- Email: [email protected]
Experimental evaluation of damage effect on dynamic characteristics of concrete encased composite column-beam connections
Engineering Failure Analysis 91 (2018) 129–150
Contents lists available at ScienceDirect
Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
Experimental evaluation of damage effect on dynamic characteristics of concrete encased composite column-beam connections Metin Hüsem, Mohammad Manzoor Nasery, Fatih Yesevi Okur, Ahmet Can Altunişik
T
⁎
Karadeniz Technical University, Department of Civil Engineering, 61080 Trabzon, Turkey
A R T IC LE I N F O
ABS TRA CT
Keywords: Ambient vibration test Concrete encased composite column Damage Dynamic characteristic
A major defect in determining the structural response for concrete encased composite columns using finite element is due to the uncertainties and assumption associated with the modeling process. Therefore, experimental measurements should be performed in order to validate the numerical results as well as obtain the real structural response. This paper seeks to address an experimental study about the evaluation of damage effect on the dynamic vibration characteristics of concrete encased composite columns (CECC) considering different column-beam connection types using ambient vibration tests. In an attempt to do so, four half scale concrete encased composite column-steel beam were built and tested in the laboratory with different column-beam connection types abbreviated as CECC-A, CECC-B, CECC-C and CECC-D without any changes in geometrical configuration of specimens and test setup. Cyclic loading tests were conducted in order to assess the post damage condition, while ambient vibration test were performed to extract the experimental dynamic characteristics such as natural frequencies, mode shapes and damping ratios using Enhanced Frequency Domain Decomposition (EFDD) and Stochastic Subspace Identification (SSI) methods for both undamaged (intact) and damaged conditions. The natural frequencies have decreased distinctly and mode shapes were broken with damages. These tests revealed that ambient vibration tests are enough to identify the dynamic characteristics of engineering structures for different conditions. The maximum differences in natural frequencies were calculated between 21.72% and 39.96%. A good agreement was noted for undamaged condition, the mode shapes were identical and the Modal Assurance Criterion (MAC) had value of 1.0. In contrast, there was not a good agreement for post damage condition. The mode shapes were different and MAC values were close to zero. Lastly, as the experimental damping ratios were examined, the results supported the idea that there were some differences and the values did not correlate favorably. These results are the compatible with the literature, but it is thought that the differences typically indicate that higher excitation levels are required to accurately capture the damping ratios.
1. Introduction The application of steel and concrete in the same section as structural elements which are also known as composite structural members goes to decades ago. Composite structural elements provide the combined behavior of two or more structural materials such
⁎
Corresponding author. E-mail addresses: [email protected] (M. Hüsem), [email protected] (M.M. Nasery), [email protected] (F.Y. Okur), [email protected] (A.C. Altunişik).
https://doi.org/10.1016/j.engfailanal.2018.04.030 Received 12 March 2018; Received in revised form 21 April 2018; Accepted 22 April 2018 Available online 23 April 2018 1350-6307/ © 2018 Elsevier Ltd. All rights reserved.
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
as steel and concrete. Due to sufficient structural performances such as high stiffness, enough strength, good ductility and large energy absorption capacity, composite steel-concrete structural members are used in vertical and horizontal structural elements widely in recent years [19]. Concrete encased composite columns constructed from the encasing of I, H, pipe or tube steel profile in reinforced concrete columns are one of the best composite structural elements in vertical bearing systems. When the encased steel composite section is tube or circular hollow, it is named as concrete encased concrete filled steel tube section. The main advantages of these types of columns are, having high stiffness and enough strength with small cross section in comparison to other types. In addition, the fire resistance of these types of columns is higher than conventional steel and concrete columns, because the encased profile is fully covered with a concrete layer [36]. Another remarkable advantage which should not be ignored is the corrosion resistance of these columns. Using such types of columns is preferred in ashore areas. On the other hand, the big disadvantage of the concrete encased composite column is the difficulty associated with its construction. Determination of structural response for concrete encased composite column using finite element analysis is very difficult because of many uncertainties and assumptions in the modeling process. Therefore, non-destructive experimental measurements should be performed to validate the numerical results and/or to obtain the real structural behavior. Ambient vibration based operational modal analyses method is one of the most efficient and useful tool which can be easily applied for these sort of structures. Many numerical and experimental researches have been conducted in order to evaluate the structural performance of concrete encased composite columns. An and Han [5] studied the behavior of concrete-encased CFST columns under combined compression and bending. It is inferred that the column-beam connection zones show best performance in lateral dynamic loads such as earthquake, wave or wind. This phenomenon is studied by many researchers with respect to cyclic loading considering different types of connection details [8,10,15,16,24,28–30,34]. Ren et al. [26] conducted an experimental study to determine the performance of concrete encased concrete filled steel tube columns under compression and torsion at the same time. Chen et al. [9] presented a detailed study related to depth ratio of encased profile which is directly affecting the flexural ductility of the composite columns. A series of experimental tests were carried out for post fire behavior of composite structural elements to investigate the efficiency of exterior concrete cover [32,33]. Although a great deal of studies related to concrete encased composite columns, beams and connection types are available, there are no enough researches regarding evaluation of the dynamic characteristic of composite structures based on ambient vibration test results. The rest of the paper is laid out as follows: Section 2 summarizes the basic formulation of ambient vibration test using Enhanced Frequency Domain Decomposition (EFDD) and Stochastic Subspace Identification methods (SSI) with Modal Assurance Criterion (MAC) value. Section 3 presents a brief test setup, column-beam connection type and loading protocol with its geometric configuration. Section 4 presents the modal characterization of the composite columns by means of experimental measurements. This section is also devoted to discussion and comparison of results. Finally, Section 5 draws the main conclusions of this study. 2. Formulation 2.1. Ambient vibration test 2.1.1. Enhanced frequency domain decomposition (EFDD) method EFDD method is an extension of FDD technique. In this method, modes are simply picked locating the peaks in singular value decomposition plots calculated from the spectral density spectra of the responses. In EFDD, the single degree of freedom (SDOF) Power Spectral Density (PSD) function that is identified around a peak of resonance, is taken back to the time domain using the Inverse Discrete Fourier Transform. The natural frequency is obtained by determining the number of zero-crossing as a function of time, and the damping by the logarithmic decrement of the corresponding SDOF normalized auto correlation function [17]. In EFDD method, the relationship between the unknown input and the measured responses can be expressed as [7,17]:
[Gyy (ω)] = [H (ω)] [Gxx (ω)] [H (ω)]T
(1)
where Gxx is the r×r Power Spectral Density (PSD) matrix of the input of which r is the number of inputs, Gyyis the m×m PSD matrix of the responses of which m is the number of responses, H(ω) is the m×r Frequency Response Function (FRF) matrix, and H (ω) and superscript T denotes complex conjugate and transpose, respectively. Solution of Eq. (1) is given in details in the literature [14,21,25,27]. 2.1.2. Stochastic subspace identification (SSI) method SSI is an output-only time domain method that directly works with time data, without the need to convert them to correlations or spectra. The model of vibration structures can be defined by a set of linear, constant coefficient and second-order differential equation [22]:
[M ]{U¨ (t )} + [C2 ]{U̇ (t )} + [K ]{U (t )} = {F (t )} = [B2 ]{u (t )}
(2)
where [M], [C2], [K] are the mass, damping and stiffness matrices, {F(t)} is the excitation force vector, and {U(t)} is the displacement vector depending on time t. Observe that the force vector {F(t)} is factorized into a matrix [B2], describing the inputs in space, and a vector {u(t)}. Solution of Eq. (2) is given in detail in the literature [11,18,35]. 130
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 1. Three dimensional view of the test setup including geometrical dimensions.
2.2. Modal assurance criterion (MAC) MAC is defined as a scalar constant relating the degree of consistency (linearity) between one modal and another reference modal vector [2] as follows,
MAC =
|{ϕai }T {ϕej }|2 {ϕai }T {ϕai }{ϕej }T {ϕej }
(3)
where {ϕai} and {ϕej} are ith and jth modal vectors obtained from different methods. 3. Laboratory models and studies 3.1. Test setup In this paper, it is aimed to determine the damage effect on dynamic vibration characteristics of concrete encased composite columns considering different column-beam connection types. For experimental investigation, four half scale composite column-steel beam specimens were constructed in laboratory. Fig. 1 shows three dimensional view of the test setup including geometrical dimensions. Each structural part was designed according to the AISC 360-10 [1]. The steel used in the study was Grade-S235, having yield strength of 235 MPa. The CECC had a square cross section of 250 × 250 mm with HEA-120 encased steel profile and the steel beam section was chosen as IPN140. The axial link elements have been manufactured from HEB200 steel profile. The dimensions 131
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 2. CECC-A connection detail including (a) 3D perspective (b) side and (c) front view.
