Load testing of a large viaduct
Luiacute;s Oliveira SANTOS
Research Officer LNEC
Lisbon, Portugal
Luiacute;s Oliveira Santos, born 1964, received his civil eng. degree (1987), his MSc (1992) and his PhD (2001) from the Tech. Univ. of Lisbon. He has been involved in load testing and monitoring of time-dependent behaviour of PSC bridges.
Jorge RODRIGUES
Research Assistant LNEC
Lisbon, Portugal
Jorge Rodrigues, born 1966, received his civil eng. degree (1989) from the Tech. Univ. of Lisbon. His research activities have been dedicated to finite element analysis, dynamic testing of civil eng. structures and output-only modal id. methods.
MinXU PhD LNEC
Lisbon, Portugal
Xu Min, born 1964, received his naval engineering degree from the Univ. of Jiao Tong of Shanghai and his civil eng. PhD from Tech. Institute of Civil Eng. of Kiev.
Summary
This paper presents the static and dynamic testing of the Loureiro Viaduct, located at the A10 highway, near Lisbon. Some innovative procedures were used during these tests, like an upgraded hydrostatic levelling system to measure vertical displacements or accelerometers to measure rotations. The dynamic tests allowed the identification of 25 modes of the natural vibration of the structure, using the technique of output-only modal identification. The experimental results are compared with the analytical values computed by the FE model developed.
Keywords:load testing, hydrostatic levelling system, monitoring, dynamic tests, modal identification, prestressed concrete bridge.
Introduction
The Loureiro Viaduct, located at the A10 highway, near Lisbon, was subjected to static and dynamic tests in July 2003, in order to perform an evaluation of the static behaviour of the bridge and to identify its dynamic characteristics (as vibration frequencies, mode shapes and damping ratios).
Loureiro Viaduct is a prestressed concrete structure, 1050 m long, with five major spans of 100 m, besides other 11 spans. The viaduct is curved in plan and the maximum height of the columns is 95 m [1]. The deck is a box-girder with 8.00 m width and a maximum height of 5.55 m. The piers of the major spans are rectangular with 8.00 times; 6.00 m, from the foundations up to 30 m below the deck, but, from this height up, the piers are made by two concrete plates with 8.00 times; 0.80 m, monolithic with the deck. Fig. 1 shows the elevation and plan views of the viaduct. General views are presented in Fig. 2.
This paper presents the finite element model used and the experimental procedures adopted. Both static and dynamic experimental results are compared with the analytical values computed by the finite element model.
Analytical model
A three dimensional, linear, elastic numerical model of the viaduct was developed in SAP2000 [2] to evaluate its response to static tests and its dynamic characteristics.
Shell elements were used for modelling the deck and the piers. Bearings at piers P8 to P15 and abutments were modelled by link elements.
Before the load tests the preliminary FE model was used to estimate the deformation of the structure on static loads and the shapes of the natural vibration modes. After the tests, the FE model was calibrated with the results from the static load tests, and then adjusted to the dynamic characteristics identified with the dynamic tests.
A9 CREL CARREGADO
40 60
100 100
100 100 100 60 50 50 50 50 50 50 50 40
E1
P1
P2
P3
P4 P5 P6
P8 P9 P10 P11 P12
P7
P13 P14
P15 E2
Fig. 1 Elevation and plan view of Loureiro Viaduct
Fig. 2 General views of Loureiro Viaduct
Static test
-
- Testing procedure
The static test was performed with twelve loaded lorries with a total weight of 3527 kN (Fig. 3). These loads were placed in 27 positions, in accordance to the load plan that maximizes the most important effects in the structure, however without inducing unwanted situations of early cracking in the structure.
During the test, vertical displacements, rotations and strains were measured at several sections. In order to
Fig. 3 Loaded lorries in use as a load test
measure the most reliable and redundant data, different types of sensors were installed.
Vertical displacements were measured at mid spans of major spans by an upgraded hydrostatic levelling system associated to pressure cells. Some of these sections were also instrumented with traditional mechanical apparatus as deflectographs, which were also used at mid span and at the supports of the other spans.
