OMA of GFRP Retrofitted Model Concrete Structure Using Ambient Vibration

Nowadays, there are a great number of various structures that have been retrofitted by using different FRP Composites (AFRP, BFRP, CFRP, GFRP, etc.). Due to this, more researches need to be conducted to know more the dynamic characteristics of these structures, not only that but also a comparison among them before and after the retrofitting is needed. In this study, a model concrete structure is tested using ambient vibration data, in order to get the dynamic behaviors. Slab of the model concrete structure are then retrofitted by using GFRP composite, and then tested by using the ambient vibration data. At this stage, it is needed to evaluate the dynamic behaviors (frequencies, mode shapes and damping ratios) of the retrofitted model concrete structure. Various types of methods of OMA, such as FDD, EFDD, SSI, etc. are used to take action in the ambient responses. Having a purpose to learn more about the effects of FRP composite, experimental model analysis of both types (retrofitted and no-retrofitted models) is conducted to evaluate their dynamic behaviors. There is a provision of ambient excitation to the structure by using ambient vibration data on ground level. Furthermore, the Enhanced Frequency Domain Decomposition is used through output-only modal identification. At the end of this study, moderate correlation is obtained between mode shapes, periods and damping ratios. The aim of this research is to show and determine the effects of GFRP Composite implementation on structural responses of the model concrete structure, in terms of changing its dynamical behaviors. The frequencies for model concrete structure and the retrofitted model concrete structure are shown to be 11.41% in average difference. Finally, it is shown that, in order to evaluate the period and rigidity of retrofitted structures, OMA might be used.


