9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

DYNAMIC TESTING OF CABLE STRUCTURES

E Caetano & A. Cunha Faculty of Engineering of the University of Porto, Portugal E-mail: ecaetano@fe.up.pt Website: www.fe.up.pt/vibest

Abstract The paper presents a brief state-of-art of experimental techniques available for identification of modal properties in Civil Engineering structures and describes the tools developed and applied at the Laboratory of Vibrations and Monitoring (VIBEST, www.fe.up.pt/vibest) of the Civil Engineering Department of the Faculty of Engineering of the University of Porto (FEUP), discussing the type and quality of relevant information that can be achieved. Practical applications to several cable structures are presented, illustrating particular problems of these structures: cablestructure interaction is discussed for the Jindo Bridge physical model; the identification for validation of numerical models is presented for Vasco da Gama Bridge and for the new Braga stadium suspended roof; the assessment of cable force for fatigue studies is referred for the Salgueiro Maia Bridge; and a monitoring program for assessment of cable vibrations is described for the International Guadiana Bridge. It is shown that dynamic testing of Civil Engineering structures has become an accurate, low cost and most valuable tool.
Keywords: vibrations; cable-stayed bridges; dynamic testing; monitoring 1. Introduction The basis for modern Experimental Stress Analysis was established in the 1940s with the introduction of strain gauges [1]. Experimental analysis on physical models constructed at reduced scales was then used to justify audacious prototypes whose structural behaviour was difficult to evaluate based on the existing analytical models, or which were important in terms of involved investment and risk of failure. Several laboratories in Europe, like the Portuguese National Laboratory of Civil Engineering, LNEC, developed expertise in model testing and instrumentation, constructing sensors and devices for specific applications. The advance in computer methods of analysis during the 1960s and 1970s conjugated with experimental observation on reduced models allowed a better insight into the behavior of structures. In particular, the construction of cable-stayed bridges, which established as a new category among the classical types in the 1960s, showed a very deep evolution and quickly spread all over the world. Due to the considerable time and cost associated with experimental observation of reduced scale structures by comparison with numerical analysis, tests on these models have been gradually abandoned, or at least employed in a narrower range of structures. As a counterpart, the development of instrumentation devices and acquisition systems allowed the increase in the instrumentation of prototypes, which provide more valuable information on the structural behavior in terms of material properties, boundary conditions, stress distribution, deformation, vibration and damping. Some examples of the former dynamic studies on prototypes can

9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

be found in [2-8]. These studies respect essentially the identification of fundamental frequencies and, eventually, of the configuration of the fundamental vibration mode of large structures, from forced, free or ambient vibrations, based on seismographs, mechanical accelerometers or strain gages. The former instrumentation systems were very limited in terms of precision and reliability, therefore they were used essentially as an accessory support to analytical and/or numerical analyses. It is however interesting to note that the basis for current dynamic testing was established in these tests. In effect, the resonance tests conducted at the Encino dam in 1961, based on mechanical exciters [4], have been conducted at the Tatara Bridge in 1999, using a more powerful hydraulic shaker, acquisition and measurement systems [9]. Also the vibration response of an arch bridge associated with the passage of loaded trucks over a wooden plank registered by a series of specially constructed strain gages during a dynamic test developed by the LNEC in 1954 [10] (Figure 1), has been employed during the Vasco da Gama tests in 1998, again with the purpose of increasing the level of response and characterising the dynamic amplification factors [11] (Figure 2). The free vibration test based on the impulsive application of a vertical force by breaking a calibrated bar that was developed in 1969 during the commissioning tests of the Tagus suspension bridge [6] were repeated, with some setup differences, at the Normandy Bridge, in 1995 [12], and later in 1998, at the Vasco da Gama Bridge [11], the purpose being an accurate identification of damping ratios. And finally, the technique adopted in the ambient vibration test of the Newport Bridge in 1971 [7], is almost the same that is currently used in the ambient vibration test of large structures, like the Vasco da Gama Bridge [11, 17].

Figure 1. Dynamic testing of the Bridge over river Sousa [10]: general view during construction; dynamic testing; strain gage; response to passage of loaded truck over wooden plank.

Figure 2. Dynamic testing of the Vasco da Gama Bridge: (a) passage of loaded truck over wooden plank; (b) view of section instrumentation. But even though the dynamic testing techniques have been established 40 years ago, the developments in computers, numerical methods of analysis, transducers and communication systems have completely modified the range of applications, the purpose of the tests, the amount of extracted data, the corresponding accuracy and the degree of intervention of the various entities involved in the design, construction and maintenance of structures.

9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

The present paper aims to systematize the current state-of-art in Dynamic Testing of Cable Structures, focusing in particular on possible purposes of testing, on the most efficient testing techniques, and on the current possibilities and limitations of experimental analysis. The topics covered are illustrated with practical applications that have been developed at the Laboratory of Vibrations and Monitoring of the Faculty of Engineering of the University of Porto.