Transversal and longitudinal rotations were measured by electric clinometers located at the top of piers P2, P4 and P6. Quite interesting was the use of force balance accelerometers during static
Fig. 4 Pressure cell used in hydrostatic levelling system
test to measure rotations. As a matter of fact, during a static test, a uniaxial accelerometer measuring horizontal acceleration, measures that direction component of the gravity, which changes when that
point rotates. These accelerometers were used at the top of piers P4 and P5 and at several sections of the deck, as presented in Fig. 5. To increase the redundancy of measured data, rotations at the top of pier P2 were also measured by mechanical air-bubble clinometers, in both longitudinal and transversal directions.
Strains were measured by inductance strain meters at the mid span P4-P5.
60 brvbar;00 brvbar;00 brvbar;00 brvbar;00 brvbar;00
P1
P2
P3
Accelerometer
P4
Air-buble clinometer
P5
Electric clinometer
P7
P6
Pressure cell
Fig. 5 Equipment used at major spans
Three automated data-acquisition systems DataTaker DT515 were used to read data from hydrostatic levelling systems and from electr
大型高架桥荷载试验
Luiacute;s Oliveira SANTOS
高级研究员
土木工程
国家实验室
里斯本，葡萄牙
Luiacute;s Oliveira Santos，1964年出生，1987年获得土木工程学位，1992年和2001年在里斯本理工大学分别获得理学硕士和博士学位。参与了PSC桥梁的荷载测试和时变特性检测。
Jorge RODRIGUES
研究助手
土木工程
国家实验室
里斯本，葡萄牙
Jorge Rodrigues, 1966年出生，1989年在里斯本理工大学获得土木工程学位。他的研究致力于有限元分析、土木结构的动态测试和只靠输出模态来识别的方法。
MinXU 教授
土木工程
国家实验室
里斯本，葡萄牙
Xu Min, 1964年出生，上海交通大学海军工程学位，基辅土木工程技术研究所土木工程博士。
摘要
本文介绍了位于里斯本附近的A10高速公路上Loureiro高架桥的静载试验和动载试验。在这些试验中，采取了一些创新的程序，比如升级之后的液压调平系统来测量垂直方向位移、加速度计来测量旋转。动载试验对结构的25个固有振动模态采用了仅用输出模态来识别的技术。实验结果与建立的有限元模型计算分析值进行了比较。
关键词：荷载实验；静液平衡系统；监测；动载实验；模态识别；预应力混凝土桥梁。
导言
Loureiro高架桥位于里斯本附近的A10高速公路上，2003年7月进行了静载和动载试验，以评估桥梁的静态性能并确定其动态特性（如振动频率、振型和阻尼比）。
Loureiro高架桥是预应力混凝土结构，长1050米，5个100米长的主跨之外还有11个分跨。高架桥平面呈曲线，立柱最大高度为95米。桥面板为箱梁，宽8米，最大高度5.55米。主跨桥墩为矩形，尺寸为8times;6米，基础到上方30m的板处的桥墩由两块8times;0.8米的混凝土板组成，与桥面板一致。图1显示了高架桥的立面图和平面图。一般视图如图2所示。
本文介绍了试验所采用的有限元模型和试验采集程序。将静载和动载试验结果与有限元模型计算的分析值进行了比较。