(length and section) of the column and steel profiles have not been changed throughout the process and were considered as constant values for all connection details. As seen in Fig. 1 that the system has three hinges to transfer the cyclic loads to the connection zone [15,16]. In order to prevent the out of plane motion and eccentric loading at column-beam region during the experimental test, column and axial link elements have been separately limited with lateral brace elements using UPN100 steel profile. 3.2. Column-beam connection types In this study, four half scale concrete encased composite columns were formed considering different column-steel beam connection details. Each detail is summarized and abbreviated in below as: CECC-A: Conventional concrete encased composite column-steel beam connection. (conventional connection) CECC-B: Welded steel through-beam concrete encased composite column connection. (welded steel through-beam connection) CECC-C: Bolted steel through-beam concrete encased composite column connection. (bolted steel through-beam connection) CECC-D: Concrete encased composite column-haunched steel beam connection. (haunched connection) CECC-A is widely used conventional steel column-beam connection detail. In this detail, concrete encased profile is continuous and steel beam is connected to the flange of encased profile by stiffness plate and four bolts. Fig. 2 illustrates the three dimensional perspective in addition to side and front views for CECC-A. CECC-B is a new connection detail in which the encased profile of column is discontinuous and the steel through-beam separates the encased profile. Plates are welded horizontally and vertically to increase the stiffness of joint on both up and down flanges of the beam. Due to the fact that all of the contact surfaces are welded, this connection type can be called as welded connection. Fig. 3 shows the three dimensional perspective as well as to side and front views for CECC-B. CECC-C is similar to connection detail CECC-B with a difference of using bolts instead of welding at the encased profile and steel beam connection point. Thus, this connection type can be called as bolted connection where the bolts are parallel to the column and work for shear forces in the case of lateral loading. Fig. 4 shows the three dimensional perspective in addition to side and front views
Fig. 3. CECC-B connection detail including (a) 3D perspective (b) side and (c) front view. 132
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 4. CECC-C connection detail including (a) 3D perspective (b) side and (c) front view.
for CECC-C. CECC-D is the improved version of widely used conventional (CECC-A) connection detail. In addition to CECC-A, CECC-D is a haunch enforced connection type and is bolted to both concrete encased composite column and steel beam. In this connection, the encased profile is continuous and steel beam is welded to the stiffness plate and connected to encased profile flange using six bolts. The steel beam is enforced by haunch at the down flange. The width of the haunch flange is equal to the steel beam flange. Fig. 5 show the three dimensional perspective in addition to side and front views for CECC-D.
3.3. Loading protocol During the cyclic loading tests for all specimens, FEMA 461 loading history which is given in Fig. 6 was used. It was seen that the cycles started from the lowest damage state and six cycles were executed in the first step. Each cycle was executed twice and the increment coefficient of the displacement amplitude was chosen as 1.4 for each step.
4. Ambient vibration test Ambient vibration tests were carried out to extract the dynamic characteristics such as natural frequencies, mode shapes and damping ratios. In the ambient vibration tests, B&K 3560 type data acquisition system with 17 channels, B&K 4506-B003 type triaxial accelerometers which have 500 mV/g sensitivity and 0.3-2000 Hz frequency range, signal cables, PULSE [23] and OMA [20] software were used as the test equipment. Ambient vibration tests were carried out during 15 min considering 0-100 Hz frequency range. The signals obtained from the accelerometers were accumulated in data acquisition system and then transferred into the PULSE and OMA software's for signal processing. Then, dynamic characteristics were extracted using Enhanced Frequency Domain Decomposition (EFDD) and Stochastic Subspace Identification (SSI) methods. Determination of the accelerometer number and its placements are very important steps to extract the correct vibration parameters. For this purpose, initial finite element model was constituted in special software and modal analysis was done. A representative model generated in PULSE software and accelerometers layout along with their directions are presented in Fig. 7.
Fig. 5. CECC-D connection detail including (a) 3D perspective (b) side and (c) front view. 133
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 6. FEMA 461 loading history for the cyclic loading tests.
Fig. 7. Accelerometer locations during ambient vibration tests.
4.1. CECC-A: conventional connection (extended end plate joint) The ambient vibration measurements were performed on concrete encased composite columns (CECC) with CECC-A type connection detail for undamaged condition. Fig. 8 presents the view of test specimen and accelerometer placements. Singular values of spectral density matrices (SVSDM) and the average of auto spectral densities (AASD) of the data set that were obtained by EFDD method for undamaged condition are given in Fig. 9. The dynamic characteristics were identified using SSI method as well. The stabilization diagram for the first four modes is given in Fig. 10. The natural frequencies, mode shapes and damping ratios are calculated as stabile values. Fig. 11 illustrates the first four mode shapes. Dynamic characteristics are summarized for undamaged condition in Table 1.
Fig. 8. CECC with CECC-A connection detail including accelerometer placements. 134
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 9. SVSDM and AASD of the data set for undamaged condition.
Fig. 10. The stabilization diagram for undamaged condition.
Fig. 11. The mode shapes obtained by EFDD and SSI methods for undamaged condition. Table 1 Dynamic characteristics of the undamaged CECC with CECC-A connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
15.10 19.78 50.83 57.02
3.69 1.85 1.45 1.06
14.89 19.31 52.79 55.88
5.04 – 1.96 0.96
Similar to the earthquake hazards, lateral forces are applied on composite model to obtain the cracks and damages, especially in column-beam connections. Fig. 12 shows the some photos from damaged CECC with CECC-A connection details. It can be seen that joint failure was occurred after cyclic loading test implementation. This means that concrete parts of connection zone was firstly collapsed while welding damage was appeared in the last stages with ending of stiffness plate. The ambient vibration measurements are conducted on damaged models using same accelerometer location and measurement properties to evaluate the changes of dynamic characteristics. EFDD and SSI results are given in Figs. 13 and 14. Likewise, the mode shapes are presented in Fig. 15 and dynamic characteristics are summarized for damaged condition in Table 2. Although interests on the structural identification methods have been increasing, there are still some debates on whether the measured or calculated dynamic characteristics are good enough as damage indicators. Therefore, some derived coefficients such as the Modal Assurance Criterion (MAC) and the Coordinated Modal Assurance Criterion (COMAC) can also be useful for monitoring the
135
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 12. Some photos from damaged CECC with CECC-A connection detail.
Fig. 13. SVSDM and AASD of the data set for damaged condition.
Fig. 14. The stabilization diagram for damaged condition.
136
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 15. The mode shapes obtained by EFDD and SSI methods for damaged condition. Table 2 Dynamic characteristics of the damaged CECC with CECC-A connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
11.82 20.10 48.18 49.55
3.25 1.55 1.14 2.36
10.52 17.97 48.16 49.47
– 2.75 0.92 2.10
structures [4,6]. MAC is the most common method to establish the correlation between measured and calculated results [3]. It can be accepted that MAC values greater than 0.9 well represent the correlated mode shapes well [12,13,31]. MAC factor is derived from the measured mode shapes considering the undamaged and damaged conditions, and is shown in Fig. 16. According to Fig. 16, a decreasing trend with the progressive damage conditions is observed. The non-monotonic behavior of MAC values is because this index is very sensitive to measurement. 4.2. CECC-B: welded steel through-beam connection The ambient vibration measurements were performed on concrete encased composite columns (CECC) with CECC-B type connection detail for undamaged condition. Fig. 17 presents the view of test specimen and accelerometer placements. Singular values of spectral density matrices (SVSDM) and the average of auto spectral densities (AASD) of the data set obtained by EFDD method for undamaged condition are given in Fig. 18. The dynamic characteristics are also identified using SSI method. The stabilization diagram for the first four modes is illustrated in Fig. 19. The natural frequencies, mode shapes and damping ratios are calculated as stabile values. Fig. 20 illustrates the first four mode shapes. Dynamic characteristics are summarized for undamaged condition in Table 3. Similar to the earthquake hazards, lateral forces were applied on composite model in order to obtain the cracks and damages, especially in column-beam connections. Fig. 21 shows the some of the photos from damaged CECC with CECC-B connection detail. It can be seen that the concrete cover of connection zone has cracked but did not collapse as a whole. There was not any damage
Fig. 16. MAC values between the undamaged and damaged measured mode shapes. 137
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 17. CECC with CECC-B connection detail including accelerometer placements.
Fig. 18. SVSDM and AASD of the data set for undamaged condition.
Fig. 19. The stabilization diagram for undamaged condition.
Fig. 20. The mode shapes obtained by EFDD and SSI methods for undamaged condition.
observed in the welding of joint. Local buckling occurred in the beam which is the great advantage of this type of connection detail. According to the earthquake resistance principles, the damages and plastic hinges must be accumulated on beams instead of columns. The ambient vibration measurements were conducted on damaged model using same accelerometer location and measurement properties to evaluate the changes of dynamic characteristics. EFDD and SSI results are given in Figs. 22 and 23. Moreover, the mode shapes are presented in Fig. 24. Dynamic characteristics are summarized for damaged condition in Table 4. MAC factor was derived from the measured mode shapes considering the undamaged and damaged conditions, and is shown in Fig. 25. According to Fig. 25, a decreasing trend with the progressive damage conditions can be seen. The non-monotonic behavior of 138
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Table 3 Dynamic characteristics of the undamaged CECC with CECC-B connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
15.59 22.63 53.93 62.88
3.52 1.09 0.75 0.96
15.78 23.38 53.63 62.66
2.66 3.41 1.16 1.11
Fig. 21. Some photos from damaged CECC with CECC-B connection detail.
Fig. 22. SVSDM and AASD of the data set for damaged condition.
139
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 23. The stabilization diagram for damaged condition.
Fig. 24. The mode shapes obtained by EFDD and SSI methods for damaged condition. Table 4 Dynamic characteristics of the damaged CECC with CECC-B connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
11.91 19.94 39.18 44.52
1.23 0.71 1.96 2.58
10.26 20.12 32.20 44.24
0.85 – 1.62 2.23
Fig. 25. MAC values between the undamaged and damaged measured mode shapes.
MAC values is due to this index's high rate of sensitivity to the measurement. 4.3. CECC-C: bolted steel through-beam connection The ambient vibration measurements were performed on concrete encased composite columns (CECC) with CECC-C type 140
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 26. CECC with CECC-C connection detail including accelerometer placements.
Fig. 27. SVSDM and AASD of the data set for undamaged condition.
Fig. 28. The stabilization diagram for undamaged condition.
