Investigation of displacements of road bridges under test loads using radar interferometry case study

Transcription

1 Bridge Maintenance, Safety, Management, Resilience and Sustainability Biondini & Frangopol (Eds) 2012 Taylor & Francis Group, London, ISBN Investigation of displacements of road bridges under test loads using radar interferometry case study T. Owerko, Ł. Ortyl, R. Kocierz & P. Kuras AGH University of Science and Technology, Kraków, Poland M. Salamak Silesian University of Technology, Gliwice, Poland ABSTRACT: Bridges, before putting into use, are subjected to test loads in order to their acceptance. According to Polish Standards, the aim of loads is to verify the computational model designed for the construction and to confirm the safety margins. Such tests are designed and executed under static and dynamic loads. In the case of static loads conventional surveying techniques are widely applied, while for dynamic loads the measurements using accelerometers are typically provided. These methods are characterized by significant limitations which include i.a. low sampling rate (measurements performed by total stations, digital levels, GPS receivers) and discrete (limited to points) nature of measurement (a set of accelerometers). The authors present the results of comparing the measurements carried out during test loads of a cable-stayed bridge, performed using classical measurement methods and ground-based radar interferometer. This innovative measuring device operates on the basis of stepped-frequency continuous wave modulation and microwave interferometry technique. It allows to determine deformation and vibration frequency of the object through a noninvasive observation of its behaviour at a frequency of up to 200 Hz. 1 INTRODUCTION Diagnosis of bridges requires to develop a reliable assessment of the bridge condition. It is a complex process which requires extensive knowledge of any tested structure. Among many sources of information used in the engineering diagnosis, the results of load testing are of particular importance (Ryall 2010). They are obtained by performing load tests or tests during operation. Static and dynamic load tests are performed during commissioning. Studies are carried out in accordance to the design load, using a controlled load precisely located on the structure. On the other hand, tests during operation are conducted as short-term extemporary studies or long-term monitoring studies with randomly changing operational loads. The results of bridges under static loads are a tool to assess the correctness of work and condition of structures (Lazinski & Salamak 2011). Regular repetition of tests during operation allows the comprehensive assessment of changes in the static properties which indicate the appearance of damages, often difficult to detect by other methods. A separate problem is the study of structures under dynamic loads. On the basis of changes in dynamic characteristics of structures many types of defects may be detected, including structure deformation, destruction and loss of material, loss of material continuity and change of position. Progressive in time parameters that allow to detect defects are: changes of vibration frequencies, mode shapes and damping characteristics. In addition to detecting defects, studies are being conducted on the possibility of locating defects based on the results of dynamic tests (e.g. Maia et al. 2003). Procedures for systematic research of bridge structures under operational loads are usually designed individually for each object as a comprehensive system to monitor its condition. The concept of each system must include the importance of the monitored structure, structure type, operational conditions, and the observed rate of degradation processes. For this reason, techniques used to observe the response of structure under load are very diverse. Ko & Ni (2005) present the wide range of monitoring systems installed on bridges. Typically, a basic group of information about the behaviour of bridge structures are the results of displacement and frequency measurements. Surveying techniques are commonly used: both electro-optical and satellite (Watson et al. 2007). Techniques using displacement transducers are also applied (Paultre et al. 1995). During measurements of velocity and acceleration of vibration, accelerometers and strain gauges are often used (Paultre et al. 1995), as well as displacement transducers. 181

