deflections by spans in the static load tests. 29 Figure 2-19 Temperature time history. 30 Figure 2-20 Temperature and normal stress relation of an upper chord in the 6th span. 31 Figure 2-21 Normal stress time history of an upper chord in 6th.

32 Figure 2-22 Dynamic component stress history. 32 Figure 2-23 Axial stress and thermal changes correlation. 33 Figure 2-24 Typical frequency spectra for the truck passing (a) and the truck left (b). 35 Figure 3-1 Arbitrary cross-sections.

42 Figure 3-2 Element types used in 3D-FE model. 42 Figure 3-3 The studied bridge structure and its restraints. 44 Figure 3-4 Gerber hinge modeled by 2-node links, SP2TR, SP2RO. 45 Figure 3-5 Deck system (Left) and computer simulation (Right).

46 Figure 3-6 Bearing support computer simulation. 46 Figure 3-7 Typical loading simulation of static load test in the first span. 47 Figure 3-8 Division size and output values correlations. 49 Figure 3-9 Meshed finite element model.

50 Figure 3-10 Model validation procedure. 53 x Figure 3-11 Relative deviation (%) between test and analytical results of each static load case acting on corresponding members. 57 Figure 3-12 Maximum vertical displacement for 13 static load cases. 58 Figure 3-13 Positions of test truck in dynamic tests.

60 Figure 3-14 Influence lines of axial force of upper chords. 62 Figure 4-1 Canadian, Australian, and American DLF versus First flexural frequency. 67 Figure 4-2 Flowchart of dynamic load allowance identification. 70 Figure 4-3 Typical PSD in case of without vehicle presence.

72 Figure 4-4 Typical PSD in case of vehicle presence. 73 Figure 4-5 Typically filtered dynamic and static member strain response. 74 Figure 4-6 Instrumentation configuration. 75 Figure 4-7 Impact factor based on strain.

76 Figure 4-8 Impact factor based on displacement. 77 Figure 4-9 Impact factor by measurement type. 78 Figure 4-10 Typical dynamic load factor versus maximum strain in U1. 79 Figure 4-11 Dynamic load factor versus maximum strain for total samples.

79 Figure 4-12 A typical histogram of dynamic load allowance versus strain response in U4. 80 Figure 4-13 A typical histogram of dynamic load allowance versus strain response in U9. 81 Figure 4-14 Typical dynamic load factor versus maximum deflection in span 1. 82 xi Figure 4-15 A typical histogram of dynamic load allowance versus maximum deflection in span 1.

83 Figure 4-16 Dynamic load factor versus deflections for total samples. 83 Figure 5-1 AASHTO MBE load rating procedure. 87 Figure 5-2 HL-93 truck. 90 Figure 5-3 Rated members in the truss.

91 xii List of Tables Table 2-1 Experimental installation details. 18 Table 2-2 Test Vehicle Weights. 19 Table 2-3 Details of test runs for field tests. 34 Table 3-1 Input parameters utilized in FE model of the bridge.

41 Table 3-2 Material density of dead load effects. 48 Table 3-3 Bearing supports as originally designed. 54 Table 3-4 Gerber hinges as originally designed. 55 Table 3-5 Proposed hinge stiffness in the calibrated model.

56 Table 3-6 Proposed bearing types in the calibrated model. 57 Table 3-7 Input parameters used in analyzing thermal load. 59 Table 3-8 Temperature change induced axial stress (MPa). 59 Table 3-9 Values of k.

61 Table 3-10 Natural frequency (Hz). 63 Table 4-1 AASHTO LRFD Table 3. 65 Table 4-2 Length varying DLF (Eq. 66 Table 5-1 Values utilized in AASHTO MBE load ratings.

88 Table 5-2 Load effect stress and Load rating factors. 92 xiii Acknowledgments I would like to take this space to thank everyone that has made this research a success. First of all, I would like to express my appreciation to Associate Professor. Takeshi MIYASHITA for his very kind guidance, assistance, and support throughout the academic research process.

I could go to ask the essential questions and make me accountable for my work. My thesis and key skills have been improved as a result. I would like to thank and acknowledge Professor Eiji IWASAKI for valuable advice, additional assistance, and useful comments. We would like to express the thankfulness for the invaluable contributions to the research from the Nagaoka Regional Promotion Bureau, Regional Management Department, and Maintenance Division Authority.

The Kozogiken Niigata Corporation and Tokyo Sokki Kenkyujyo Co., who supported to conduct the field load tests, are gratefully appreciated. Last but not least, I would like to thank my families and my friends, my girlfriend for their love and constant support throughout the 2-year research in Japan. Nagaoka July 2015 TRAN DUY KHANH xiv Chapter 1 Introduction 1.