Fig. 29. The mode shapes obtained by EFDD and SSI methods for undamaged condition.
connection detail for undamaged condition. Fig. 26 presents the view of test specimen and accelerometer placements. Singular values of spectral density matrices (SVSDM) and the average of auto spectral densities (AASD) of the data set obtained by EFDD method for undamaged condition are given in Fig. 27. The dynamic characteristics are identified using SSI method as well. The stabilization diagram for the first four modes is given in Fig. 28. The natural frequencies, mode shapes and damping ratios are calculated as stabile values. Fig. 29 illustrates the first four mode shapes. Dynamic characteristics are summarized for undamaged 141
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Table 5 Dynamic characteristics of the undamaged CECC with CECC-C connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
12.41 18.17 39.88 51.35
3.12 0.78 1.28 1.49
12.46 21.90 39.77 51.04
5.03 – 1.79 1.51
Fig. 30. Some photos from damaged CECC with CECC-C connection detail.
condition in Table 5. Lateral forces similar to the earthquake hazards were applied on composite model in order to obtain the cracks and damages, especially in column-beam connections. Fig. 30 shows some of the photos from damaged CECC with CECC-C connection detail. It can be seen that this new type connection has also shown good performances. After the concrete cover failure, local buckling has occurred in external parts of the beam which was not inside the concrete encased composite column. There was not any damage in the steel beam connection with encased profile in the column. The ambient vibration measurements are conducted on damaged model using same accelerometer location and measurement properties to evaluate the changes of dynamic characteristics. EFDD and SSI results are given in Figs. 31 and 32. Likewise, the mode shapes are presented in Fig. 33. Dynamic characteristics for damaged condition are summarized in Table 6.
Fig. 31. SVSDM and AASD of the data set for damaged condition. 142
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 32. The stabilization diagram for damaged condition.
Fig. 33. The mode shapes obtained by EFDD and SSI methods for damaged condition. Table 6 Dynamic characteristics of the damaged CECC with CECC-C connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
9.46 18.50 39.06 43.32
6.72 2.25 2.05 0.94
8.61 14.92 36.82 38.80
2.67 – 3.89 1.54
Fig. 34. MAC values between the undamaged and damaged measured mode shapes.
MAC factor was derived from the measured mode shapes considering the undamaged and damaged conditions, and is shown in Fig. 34. According to Fig. 34, a decreasing trend with the progressive damage conditions was observed. The non-monotonic behavior of MAC values is because of this index's sensitivity to the measurement.
143
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 35. CECC with CECC-D connection detail including accelerometer placements.
4.4. CECC-D: haunched connection The ambient vibration measurements have been performed on concrete encased composite columns (CECC) with CECC-D type connection detail for undamaged condition. Fig. 35 presents the view of test specimen and accelerometer placements. Singular values of spectral density matrices (SVSDM) and the average of auto spectral densities (AASD) of the data set obtained by EFDD method for undamaged condition are given in Fig. 36. The dynamic characteristics were also identified using SSI method. The stabilization diagram for the first four modes is given in Fig. 37. The natural frequencies, mode shapes and damping ratios were calculated as stabile values. Fig. 38 illustrates the first four mode shapes. Dynamic characteristics for undamaged condition are summarized in Table 7. Lateral forces similar to the earthquake hazards were applied on composite model in order to obtain the cracks and damages, specifically in column-beam connections. Fig. 39 shows some of the photos from damaged CECC with CECC-D connection detail. It can be seen that the biggest collapsed region in concrete was obtained in this connection type because of the fact that the beam was enforced with haunch and distributed the big area in this region. After the failure of concrete part, welding damage was observed at upper and lower sides of the connection. Especially, welding between the haunch and stiffness plate was completely split. The stiffness plate has bended in the connection region as well. The ambient vibration measurements were conducted on damaged model using same accelerometer location and measurement properties to evaluate the changes of dynamic characteristics. EFDD and SSI results are given in Figs. 40 and 41. Besides, the mode shapes are presented in Fig. 42. Dynamic characteristics for damaged condition are summarized in Table 8. MAC factor was derived from the measured mode shapes considering the undamaged and damaged conditions, and is shown in Fig. 43. According to Fig. 43, a decreasing trend with the progressive damage conditions is being observed. The non-monotonic behavior of MAC values is because of this index's sensitivity to the measurement. Comparison of all identified natural frequencies and damping ratios for undamaged and damaged models obtained in experimental manner considering different column-beam connection types are presented in Tables 9 and 10. A good agreement can be seen between experimentally identified natural frequencies for all connection types using EFDD and SSI methods. The maximum differences were obtained below 5% except for the second mode of CECC-C and CECC-D type connections. But, there was no enough correlation between damping ratios. These results are the compatible with the literature, but it is thought that the differences typically indicate that higher excitation levels are essential in order to accurately capture the damping ratios. It can be seen that CECC-D type connection ensures the most rigid column-beam intersection regions when the results were compared with conventional connection type (CECC-A). When cracks and damages were formed in the column-beam connections, the
Fig. 36. SVSDM and AASD of the data set for undamaged condition. 144
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 37. The stabilization diagram for undamaged condition.
Fig. 38. The mode shapes obtained by EFDD and SSI methods for undamaged condition. Table 7 Dynamic characteristics of the undamaged CECC with CECC-D connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
17.62 21.51 56.16 70.44
1.58 1.29 0.96 1.13
18.01 23.74 56.02 70.20
2.83 3.48 1.60 1.25
natural frequencies have been strongly affected. They have decreased due to a decrement in the flexural rigidity at the cracked sections. The decrease in natural frequencies due to damages was not monotonic. The maximum reduction percentages were calculated as 21.72%–29.35% for CECC-A, 29.20%–39.96% for CECC-B, 23.77%–31.87% for CECC-C and 31.44%–36.53% for CECC-D connection types by EFDD and SSI methods, respectively. Additional comparisons are presented in Fig. 44 for better understanding. The first four experimental mode shapes are plotted and compared in Figs. 16, 25, 34, and 43 to display the damage effects on the structural response. As can be clearly seen that the mode shapes are broken with damages and there is not a good agreement between mode shapes after damaged conditions. According to these figures, a decreasing trend with the progressive damage conditions is being observed. The non-monotonic behavior of MAC values is because of this index's sensitivity to the measurement. 5. Conclusions This paper is the first one that presents vibration based modal parameter identification of concrete encased composite columns considering different column-beam connection types for undamaged and damaged conditions using ambient vibration tests. Our work has led us to conclude the following noteworthy points:
• The natural frequencies obtained as a result of the ambient vibration tests applied to the undamaged state of the concrete encased • •
composite columns using EFDD and SSI methods were very close to each other. This confirms that the results obtained for both methods are in harmony since the maximum differences were obtained below the 5% except for the second mode of CECC-C and CECC-D type connections. Remarkably, in the conventional connection (CECC-A), joint failure has occurred after cyclic loading test implementation. A concrete part of connection zone has firstly collapsed while welding damages have appeared in the last stages with ending of stiffness plate. In the welded steel through-beam connection (CECC-B), concrete cover of connection zone has cracked but did not collapse as a whole. There was not any damage in the joint welding. Local buckling has occurred in the beam which is the critical advantage of 145
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 39. Some photos from damaged CECC with CECC-D connection detail.
Fig. 40. SVSDM and AASD of the data set for damaged condition.
Fig. 41. The stabilization diagram for damaged condition.
this type of connection detail.
• In the bolted steel through-beam connection (CECC-C), it can be seen that this new type connection has also shown good performances. After the concrete cover failure, local buckling has occurred in external parts of the beam which was not inside the concrete encased composite column. This substantiates the fact that the new type connection has given away good performance. 146
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 42. The mode shapes obtained by EFDD and SSI methods for damaged condition. Table 8 Dynamic characteristics of the damaged CECC with CECC-D connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
12.08 20.90 49.06 53.86
5.71 0.52 0.58 3.80
11.43 21.01 48.94 53.52
1.51 1.88 1.59 2.30
Fig. 43. MAC values between the undamaged and damaged measured mode shapes.
Table 9 Comparison of natural frequencies before and after damage condition for all column-beam connection types. Mode number
1 Undamaged condition 2 3 4 Max. Decrease (%) 1 Damaged condition 2 3 4
CECC-A
CECC-B
CECC-C
CECC-D
EFDD (Hz)
Dif. (%)
SSI (Hz)
EFDD (Hz)
Dif. (%)
SSI (Hz)
EFDD (Hz)
Dif. (%)
SSI (Hz)
EFDD (Hz)
Dif. (%)
SSI (Hz)
15.10 19.78 50.83 57.02 21.72
1.39 2.38 3.86 2.00 –
14.89 19.31 52.79 55.88 29.35
15.59 22.63 53.93 62.88 29.20
1.21 3.31 0.56 0.35 –
15.78 23.38 53.63 62.66 39.96
12.41 18.17 39.88 51.35 23.77
0.40 20.53 0.28 0.60 –
12.46 21.90 39.77 51.04 31.87
17.62 21.51 56.16 70.44 31.44
2.21 10.36 0.25 0.34 –
18.01 23.74 56.02 70.20 36.53
11.82 20.10 48.18 49.55
10.99 10.59 0.04 0.16
10.52 17.97 48.16 49.47
11.91 19.94 39.18 44.52
13.85 0.90 17.81 0.63
10.26 20.12 32.20 44.24
9.46 18.50 39.06 43.32
8.98 19.35 5.73 10.43
8.61 14.92 36.82 38.80
12.08 20.90 49.06 53.86
5.38 0.53 0.24 0.63
11.43 21.01 48.94 53.52
Bold, italics and underline are used to emphasize the differences between EFDD and SSI, also after damaged condition.
