Dynamic testing of a four-storey building with reinforced concrete and unreinforced masonry walls: prediction, test results and data set

Abstract

This paper presents the results of a series of shake-table tests on a half-scale, four-storey building with reinforced concrete and unreinforced masonry walls. Due to the lack of reference tests, the seismic behaviour of such mixed structures is poorly understood. The test unit was subjected to several runs of increasing intensity yielding performance states between minor damage and near collapse. Before the test, the expected peak table accelerations leading to different limit states were estimated using the capacity spectrum method, and the predicted values corresponded rather well to actual sustained accelerations. Next to these analyses, the paper describes the test unit, instrumentation and input motion, and comments on the response of the mixed structure in terms of damage evolution and global response quantities, such as force–displacement response and drift and acceleration profiles. The raw and post-processed data sets are made publically available, and all relevant information with regard to data organisation and post-processing procedure is described in an appendix to this paper. The test serves therefore as a benchmark for the validation of numerical models of such mixed structures. The project aims at providing a foundation for the development of seismic design and assessment methods of mixed structures, which are currently not covered by structural codes, including Eurocode 8.

Appendices

Appendices

The objective of the appendices is to share the data gained from the shake-table test by documenting and post-processing in such a way that it can be easily used by third parties. The appendices summarise the geometry of the test unit (Appendix 1) and material properties of all construction materials, i.e., mortar, brick, masonry, concrete and reinforcing bars (Appendix 2). Appendix 3 describes the instrumentation of the test unit. Appendix 4 details the two sets of data that can be downloaded via the following link: www.zenodo.org using the doi: 10.5281/zenodo.11578. The two sets of data correspond to (1) the unprocessed data and (2) a set of post-processed data and derived data where the data from the conventional and optical measurement systems employed in the test are synchronised.

Appendix 1: Test unit

2.1 Geometry of the test unit

The test unit was a four-storey structure with RC and URM walls and was built at half-scale (Fig. 26). The test unit had a rectangular footprint, and the walls—six URM walls and two RC walls—were arranged along the perimeter of the building (Fig. 27).

Fig. 26

Views of the test specimen: north side (a), west side (b) south side (c) and definition of the global reference system for the “Post-processed Data” (d). All measures are in mm

Fig. 27

Test specimen: plan view; all measures are in mm

The testing at reduced scale was conducted following the “Artificial Mass Simulation” law (Krawinkler 1979). This scaling law requires that the stiffness, strength and deformation capacities of the construction materials at reduced scale are the same as those at full-scale and that the density is increased by the scaling factor. However, changing the density without influencing the mechanical properties is typically not feasible, and additional masses were added in the form of concrete blocks casted on top of each slab. Plastic sheets were placed between the additional masses and the slab in order to minimise the blocks’ contribution to the stiffness and strength of the floor slabs. Table 3 summarises the masses of the test unit and the shake-table.

Table 3 Summary of structural masses

The URM walls were constructed using half-scale clay brick units that were specifically produced for the project; the dimension of the brick units were 150 × 95 × 95 mm (L × W × H). The walls were assembled using standard mortar of class M15; the thickness of the horizontal and vertical mortar joints was also scaled down by a factor of two and was therefore 5 mm thick. The brick units employed for the construction of the URM walls were selected after an experimental campaign that compared the behaviour of URM walls at full-scale and half-scale (Petry and Beyer 2014a). The experimental campaign included material tests as well as quasi-static cyclic tests on half- and full-scale URM walls. It showed that the half-scale masonry leads to very similar values as the prototype masonry with regard to lateral stiffness, strength and deformation at peak strength but has an ultimate drift capacity (associated with a 20 % drop in horizontal strength) which exceeds that of the full-scale masonry by approximately 10 % (Petry and Beyer 2014a). At axial load failure, the drift capacity of the half-scale masonry was, on average, approximately 40 % larger than that of the full-scale masonry.