2 In Poland, test loads are carried out according to the following standards: PN-S-10030:1985. Bridges. Loads, PN-S-10052:1982. Bridges. Steel structures. Design, PN-S-10050:1989. Bridges. Steel structures. Requirements and testing, PN-S-10042:1991. Bridges. Concrete, reinforced concrete and prestressed concrete structures. Design, PN-S-10040:1999. Bridges. Concrete, reinforced concrete and prestressed concrete structures. Requirements and testing, Ordinance No. 35 (12 August 2008) of General Director for National Roads and Motorways. Recommendation for testing of road bridges under test load. Well known Eurocodes have also been used since they were introduced in Poland with the national Annexes, according to the Annex 1 to the Decree of the Minister of Infrastructure (12 March 2009) regarding the regulation on the technical conditions to be met by buildings and their location. Test loads are conducted by the independent company of bridge contractor. This introduces additional control of the structure before it is put into operation. Results are passed to the principal recipient (General Directorate for National Roads and Motorways), which is the governmental agenda responsible for the development of motorways and national roads. Type of test measurements performed during load tests depends on the type of construction. According to the type of supporting structure the following values may be measured: beam deflection by dial indicator, bridge support subsidence by precise levelling, pylon deflection by theodolite, horizontal displacement by dial indicator. Dynamic measurements are performed by determination: time series of vertical displacements of the bridge girders by induction sensors, accelerations in vertical and horizontal directions by accelerometers. Spectrum of acceleration measurements is examined in terms of compatibility with modal analysis of structure. Measurements of deflection using sensors and levelling enable to verify the real load capacity of bridge by comparison with deflections values obtained from the design. Plenty examples of the usage of GPS/GNSS technology to load testing and bridge monitoring can be found in literature (e.g. Baraka & El-Shazly 2005, Kapović et al. 2005). A significant disadvantage of above mentioned measurement methods is the necessity of direct access to the structure in order to install sensors and other additional devices, especially in cases of difficult access (bridges over rivers, motorways and railways during traffic). Moreover, these methods provide only discrete information on the current position of the object. To get full information about behaviour of the structure, it is necessary to install a number of sensors or make measurements at many points. This paper presents the results of comparing the measurements carried out during test loads of a cable-stayed bridge, performed using classical measurement methods and ground-based radar interferometer. The advantage of radar interferometer relating to the commonly used methods to assess the behaviour of bridge structures is a non-contact, direct and multi-point measurement of displacements of structure. An important fact is measuring the displacement with the appropriate frequency, which allows for dynamic analysis and determination of displacement amplitude without integrating acceleration measured by common accelerometers. Radar interferometry is a technique used for imaging changes of the earth's surface by satellites (Gens & van Genderen 1996). Recently it is also being used in ground-based imaging. It can provide information about the movements of surface such as slopes or landslides (Pieraccini et al. 2006) and also enables the static and dynamic tests of engineering structures such as bridges (Gentile 2010). 2 SUBJECT OF MEASUREMENT Structure, for which the static and dynamic load tests were carried out, is the bridge over the Skawa River in Zembrzyce, Poland (Fig. 1). It is a cable-stayed structure. Its construction consists of four continuous spans with one nonsymmetrical pylon and steel cables, formed in fan shape. Spans have prestressed, slab-and-girder construction. The designed spans are: = m (Fig. 2). Figure 1. Brigde on the Skawa River under construction 182

3 Figure 2. Longitudinal section of the structure The bridge has the total width of m. A twoway roadway of 8.00 m in width (between kerbstones) runs on the structure. The remaining part is occupied by pavement surfaces with railings and crash barriers. The supporting structure consists of two girders of rectangular cross-section, connected by plate. The girders are 2.03 m high and 0.80 m wide (at the bottom). The deck is 5.50 m wide and has a variable thickness from 30 cm in the middle to 38 cm in the fixing point. Two-sided cantilevers have 3.02 m in length. Diaphragms are designed in the axes of bridge supports and in spans, including places of cables anchorage. Spans no. I and II were designed of prestressed concrete (C35/C45). Spans no. III and IV are cable-supported. Cables were made of bundles of strands 0.6 protected against corrosion by means of the individual polyethylene shields for each strand and cement injections inside casing pipes. At the ends of structure monolithic reinforced concrete abutments were designed, just as two supports and the pylon. The pylon is m high and consists of two asymmetrical arms leaving the stems (C40/C50 concrete). At the top of the pylon, its arms are connected by beams. All supports are located on the piles of 120 cm diameter and 8 m long. 3 MEASUREMENT METHODS To perform measurements during the load tests, the following methods were applied: radar interferometry (IBIS-S system), satellite positioning (GPS/GNSS receivers) and levelling (precise digital level). 