147
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Table 10 Comparison of damping ratios before and after damage condition for all column-beam connection types. Mode number
1 2 3 4 1 2 3 4
Undamaged condition
Damaged condition
CECC-A
CECC-B
CECC-C
CECC-D
EFDD (Hz)
Dif. (%)
SSI (Hz)
EFDD (Hz)
Dif. (%)
SSI (Hz)
EFDD (Hz)
Dif. (%)
SSI (Hz)
EFDD (Hz)
Dif. (%)
SSI (Hz)
3.69 1.85 1.45 1.06 3.25 1.55 1.14 2.36
19.84 – 35.17 9.43 – 77.42 19.30 11.02
5.04 – 1.96 0.96 – 2.75 0.92 2.10
3.52 1.09 0.75 0.96 1.23 0.71 1.96 2.58
24.43 212.84 54.7 15.6 30.89 – 17.34 13.57
2.66 3.41 1.16 1.11 0.85 – 1.62 2.23
3.12 0.78 1.28 1.49 6.72 2.25 2.05 0.94
61.20 – 39.80 1.34 60.27 – 89.80 63.80
5.03 – 1.79 1.51 2.67 – 3.89 1.54
1.58 1.29 0.96 1.13 5.71 0.52 0.58 3.80
79.10 170 66.70 10.60 73.56 262.00 174.00 39.47
2.83 3.48 1.60 1.25 1.51 1.88 1.59 2.30
Bold, italics and underline are used to emphasize the differences between EFDD and SSI, also after damaged condition.
Fig. 44. Graphical presentations of natural frequencies changing before and after damage condition using EFDD and SSI methods.
148
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
• In the haunched connection (CECC-D), the biggest collapsed region was obtained in concrete because of the fact that the beam was • • • • •
enforced with haunch and distributed the big area in this region. After the failure of concrete part, welding damage was observed at upper and lower parts of the connection. Especially, welding between the haunch and stiffness plate was completely split. The stiffness plate has bended in the connection region as well. It is fundamental to note that CECC-D type connection ensures the most rigid column-beam intersection regions when the results are compared with conventional connection type (CECC-A). When cracks and damages were formed in the column-beam connections, the natural frequencies have been strongly affected. We believe that these frequencies decrease due to a decrement in the flexural rigidity at the cracked sections. It is crucial to note that the decrease in natural frequencies due to damages is not monotonic. The maximum differences in natural frequencies were calculated between 21.72% and 39.96%. MAC is the most common method to establish the correlation between measured and calculated results. It can be accepted that MAC values greater than 0.9 well represent the correlated mode shapes. MAC factor is derived from the measured mode shapes considering the undamaged and damaged conditions. It can be clearly seen that the mode shapes are broken with damages and there is not a good agreement between mode shapes after damaged conditions. According to these figures, a decreasing trend with the progressive damage conditions is being observed. The non-monotonic behavior of MAC values is because of this index's sensitivity to the measurement.
To validate the experimental results, finite element analysis should be done and calculated differences should be minimized with model updating procedure to reflect the current response of each specimen. It is well-known MAC values can provide the information about the overall stiffness change due to damages, contrary to expectation, they cannot predict the location of the damage. COMAC criterion indicates consistency, but not validity or orthogonality. With COMAC, any changes in the stiffness at each node can be detected. References [1] AISC 360-10, Specification for structural steel building, ANSI/AISC, Chicago, USA, 2010. [2] R.J. Allemang, The modal assurance criterion: twenty years of use and abuse, Sound and Vibration 37 (2003) 14–23. [3] R.J. Allemang, D.L. Brown, A correlation coefficient for modal vector analysis, Proceedings of the 1st International Modal Analysis Conference, 1982, pp. 110–116. [4] A.C. Altunışık, F.Y. Okur, V. Kahya, Modal parameter identification and vibration based damage detection of a multiple cracked cantilever beam, Eng. Fail. Anal. 79 (2017) 154–170. [5] Y.F. An, L.H. Han, Behaviour of concrete-encased CFST columns under combined compression and bending, J. Constr. Steel Res. 101 (2014) 314–330. [6] N. Baghiee, R.M. Esfahani, K. Moslem, Studies on damage and FRP strengthening of reinforced concrete beams by vibration monitoring, Eng. Struct. 31 (2009) 875–893. [7] J.S. Bendat, A.G. Piersol, Random Data: Analysis and Measurement Procedures, John Wiley and Sons, USA, 2004. [8] C. Campian, Z. Nagy, M. Pop, Behavior of fully encased steel-concrete composite columns subjected to monotonic and cyclic loading, Procedia Engineering 117 (2015) 439–451. [9] C.C. Chen, C.C. Chen, J.H. Shen, Effects of steel-to-member depth ratio and axial load on flexural ductility of concrete-encased steel composite columns, Eng. Struct. 155 (2018) 157–166. [10] F.X. Ding, G.A. Yin, L.Z. Jiang, Y. Bai, Composite frame of circular CFST column to steel-concrete composite beam under lateral cyclic loading, Thin-Walled Struct. 122 (2018) 137–146. [11] D.J. Ewins, Modal Testing: Theory and Practice, Research Studies Press, New York, Letchworth, Hertfordshire, England, 1984. [12] D.J. Ewins, Modal Testing: Theory and Practice, John Wiley &Sons, Inc., New York, 1995. [13] C.R. Farrar, T.A. Duffey, Bridge modal properties using simplified finite element analysis, J. Bridg. Eng. 3 (1) (1998) 38–46. [14] A.J. Felber, Development of Hybrid Bridge Evaluation System, University of British Columbia, Vancouver, Canada, 1993 (PhD Thesis). [15] H.J. Hwang, T.S. Eom, H.G. Park, S.H. Lee, H.S. Kim, Cyclic loading test for beam-column connections of concrete-filled U-shaped steel beams and concreteencased steel angle columns, Journal of Structural Engineering ASCE 141 (2015) 11 (04015020-1-12). [16] H.J. Hwang, T.S. Eom, H.G. Park, S.H. Lee, H.S. Kim, Cyclic loading test for beam-column connections of concrete-filled U-shaped steel beams and concreteencased steel angle columns, J. Struct. Eng. 141 (11) (2015) 04015020. [17] N.J. Jacobsen, P. Andersen, R. Brincker, Using enhanced frequency domain decomposition as a robust technique to harmonic excitation in operational modal analysis, Proceedings of ISMA2006: International Conference on Noise & Vibration Engineering, Leuven, Belgium, 2006. [18] J.N. Juang, Applied System Identification, Prentice-Hall Inc., Englewood Cliffs (NJ), 1994. [19] Q.Q. Liang, Analysis and Design of Steel and Composite Structures, CRC Press, 2014. [20] OMA, Software: Operational modal analysis, Release 4.0. Structural Vibration Solution A/S, Denmark, 2006. [21] B. Peeters, System Identification and Damage Detection in Civil Engineering, K.U, Leuven, Belgium, (2000) (PhD Thesis). [22] B. Peeters, G. De Roeck, Reference based stochastic subspace identification in civil engineering, Proceedings of the 2nd International Conference on Identification in Engineering Systems, Swansea, UK, 1999. [23] PULSE, Analyzers and Solutions, Release 11.2. Bruel and Kjaer, Sound and Vibration Measurement A/S, Denmark, (2006). [24] W.W. Qian, W. Li, L.H. Han, X.L. Zhao, Analytical behavior of concrete-encased CFST columns under cyclic lateral loading, J. Constr. Steel Res. 120 (2016) 206–220. [25] C. Rainieri, G. Fabbrocino, E. Cosenza, G. Manfredi, Implementation of OMA procedures using labview: Theory and application, 2nd International Operational Modal Analysis Conference, Copenhagen, Denmark, 2007. [26] Q.X. Ren, L.H. Han, C. Hou, Z. Tao, S. Li, Concrete-encased CFST columns under combined compression and torsion: experimental investigation, J. Constr. Steel Res. 138 (2017) 729–741. [27] W.X. Ren, T. Zhao, I.E. Harik, Experimental and analytical modal analysis of steel arch bridge, Journal of Structural Engineering ASCE 130 (7) (2004) 1022–1031. [28] B. Sevim, Dynamic and static behavior of a steel covered column, Journal of Structural Engineering and Applied Mechanics 1 (1) (2018) 22–27. [29] K. Wang, X.F. Lu, S.F. Yuan, D.F. Cao, Z.X. Chen, Analysis on hysteretic behavior of composite frames with concrete-encased CFST columns, J. Constr. Steel Res. 135 (2017) 176–186. [30] W. Wang, X. Qian, B. Yang, Fracture in concrete-filled square hollow section to H-section joints under low-cycle loading, J. Constr. Steel Res. 101 (2014) 363–372.
149
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
[31] X. Wang, J.A. Swanson, A.J. Helmicki, V.J. Hunt, Development of dynamic response-based objective functions for finite-element modeling of bridges, J. Bridg. Eng. 12 (5) (2007) 552–561. [32] K. Xiang, Y.C. Pan, G.H. Wang, Axial compression analysis of square concrete-encased concrete filled steel tubular stub columns after fire exposure, Procedia Engineering 211 (2018) 1151–1159. [33] T. Yi Soang, L. Hai Han, Post-fire behaviour of concrete-filled steel tubular column to axially and rotationally restrained steel beam joint, Fire Saf. J. 69 (2014) 147–163. [34] F. Yu Liao, L. Hai Han, Z. Tao, Behaviour of composite joints with concrete encased CFST columns under cyclic loading: experiments, Eng. Struct. 59 (2014) 745–764. [35] D.J. Yu, W.X. Ren, EMD-based stochastic subspace identification of structures from operational vibration measurements, Eng. Struct. 27 (12) (2005) 1741–1751. [36] J. Zhang, Design of Composite Columns, Helsinki University of Technology Department of Civil and Environmental Engineering, 2004.