Past experience on the scaling of RC members showed that testing at half-scale leads to reasonable results if the bar diameter and the maximum aggregate size are scaled accordingly (Haroun et al. 1998; Ohtaki 2000). Typical, standard concrete has a maximum aggregate size of 32 mm. For the casting of half-scale members, the maximum aggregate size should therefore be 16 mm. To ease the pumping of the concrete, a maximum aggregate size of 8 mm was used instead. It is expected to have no impact on the obtained results, as the RC walls underwent only very limited ductility demands, the crack widths were small and the walls did not fail.

The two RC walls had the same geometry and reinforcement layout (Fig. 28). The RC slabs were 150 mm thick. The top and bottom reinforcement of the slab consisted of a steel net with D10 mm bars at 100 mm spacing.

Fig. 28

Reinforcement layout of RC walls: elevation (left) and cross sections (right); all measures are in mm

2.2 Construction and transportation

The test unit was built between September and October 2012 outside the TREES laboratory of EUCENTRE. The construction started with the casting of the RC ring foundation. For each of the four stories, first the six URM walls were built. Then, the reinforcement cages of the RC walls and slab were prepared, and finally, the walls and slabs were casted simultaneously.

The transportation of the specimen from the construction area to the shake-table was performed 28 days after the last casting of concrete. The structure was lifted by means of four hydraulic jacks and then moved using a slider system. In order to prevent cracking during the transportation, all URM walls were post-tensioned by two high-strength steel bars connecting the foundation to a steel beam at the top of each URM wall. The test unit was placed onto the shake-table, the foundation ring was post-tensioned to the shake-table and the post-tensioning system of the URM walls was removed. A visual inspection and a comparison of ambient noise vibration measurements before and after the transportation showed that the structure had not suffered any damage during the transportation. Before instrumenting the structure, the additional masses were casted with an uniform thickness equal to 430 mm; the layout of the additional masses over the RC slabs is presented in Fig. 29.

Fig. 29

Additional masses: plan view of the layout; all measures are in mm

Appendix 2: Material properties

3.1 RC elements

RC walls and slabs were casted using a concrete of class C28/35 and steel reinforcement of class B450C according to the classification of CEN (2004a). The RC foundation was casted using the same steel class and concrete of class C40/50. To investigate the mechanical properties of the concrete and steel, compression tests on cubic and cylindrical concrete samples and tensile tests on steel bars were conducted.

During each cast, nine concrete cubic samples were collected. The cubes with side lengths of 150 mm were used to determine the cubic strength (f _cm,cube ) according to UNI EN 12390-3 (2003). They were tested approximately five months after the casting, i.e., approximately two months after the testing of the building. Additionally, when casting the second storey, five cylindrical samples with a diameter of 160 mm and height of 320 mm were collected. Three of the cylinders were used to determine the elastic modulus of the concrete (E _cm ) according to UNI 6556 (1976) and the cylindrical strength (f _cm ). The remaining two cylinders were used for determining the concrete tensile strength (f _ctm ) by means of “double-punch” tests. These tests are carried out on cylinder samples which had been cut at half-height (Chen 1970). All the compression tests were performed in the laboratory of the Department of Civil Engineering and Architecture of the University of Pavia, while the “double-punch” tests were performed in the laboratory of the Institute of Civil Engineering at EPFL. Table 4 reports the results from the concrete tests; for each parameter, the mean value and the coefficient of variation are given.

Table 4 Results from the compression tests and from double-punch tests on concrete samples

For the construction of RC walls, only bars of 6 and 8 mm diameter were used. For each diameter, five bar samples—500 mm long—were subjected to tensile tests to determine their yield strength (f _y ) and tensile strength (f _t ). The same test was performed on five samples of the steel net used for the reinforcement of the RC slabs. The tensile test on straight samples was performed according to UNI EN ISO 15630-1 (2010), while the test on the steel net samples was performed following the Standard UNI EN ISO 15630-2 (2002). All tests on steel samples were performed in the laboratory of the Department of Civil Engineering and Architecture of the University of Pavia, and the results of these tests are summarised in Table 5.