3.1 IBIS system IBIS system (Image by Interferometric Survey) was developed by the Italian company IDS to monitor the movements of masses of land and/or engineering structures. The IBIS-S version is used to measure displacements of buildings, of which one dimension is significantly larger than the others, such as towers or bridges (Fig. 3). The IBIS-S system includes: radar unit active radar that generates, transmits and receives electromagnetic waves from the scope of the K u band (the wave frequency is around 17 GHz). notebook with the software controlling operation of the radar, communicating with radar via USB interface, transmitting and receiving antennas ( horns ) of defined characteristics of radiation (Fig. 4), 12 V battery pack, enabling work in field. Figure 3. IBIS-S radar during work Figure 4. Plane patterns of antenna of 23.5 dbi maximum gain used during measurement: a. horizontal, b. vertical The bandwidth range used by the radar is B = 300 MHz, which allows the maximum resolution of R = 0.5 m. The concept of resolution should be understood as the minimum distance between two points on the object at which they may be considered as different points. This means the opportunity to observe each object point separated by not less than 50 cm from another point along the radial direction, i.e. the direction of wave propagation (Fig. 5). If the distance is less than R, the points will be treated as one. The distance is taken along the direction of wave propagation. Parameters that allow to use the IBIS-S radar in displacements measurements are the following: recording frequency of all observed points position: up to 200 Hz (Gentile & Bernardini 2008), measurement range: up to 1000 m (Gentile 2010), accuracy of the radial component of displacement: 0.1 mm (Pieraccini et al. 2004). 183

4 Figure 5. Diagram of radial resolution Resolution, recording frequency and measurement accuracy are mutually dependent and it is not possible to work with maximum values of these parameters simultaneously. IBIS-S can be used to measure short-term fast movements as well as slow static movements. Radar allows also to measure movements in real time. It does not require direct access to the structure, however, identification of the specific points on the radar profile can be supported by installing reflectors on the structure, which scatter the radar beam. The IBIS-S system is based on two radar techniques: microwave interferometry, stepped-frequency continuous wave modulation. Microwave interferometry technique allows to achieve high accuracy measurement of displacement. Point displacement is calculated based on the phase difference of waves received by the receiver at different times. Movement of point in the direction of electromagnetic wave propagation induces a phase shift between the signals reflected from the surface of the object. The value of displacement d along the direction of wave propagation can be written as: d = (λ φ) / (4π), where λ is the length of electromagnetic wave, and φ is the phase shift. Stepped frequency continuous wave technique is used to avoid installation of multiple units of measurement on the structure. The corresponding output signal allows to obtain an image displacement of many points (virtual sensors) on the object. In fact, small inhomogeneities of the object, on which the wave is scattered, are measured. Signal used in radar has the form of a short pulse. The shorter the pulse duration τ, the higher measurement resolution R can be obtained. This follows from the relation R = c τ/2, where c is the speed of light in vacuum. The relationship between pulse duration τ and bandwidth B used of the microwave range can be written as τ B = 1, and therefore the resolution of the radar is R = c/(2b). Increasing the resolution of measurement is achieved by reducing the value of τ or increasing the value of B. Radars SFCW, instead of using short pulses, receive bandwidth through a wide range of discrete, linear increase in the frequency of discrete values of f. Bandwidth can be expressed as: B = (N 1) f, where N is the number of different frequencies which are within the bandwidth B. IBIS-S is a one-dimensional instrument. It is best suited for observation of the high bridges for which the projection factor is beneficial. Frequency of data recording that reaches up to 200 Hz can be successfully used for dynamic monitoring. Not without significance is the fact that at one time it is possible to observe several points representing the object. These points, however, must be separated by at least 0.5 m (radially from the instrument) and have favourable characteristics in terms of the reflection signal. In particular cases e.g. for truss bridges significant problems in obtaining the reflection signal from a given element may occur, even when using corner reflectors. 