2. Dynamic assessment of cable structures Cable structures are typically highly hyperstatic structures, whose equilibrium is ensured entirely or partially by groups of cables. The high flexibility and low damping of these structural elements are responsible for their proneness to vibrations. Many cable structures, like cable-stayed bridges, are formed by an association of two different substructures, here designated as primary and secondary sub-systems. The primary sub-system (the deck and towers, in the case of a cable-stayed bridge) is the part of the structure characterised by higher stiffness and massiveness, while the secondary system, represented by the group of cables, has a much smaller stiffness and mass. The two subsystems interact, presenting a complex dynamic behaviour, which is nevertheless dominated by the primary sub-systems characteristics. The possibility to numerically model more and more complex structures and the use of high strength steel have allowed an optimisation of the shape and a more efficient use of materials, the consequence being progressively lighter and larger structures. On another hand, economical and technical advantages and the aesthetical quality of such structures are responsible for a boom in their construction. This spread of construction all over the world has also been accompanied by a growing number of reported vibrations, which in most cases were not expected or are not completely understood at the current state-of-art. Vibration of cable structures can occur as a consequence of the most diverse actions, like wind, earthquakes, traffic and man induced loads. Wind actions in particular are responsible for different types of vibrations and instability phenomena, like direct effects (buffeting) and aeroelastic effects (lock-in, galloping, rain-wind or wake vibrations). Vibrations can still occur as the result of parametric resonance phenomena within the structure. The difficulty in characterising the actions associated with these vibrations, enhanced by the fact that various sources are in many cases simultaneously present, have motivated the development of many experimental studies in recent years, both at the research and application levels. Possible purposes of such studies are listed below and are illustrated in the following sections with applications that have been developed in the context of the authors activity at FEUP: - Calibration of numerical models: Identification of natural frequencies and modal shapes for validation of numerical models; - Damping assessment: Identification of damping ratios to use in numerical studies; - Assessment of vibrations: Identification and separation of possible phenomena of vibration; - Control of vibrations: Evaluation of the effectiveness of control measures, like passive dampers installed in cables; - Characterisation of input loads: Characterisation of traffic loads from measurements of dynamic amplification factors. Characterisation of wind loads from in-situ measurements of wind; - Fatigue assessment: Evaluation of residual life of cables submitted to frequent vibrations; - Health monitoring: Permanent characterisation of structure condition. Identification and localization of damage based on the time variation of modal properties.

3. Testing techniques 3.1 Input-Output techniques The former experiments of Civil Engineering structures were based on the conventional inputoutput techniques directly imported from Electrical and Mechanical Engineering. These comprehend in particular the application of a controlled force, the measurement of the response, and the application of System Identification algorithms to the set of frequency response functions, impulse responses, or time series. The so-called conventional Modal Testing technique is considered the most precise and reliable dynamic testing technique and can easily be applied to small structures, like physical models of cable structures. The transposition of this technique to large Civil Engineering structures meets however the difficulty associated with the required power of the excitation devices, in order to achieve a response higher than ambient response.

9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

Equipment and procedures The excitation of small and medium size Civil Engineering structures can be induced by a hammer (Fig. 3a), similar to those currently used in Mechanical Engineering. This device has the advantage of providing a wide-band input, able to stimulate different modes of vibration. The main drawbacks are the relatively low spectral estimates frequency resolution, which can preclude the accurate estimation of modal damping factors, and the lack of energy to excite some relevant modes of vibration. Due to this last factor, some laboratories have built special impulse devices more specifically designed to excite bridges (Fig. 3b). An alternative, also derived from Mechanical Engineering, is the use of large electrodynamic shakers (Fig. 3c), which can apply a large variety of input signals (random, multi-sine, etc.), when duly controlled both in frequency and amplitude using a signal generator and a power amplifier. The shakers have capacity to excite structures in a lower frequency range and higher frequency resolution can be attained. The possibility of application of sinusoidal forces allows for the excitation of the structure in resonance and, consequently, for a direct identification of the mode shape.

(a) (b) (c) (d) Figure 3. a) Impulse hammer; b) Impulse excitation device for bridges (K.U.Leuven); c)Electrodynamic shaker over three load cells; d) Eccentric mass vibrator. The controlled excitation of large Civil Engineering structures requires the use of heavy equipment. One option frequently used in the past in dynamic testing of dams was the eccentric mass vibrator (Fig. 3d), which enables the application of sinusoidal forces with variable frequency and amplitude. The main drawbacks of this technique are low force amplitude induced at low frequencies and some difficulty to measure the applied force and ensure no relative movement of the vibrator with regard to the structure. A better alternative to provide a wide-band excitation over the most interesting frequency range for large Civil Engineering structures is the servo-hydraulic shaker. Figure 4 shows, two shakers of this type built at EMPA (www.empa.ch) to excite bridges or dams, vertically and laterally. It shows also an image of the field vibration tests performed on the Tatara Bridge [13], where two exciters with a total weight of 370kN each were used, and a maximum vibration amplitude of 30.5cm was measured in correspondence with the first vertical vibration mode, at a natural frequency of 0.23Hz. The cost of such exciters and corresponding operation, the need of adequate transportation and of particular safety measures, make this test technique extremely expensive and difficult to apply. For this reason, and also because some alternative techniques can currently be applied with good results, only in very few structures have this type of tests been conducted. The dynamic response of the structure is usually measured with accelerometers (piezoelectric, piezoresistive, capacitive or force balance), due to their relatively low cost and high sensitivity. A particular characteristic of piezoelectric accelerometers is that they dont need specific power supply and operate well over a wide frequency range. However, most of them are not suited to low frequency applications. On the contrary, piezoresistive, capacitive and force balance accelerometers can provide DC or low frequency response capability. The electrical signals captured by these transducers are usually rather low and so must be amplified by conditioning units, that may also provide anti-aliasing low-pass filtering (allowing lower sampling rates) and analogue integration to velocities or displacements. The data acquisition and storage of measurement data involves the use of an analogue-to-digital (A/D) converter inserted in a digital computer. The digital raw data must be preliminary analysed and processed, considering operations of scale conversion, trend-removal and decimation. Afterwards, depending on the domain of the identification algorithm, different approaches can be followed. Considering the mostly applied frequency domain algorithms, the acceleration time series can be subdivided in different blocks, windowed, for reducing leakage effects, and then used in the evaluation

9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

of average auto and cross spectra estimates, based on the FFT algorithm. At last, estimates of FRFs can be obtained using estimators H1 or H2 [14]. The automatic evaluation of FRFs requires appropriate software for analysis and signal processing, which is already available in commercial Fourier analyzers. These analyzers are sometimes simply materialized through the insertion of a specific PCMCIA card into a laptop, allowing either the acquisition of data through input channels or the control of a shaker through an output channel.