150
Contents lists available at ScienceDirect
Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
Experimental evaluation of damage effect on dynamic characteristics of concrete encased composite column-beam connections Metin Hüsem, Mohammad Manzoor Nasery, Fatih Yesevi Okur, Ahmet Can Altunişik
T
⁎
Karadeniz Technical University, Department of Civil Engineering, 61080 Trabzon, Turkey
A R T IC LE I N F O
ABS TRA CT
Keywords: Ambient vibration test Concrete encased composite column Damage Dynamic characteristic
A major defect in determining the structural response for concrete encased composite columns using finite element is due to the uncertainties and assumption associated with the modeling process. Therefore, experimental measurements should be performed in order to validate the numerical results as well as obtain the real structural response. This paper seeks to address an experimental study about the evaluation of damage effect on the dynamic vibration characteristics of concrete encased composite columns (CECC) considering different column-beam connection types using ambient vibration tests. In an attempt to do so, four half scale concrete encased composite column-steel beam were built and tested in the laboratory with different column-beam connection types abbreviated as CECC-A, CECC-B, CECC-C and CECC-D without any changes in geometrical configuration of specimens and test setup. Cyclic loading tests were conducted in order to assess the post damage condition, while ambient vibration test were performed to extract the experimental dynamic characteristics such as natural frequencies, mode shapes and damping ratios using Enhanced Frequency Domain Decomposition (EFDD) and Stochastic Subspace Identification (SSI) methods for both undamaged (intact) and damaged conditions. The natural frequencies have decreased distinctly and mode shapes were broken with damages. These tests revealed that ambient vibration tests are enough to identify the dynamic characteristics of engineering structures for different conditions. The maximum differences in natural frequencies were calculated between 21.72% and 39.96%. A good agreement was noted for undamaged condition, the mode shapes were identical and the Modal Assurance Criterion (MAC) had value of 1.0. In contrast, there was not a good agreement for post damage condition. The mode shapes were different and MAC values were close to zero. Lastly, as the experimental damping ratios were examined, the results supported the idea that there were some differences and the values did not correlate favorably. These results are the compatible with the literature, but it is thought that the differences typically indicate that higher excitation levels are required to accurately capture the damping ratios.
1. Introduction The application of steel and concrete in the same section as structural elements which are also known as composite structural members goes to decades ago. Composite structural elements provide the combined behavior of two or more structural materials such
⁎
Corresponding author. E-mail addresses: [email protected] (M. Hüsem), [email protected] (M.M. Nasery), [email protected] (F.Y. Okur), [email protected] (A.C. Altunişik).
https://doi.org/10.1016/j.engfailanal.2018.04.030 Received 12 March 2018; Received in revised form 21 April 2018; Accepted 22 April 2018 Available online 23 April 2018 1350-6307/ © 2018 Elsevier Ltd. All rights reserved.
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
as steel and concrete. Due to sufficient structural performances such as high stiffness, enough strength, good ductility and large energy absorption capacity, composite steel-concrete structural members are used in vertical and horizontal structural elements widely in recent years [19]. Concrete encased composite columns constructed from the encasing of I, H, pipe or tube steel profile in reinforced concrete columns are one of the best composite structural elements in vertical bearing systems. When the encased steel composite section is tube or circular hollow, it is named as concrete encased concrete filled steel tube section. The main advantages of these types of columns are, having high stiffness and enough strength with small cross section in comparison to other types. In addition, the fire resistance of these types of columns is higher than conventional steel and concrete columns, because the encased profile is fully covered with a concrete layer [36]. Another remarkable advantage which should not be ignored is the corrosion resistance of these columns. Using such types of columns is preferred in ashore areas. On the other hand, the big disadvantage of the concrete encased composite column is the difficulty associated with its construction. Determination of structural response for concrete encased composite column using finite element analysis is very difficult because of many uncertainties and assumptions in the modeling process. Therefore, non-destructive experimental measurements should be performed to validate the numerical results and/or to obtain the real structural behavior. Ambient vibration based operational modal analyses method is one of the most efficient and useful tool which can be easily applied for these sort of structures. Many numerical and experimental researches have been conducted in order to evaluate the structural performance of concrete encased composite columns. An and Han [5] studied the behavior of concrete-encased CFST columns under combined compression and bending. It is inferred that the column-beam connection zones show best performance in lateral dynamic loads such as earthquake, wave or wind. This phenomenon is studied by many researchers with respect to cyclic loading considering different types of connection details [8,10,15,16,24,28–30,34]. Ren et al. [26] conducted an experimental study to determine the performance of concrete encased concrete filled steel tube columns under compression and torsion at the same time. Chen et al. [9] presented a detailed study related to depth ratio of encased profile which is directly affecting the flexural ductility of the composite columns. A series of experimental tests were carried out for post fire behavior of composite structural elements to investigate the efficiency of exterior concrete cover [32,33]. Although a great deal of studies related to concrete encased composite columns, beams and connection types are available, there are no enough researches regarding evaluation of the dynamic characteristic of composite structures based on ambient vibration test results. The rest of the paper is laid out as follows: Section 2 summarizes the basic formulation of ambient vibration test using Enhanced Frequency Domain Decomposition (EFDD) and Stochastic Subspace Identification methods (SSI) with Modal Assurance Criterion (MAC) value. Section 3 presents a brief test setup, column-beam connection type and loading protocol with its geometric configuration. Section 4 presents the modal characterization of the composite columns by means of experimental measurements. This section is also devoted to discussion and comparison of results. Finally, Section 5 draws the main conclusions of this study. 2. Formulation 2.1. Ambient vibration test 2.1.1. Enhanced frequency domain decomposition (EFDD) method EFDD method is an extension of FDD technique. In this method, modes are simply picked locating the peaks in singular value decomposition plots calculated from the spectral density spectra of the responses. In EFDD, the single degree of freedom (SDOF) Power Spectral Density (PSD) function that is identified around a peak of resonance, is taken back to the time domain using the Inverse Discrete Fourier Transform. The natural frequency is obtained by determining the number of zero-crossing as a function of time, and the damping by the logarithmic decrement of the corresponding SDOF normalized auto correlation function [17]. In EFDD method, the relationship between the unknown input and the measured responses can be expressed as [7,17]:
[Gyy (ω)] = [H (ω)] [Gxx (ω)] [H (ω)]T
(1)
where Gxx is the r×r Power Spectral Density (PSD) matrix of the input of which r is the number of inputs, Gyyis the m×m PSD matrix of the responses of which m is the number of responses, H(ω) is the m×r Frequency Response Function (FRF) matrix, and H (ω) and superscript T denotes complex conjugate and transpose, respectively. Solution of Eq. (1) is given in details in the literature [14,21,25,27]. 2.1.2. Stochastic subspace identification (SSI) method SSI is an output-only time domain method that directly works with time data, without the need to convert them to correlations or spectra. The model of vibration structures can be defined by a set of linear, constant coefficient and second-order differential equation [22]:
[M ]{U¨ (t )} + [C2 ]{U̇ (t )} + [K ]{U (t )} = {F (t )} = [B2 ]{u (t )}
(2)
where [M], [C2], [K] are the mass, damping and stiffness matrices, {F(t)} is the excitation force vector, and {U(t)} is the displacement vector depending on time t. Observe that the force vector {F(t)} is factorized into a matrix [B2], describing the inputs in space, and a vector {u(t)}. Solution of Eq. (2) is given in detail in the literature [11,18,35]. 130
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 1. Three dimensional view of the test setup including geometrical dimensions.
2.2. Modal assurance criterion (MAC) MAC is defined as a scalar constant relating the degree of consistency (linearity) between one modal and another reference modal vector [2] as follows,
MAC =
|{ϕai }T {ϕej }|2 {ϕai }T {ϕai }{ϕej }T {ϕej }
(3)
where {ϕai} and {ϕej} are ith and jth modal vectors obtained from different methods. 3. Laboratory models and studies 3.1. Test setup In this paper, it is aimed to determine the damage effect on dynamic vibration characteristics of concrete encased composite columns considering different column-beam connection types. For experimental investigation, four half scale composite column-steel beam specimens were constructed in laboratory. Fig. 1 shows three dimensional view of the test setup including geometrical dimensions. Each structural part was designed according to the AISC 360-10 [1]. The steel used in the study was Grade-S235, having yield strength of 235 MPa. The CECC had a square cross section of 250 × 250 mm with HEA-120 encased steel profile and the steel beam section was chosen as IPN140. The axial link elements have been manufactured from HEB200 steel profile. The dimensions 131
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 2. CECC-A connection detail including (a) 3D perspective (b) side and (c) front view.