3.2 GPS/GNNS system in RTK mode Second measurement technology, which allows to obtain position (3D coordinates), is GNSS RTK method (Global Navigation Satellite System, which include NAVSTAR GPS system, in Real Time Kinematic mode). To determine the position in this technology it is necessary to observe at least 5 satellites. GNSS receivers follows phase of the satellite signals at L1 and L2 frequencies. In addition, to calculate the position constant communication with the reference station must be assured. It is necessary to receive corrections to the pseudodistance or observation (now mostly through the GSM network, in RTCM or CMR format). Obviously, the longer distance between rover and reference station, the lower accuracy of positioning. A kind of development of the RTK is the RTN (Real Time Network) mode. The principles of work are the same but determining the position is more accurate. Real time corrections are calculated on the basis of a few nearest reference stations. In order to enable precise satellite positioning on wide area (e.g. entire country), the CORS (Continuously Operating Reference Station) networks are established (Grejner-Brzezinska et al., 2007). RTK GPS systems are widely used in monitoring the behaviour of bridge pylons. They provide continuous reliable information at the level of accuracy of 2 3 cm. Limitations associated with this technology are related to the potentially occurring cover of the horizon and the signal multipath. These problems will occur especially if we want to use these instruments to observe the bridge spans. The standard frequency of data recording is up to 20 Hz, but measurement accuracy significantly reduces application of this technology for vibration measurements. 3.3 Precise digital levelling Digital level performs reading of position of horizontal axis on a rod with optoelectronic method. In the telescope an electronic transducer of rod image is mounted. Image of rod has a form of special bar- 184

5 code, which is composed of alternate light and dark ranges of varying thickness. Reading of position of horizontal axis on a rod is done by comparing two images a real image of rod projected by optical system on CCD sensor and the pattern of road inserted on a chip by manufacturer. Comparing is performed with the correlation method of measured signal and reference signal. The greatest advantage of digital level is the automation of measuring and recording process. Measuring process and recording the necessary data begin after aiming at a rod and triggering the reading. Digital levels are equipped with port for communication with external device such as PC or PDA. The connection can be done using a cable for data transmission (RS232 or USB) and also via Bluetooth using appropriate adapters. Two-way communication between devices enables external management of digital level working. However, it is necessary to programme individually the communication process between PC/PDA and digital level. The accuracy of single reading with the use of fiberglass rod equals about 0.1 mm (0.03 mm with the use of invar rod) for a distance of 25 m. For digital level the effectiveness of performing measurement depends on external condition stronger than for classical level. Atmospheric condition (wind, insolation, precipitation) should be considered when planning the measurement session. intervals, loadings of supports in 30-minutes intervals and asymmetrical loadings in 15-minutes intervals. Dynamic loadings were carried out by trucks which were passing at different speeds: 10, 30, 50, 70 km/h. Trucks were passing in two different directions in some cases single truck, in another two trucks simultaneously. In addition, at the speed of 50 km/h a stronger forcing of bridge reaction was carried out by crossing the threshold and emergency braking. Time series of vertical displacement of bridge girders were recorded in 1/2 and 3/4 length of span no. IV during trucks passing. Four induction displacement sensors fixed to girders were used in tests. Measuring techniques, described in section 3 of this paper, were applied during static load tests. The task of IBIS-S system was to measure displacement of span no. III in first three load schemes (Fig. 6). 