I.
INTRODUCTION: Most of structures located in regions prone to earthquake hazards suffer from various types of destruction caused by seismic loads. Under such earthquake occurring, the parts (especially the columns) of building structures suffer damage. Looking on the other side, especially considering the performance of such buildings under seismic occurrence, there is a great need to strengthen the columns without changing their building masses; this clearly shows that there is a need to investigate the connection between technical repairing or strengthening procedures and the column capacity. In this understanding, more researches are being conducted to get required performance of structures under seismic loading, by means of looking at different point of view and directions.
Recently, application of fiber reinforced plastic composite system by gluing them to external part of the reinforced concrete structures is gradually becoming popular for the aim of repairing and strengthening (Yang et al. 2017), Keykha (2017), (Smyrou et al. 2015), Elwan and Omar (2014). Fibers to be used, as they have required characteristics include: glass, aramid, basalt, carbon, etc. The production of these fibers is done in two ways: either as plates (covered by thin fibers) or as tissues (knitted in one and two directions). The behavior of the system that is covered with external FRP composite is related to the type of the element covered. Generally FRPs have been separated into three categories: bending strengthening, shear strengthening and envelope scripts.
In order to strengthen reinforced concrete structures, the prevention of severe bending and shearing is realized by covering beams by FRP composite. Increasing the resistance and ductility of the system under lateral seismic loads the main goal of this covering. Li and Sung in (2003) they had presented lot of analytical and experimental tests on benchmark and on reinforced concrete damaged circular bridge column. In the benchmark column is a 40% scale reinforced concrete circular bridge column damaged because of shear failure during a cyclic-loading test. Then the column repaired by epoxy and nonshrinkage mortar and rehabilitated by (CFRP) carbon fiber reinforced plastic after the cyclic-loading test. Experimental result could be predicted accurately by the analytical lateral force-displacement relationship of the bridge columns, especially in the nonlinear regions. In their study, for circular reinforced concrete bridge column, the result has been reached so that for a true repair; a change of the shear-failure mode of bridge column to the bending-failure refraction occurs, in other words this increases the seismic performance the analytical and experimental by (Montoya et al. 2004) are fitted with the numerical results of nonlinear finite element evaluation for the behavior of steel and FRP contained concrete columns which formulated and implemented. The performance of reinforced concrete column which was covered with carbon FRP was determined under uniaxial compression load Cole (2001). When Strengthened with CF-130 carbon fiber laminates, the experimental result for five circular columns and three rectangular columns were tested in pure compression shows that ±45 degrees CFRP laminate can effectively be used to provide columns ductility performance. When the main goal is to boost the load capacity, a unidirectional FRP laminate might be more effective according to Paretti and Nanni (2002). According to Parvin and Wang study (2002), they talk over the effect of strain gradient and FRP thickness on square concrete columns reinforced with FRP wraps. The results for nine square concrete columns were tested under eccentric load and two different levels of eccentricity. it was shown that the chosen eccentricity values were small enough to produce any longitudinal tension in the wrap. The aim of this study is to evaluate the performance of reinforced concrete column, which has rectangular cross-section, under axial static compression load by using analytical, numerical and experimental evaluations and also to increase the source of statistics with a comparison target on this field. It has been shown that beams of existing structures suffer too much during seismic loading. Reinforced concrete rectangular cross-section column was used to evaluate their performance under axial static compression load by using analytical, numerical and experimental evaluations and also to increase the source of statistics with a comparison target on this field. It has been shown that beams of existing structures suffer too much during seismic loading. analytical and experimental results by testing "T" cross section reinforced concrete beam, the beams strengthen with carbon fiber reinforced plastic composite (CFRP), the results show that tension increased at the negative moment region approximately 40% according to (Namboorimadathil et al. 2002) study. The distance from support to CFRP origin and effect of cross-section beam and its behavior have been studied in (Ahmed et al. 2001) study, when it was strengthened with CFRP composite at the tensile region of reinforced concrete beam. Computation formula has been composed related to experimental results, to guess the design load that is equal to the limit position of beam. In this examination original shear stress and slight effect have been taken into consideration. The performance of partial bridge strengthened by CFRP composite has been tested in (Ramos et al. 2004) study. On partial scaled and full-scaled specimen, partial beams experiments were conducted. Bond scaled experiment has been shown as alternative for characterizing repair and strengthening the partial structures with CFRP composite. For pre-stressed three reinforced concrete girder bridge that suffered damage which repairstrengthening with CFRP composite. Experimental results before and after repairing was presented by (Klaiber et al. 2003) study, the results shown that using of CFRP is productive. The girder bending displacements have been decreased more than 20% when CFRP was used. When Strengthened with CF-130 carbon fiber laminates, fifteen rectangular beams were tested in pure compression. The experimental result shows that CFRP laminate can effectively be used to provide beams ductility performance. The effect of FRP wrapping number to the maximum axial capacity has been evaluated Kasimzade and Tuhta (2012).
As already known, forced (shaker, impact, pull back or quick release tests) and ambient vibration techniques are available for vibration testing of large structures. Force vibration methods are tougher and generally not cheap compared to ambient vibration tests. So the later (ambient vibration testing, also These structure response characteristics give a general idea of the preferred quantity to be measured. A few studies on the analysis of ambient vibration measurements of buildings from 1982 until 1996 were discussed in Ventura and Schuster (1996). Last ten years Output-Only Model Identification studies of buildings are given in appropriate references as structural vibration solutions. For the modal updating of the structure it is necessary to estimate sensitivity of reaction  Sestieri and Ibrahim (1994), ). The solution for algebraic Riccati equation matrix and orthogonality projection, which is more intensively and inevitably used in system identification, was deeply investigated in works of Aliev (1998). In engineering structures there are three types of identification that are used: modal parameter identification; structural-modal parameter identification and control-model identification methods. In the frequency domain the identification is based on the singular value decomposition of the spectral density matrix and it is denoted as Frequency Domain Decomposition (FDD) and its further development Enhanced Frequency Domain Decomposition (EFDD). In the time domain there are three different implementations of the Stochastic Subspace Identification (SSI) technique: Unweighted Principal Component (UPC); Principal component (PC); Canonical Variety Analysis (CVA) which were used for the modal updating of the structure Friswell and Mottershead (1995), Marwala (2010). It is required to estimate the sensitivity of reaction of examined system to change of random or fuzzy parameters of a structure. Investigated measurement noise perturbation influences to the identified system modal and physical parameters, estimated measurement noise border, for which identified system parameters are acceptable for validation of finite element model of examined System identification is realized by observer Kalman filter (Juang et al. 1993) and Subspace Overschee with De Moor (1996) algorithms. For some specialties, the observer gain coincides with the Kalman gain. Stochastic state-space model of the structure is simulated by Monte-Carlo method. The dangers and difficulties of shaking the structures with forced vibrations, the large size of the vibrators, the possibility of destruction of the structure, the disruption of daily life, the cost of the experiment is taken into consideration, the advantages of operational modal analysis are emphasized according to the experimental modal analysis and therefore the operational method is more widely used. Operational modal analysis technique it has important advantages such as the full scale of the work being done, no damage, no external load impact on the structure. For this purpose, experimental modal analysis of a model concrete structure for dynamic characteristics was evaluated. Then, retrofitted model concrete structure for dynamic characteristics was also evaluated. Ambient excitation was provided ambient vibration data on ground level. The Enhanced Frequency Domain Decomposition is used for the output-only modal identification.