(length and section) of the column and steel profiles have not been changed throughout the process and were considered as constant values for all connection details. As seen in Fig. 1 that the system has three hinges to transfer the cyclic loads to the connection zone [15,16]. In order to prevent the out of plane motion and eccentric loading at column-beam region during the experimental test, column and axial link elements have been separately limited with lateral brace elements using UPN100 steel profile. 3.2. Column-beam connection types In this study, four half scale concrete encased composite columns were formed considering different column-steel beam connection details. Each detail is summarized and abbreviated in below as: CECC-A: Conventional concrete encased composite column-steel beam connection. (conventional connection) CECC-B: Welded steel through-beam concrete encased composite column connection. (welded steel through-beam connection) CECC-C: Bolted steel through-beam concrete encased composite column connection. (bolted steel through-beam connection) CECC-D: Concrete encased composite column-haunched steel beam connection. (haunched connection) CECC-A is widely used conventional steel column-beam connection detail. In this detail, concrete encased profile is continuous and steel beam is connected to the flange of encased profile by stiffness plate and four bolts. Fig. 2 illustrates the three dimensional perspective in addition to side and front views for CECC-A. CECC-B is a new connection detail in which the encased profile of column is discontinuous and the steel through-beam separates the encased profile. Plates are welded horizontally and vertically to increase the stiffness of joint on both up and down flanges of the beam. Due to the fact that all of the contact surfaces are welded, this connection type can be called as welded connection. Fig. 3 shows the three dimensional perspective as well as to side and front views for CECC-B. CECC-C is similar to connection detail CECC-B with a difference of using bolts instead of welding at the encased profile and steel beam connection point. Thus, this connection type can be called as bolted connection where the bolts are parallel to the column and work for shear forces in the case of lateral loading. Fig. 4 shows the three dimensional perspective in addition to side and front views
Fig. 3. CECC-B connection detail including (a) 3D perspective (b) side and (c) front view. 132
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 4. CECC-C connection detail including (a) 3D perspective (b) side and (c) front view.
for CECC-C. CECC-D is the improved version of widely used conventional (CECC-A) connection detail. In addition to CECC-A, CECC-D is a haunch enforced connection type and is bolted to both concrete encased composite column and steel beam. In this connection, the encased profile is continuous and steel beam is welded to the stiffness plate and connected to encased profile flange using six bolts. The steel beam is enforced by haunch at the down flange. The width of the haunch flange is equal to the steel beam flange. Fig. 5 show the three dimensional perspective in addition to side and front views for CECC-D.
3.3. Loading protocol During the cyclic loading tests for all specimens, FEMA 461 loading history which is given in Fig. 6 was used. It was seen that the cycles started from the lowest damage state and six cycles were executed in the first step. Each cycle was executed twice and the increment coefficient of the displacement amplitude was chosen as 1.4 for each step.
4. Ambient vibration test Ambient vibration tests were carried out to extract the dynamic characteristics such as natural frequencies, mode shapes and damping ratios. In the ambient vibration tests, B&K 3560 type data acquisition system with 17 channels, B&K 4506-B003 type triaxial accelerometers which have 500 mV/g sensitivity and 0.3-2000 Hz frequency range, signal cables, PULSE [23] and OMA [20] software were used as the test equipment. Ambient vibration tests were carried out during 15 min considering 0-100 Hz frequency range. The signals obtained from the accelerometers were accumulated in data acquisition system and then transferred into the PULSE and OMA software's for signal processing. Then, dynamic characteristics were extracted using Enhanced Frequency Domain Decomposition (EFDD) and Stochastic Subspace Identification (SSI) methods. Determination of the accelerometer number and its placements are very important steps to extract the correct vibration parameters. For this purpose, initial finite element model was constituted in special software and modal analysis was done. A representative model generated in PULSE software and accelerometers layout along with their directions are presented in Fig. 7.
Fig. 5. CECC-D connection detail including (a) 3D perspective (b) side and (c) front view. 133
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 6. FEMA 461 loading history for the cyclic loading tests.
Fig. 7. Accelerometer locations during ambient vibration tests.
4.1. CECC-A: conventional connection (extended end plate joint) The ambient vibration measurements were performed on concrete encased composite columns (CECC) with CECC-A type connection detail for undamaged condition. Fig. 8 presents the view of test specimen and accelerometer placements. Singular values of spectral density matrices (SVSDM) and the average of auto spectral densities (AASD) of the data set that were obtained by EFDD method for undamaged condition are given in Fig. 9. The dynamic characteristics were identified using SSI method as well. The stabilization diagram for the first four modes is given in Fig. 10. The natural frequencies, mode shapes and damping ratios are calculated as stabile values. Fig. 11 illustrates the first four mode shapes. Dynamic characteristics are summarized for undamaged condition in Table 1.
Fig. 8. CECC with CECC-A connection detail including accelerometer placements. 134
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 9. SVSDM and AASD of the data set for undamaged condition.
Fig. 10. The stabilization diagram for undamaged condition.
Fig. 11. The mode shapes obtained by EFDD and SSI methods for undamaged condition. Table 1 Dynamic characteristics of the undamaged CECC with CECC-A connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
15.10 19.78 50.83 57.02
3.69 1.85 1.45 1.06
14.89 19.31 52.79 55.88
5.04 – 1.96 0.96
Similar to the earthquake hazards, lateral forces are applied on composite model to obtain the cracks and damages, especially in column-beam connections. Fig. 12 shows the some photos from damaged CECC with CECC-A connection details. It can be seen that joint failure was occurred after cyclic loading test implementation. This means that concrete parts of connection zone was firstly collapsed while welding damage was appeared in the last stages with ending of stiffness plate. The ambient vibration measurements are conducted on damaged models using same accelerometer location and measurement properties to evaluate the changes of dynamic characteristics. EFDD and SSI results are given in Figs. 13 and 14. Likewise, the mode shapes are presented in Fig. 15 and dynamic characteristics are summarized for damaged condition in Table 2. Although interests on the structural identification methods have been increasing, there are still some debates on whether the measured or calculated dynamic characteristics are good enough as damage indicators. Therefore, some derived coefficients such as the Modal Assurance Criterion (MAC) and the Coordinated Modal Assurance Criterion (COMAC) can also be useful for monitoring the
135
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 12. Some photos from damaged CECC with CECC-A connection detail.
Fig. 13. SVSDM and AASD of the data set for damaged condition.
Fig. 14. The stabilization diagram for damaged condition.
136
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 15. The mode shapes obtained by EFDD and SSI methods for damaged condition. Table 2 Dynamic characteristics of the damaged CECC with CECC-A connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
11.82 20.10 48.18 49.55
3.25 1.55 1.14 2.36
10.52 17.97 48.16 49.47
– 2.75 0.92 2.10
structures [4,6]. MAC is the most common method to establish the correlation between measured and calculated results [3]. It can be accepted that MAC values greater than 0.9 well represent the correlated mode shapes well [12,13,31]. MAC factor is derived from the measured mode shapes considering the undamaged and damaged conditions, and is shown in Fig. 16. According to Fig. 16, a decreasing trend with the progressive damage conditions is observed. The non-monotonic behavior of MAC values is because this index is very sensitive to measurement. 4.2. CECC-B: welded steel through-beam connection The ambient vibration measurements were performed on concrete encased composite columns (CECC) with CECC-B type connection detail for undamaged condition. Fig. 17 presents the view of test specimen and accelerometer placements. Singular values of spectral density matrices (SVSDM) and the average of auto spectral densities (AASD) of the data set obtained by EFDD method for undamaged condition are given in Fig. 18. The dynamic characteristics are also identified using SSI method. The stabilization diagram for the first four modes is illustrated in Fig. 19. The natural frequencies, mode shapes and damping ratios are calculated as stabile values. Fig. 20 illustrates the first four mode shapes. Dynamic characteristics are summarized for undamaged condition in Table 3. Similar to the earthquake hazards, lateral forces were applied on composite model in order to obtain the cracks and damages, especially in column-beam connections. Fig. 21 shows the some of the photos from damaged CECC with CECC-B connection detail. It can be seen that the concrete cover of connection zone has cracked but did not collapse as a whole. There was not any damage
Fig. 16. MAC values between the undamaged and damaged measured mode shapes. 137
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 17. CECC with CECC-B connection detail including accelerometer placements.
Fig. 18. SVSDM and AASD of the data set for undamaged condition.
Fig. 19. The stabilization diagram for undamaged condition.
Fig. 20. The mode shapes obtained by EFDD and SSI methods for undamaged condition.
observed in the welding of joint. Local buckling occurred in the beam which is the great advantage of this type of connection detail. According to the earthquake resistance principles, the damages and plastic hinges must be accumulated on beams instead of columns. The ambient vibration measurements were conducted on damaged model using same accelerometer location and measurement properties to evaluate the changes of dynamic characteristics. EFDD and SSI results are given in Figs. 22 and 23. Moreover, the mode shapes are presented in Fig. 24. Dynamic characteristics are summarized for damaged condition in Table 4. MAC factor was derived from the measured mode shapes considering the undamaged and damaged conditions, and is shown in Fig. 25. According to Fig. 25, a decreasing trend with the progressive damage conditions can be seen. The non-monotonic behavior of 138
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Table 3 Dynamic characteristics of the undamaged CECC with CECC-B connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
15.59 22.63 53.93 62.88
3.52 1.09 0.75 0.96
15.78 23.38 53.63 62.66
2.66 3.41 1.16 1.11
Fig. 21. Some photos from damaged CECC with CECC-B connection detail.
Fig. 22. SVSDM and AASD of the data set for damaged condition.
139
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 23. The stabilization diagram for damaged condition.
Fig. 24. The mode shapes obtained by EFDD and SSI methods for damaged condition. Table 4 Dynamic characteristics of the damaged CECC with CECC-B connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
11.91 19.94 39.18 44.52
1.23 0.71 1.96 2.58
10.26 20.12 32.20 44.24
0.85 – 1.62 2.23
Fig. 25. MAC values between the undamaged and damaged measured mode shapes.
MAC values is due to this index's high rate of sensitivity to the measurement. 4.3. CECC-C: bolted steel through-beam connection The ambient vibration measurements were performed on concrete encased composite columns (CECC) with CECC-C type 140
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 26. CECC with CECC-C connection detail including accelerometer placements.
Fig. 27. SVSDM and AASD of the data set for undamaged condition.
Fig. 28. The stabilization diagram for undamaged condition.