(a) (b) (c) Figure 4. Servo-hydraulic shakers to excite (a) bridges (vertically), (b) dams (laterally) (EMPA) (c) Shakers on the Tatara Bridge (www.hsbp.go.jp/bridge/tata_inf.htm)

Input-Output Modal Identification methods There is presently a wide variety of input-output modal identification methods, whose application relies either on estimates of a set of FRFs or on the corresponding Impulse Response Functions (IRFs), which can be obtained through the inverse Fourier Transform. These methods try to perform some fitting between measured and theoretical functions and employ different optimization procedures and different levels of simplification. They are usually classified according to the following criteria: (i) Domain of application (Time or Frequency); (ii) Type of formulation (Indirect or Modal and Direct); (iii) Number of modes analysed (SDOF or MDOF); (iv) Number of inputs and type of estimates (SISO, SIMO, MIMO, MISO). The former methods of identification were developed in the frequency domain. In the simpler SDOF formulations (e.g. Peak Amplitude, Circle-Fit, Inverse methods), a fitting between a measured and a theoretical FRF of a SDOF system in the vicinity of each resonant frequency is developed, neglecting the contribute of resonant modes. In the more sophisticated MDOF methods (e.g. Rational Fraction Polynomial (RFP), Complex Exponential Frequency Domain (CEFD), Polyreference Frequency Domain (PRFD)), the fitting between measured and theoretical FRFs is made globally in a wide range of frequencies. Time domain methods, which tend to provide the best results when a large frequency range or a large number of modes exist in the data, began to be developed as consequence of some limitation in terms of spectral estimates frequency resolution, as well as leakage errors in the estimates. The most widely known methods are either Indirect (e,g. Complex Exponential (CE), Least-Squares Complex Exponential (LSCE), Polyreference Complex Exponential (PRCE), Ibrahim Time Domain (ITD), Eigensystem Realization Algorithm (ERA)) or Direct (e.g. Autoregressive Moving-Average (ARMA)). The gradual development of all these methods, which are extensively described by Maia et al. [14], shows a tendency to completely automated systems of acquisition, analysis, processing and identification, instead of some initial trend to the use of interactive programs. Beyond that, the bestperforming methods have been implemented in robust Modal Analysis software [15]. A special class of modal identification methods, called Tuned-Sinusoidal methods (e.g. Asher, Mau) corresponds to the particular type of tests that are based on the application of a sinusoidal excitation at each natural frequency, as can happen using eccentric mass vibrators. Practical applications The limitations above referred make applications of the forced vibration tests rare in large Civil Engineering structures. They are however of great interest in the testing of physical models. Figure 5 shows the physical model of Jindo Bridge (South Korea), which was extensively tested to analyze the importance of dynamic cable-structure interaction in terms of seismic response analysis [16]. The model, described in [11], is constructed at a scale of 1:150, has a total length of 3.23m, and is classified as distorted small-scale model with artificial mass simulation. Considering the purpose referred, several forced vibration tests were performed either using an electro-dynamic shaker or two different shaking tables (at Univ. Bristol and ISMES), and considering two alternative configurations for

9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

the model. First, additional masses were distributed along the cables, according to the similitude theory, in order to idealize the cables mass and consider the lateral cables vibration (Fig. 5b)). In a second phase, no distributed additional mass was introduced along the cables, but equivalent masses have been concentrated at their extremities (Fig. 5c)). Referring to the test technique, the electrodynamic shaker shown in Figure 6a) applied an excitation at successive locations along the bridge deck, the input force being measured through interposition of a force sensor between the shaker and the model, and the response being measured by a small piezoelectric accelerometer successively moved along the deck and towers (Fig. 6b). The fitting of collected FRFs based on the Rational Fraction Polynomial Method provided extremely accurate estimates of natural frequencies and modal configurations, as can be observed in Figure 7 in the fitting of an FRF and in the numerical /identified modal configurations of two closely spaced modes associated with different motions of cables.

(a) (b) (c) Figure 5. (a) Jindo cable-stayed bridge; (b) Physical model on shaking table (EERC, Univ. Bristol); (c) Physical model on shaking table (ISMES).

(a) (b) (c) Figure 6. (a) Force application with electro-dynamic shaker; (b) Response measurement with piezoelectric accelerometer; (c) Measurement of cables tensions.
Amplitude FRF ((m/s2)/N) 1.00E+01
Shaker

1.00E+00 1.00E-01 1.00E-02 8 9 10 Frequency (Hz) 11 12 Ajust. Exp.

(a) (b) Figure 7. (a) Amplitude of FRF relating vertical acceleration at 1/3 span with the vertical force applied at the opposite 1/3 span; (b) Identified pattern of a set of multiple modes. As for the shaking table tests, the motion of the shaking table according to a specified law allowed for the construction of frequency response functions relating the bridge acceleration at several locations with the measured base acceleration. The fitting of these functions based on the identification algorithm above referred provided higher quality estimates of modal properties, than the single input test. This is because the input energy is better distributed along the structure, therefore inducing higher contributions of mode shapes in the bridge response. This study permitted to identify the existence of different sets of multiple modes, some being pure cable modes, but others coupled modes. Each of these sets present a common shape for the deck