Table 5 Results from the tensile test on steel samples

3.2 URM elements

The characterisation of the URM was comprised different types of material tests. First, mortar samples were collected during the construction of both the shake-table test unit and the masonry wallets for material tests. Second, tests on masonry wallets and triplets were carried out to determine the masonry properties. More than one hundred mortar samples were collected during the construction of the URM walls; these samples in the form of prisms (160 × 40 × 40 mm) were subjected to three-point bending tests for the definition of the flexural strength f _tm , and the two fragments were then used for compression tests to determine the cube mortar strength f _m . The tests were performed in the laboratory of the Department of Civil Engineering and Architecture of the University of Pavia according to UNI EN 1015-11 (1999). The results of these mortar tests are summarised in Table 6.

Table 6 Results from three-point bending test and compression test on mortar samples from test unit

As outlined in Sect. 5.1, prior to the construction of the shake-table test unit, a separate campaign studied the effect of scaling on masonry (Petry and Beyer 2014a). Next to quasi-static cyclic tests, the following material tests were performed at the laboratory of the Institute of Civil Engineering at EPFL (Fig. 30): compression tests (CEN 2002), diagonal compression tests (RILEM 1991) and shear triplet tests (CEN 2007). The results are reported in Tables 7, 8 and 9 to complement the mortar tests performed at the University of Pavia. The URM wallets for these material tests were built using half-scale model bricks from the same batch as those used for the construction of the shake-table test unit. For logistical reasons, the mortar was obtained from a different producer but was also classified as M15. For comparison, the compression strength of the mortar used for the construction of the wallets is reported in Table 10. The obtained mortar-strength values correspond well to the mortar strengths of the first- and second-storey URM walls of the shake-table test unit, i.e., to those two stories that underwent the largest inelastic deformations (Table 6).

Fig. 30

URM wallets for: a compression test, b diagonal compression test and c shear test (Petry and Beyer 2014a)

Table 7 Compressive strength, Poisson’s ratio and E-modulus of masonry from compression tests (Petry and Beyer 2014a)

Table 8 Tensile strength of masonry from diagonal compression test (Petry and Beyer 2014a)

Table 9 Peak strength of mortar-brick interface from shear triplet tests (Petry and Beyer 2014a)

Table 10 Mortar strength of mortar used for the construction of the masonry wallets (Petry and Beyer 2014a)

The compression tests on masonry wallets were used to determine the compression strength (f _u ), the Poisson’s ratio ν and the elastic modulus E of the masonry for vertical compression. The latter was evaluated for different stress ranges. The Poisson’s ratio was evaluated between 0 and 1/3 of the peak strength.

Appendix 3: Instrumentation

The instrumentation of the test unit was as follows: 20 accelerometers, 8 wire potentiometers, 41 potentiometers, 24 omega gages and an optical measurement system that recorded the displacements of 492 markers on the URM walls of the west face. In the following, the instrumentation setup is described, and figures show the location and orientation of all instruments. Accelerometers are annotated by “Acc”, potentiometers by “Pot”, and omega gages by “OG”. The data recorded by these conventional instruments was recorded with a sampling frequency of 1024 Hz, while the optical measurement system recorded data with a sampling frequency of 60 Hz. The two sets of data were recorded with two independent systems and were therefore not synchronous as raw data sets but were synchronised when post-processing the data.