Fig. 29. The mode shapes obtained by EFDD and SSI methods for undamaged condition.
connection detail for undamaged condition. Fig. 26 presents the view of test specimen and accelerometer placements. Singular values of spectral density matrices (SVSDM) and the average of auto spectral densities (AASD) of the data set obtained by EFDD method for undamaged condition are given in Fig. 27. The dynamic characteristics are identified using SSI method as well. The stabilization diagram for the first four modes is given in Fig. 28. The natural frequencies, mode shapes and damping ratios are calculated as stabile values. Fig. 29 illustrates the first four mode shapes. Dynamic characteristics are summarized for undamaged 141
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Table 5 Dynamic characteristics of the undamaged CECC with CECC-C connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
12.41 18.17 39.88 51.35
3.12 0.78 1.28 1.49
12.46 21.90 39.77 51.04
5.03 – 1.79 1.51
Fig. 30. Some photos from damaged CECC with CECC-C connection detail.
condition in Table 5. Lateral forces similar to the earthquake hazards were applied on composite model in order to obtain the cracks and damages, especially in column-beam connections. Fig. 30 shows some of the photos from damaged CECC with CECC-C connection detail. It can be seen that this new type connection has also shown good performances. After the concrete cover failure, local buckling has occurred in external parts of the beam which was not inside the concrete encased composite column. There was not any damage in the steel beam connection with encased profile in the column. The ambient vibration measurements are conducted on damaged model using same accelerometer location and measurement properties to evaluate the changes of dynamic characteristics. EFDD and SSI results are given in Figs. 31 and 32. Likewise, the mode shapes are presented in Fig. 33. Dynamic characteristics for damaged condition are summarized in Table 6.
Fig. 31. SVSDM and AASD of the data set for damaged condition. 142
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 32. The stabilization diagram for damaged condition.
Fig. 33. The mode shapes obtained by EFDD and SSI methods for damaged condition. Table 6 Dynamic characteristics of the damaged CECC with CECC-C connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
9.46 18.50 39.06 43.32
6.72 2.25 2.05 0.94
8.61 14.92 36.82 38.80
2.67 – 3.89 1.54
Fig. 34. MAC values between the undamaged and damaged measured mode shapes.
MAC factor was derived from the measured mode shapes considering the undamaged and damaged conditions, and is shown in Fig. 34. According to Fig. 34, a decreasing trend with the progressive damage conditions was observed. The non-monotonic behavior of MAC values is because of this index's sensitivity to the measurement.
143
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 35. CECC with CECC-D connection detail including accelerometer placements.
4.4. CECC-D: haunched connection The ambient vibration measurements have been performed on concrete encased composite columns (CECC) with CECC-D type connection detail for undamaged condition. Fig. 35 presents the view of test specimen and accelerometer placements. Singular values of spectral density matrices (SVSDM) and the average of auto spectral densities (AASD) of the data set obtained by EFDD method for undamaged condition are given in Fig. 36. The dynamic characteristics were also identified using SSI method. The stabilization diagram for the first four modes is given in Fig. 37. The natural frequencies, mode shapes and damping ratios were calculated as stabile values. Fig. 38 illustrates the first four mode shapes. Dynamic characteristics for undamaged condition are summarized in Table 7. Lateral forces similar to the earthquake hazards were applied on composite model in order to obtain the cracks and damages, specifically in column-beam connections. Fig. 39 shows some of the photos from damaged CECC with CECC-D connection detail. It can be seen that the biggest collapsed region in concrete was obtained in this connection type because of the fact that the beam was enforced with haunch and distributed the big area in this region. After the failure of concrete part, welding damage was observed at upper and lower sides of the connection. Especially, welding between the haunch and stiffness plate was completely split. The stiffness plate has bended in the connection region as well. The ambient vibration measurements were conducted on damaged model using same accelerometer location and measurement properties to evaluate the changes of dynamic characteristics. EFDD and SSI results are given in Figs. 40 and 41. Besides, the mode shapes are presented in Fig. 42. Dynamic characteristics for damaged condition are summarized in Table 8. MAC factor was derived from the measured mode shapes considering the undamaged and damaged conditions, and is shown in Fig. 43. According to Fig. 43, a decreasing trend with the progressive damage conditions is being observed. The non-monotonic behavior of MAC values is because of this index's sensitivity to the measurement. Comparison of all identified natural frequencies and damping ratios for undamaged and damaged models obtained in experimental manner considering different column-beam connection types are presented in Tables 9 and 10. A good agreement can be seen between experimentally identified natural frequencies for all connection types using EFDD and SSI methods. The maximum differences were obtained below 5% except for the second mode of CECC-C and CECC-D type connections. But, there was no enough correlation between damping ratios. These results are the compatible with the literature, but it is thought that the differences typically indicate that higher excitation levels are essential in order to accurately capture the damping ratios. It can be seen that CECC-D type connection ensures the most rigid column-beam intersection regions when the results were compared with conventional connection type (CECC-A). When cracks and damages were formed in the column-beam connections, the
Fig. 36. SVSDM and AASD of the data set for undamaged condition. 144
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 37. The stabilization diagram for undamaged condition.
Fig. 38. The mode shapes obtained by EFDD and SSI methods for undamaged condition. Table 7 Dynamic characteristics of the undamaged CECC with CECC-D connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
17.62 21.51 56.16 70.44
1.58 1.29 0.96 1.13
18.01 23.74 56.02 70.20
2.83 3.48 1.60 1.25
natural frequencies have been strongly affected. They have decreased due to a decrement in the flexural rigidity at the cracked sections. The decrease in natural frequencies due to damages was not monotonic. The maximum reduction percentages were calculated as 21.72%–29.35% for CECC-A, 29.20%–39.96% for CECC-B, 23.77%–31.87% for CECC-C and 31.44%–36.53% for CECC-D connection types by EFDD and SSI methods, respectively. Additional comparisons are presented in Fig. 44 for better understanding. The first four experimental mode shapes are plotted and compared in Figs. 16, 25, 34, and 43 to display the damage effects on the structural response. As can be clearly seen that the mode shapes are broken with damages and there is not a good agreement between mode shapes after damaged conditions. According to these figures, a decreasing trend with the progressive damage conditions is being observed. The non-monotonic behavior of MAC values is because of this index's sensitivity to the measurement. 5. Conclusions This paper is the first one that presents vibration based modal parameter identification of concrete encased composite columns considering different column-beam connection types for undamaged and damaged conditions using ambient vibration tests. Our work has led us to conclude the following noteworthy points:
• The natural frequencies obtained as a result of the ambient vibration tests applied to the undamaged state of the concrete encased • •
composite columns using EFDD and SSI methods were very close to each other. This confirms that the results obtained for both methods are in harmony since the maximum differences were obtained below the 5% except for the second mode of CECC-C and CECC-D type connections. Remarkably, in the conventional connection (CECC-A), joint failure has occurred after cyclic loading test implementation. A concrete part of connection zone has firstly collapsed while welding damages have appeared in the last stages with ending of stiffness plate. In the welded steel through-beam connection (CECC-B), concrete cover of connection zone has cracked but did not collapse as a whole. There was not any damage in the joint welding. Local buckling has occurred in the beam which is the critical advantage of 145
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 39. Some photos from damaged CECC with CECC-D connection detail.
Fig. 40. SVSDM and AASD of the data set for damaged condition.
Fig. 41. The stabilization diagram for damaged condition.
this type of connection detail.
• In the bolted steel through-beam connection (CECC-C), it can be seen that this new type connection has also shown good performances. After the concrete cover failure, local buckling has occurred in external parts of the beam which was not inside the concrete encased composite column. This substantiates the fact that the new type connection has given away good performance. 146
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Fig. 42. The mode shapes obtained by EFDD and SSI methods for damaged condition. Table 8 Dynamic characteristics of the damaged CECC with CECC-D connection detail. Mode number
1 2 3 4
EFDD
SSI
Frequency (Hz)
Damping ratio (%)
Frequency (Hz)
Damping ratio (%)
12.08 20.90 49.06 53.86
5.71 0.52 0.58 3.80
11.43 21.01 48.94 53.52
1.51 1.88 1.59 2.30
Fig. 43. MAC values between the undamaged and damaged measured mode shapes.
Table 9 Comparison of natural frequencies before and after damage condition for all column-beam connection types. Mode number
1 Undamaged condition 2 3 4 Max. Decrease (%) 1 Damaged condition 2 3 4
CECC-A
CECC-B
CECC-C
CECC-D
EFDD (Hz)
Dif. (%)
SSI (Hz)
EFDD (Hz)
Dif. (%)
SSI (Hz)
EFDD (Hz)
Dif. (%)
SSI (Hz)
EFDD (Hz)
Dif. (%)
SSI (Hz)
15.10 19.78 50.83 57.02 21.72
1.39 2.38 3.86 2.00 –
14.89 19.31 52.79 55.88 29.35
15.59 22.63 53.93 62.88 29.20
1.21 3.31 0.56 0.35 –
15.78 23.38 53.63 62.66 39.96
12.41 18.17 39.88 51.35 23.77
0.40 20.53 0.28 0.60 –
12.46 21.90 39.77 51.04 31.87
17.62 21.51 56.16 70.44 31.44
2.21 10.36 0.25 0.34 –
18.01 23.74 56.02 70.20 36.53
11.82 20.10 48.18 49.55
10.99 10.59 0.04 0.16
10.52 17.97 48.16 49.47
11.91 19.94 39.18 44.52
13.85 0.90 17.81 0.63
10.26 20.12 32.20 44.24
9.46 18.50 39.06 43.32
8.98 19.35 5.73 10.43
8.61 14.92 36.82 38.80
12.08 20.90 49.06 53.86
5.38 0.53 0.24 0.63
11.43 21.01 48.94 53.52
Bold, italics and underline are used to emphasize the differences between EFDD and SSI, also after damaged condition.