9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

and towers and different cables motion, the corresponding natural frequencies being very close, always in the vicinity of a global mode of the primary system (Figure 7). 3.2 Output-only techniques The impracticability of developing forced vibration tests on very large structures has led to the search of alternative testing techniques. Ambient vibration tests, in which the modal parameters are identified directly from the analysis of the structure response induced by ambient excitation (wind or traffic), have been for a long time applied in the test of flexible structures, like tall buildings [5, 8] and suspension bridges [3, 7]. However, the lack of sensitivity of transducers and the limitations in A/D converters have strongly compromised the type and quality of extracted data. This panorama changed in recent years, with the significant technological developments registered in the fields of transducers and A/D converters that provide extremely accurate measurements. Moreover, the progressive use of this type of tests has stimulated the development of more powerful identification methods. Equipment and procedures Modern force balance accelerometers (Fig. 2b), specially conceived for measurements in the range DC-200Hz and virtually insensitive to high frequency vibrations, have contributed very significantly to the success of ambient vibration tests. In such tests, structural ambient response is captured by one or more reference sensors, at fixed positions, together with a set of roving sensors, placed at different measurement points along the structure, in different setups. The number of points used is conditioned by the spatial resolution needed to characterize appropriately the shape of the most relevant modes of vibration (according to preliminary finite element modeling), while the reference points must be conveniently far from the corresponding nodal points. Force balance accelerometers require appropriate power supply, and their analogue signals (that may be locally amplified) are usually transmitted to a data acquisition system with an A/D conversion card, of at least 16 bit, through relatively long electrical cables. This system can be based on a normal PC, although some data acquisition and processing systems, specifically conceived for the performance of ambient vibration tests are already available (Fig. 8a), playing a role similar to the Fourier analyzers in the context of classical Experimental Modal Analysis. Although most of the output-only modal identification tests in large Civil Engineering structures have been based worldwide on the use of long electrical cables, the implementation of this type of solution is rather cumbersome and time consuming. Therefore, there is presently a natural tendency to develop wireless architectures or, at least reduce drastically the cables length, by introducing local digitization and single cable signal transmission. A very efficient and comfortable alternative has been intensively used at FEUP[17] and LNEC[18], based on tri-axial strong motion recorders duly synchronized through GPS antenna (Fig. 8b). With respect to cable response measurements, both accelerometers or load cells (Fig. 8c) can be employed.

(a) (b) (c) Figure 8. (a) Multi-channel data acquisition and processing system for ambient vibration tests; (b) Strong motion tri-axial seismograph with GPS antenna; (c) Load cell installed in strand. Output-Only Modal Identification methods The ambient excitation has commonly a multiple input nature and a wide-band frequency content, stimulating a significant number of modes of vibration. For simplicity, output-only modal identification methods assume the excitation input as a zero mean Gaussian white noise, meaning that the real excitation can be interpreted as the output of a suitable filter excited with that white noise input. There are two main groups of output-only modal identification methods: non-parametric methods, essentially developed in the frequency domain, and parametric methods, developed in the time domain.

9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

The basic frequency domain method (Peak-Picking, PP) that was applied in former tests [5, 7, 8] was only conveniently systematized by Felber [19] about twelve years ago. This approach, which leads in fact to estimates of operational mode shapes, is based on the construction of average normalized power spectral densities (ANPSDs) and ambient response transfer functions involving all the measurement points, and allowed the development of software for modal identification and visualization used at UBC and EMPA [19]. The frequency domain approach was subsequently improved [20, 21] by performing a single value decomposition of the matrix of response spectra, so as to obtain power spectral densities of a set of SDOF systems. This method (Frequency Domain Decomposition (FDD)) was better detailed and systematized by Brincker et al. [22], and subsequently enhanced [23] in order to extract modal damping factors estimates. In this last approach (EFDD) these estimates are obtained through inspection of the decay of auto-correlation functions, evaluated by performing the inverse Fourier transform of the SDOF systems power spectral densities. Time domain parametric methods involve the choice of an appropriate mathematical model to idealize the dynamic structural behaviour (usually time discrete state-space stochastic models, ARMAV or ARV models) and the identification of the values of the modal parameters so as that model fits as much as possible the experimental data, following some appropriate criterion. These methods can be directly applied to discrete response time series or, alternatively, to response correlation functions. The evaluation of these functions can be made based on their definition, using the FFT algorithm [24] or applying the Random Decrement method (RD) [25]. A peculiar aspect of output-only modal identification based on the fitting of response correlation functions is the possibility to use methods that stem from classical input-output identification methods, based on impulse response functions. Some of these methods are the Ibrahim Time Domain (ITD) [26], the Multiple Reference Ibrahim Time Domain (MRITD) [27], the Least-Squares Complex Exponential (LSCE) [28], the Polyreference Complex Exponential (PRCE) [29] or the Covariance-Driven Stochastic Subspace Identification (SSI-COV) [30]. An alternative method, that allows direct application to the response time series is the Data-Driven Stochastic Subspace Identification (SSI-DATA) [31]. Its still worth noting that the Random Decrement technique, usually associated to the application of time domain methods like Ibrahims, can be also the basis for the application of frequency domain methods, like PP, FDD or EFDD, as it leads to free vibration responses, from which power spectral densities can be evaluated using the FFT algorithm [32], reducing noise effect (methods RD-PP, RD-FDD and RD-EFDD). These methods have been recently systematized, applied and compared by Rodrigues [18]. Very recently, the new operational Polymax parameter estimation method was introduced by LMS (www.lms.be) [33]. This method operates on spectra or half spectra (i.e. the Fourier transforms of the positive time lags of the correlation functions) and its main advantage consists in yielding extremely clear stabilization diagrams, making an automation of the parameter identification process rather straightforward, which may enable the continuous monitoring of structural dynamic properties. Examples of ambient vibration tests Ambient vibration tests have been performed with great success in large buildings, bridges and special structures, allowing the creation of high quality experimental databases that have been used with different purposes. Three applications to structures constructed in Portugal are focused in this context, namely the Vasco da Gama Bridge, located in Lisbon, the International Guadiana Bridge, in the South, next to the border with Spain, and the New Braga stadium suspended roof, constructed recently in Braga, North of Portugal. The ambient vibration tests at the Vasco da Gama Bridge were developed by FEUP at the commissioning stage, following the request of the owner of the bridge to measure natural frequencies, vibration modes and damping ratios. The crossing of the river Tagus has a total extension of 12300m, including five viaducts and a cable-stayed bridge over the main navigational channel, with a center span of 420m and a total length of 829.2m (Figure 9). This cable-stayed bridge is of total suspension type, has a 31m wide deck formed by two longitudinal prestressed concrete girders connected by a cast in-situ slab 0.25m thick and by transversal I-girders every 4.42m. The two H-shaped towers are 147m above a massive zone used as protection against ship collision. The ambient vibration tests were conducted on 5 km of the viaducts and on the main bridge. Focusing on the cable-stayed bridge test results, the ambient structural response was measured during periods of 16 minutes at 58 points along the deck and towers (upstream and downstream) using a wireless system based on 6 tri-axial 16-bit seismographs duly synchronized by a laptop. The identification of a significant number of lateral, vertical and torsion modes in the relevant frequency range of 0-1Hz was performed, in a first instance [17], using the PP method. Subsequently, SSI and FDD methods were also applied [34,35] and compared using the software MACEC [30] and ARTEMIS [36], leading also to estimates of modal damping factors (although very accurate damping estimates