分析模型
使用SAP2000建立了高架桥的三维线性弹性数值模型，以评价高架桥在静载试验中的响应和动力特性。
使用壳单元模拟板和桥墩。桥墩P8到P15桥台处的支座用连接单元来建模。
在荷载试验之前，用有限元模型估算结构在静载作用下的变形和自振模态。试验后，利用静载试验结果对有限元模型进行校准，并将其调整到动态试验所确定的动态特性。
A9 CREL CARREGADO
40 60
100 100
100 100 100 60 50 50 50 50 50 50 50 40
E1
P1
P2
P3
P4 P5 P6
P8 P9 P10 P11 P12
P7
P13 P14
P15 E2
图1.Loureiro高架桥立面图和平面图
图2.Loureiro高架桥总视图
静载试验
-
- 试验过程
静载试验使用12辆总重量为3527千牛的载重卡车进行。根据试验方案，这些荷载要放置在27个位置，最大限度地对结构产生效应，但不会导致结构出现不必要的早期开裂情况。
在试验过程中，会在几个不同截面上测量垂直位移、旋转和应变。 为了测量最可靠和不需要的数据，安装了不同类型的传感器。
图3.荷载试验所用的装载卡车
通过与压力室相关的升级液压水准系统测量主跨跨中的垂直位移。其中一些节段还配备了传统的机械装置作为偏转仪，这些装置也用于其他跨段的中跨支撑。
通过位于P2、P4和P6顶部的电子测角仪测量横向和纵向旋转。有趣的是在静载试验中使用力平衡加速度计来测量旋转。
图4.液压水准系统中使用的压力传感器
事实上，在静载试验期间，用测量水平加速度的单轴加速度计测量重力方向在该点旋转而发生变化的分量。如图5所示，这些加速度计用于桥墩P4和P5的顶部以及桥面的几个部分。为了增加测量数据的多余程度，还采用了机械气泡测力计在纵向和横向测量了桥墩P2顶部的旋转。
应变测量采用的电感应变计布置在跨中P4-P5处。
60 brvbar;00 brvbar;00 brvbar;00 brvbar;00 brvbar;00
P1
P2
P3
Accelerometer
P4
Air-buble clinometer
P5
Electric clinometer
P7
P6
Pressure cell
图5.大跨度桥梁使用的设备
三个自动数据采集系统——数据采集DT515被用来从液压水准系统和电子测斜仪中读取数据。加速度计和电感应变计的读数由动载试验中常用的采集系统完成。两个系统都能实时有效地控制试验数据。
-
- 主要结果
图6.记录仪器
静载试验中获得了大量的试验数据。在这些所有数据中，具有说明性的结果如图7所示。这些图表显示出了实验值和计算的结构变形有良好的一致性。液压水准系统测量值和偏转仪测量值之间也有良好的一致性。
Dz (mm)
-20
0
20 E1 P1 P2 P3 P4 P5 P6 P7 P8
40
Dz (mm)
-20
0
20 E1 P1 P2 P3 P4 P5 P6 P7 P8
40
Dz (mm)
-20
0
20 E1 P1 P2 P3 P4 P5 P6 P7 P8
40
Hydrostatic levelling system Deflectographs Computed
图7.垂直方向位移的测量值和计算值
能够实现良好的相关性也与电测斜仪、机械气泡测斜仪、加速度计和数值模型计算相关。
动载试验
-
- 试验过程
进行动载试验是为了获得结构的动态特性（振动频率、振型和阻尼比）。在试验过程中，使用了14个运动测量单轴型传感器(ES-U)加速度计（图8）以及在LNEC科学仪器中心开发的信号调节设备，用国家仪器数据采集设备（DAQ板AI-16XE-50和SCXI-1000DC底盘以及SCXI-1140采集板）测量环境激励（主要是风）引起的加速度。
环境振动试验在九个装置中进行。在这些装置安装过程中，在高架桥的29个路段测量了垂直和横向加速度，其中两个路段固定了五个传感器，而其他路段则在移动中设置。利用输出模态识别方法，对结构的几种固有振动模态进行了识别，并对各模态的特性进行了估计。加速度计的定位如图9所示。固定设备位于第33、34、43和44点，如图9所示。
图8.加速度计
在每次动态试验设置中，使用1000赫兹的采集频率在约30分钟的时间内获取环境振动数据。这样获得的数据纪录之后，使用8级Butterworth滤波器在6.25赫兹的低频滤波进行预处理，并抽取到15.625赫兹的采样频率。
A9 CREL CARREGADO
brvbar;00 brvbar;00
brvbar;00
brvbar;00
brvbar;00
25 25 brvbar;6 brvbar;6 brvbar;8
S1 S2 S3 S4 S5 S6