147
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
Table 10 Comparison of damping ratios before and after damage condition for all column-beam connection types. Mode number
1 2 3 4 1 2 3 4
Undamaged condition
Damaged condition
CECC-A
CECC-B
CECC-C
CECC-D
EFDD (Hz)
Dif. (%)
SSI (Hz)
EFDD (Hz)
Dif. (%)
SSI (Hz)
EFDD (Hz)
Dif. (%)
SSI (Hz)
EFDD (Hz)
Dif. (%)
SSI (Hz)
3.69 1.85 1.45 1.06 3.25 1.55 1.14 2.36
19.84 – 35.17 9.43 – 77.42 19.30 11.02
5.04 – 1.96 0.96 – 2.75 0.92 2.10
3.52 1.09 0.75 0.96 1.23 0.71 1.96 2.58
24.43 212.84 54.7 15.6 30.89 – 17.34 13.57
2.66 3.41 1.16 1.11 0.85 – 1.62 2.23
3.12 0.78 1.28 1.49 6.72 2.25 2.05 0.94
61.20 – 39.80 1.34 60.27 – 89.80 63.80
5.03 – 1.79 1.51 2.67 – 3.89 1.54
1.58 1.29 0.96 1.13 5.71 0.52 0.58 3.80
79.10 170 66.70 10.60 73.56 262.00 174.00 39.47
2.83 3.48 1.60 1.25 1.51 1.88 1.59 2.30
Bold, italics and underline are used to emphasize the differences between EFDD and SSI, also after damaged condition.
Fig. 44. Graphical presentations of natural frequencies changing before and after damage condition using EFDD and SSI methods.
148
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
• In the haunched connection (CECC-D), the biggest collapsed region was obtained in concrete because of the fact that the beam was • • • • •
enforced with haunch and distributed the big area in this region. After the failure of concrete part, welding damage was observed at upper and lower parts of the connection. Especially, welding between the haunch and stiffness plate was completely split. The stiffness plate has bended in the connection region as well. It is fundamental to note that CECC-D type connection ensures the most rigid column-beam intersection regions when the results are compared with conventional connection type (CECC-A). When cracks and damages were formed in the column-beam connections, the natural frequencies have been strongly affected. We believe that these frequencies decrease due to a decrement in the flexural rigidity at the cracked sections. It is crucial to note that the decrease in natural frequencies due to damages is not monotonic. The maximum differences in natural frequencies were calculated between 21.72% and 39.96%. MAC is the most common method to establish the correlation between measured and calculated results. It can be accepted that MAC values greater than 0.9 well represent the correlated mode shapes. MAC factor is derived from the measured mode shapes considering the undamaged and damaged conditions. It can be clearly seen that the mode shapes are broken with damages and there is not a good agreement between mode shapes after damaged conditions. According to these figures, a decreasing trend with the progressive damage conditions is being observed. The non-monotonic behavior of MAC values is because of this index's sensitivity to the measurement.
To validate the experimental results, finite element analysis should be done and calculated differences should be minimized with model updating procedure to reflect the current response of each specimen. It is well-known MAC values can provide the information about the overall stiffness change due to damages, contrary to expectation, they cannot predict the location of the damage. COMAC criterion indicates consistency, but not validity or orthogonality. With COMAC, any changes in the stiffness at each node can be detected. References [1] AISC 360-10, Specification for structural steel building, ANSI/AISC, Chicago, USA, 2010. [2] R.J. Allemang, The modal assurance criterion: twenty years of use and abuse, Sound and Vibration 37 (2003) 14–23. [3] R.J. Allemang, D.L. Brown, A correlation coefficient for modal vector analysis, Proceedings of the 1st International Modal Analysis Conference, 1982, pp. 110–116. [4] A.C. Altunışık, F.Y. Okur, V. Kahya, Modal parameter identification and vibration based damage detection of a multiple cracked cantilever beam, Eng. Fail. Anal. 79 (2017) 154–170. [5] Y.F. An, L.H. Han, Behaviour of concrete-encased CFST columns under combined compression and bending, J. Constr. Steel Res. 101 (2014) 314–330. [6] N. Baghiee, R.M. Esfahani, K. Moslem, Studies on damage and FRP strengthening of reinforced concrete beams by vibration monitoring, Eng. Struct. 31 (2009) 875–893. [7] J.S. Bendat, A.G. Piersol, Random Data: Analysis and Measurement Procedures, John Wiley and Sons, USA, 2004. [8] C. Campian, Z. Nagy, M. Pop, Behavior of fully encased steel-concrete composite columns subjected to monotonic and cyclic loading, Procedia Engineering 117 (2015) 439–451. [9] C.C. Chen, C.C. Chen, J.H. Shen, Effects of steel-to-member depth ratio and axial load on flexural ductility of concrete-encased steel composite columns, Eng. Struct. 155 (2018) 157–166. [10] F.X. Ding, G.A. Yin, L.Z. Jiang, Y. Bai, Composite frame of circular CFST column to steel-concrete composite beam under lateral cyclic loading, Thin-Walled Struct. 122 (2018) 137–146. [11] D.J. Ewins, Modal Testing: Theory and Practice, Research Studies Press, New York, Letchworth, Hertfordshire, England, 1984. [12] D.J. Ewins, Modal Testing: Theory and Practice, John Wiley &Sons, Inc., New York, 1995. [13] C.R. Farrar, T.A. Duffey, Bridge modal properties using simplified finite element analysis, J. Bridg. Eng. 3 (1) (1998) 38–46. [14] A.J. Felber, Development of Hybrid Bridge Evaluation System, University of British Columbia, Vancouver, Canada, 1993 (PhD Thesis). [15] H.J. Hwang, T.S. Eom, H.G. Park, S.H. Lee, H.S. Kim, Cyclic loading test for beam-column connections of concrete-filled U-shaped steel beams and concreteencased steel angle columns, Journal of Structural Engineering ASCE 141 (2015) 11 (04015020-1-12). [16] H.J. Hwang, T.S. Eom, H.G. Park, S.H. Lee, H.S. Kim, Cyclic loading test for beam-column connections of concrete-filled U-shaped steel beams and concreteencased steel angle columns, J. Struct. Eng. 141 (11) (2015) 04015020. [17] N.J. Jacobsen, P. Andersen, R. Brincker, Using enhanced frequency domain decomposition as a robust technique to harmonic excitation in operational modal analysis, Proceedings of ISMA2006: International Conference on Noise & Vibration Engineering, Leuven, Belgium, 2006. [18] J.N. Juang, Applied System Identification, Prentice-Hall Inc., Englewood Cliffs (NJ), 1994. [19] Q.Q. Liang, Analysis and Design of Steel and Composite Structures, CRC Press, 2014. [20] OMA, Software: Operational modal analysis, Release 4.0. Structural Vibration Solution A/S, Denmark, 2006. [21] B. Peeters, System Identification and Damage Detection in Civil Engineering, K.U, Leuven, Belgium, (2000) (PhD Thesis). [22] B. Peeters, G. De Roeck, Reference based stochastic subspace identification in civil engineering, Proceedings of the 2nd International Conference on Identification in Engineering Systems, Swansea, UK, 1999. [23] PULSE, Analyzers and Solutions, Release 11.2. Bruel and Kjaer, Sound and Vibration Measurement A/S, Denmark, (2006). [24] W.W. Qian, W. Li, L.H. Han, X.L. Zhao, Analytical behavior of concrete-encased CFST columns under cyclic lateral loading, J. Constr. Steel Res. 120 (2016) 206–220. [25] C. Rainieri, G. Fabbrocino, E. Cosenza, G. Manfredi, Implementation of OMA procedures using labview: Theory and application, 2nd International Operational Modal Analysis Conference, Copenhagen, Denmark, 2007. [26] Q.X. Ren, L.H. Han, C. Hou, Z. Tao, S. Li, Concrete-encased CFST columns under combined compression and torsion: experimental investigation, J. Constr. Steel Res. 138 (2017) 729–741. [27] W.X. Ren, T. Zhao, I.E. Harik, Experimental and analytical modal analysis of steel arch bridge, Journal of Structural Engineering ASCE 130 (7) (2004) 1022–1031. [28] B. Sevim, Dynamic and static behavior of a steel covered column, Journal of Structural Engineering and Applied Mechanics 1 (1) (2018) 22–27. [29] K. Wang, X.F. Lu, S.F. Yuan, D.F. Cao, Z.X. Chen, Analysis on hysteretic behavior of composite frames with concrete-encased CFST columns, J. Constr. Steel Res. 135 (2017) 176–186. [30] W. Wang, X. Qian, B. Yang, Fracture in concrete-filled square hollow section to H-section joints under low-cycle loading, J. Constr. Steel Res. 101 (2014) 363–372.
149
Engineering Failure Analysis 91 (2018) 129–150
M. Hüsem et al.
[31] X. Wang, J.A. Swanson, A.J. Helmicki, V.J. Hunt, Development of dynamic response-based objective functions for finite-element modeling of bridges, J. Bridg. Eng. 12 (5) (2007) 552–561. [32] K. Xiang, Y.C. Pan, G.H. Wang, Axial compression analysis of square concrete-encased concrete filled steel tubular stub columns after fire exposure, Procedia Engineering 211 (2018) 1151–1159. [33] T. Yi Soang, L. Hai Han, Post-fire behaviour of concrete-filled steel tubular column to axially and rotationally restrained steel beam joint, Fire Saf. J. 69 (2014) 147–163. [34] F. Yu Liao, L. Hai Han, Z. Tao, Behaviour of composite joints with concrete encased CFST columns under cyclic loading: experiments, Eng. Struct. 59 (2014) 745–764. [35] D.J. Yu, W.X. Ren, EMD-based stochastic subspace identification of structures from operational vibration measurements, Eng. Struct. 27 (12) (2005) 1741–1751. [36] J. Zhang, Design of Composite Columns, Helsinki University of Technology Department of Civil and Environmental Engineering, 2004.
150