9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

require longer measurement periods). Figure 10 presents plots of some fundamental identified modes. A very good correlation between numerical and experimental results has been obtained. Also it has been concluded that the three identified methods, PP, FDD and SSI, provide very similar estimates. It is still worth mentioning that ambient vibration measurements of several cables response were developed on the bridge, based both on accelerometers and on a laser Doppler sensor that was simply laid on the floor under each stay cable (Fig. 11). These measurements were used to identify the installed force, based on the vibrating chord theory, and permitted also to confirm the efficiency and accuracy of the laser sensor for cable measurements.

Figure 9. View of Vasco da Gama Bridge.

Figure 10. Identified first (left) and second (right) bending (up) and torsion (down) modes.

1.00E+00 Acceleration PSD (m^2/s^4) 1.00E-02 1.00E-04 1.00E-06 1.00E-08 1.00E-10 1.00E-12 0 0.5 1 1.5 2 2.5 3 3.5 4 Frequency (Hz) Accelerometer Laser

(a) (b) Figure 11. (a) Measurement of vibrations in cable using an accelerometer or Laser head placed on the deck surface, hitting the stay-cable with a vertical laser beam (at night); (b) power spectral estimate based on accelerometer/ Laser. Similar to the Vasco da Gama Bridge, a complete ambient survey was recently developed on the International Guadiana Bridge [37]. This concrete cable-stayed bridge (Fig. 12a)), with a centre span

9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

of 324m, two lateral spans of 135m and two transition spans of 36m, was constructed in 1991. Despite the generally good performance under normal traffic and ambient conditions, the bridge exhibits significant and systematic cable vibrations, which were actually reported during construction and during the commissioning tests. These vibrations are in a great deal associated with the absence of the protection with a pipe of the individual parallel strands that form each stay cable, which are simply clamped at the thirds of length or at mid-length, depending on the size of the cable. This type of design is not in use now, but it was actually introduced in the late 1980s by Freyssinet, associated with the monostrand technology, aiming to improve the tensioning process and the protection of the strand against corrosion, and also to facilitate the replacement of strands. Many cable-stayed bridges constructed in the late 1980s and beginning of the 1990s employed this technology. But vortices associated with the action of the wind motivate oscillation of the individual strands inside the bundle, which start hitting against each other. This effect, eventually amplified by internal resonance effects, leads to a vibration of the bundle of cables that attains significant amplitudes, of the order of 1m peakto-peak, and results in damages of the clamps, and of the deviator guides and neoprene rings close to the anchorages, which eventually end up affecting the strand protection (Fig. 12b)).

(a)
1E-02
Amplitude 1E-02 1E-03

(b)
PSD - half-sum of vertical aceleration S1 - v=2m/s 1E-03 S2 - v=2m/s S3 - v=2m/s 1E-04 S4 - v=9m/s S5 - v=9m/s S6 - v=9m/s 1E-05 S7 - v=14m/s

Power Spectral Density functions at reference section (half-sum of vertical acceleration)

1E-04 1E-05 1E-06 0 0.2 0.4 0.6 0.8

wind - 2 m/s
1 1.2

wind - 10 m/s
2 2.2 2.4

wind - 14 m/s
2.6 2.8 3

1.4 1.6 1.8 Frequency (Hz)

1E-06 0.25

0.35

0.45 Frequency (Hz)

0.55

0.65

(d) I Figure 12. (a) View of Guadiana Bridge; (b) Damage in strands and deviator guide; (c) PSD functions of half-sum vertical; Zoom of PSD functions at reference section, as function of mean wind speed. An ongoing research project, funded by the National Science Foundation FCT, aims to install a monitoring system on the bridge in order to investigate the patterns of cable vibration, to evaluate the risk of premature failure of strands by fatigue, as a consequence of repeated vibration occurrences, and identify the consequent damage. The so-called zero-state of the bridge has therefore been defined from the results of a complete ambient survey that was developed based on 4 triaxial 18-bit strong motion recorders duly programmed by a PC and synchronized by GPS antennas. Using the technique above referred for the Vasco da Gama Bridge, a total of 27 pairs of points along the deck and 4 additional sections along the towers were instrumented. Time series of 21 minutes were collected at each measurement section, corresponding approximately to 500 cycles of the longest vibration period of the structure. In parallel with the measurement of accelerations, the wind speed was measured at deck level using a propeller anemometer. The identification of modal parameters was then performed, based on the three identification techniques, PP, FDD and SSI. Although generally very good results were attained, and a good correlation was found with numerically obtained modal parameters [38], some difficulties were felt in the identification, which are consequence of the significant oscillation of many cables during several setups. Figures 12 c) and d) show estimates of the average power spectral densities at the reference sections during setups associated with three different levels of average wind velocity. Higher levels of wind velocity clearly stimulate the vibration of many stay cables, fact that is evident by the appearance of spectral contributes from their fundamental modes (in the range 0.6-0.9Hz) or second harmonics, that can not be interpreted as global natural frequencies of the bridge. Inspection of the spectral peaks (Fig.12 (d)) shows also the increase of modal damping with wind speed.