4 FIELD TESTS 4.1 Load testing of the bridge Load tests on the bridge on the Skawa River included static and dynamic tests. Static loads were carried out according to seven schemes: 1 S1 loading the span no. I, 2 S2 loading the span no. II, 3 S3/P3 loading the span no. III and support no. 3, 4 S4 loading the span no. IV, 5 P4 loading the pylon no. 4, 6 S3A asymmetrical loading the span no. III, 7 S3B asymmetrical loading the span no. III. Four basic span loadings (S1, S2, S3/P3 and S4) were chosen based on the condition of maximum load of bridge girders in successive spans and cables of cable-stayed spans at the same time. Two loadings of supports (S3/P3 and P4) were chosen based on the condition of maximum reaction and simultaneous large response over the support no. 3 and pylon no. 4. Two additional, asymmetrical loadings of the third span (S3A and S3B) were chosen for the assessment of torsional stiffness of the span. Throughout the static test 10 four-axle trucks were used. The total mass of single truck was about 32 tons. Span loadings were carried out in 45-minutes Figure 6. Static load schemes 4.2 Systems applied In static loads IBIS-S measurements were carried out from the left riverbank. For the span no. III located directly over the river, deflection measurements by means of dial gauges were impossible. Precise levelling from the station above the support to the rod on a span was applied. For this purpose two digital levels Leica DNA03 were used. Measurements and recording were fully automated. Computer connected to the level through the serial port transmits the appropriate commands in the GSIOnLine language. Digital level, after performing the necessary action, sends the results back to the terminal. According to the conducted tests, measurement duration takes 3-4 seconds. In comparison to other methods it is not an advantageous value, but for static load measurements entirely sufficient. All the measurements taken can be performed in real time or stored in a database and post-processed later. Level measurements were implemented in S2 and 185

6 S3/P3 schemes. They were carried out to the fiberglass rods. In S3/P3 scheme an additional measurement was performed by the GNSS receiver located on a tripod and configured to work in RTK mode. GNSS corrections were received with the GSM module with reference to the ASG-EUPOS network the nationwide network of continuously operating reference stations. The receiver was determining its consecutive positions with a frequency of 10 Hz. During the dynamic tests only the IBIS-S system was used. Its task was to measure displacement of span no. III. The measurement was carried out from the right riverbank. Layout of devices in relation to the bridge is shown in Figure 7. Range of the bridge structure observed by the radar is shown in Figure 9. Position of range bins (Rbins) are marked. These points produce strong radar return signal and represent the span no. III. They were subjected to further analysis. The graph of displacements, measured in the first three load schemes by means of the IBIS system, GNSS receiver and digital levelling, is presented in Figure 10. Figure 9. Points observed by the radar Figure 7. Layout of devices in relation to the bridge 5 RESULTS 5.1 Static measurements The IBIS-S system was operating during static load tests in schemes S1, S2 and S3/P3. Graph of reflected signal intensity in the range of whole scene observed by the radar is presented in Figure 8. Figure 8. Range profile of the observed scene The first graph of Figure 10 shows the reaction of the span no. III while loading the span no. I. The greatest value of deflection (51 mm) is obtained by the point represented by Rbin64 (the second diaphragm). After unloading this point does not return to its original position, but maintains the deflection at the level of 4 to 5 mm. Points represented by other Rbins respond similarly, but appropriately weaker. The graph also shows information about the increase of measurement noise with the distance on each Rbin. The second graph of Figure 10 shows the reaction of the span no. III while loading the span no. II. The greatest value of deflection (24 mm, but upward) is obtained by the point represented by Rbin130 (the fourth diaphragm). During this load scheme measurements with digital level N2 and an optical level, set in the vicinity of point Rbin64, were performed. In this scheme the value of maximum upward deflection of point Rbin64 is up to 17 mm with an average of 16.2 mm. Digital level determines the deflection of 14.5 mm, while the deflection obtained by the optical level is 15.4 mm (shown in Fig. 10 by dashed line and described as LT). The third graph of Figure 10 shows the reaction of the span no. III on its load. The greatest value of deflection (120 mm), with 119 mm average, is obtained by the point represented by Rbin64 (the second diaphragm). During this load scheme measurements with digital and optical levels (N2 and LT in Fig. 10, respectively) were performed. The optical level was set in the vicinity of point Rbin64. Digital level determines the deflection of 107 mm, while the deflection obtained by the optical level is 110 mm. 186