brvbar;8 brvbar;6 brvbar;6 brvbar;6 brvbar;6 brvbar;8 brvbar;8 brvbar;6 brvbar;6
S7 S8 S9 S10 S11 S12 S13 S14
brvbar;6 brvbar;6 brvbar;8 brvbar;8 brvbar;6 brvbar;6 brvbar;6 brvbar;6
brvbar;8 brvbar;8 brvbar;6 brvbar;6 25 25
S15 S16 S17 S18 S19 S20
S21 S22 S23 S24
S25 S26
S27
S28
S29
P2
data acquisition
P3 equipment and
P7
computer
P5 P6
P4
17
13 15
19 21
22
23 25
24 26
27
28
29 31 33
30 32 34
35
36
37 39 41
38 40 42
43 45
44 46
47 49 51 53
48 50 52 54
P6
55 57
56 58
11
9
7
5 12
3 8 10
1 6
18 20 P5 P7
14 16
P4
4
2 P3
P2
图9.动载试验中加速度计的位置
除了环境振动试验之外，还对卡车在控制速度下通过高架桥进行了动载试验，以评估该作用的动态效应。
-
- 模态识别
高架桥的模态识别采用输出模态识别软件ARTEMIS。该程序允许从实测响应中估算结构的振动固有频率、相关振型和模态阻尼比，使用起来方便快捷。
根据测试数据，估计除了光谱密度和相关函数。功率谱密度（PSD）矩阵由样本计算，每个样本有1024个数据点，重叠率为66.67%。采集的频率为15.625赫兹，因此频谱的频率为0.015赫兹。
采用在ARTEMIS实现的频域分析（FDD）技术对结构的固有频率和振型进行识别。在该技术手段中，功率谱密度（PSD）矩阵通过奇异值分解（SVD）在每条频率线上进行分解。奇异值（SV）图作为频率函数，由SVD估计，可用于确定模态频率。奇异值图的峰值表明了结构模态的存在。与局部最大奇异值相对应的奇异向量是响应的无符号振型。图10显示垂直加速度的PSD矩阵的前9个奇异值的光谱。在该图中，还显示了由FDD识别的固有频率。
利用FDD技术，通过对动态试验数据的分析，共识别出12个垂直振动模态、1个扭转模态和12个横向振动模态。前三个横向和垂直振型如图11和图12所示。
d B | 1 .0 / Hz
2 0
Frequenc y Domain Dec ompos ition - Peak Pic king A v erage of the Normaliz ed Singular V alues of Spec tral Dens ity Matric es of all Data Sets .
1 0
0
-1 0
-2 0
-3 0
-4 0
-5 0 0 2 4 6
Fre q u e n cy [Hz] AR TeMI S Ex t rac t or, 000-46e2-950a-640f b97f df 5, AR TX-0310E-150102H AN D Y , Ent erpris e Lic ens e Projec t : v lour_v ert . ax p
图10.FDD：PSD矩阵奇异值和意思别自然频率的光谱
f = 0.481 Hz f = 0.603 Hz f = 0.847 Hz
图11.确定横截面形状
f = 1.114 Hz f = 1.305 Hz f = 1.572 Hz
图12.确定垂直方向位移形状
-
- 与有限元模型的比较
6
Vert. modes Transv. modes
5
4
Comput. freq. (Hz)
3
2
1
0
0 1 2 3 4 5 6
Experim ental frequency (Hz)
图13.有限元模型与试验频率
利用静载试验结果验证的有限元模型对动载试验结果进行了解释。
根据动力试验的模态识别，对有限元模型进行了调整。为模拟桥台处的支座，对连接件的刚度进行了调整，以使计算的模态特性与试验确定的模态特性相匹配。
更新的有限元模型计算出的固有频率与图13所示试验确定的频率进行了比较。固有频率和振型都具有良好的一致性。
结论
静载试验和动载试验结果与有限元模型计算的分析值都有很好的相关性。
液压水准系统结合压力传感器是测量箱梁桥垂直方向位移的一种精确方法
在静载试验中使用加速度计是非常成功的成就。
关于这座高架桥结构性能的试验数据对评定其在施工结束之后和通车之前的实际
资料编号：[5649]
以上是毕业论文外文翻译，课题毕业论文、任务书、文献综述、开题报告、程序设计、图纸设计等资料可联系客服协助查找。