9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

The same measurement technique, sensors and identification algorithms have been used in other applications with great success, as was the case of the Millau Bridge tests coordinated by Grillaud and Flamand [39], and the suspended roof of the new Braga sports stadium (Fig. 13) [40]. Although being completely different in geometry and behaviour, these two structures have natural frequencies within the same range and display similar amplitudes of vibration under ambient loads, lower than 10mg.

Figure 13. (a) General view of Braga Municipal Sports stadium; (b) Free vibration response at three different points of the cable-roof after interruption of resonant excitation. Free vibration tests The accurate identification of modal damping factors is a point of major difficulty in the identification process, not only due to the considerably larger scatter associated to the estimates provided by the several methods used with regard to natural frequencies and mode shapes, but also because the viscous damping assumption does not correspond exactly to the real damping characteristics, the values of modal damping ratios increasing gradually with the levels of oscillation. However, in several circumstances the accurate identification of modal damping factors is required, for the assessment of aeroelastic instability or other vibration problems. In those cases, the performance of a free vibration test (FVT) can be of interest. Such a test consists in the application of an impulsive action into the structure, for example by sudden release of a suspended weight, or prestressed cable. This impulse originates a response that is much higher than the ambient response, and the application of identification algorithms to the free decay response provides the most accurate estimates of damping. Tests of this nature were conducted at the Normandy, Vasco da Gama and Millau Bridges. On the rd Vasco da Gama Bridge, a barge weighting 60 ton was suspended eccentrically from about the 1/3 of the span in the low tide and the cut of the suspension rod originated a maximum dynamic displacement of 25mm, and a free decay response within a period of about 8 minutes. On the Braga Municipal Sports stadium, a mechanical vibrator was used to induce a resonant response at several natural frequencies. A sudden interruption of the excitation produced a free vibration response as shown in Fig.13 (b), whose maximum amplitude was about 29mm for the second vibration mode at a natural frequency of 0.29Hz. The identified damping ratios showed in both structures extremely low values, in the range 0.2%-0.5% for the generality of the identified vibration modes. The first lateral mode of Vasco da Gama Bridge displayed however a damping ratio of 1.2%. This value higher almost an order of magnitude than the damping ratios of other modes is found in many cable-stayed bridges and is possibly associated with the dissipating characteristics of supports along the lateral direction. Assessment of traffic induced vibrations The assessment of traffic induced vibrations is another possible purpose of output-only dynamic tests. In this case, the structural response under controlled traffic is measured, in order to evaluate dynamic amplification factors. An example of application concerns the dynamic tests performed in Salgueiro Maia cable-stayed bridge (Figure 14(a)). This single-plane cable-stayed bridge, located near Santarm, has a central span of 246m and lateral spans of 78m and 42m. Both the deck and towers are completely isolated, due to the high seismic risk of the area, by interposition of appropriate isolation devices between that main part of the bridge structure and the supporting piers. The experiments, extensively described in reference [42], comprehended a preliminary ambient vibration test to support the updating and experimental validation of a finite element model for numerical simulation of the bridge under road traffic load, and involved the installation of a new measurement system including embedded strain gages in several points of the concrete deck near the mid-span, and load cells in strands of 13 stay-cables of the bridge. This measurement system was first tested during the static tests performed by LNEC at the commissioning stage, and was subsequently used for the evaluation of dynamic strains (Fig. 14(b)) and stresses, and dynamic amplification factors (DAFs) associated to the passage of heavy trucks passing along several lanes at different speeds. Maximum DAF values of about 1.2 were obtained for some cable tensions and bending effects at midspan at velocities of 90km/h. The passage over a wooden plank led to higher values of about 1.4 for

9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

the lower velocities of 30km/h. Another purpose of the instrumentation on the bridge is the monitoring of force in strands, in order to conduct fatigue studies of stay cables.
16.0 14.0 12.0 10.0 ( m/m) 8.0 6.0 4.0 2.0 0.0 -2.0 -4.0 10.0 v = 90km/h

sta dyn

12.0

14.0

16.0

18.0

20.0 time (s)