7 dar return signal and represent the span no. III. They were subjected to further dynamic analysis. The bridge response was recorded for all truck forcing. Dynamic response of the span no. III during the truck passing at the speed of 50 km/h and jumping over the threshold in the middle of the span no. IV is shown in Figure 13. Figure 12. Points observed by the radar in the dynamic mode Figure 10. Graphs of bridge displacement obtained with different methods in three first load schemes 5.2 Dynamic measurements During dynamic measurements IBIS-S system was working with 50 Hz sampling frequency. The graph of intensity of radar signal, reflected from the whole scene, is presented in Figure 11. Figure 13. Dynamic response of the span no. III Figure 11. Range profile of observed scene Range of the bridge structure observed by the radar is shown in Figure 12. Position of range bins (Rbins) are marked. These points produce strong ra- Figure 14. Natural frequency spectrum recorded at the point Rbin22 187

8 For all observed points of the span no. III the natural frequency spectrum was determined in the IBIS Data Viewer software. An example of natural frequency spectrum recorded at the point Rbin22 is presented in Figure 14. Table 1 summarizes the natural frequencies calculated on the basis of the theoretical modal analysis (MA), three points observed by IBIS-S (Rbin22, Rbin53, Rbin118) and frequencies determined on the basis of observation by induction sensors (LT). Table 1. Comparison of natural frequencies. No. of vibration MA IBIS LT mode Hz Hz Hz CONCLUSIONS Performed experiments allow to present the following conclusions. 1 Radar interferometry technique gives good results in measuring deflection of the bridge span during static and dynamic load test of the bridge. 2 For bridge span located directly above the river, where it was impossible to use conventional measurement solutions, values obtained from digital and optical levelling were consistent with radar results. 3 It is required to explain result of bridge span deflection measured by the radar after unloading, when higher values of deflection were recorded than by means of conventional techniques. 4 During dynamic measurements of span above the river the value of the maximum deflection for the test runs was determined as well as the vales of natural frequencies, which are consistent with theoretical model values. REFERENCES Baraka, M.A. & El-Shazly, A.H., Monitoring Bridge Deformations during Static Loading Tests Using GPS. In Proceedings of FIG Working Week 2005, Cairo, April Gens, R. & van Genderen, J.L., SAR interferometry: issues, techniques, applications. International Journal of Remote Sensing 17(10): Gentile, C Application of Radar Technology to Deflection Measurement and Dynamic Testing of Bridges. In Guy Kouemou (ed.), Radar Technology: Vukovar: In- Tech. Gentile, C. & Bernardini, G Output-only modal identification of a reinforced concrete bridge from radar-based measurements. NDT & E International 41(7): Grejner-Brzezinska, A.D., Kashani, I., Wielgosz, P., Smith, D.A., Spencer, P.S.J., Robertson, D.S. & Mader, G.L., Efficiency and Reliability of Ambiguity Resolution in Network-Based Real-Time Kinematic GPS. Journal of Surveying Engineering 133(2): Kapović, Z., Novaković, G. & Paar, R., Deformation monitoring of the bridges by conventional and GPS methods. In Proceedings of 5th International Scientific Conference SGEM2005, Albena, June Ko, J.M. & Ni, Y.Q., Technology developments in structural health monitoring of large-scale bridges. Engineering Structures 27(12): Lazinski, P. & Salamak, M., Identification of computational models in load carrying structures of concrete bridges on the basis of making load tests. In Proceedings of 7th Central European Congress on Concrete Engineering CCC2011, Baltonfured, September Maia, N.M.M., Silva, J.M.M., Almas, E.A.M. & Sampaio, R.P.C., Damage Detection in Structures: From Mode Shape to Frequency Response Function Methods. Mechanical Systems and Signal Processing 17(3): Paultre, P., Proulx, J. & Talbot, M., Dynamic Testing Procedures for Highway Bridges Using Traffic Loads. Journal of Structural Engineering 121(2): Pieraccini, M., Fratini, M., Parrini, F., Macaluso, G. & Atzeni, C High-speed CW step-frequency coherent radar for dynamic monitoring of civil engineering structures. Electronics Letters 40(14): Pieraccini, M., Luzi, G., Mecatti, D., Noferini, L., Macaluso, G. & Atzeni, C., Ground-based radar interferometry for monitoring unstable slopes. In Proceedings of Joint 12th FIG International Symposium on Deformation Measurements and Analysis / 3rd IAG Symposium on Geodesy for Geotechnical and Structural Engineering, Baden, May Ryall, M.J., Bridge Management. Second Edition: Amsterdam: Elsevier. Watson, C., Watson, T. & Coleman, R., Structural Monitoring of Cable-Stayed Bridge: Analysis of GPS versus Modeled Deflections. Journal of Surveying Engineering 133(1): ACKNOWLEDGEMENTS The research was supported by the MNiSW (Polish Ministry of Science and Higher Education), under grant N N Part of research was made at the request of Mota-Engil Central Europe company. 188