II. MODAL PARAMETER EXTRACTIONS
Where n is the number of modes λ k is the pole and, R k is the residue. Then Eq. (1) becomes as: Where the singular values, superscript is denotes complex conjugate and transpose. Multiplying the two partial fraction factors and making use of the Heaviside partial fraction theorem, after some mathematical manipulations, the output PSD can be reduced to a pole/residue form as fallows; Where A k is the th residue matrix of the output PSD. In the EFDD identification, the first step is to estimate the PSD matrix. The estimation of the output PSD known at discrete frequencies is then decomposed by taking the SVD (singular value decomposition) of the matrix; Where the matrix U i =[u i1 ,u i2 , … ,u im ] is a unitary matrix holding the singular vectors u ij and s ij G is a diagonal matrix holding the scalar singular values. The first singular vector u ij is an estimation of the mode shape. PSD function is identified around the peak by comparing the mode shape estimation u ij with the singular vectors for the frequency lines around the peak. From the piece of the SDOF density function obtained around the peak of the PSD, the natural frequency and the damping can then be obtained.

III. DESCRIPTION OF MODEL CONCRETE STRUCTURE
Model concrete structure is 55 cm height. Thickness of elements (shear walls and slab) is 5.5 cm. C20 concrete is used in construction. The structure dimensions are shown in Fig. 2.

IV. EXPERIMENTAL MODAL ANALYSIS OF MODEL CONCRETE STRUCTURE
The ambient excitation is provided by using recorded micro tremor data on ground level. Two accelerometers (triaxial) are used to measure ambient vibrations, one of them is allocated as reference sensor, which is always located in the slab (shown by the red arrows in Fig. 3a, b). Two accelerometers are used as roving sensors (shown by the black arrows in Fig. 3a, b). The response was measured in two data sets (Fig. 3 a, b). For two data sets, 2 and 4 degree of freedom records are used respectively (Fig. 3 a, b). Every data set (Fig. 3 a, b) is measured within 100 minutes. The selected measurement points and directions are shown in Fig. 3 a, b. The ambient excitation is provided by using recorded micro tremor data on ground level (Fig. 4 a, b). a) First setup b) Second setup Fig. 3 Accelerometers location of experimental model in the 3D view The data acquisition computer provides the ambient vibration records. During measurements, the data files from the previous setup are transferred to the computer for data analysis by using a software package. However, in case there is a display of unexpected signal drifts or unwanted noise or corrupted for some unknown reasons, the data set must be discarded and measurements be repeated.

Fig. 4 Ambient vibrations recorded by the seismometer
Before measurements the cable used to connect the sensors to the data acquisition equipment must be laid out. Following each measurement, the roving sensors are systematically located from floor to floor until the test is completed (Fig. 3 a, b). The equipment used for the measurement includes three sensebox accelerometers (with both x and y directional measurements) and güralp systems seismometer and matlab data acquisition toolbox (wincon).  The Eigen frequencies are found as the peaks of nonparametric spectrum estimates when the simple peakpicking method (PPM) is used. This frequency selection procedure becomes a subjective task in case of noisy test data, weakly excited modes and relatively close Eigen frequencies. Also for damping ratio estimation, the related half-power bandwidth method is not favorable. This why the most popular and useful algorithm to use is Frequency domain one, because of its convenience and operating speed.
Singular values of spectral density matrices, attained from vibration data using PP (Peak Picking) technique are shown in Fig. 5. Natural frequencies acquired from all measurement setup are given in Table 2. The first five mode shapes extracted from experimental modal analyses are given in Fig. 6. When all measurements are examined, it can be seen that a best accordance is found between experimental mode shapes. In addition, when both setup sets are experimentally identified modal parameters are checked with each other, it can be seen that there is a best agreement between the mode shapes in the experimental modal analyses.  Fig. 6 experimentally identified mode shapes of model

V. EXPERIMENTAL MODAL ANALYSIS OF RETROFITTED MODEL CONCRETE STRUCTURE
In the case of retrofitted beams, the following are studies made on it to check and examine the efficiency of using unidirectional GFRP composite: slab of the model concrete structure is retrofitted with one layer GFRP composite. The Unidirectional GFRP composite and its components YKS Fiber is product of YKS Corporation (Fig. 7). The properties of the dry carbon fiber composite are: E=1.350E11 N/m 2 , Poisson ratio μ=0.3, mass per unit volume ρ=15696 N/m 3 , thickness=0.000152 m.

Fig. 7 GFRP composite and using details
The steps to pass through during retrofitting are shown below in details: A thin coat of putty is applied (Fig. 7) to the slab, approximately 1hour of curing in order to prepare a surface for application of GFRP composite. Next step, bottom surface of slab is covered with GFRP composites. After these setups, ambient vibration tests are followed by curing to obtain experimental dynamic characteristics similar to previously used properties in order to obtain comparative measurements. SVSDM are shown in Fig. 8. Table 3 shows the identified natural frequencies and modal damping ratios.
It is clear that using GFRP composites seems to be very effective for strengthening concrete members along with increasing stiffness; this research aims to determine how GFRP composite implementation affects structural response of model concrete structure by changing of dynamic characteristics.   Fig. 9 experimentally identified mode shapes of retrofitted model concrete structure