22.0

24.0

26.0

28.0

30.0

(a) (b) Figure 14 (a) Lateral view of Salgueiro Maia Bridge; (b) Strain measured by embedded strain gage during dynamic tests Assessment of cable vibrations The assessment of cable vibrations is another purpose for output-only dynamic testing, as exemplified above with the case of the international Guadiana Bridge. The necessity to identify a pattern for the observed vibrations requires the performance of continuous measurements. Quantities like deck accelerations, cable accelerations or forces, temperatures, and wind velocity and direction, need to be monitored for long periods, so that correlations of vibration events can be established. For this reason, a monitoring program has been under development [37], which will comprehend the network of about 40 sensors schematically represented in Figure 15 and which will incorporate several innovative developments in order to maximise efficiency and cost. These are namely the performance of the digitisation close by the sensors, and the use of one single cable connecting all sensors and the central acquisition system, for control of acquisition, powering and data transmission. With such a system, not only a significant reduction of electric cable is achieved, but also the signal-to-noise ratio is increased and the data transmission is accelerated. A key aspect of the permanent instrumentation of a bridge located remotely, as is the case, is the possibility to remotely control the acquisition system, in order to restart the system in case of failure, change measurement parameters and access data within a limited time after measurement. The recent development in communication systems has allowed fulfilling such requirement and even providing additional information, like sending SMS or email alerts, upon the occurrence of significant level events. That possibility means the necessity to chose a series of sensors that permanently enquire the bridge and automatically store a specified amount of data at all channels upon activation of a pre-defined trigger. Other important aspects of the system are synchronisation and compression of data to save. It is of no interest to save full-day sampled time series. Therefore, an important requirement of an efficient monitoring system is the ability to transform the huge amount of sampled time histories into statistical parameters and function estimates that resume the bridge condition adequately. Such information is of utmost importance in the future identification and localisation of damage and in the fatigue assessment of cables.

SPAIN
S3

3 S2

4 S1

S4

16

17

32

PORTUGAL

3 i Uniaxial accelerometer, I channels Sonic/propeller anemometer Magneto-elastic sensor Bi-axial accelerometer Video cameras Temperature sensors Strain gages Data acquisition system

Figure 15. Monitoring plan for the International Guadiana Bridge.

9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

4. Conclusions The paper focuses on the dynamic testing of cable structures. High flexibility, low damping, a significant proneness to experience vibrations, and a complex dynamic behaviour, characterised by the presence of multiple modes closely spaced in frequency and with different natures (local or global), are characteristics of these structures that pose experimentalists particular demands: the necessity to employ very sensitive sensors in a low frequency range (DC-10Hz), combined with high resolution A/D converters (minimum 16 bit), and the requirement to collect very long time series (more than 15 minutes per section, generally), in order to obtain high resolution spectral estimates and therefore separate closely spaced modes. The various applications above described have shown that currently available sensors and measurement systems can provide extremely accurate measurements, with the most diverse applications. Moreover, considering the combination of high quality ambient vibration response signals with powerful system identification algorithms, it can be concluded that ambient vibration identification methods are established tools for the structural engineer. A new field that is now under strong evolution concerns the automation of ambient vibration tests. The necessity to monitor the dynamic behaviour of cable structures in order to assess particular aspects of their dynamic behaviour poses an additional requirement of permanence of measurements. The concept of Smart Structure, as a structure that continuously monitors its condition through the permanent and synchronous enquire of various of its components (sensors), that automatically communicates with a central operator and stores data in case of particular events, or that continuously processes the sampled data and calculates functions whose variations provide indicators of failure, can be considered the current trend in dynamic testing. And despite the initial demand for this concept has come from structural engineers, the growing requirement of society to assess security and optimise maintenance procedures have generalised its interest. References [1] Sabnis, G., Harris, H., White, R. & Mirza, M. 1983. Structural Modeling and Experimental Techniques. Prentice-Hall, Inc.. ISBN 013853960X. [2] Carder, D. S. 1936. Observed vibrations of buildings, Bulletin of the Seismological Society of America, Vol. 26, no. 4, pp 245-277; [3] Vincent, G.S. 1958. Golden Gate Bridge Vibration Study, ASCE Journal of the Structural Division, Vol, 4, ST6 . [4] Keightley, W., Housner, G. & Hudson, D. 1961. Vibration Tests of the Encino Dam InTake Tower. Report. California Institute of Technology. [5] Crawford R. & Ward, H. S. 1964. Determination of the natural pedriod of buildings, Bulletin of the Seismological Society of America, Vol. 54, no. 6, pp. 1743-1756. [6] Marecos, J., Castanheta, M. & Teixeira Trigo, J. 1969. Field Observation of Tagus River Suspension Bridge. Journal of the Structural Division, Vol. 95, No. ST4: 555-583. ASCE. [7] McLamore, V. R., Hart, G. C., Stubbs, I. R. 1971. Ambient vibration of two suspension bridges, ASCE Journal of the Structural Division, Vol. 97, No. ST10, pp. 2567-2582 [8] Trifunac, M. D. 1972. Comparisons between ambient and forced vibration experiments, Earthquake Engineering and Structural Dynamics, Vol. 1, pp. 133-150. [9] Yamaguchi, K., Manabe, Y., Sasaki, N. & Morishita, K. 1999. Field Observation and Vibration Test of the Tatara Bridge. Cable-Stayed Bridges- Past, Present and Future. IABSE Conference, 2-4 June, Malm, Sweden. [10] Marecos, J. 1954. Observation and test of the bridge over the river Sousa, LNEC Report, Proc. 145-II, Structural Division (in Portuguese). [11] Caetano, E. 2000. Dynamics of Cable-Stayed Bridges: Experimental Assessment of CableStructure Interaction, PhD Thesis, University of Porto, Portugal. [12] Bietry, J. & Jan, P. 1995. Essais Dynamiques du Pont de Normandie. Report of Mission du Pont de Normandie. EN-D 95.5 C. [13] Yamaguchi, K., Manabe, Y., Sasaki, N. & Morishita, K. 1999. Field Observation and Vibration Test of the Tatara Bridge. Cable-Stayed Bridges- Past, Present and Future. IABSE Conference, 2-4 June, Malm, Sweden. [14] Maia, N. et al. 1997. Theoretical and Experimental Modal Analysis, Research Studies Press, UK. [15] Han, M-C & Wicks, A.L. 1989. On the application of Forsythe orthogonal polynomials for global th modal parameter estimation, Proc. 7 Int. Modal Analysis Conference. [16] Caetano, E., Cunha, A. & Taylor, C. 2000. Investigation of Dynamic Cable-deck Interaction in a Physical Model of a Cable-stayed Bridge. Part I: Modal Analysis, Int. Journal Earthquake

9th International Conference on Inspection, Appraisal, Repairs and Maintenance of Structures, Fuzhou, China, October 2005

Engineering and Structural Dynamics, Vol.29, No.4, p.481-498. [17] Cunha, A, Caetano, E. & Delgado, R. 2001. Dynamic Tests on a Large Cable-Stayed Bridge. An Efficient Approach, Journal Bridge Engineering, ASCE, Vol.6, No.1, p.54-62. [18] Rodrigues, J. 2004. Stochastic Modal Identification. Methods and Applications in Civil Engineering Structures, Ph.D. Thesis (in Portug.), Univ. of Porto (FEUP/LNEC). [19] Felber, A. 1993. Development of a Hybrid Bridge Evaluation System, Ph.D. Thesis, University of British Columbia (UBC), Vancouver, Canada. [20] Prevosto, M.. 1982. Algorithmes dIdentification des Caractristiques Vibratoires de Structures Mcaniques Complexes, Ph.D. Thesis, Univ. De Rennes I, France. [21] Corra, M.R. & Campos Costa, 1992. Ensaios Dinmicos da Ponte sobre o Rio Arade, in Pontes Atirantadas do Guadiana e do Arade (in Portuguese), ed. By LNEC. [22] Brincker, R., Zhang, L. & Andersen, P. 2001. Modal Identification from Ambient Responses th using Frequency Domain Decomposition, Proc. 18 Int. Modal Analysis Conference, Kissimmee, USA. [23] Brincker, R., Ventura, C. & Andersen, P. 2000. Damping Estimation by Frequency Domain th Decomposition, Proc. 19 Int. Modal Analysis Conference, San Antonio, USA. [24] Brincker, R., Krenk, S., Kirkegaard, P.H. & Rytter, A. 1992. Identification of the Dynamical Properties from Correlation Function Estimates, Bygningsstatiske Meddelelser, Danish Society for Structural Science and Engineering, Vol.63, N.1, p.1-38. [25] Asmussen, J.C. 1997. Modal Analysis based on the Random Decrement Technique. Application to Civil Engineering Structures, Ph.D. Thesis, Univ. Aalborg. [26] Ibrahim, S.R. & Mikulcik, E.C. 1977. A Method for the Direct Identification of Vibration Parameters from the Free Response, The Shock and Vibration Bulletin, Vol.47, N.4, p.183-198. [27] Fukuzono, K. 1986. Investigation of Multiple-Reference Ibrahim Time Domain Modal Parameter Estimation Technique, M.Sc. Thesis, Univ. Cincinnati, USA. [28] Brown, D.L., Allemang, R.J. , Zimmerman, R. & Mergeay, M. 1979. Parameter Estimation Techniques for Modal Analysis, SAE Technical Paper Series, N.790221. [29] Vold, H., Kundrat, J., Rocklin, G.T. & Russel, R. 1982. A Multi-Input Modal Estimation Algorithm for Mini-Computers, SAE Technical Paper Series, N.820194. [30] Peeters, B. 2000. System Identification and Damage Detection in Civil Engineering, Ph.D. Thesis, K.U.Leuven, Belgium. [31] Van Overschee, P. & De Moor, B. 1996. Subspace Identification for Linear Systems: Theory, Implementation, Applications, Kluwer Academic Publishers, The Netherlands. [32] Rodrigues, J., Brincker, R. & Andersen, P. 2004. Improvement of Frequency Domain Outputrd Only Modal Identification from the Application of the Random Decrement Technique, Proc. 23 Int. Modal Analysis Conference, Deaborn, USA. [33] Peeters, B., Vanhollebeke, F. & Van der Auweraer, H. 2005. Operational PolyMAX for rd Estimating the Dynamic Properties of a Stadium Structure during a Football Game, Proc. 23 Int. Modal Analysis Conference, Orlando, USA. [34] Peeters, B., De Roeck, G., Caetano, E. & Cunha, 2002. Dynamic study of the Vasco da Gama Bridge, Proc. Of the International Conference on Noise and Vibration Engineering, ISMA, Leuven, Belgium. [35] Cunha, A. Caetano, E., Brincker, R. & Andersen, P. 2004. Identification from the Natural nd Response of Vasco da Gama Bridge, Proc. 22 Int. Modal Analysis Conference, Deaborn, USA. [36] ARTeMIS Extractor Pro, Structural Vibration Solutions, Aalborg, Denmark. [37] Caetano, E., Cunha, A. & Magalhes, F. 2005.Vibration assessment of the International Guadiana Bridge. Proc. Eurodyn, Paris, Sept. (accepted for publication). [38] Magalhes, F., Caetano, E. & Cunha, A. 2005. Assessment of dynamic properties of Guadiana cable-stayed bridge based on different output-only identification techniques, Proc. EVACES, Bordeaux, Oct. (submitted). [39] Flamand, O. & Grillaud, G. 2005. Identification Modale du Viaduc de Millau, Technical Report EN-CAPE 05.007 C-V0, CSTB, France. [40] Caetano, E., Cunha, A., Magalhes, F. & Furtado, R. 2005. Numerical and experimental studies of Braga Sports Stadium suspended roof, Proc. EVACES, Bordeaux, Oct. (submitted). [41] Magalhes, F., Caetano, E. & Cunha, A. 2004. Experimental Identification of Modal Damping Coefficients of the New Braga Stadium Cable-Roof, Technical Report (in Portuguese), Univ. Porto, FEUP/VIBEST. [42] Calada, R., Cunha, A. and Delgado, R. 2005. Analysis of traffic induced vibrations in a cablestayed bridge. Part I: Experimental assessment, Journal of Bridge Engineering. ASCE, Vol. 10, No. 4, pp. 370-385.