4.1 Accelerometers

Each concrete slab was instrumented with four accelerometers, two in the longitudinal direction (i.e., the direction of motion) and two in the transversal direction. In addition, three accelerometers were installed on the foundation of the building: The two on the north side recorded accelerations in the longitudinal direction (“Acc_2” and “Acc_3”) and the one on the south side in the transversal direction (“Acc_4”). One other accelerometer was installed on the shake-table, recording accelerations in the longitudinal direction (“Acc_1”). Figure 31 shows the layout of accelerometers on the foundation and shake-table as well as on the first-storey slab; the configuration on the other three slabs was identical to that of the first-storey slab. The complete arrangement of the accelerometers is presented in Fig. 32; in both figures, the arrows indicate the positive direction of the unprocessed readings.

Fig. 31

Configuration of accelerometers: plan view of the foundation (a) and 1st storey slab (b); all measures are in mm

Fig. 32

Configuration of accelerometers: northeast view (a) and southeast view (b)

4.2 Potentiometers

The potentiometers were employed for two main purposes: measuring the in-plane deformations of the east RC and URM walls of the first storey and measuring the deformations of the out-of-plane loaded URM walls at the second, third and fourth storey.

The RC wall of the first storey on the east side of the building was instrumented with four wire potentiometers (measurement range ±50 mm) measuring the diagonal deformations of the top and bottom half of the panel (Fig. 33). Two wire potentiometers at the base were connected to the foundation of the building, while the two at the top were attached to the bottom face of the first-storey slab. In addition, each short side of the wall was instrumented with four potentiometers (±12.5 mm) covering the whole height of the wall. All potentiometers had an equal base length of 350 mm, and the top and bottom instruments were connected to the foundation and first-storey slab, respectively. For all potentiometers, positive measurement values corresponded to an elongation of the instruments.

Fig. 33

Potentiometers on the first-storey RC wall on the east side of the structure: layout (a) and image of the wall (b); all measures are in mm

Both first-storey URM walls on the east side of the building were instrumented with two wire potentiometers (±50 mm) covering the two diagonals of the panels (Figs. 34, 35). As in the case of the RC walls, the potentiometers were connected to the foundation and the first-storey slab, and positive measurement values corresponded to an elongation of the instrument.

Fig. 34

Configuration of potentiometers on the first-storey URM walls on the east side of the structure; all measures are in mm

Fig. 35

Configuration of potentiometers on the first-storey URM walls on the east side of the structure: southeast wall (a) and northeast wall (b)

Two potentiometers (±12.5 mm) were installed to record possible relative displacements between the foundation of the building and the shake-table. The two potentiometers were labelled “Pot_9” and “Pot_10” and were installed on the east and west sides of the foundation. The orientation of the two instruments was such that for “Pot_9”, positive values indicated a sliding of the foundation towards the north, and for “Pot_10”, positive values indicated a sliding towards the south. One further potentiometer “Pot_49” (±250 mm) was connected to the floor of the laboratory and the shake-table to measure the displacement of the table. The values recorded by this potentiometer were positive for displacements towards the north.

Each of the out-of-plane loaded URM walls of the second, third and fourth storey was instrumented with five potentiometers. On the top and bottom row of bricks, two potentiometers (±25 mm) were installed to measure vertical displacements. From these measurements, the relative rotation of the top and bottom of the wall with respect to the slabs could be computed. Figure 36 shows as example the two potentiometers “Pot_33” and “Pot_39” measuring the internal and external vertical displacement of the bottom row of bricks at the third story on the north side of the building. One further potentiometer (±125 mm) measured the out-of-plane displacement at mid-height of the walls with respect to the bottom slab. The layout of all potentiometers employed for measuring the deformations of the out-of-plane loaded walls is presented in Figs. 37 and 38. For all potentiometers, positive readings correspond to an elongation of the instrument.

Fig. 36

Potentiometers measuring the relative rotation between slab and out-of-plane loaded wall: “Pot_33” (a) and “Pot_39” (b) measuring base rotation of the third-storey URM walls on the north face of the building

Fig. 37

Potentiometers’ configuration for the measure of deformations of the out-of-plane loaded URM walls of the second, third and fourth stories

Fig. 38

Potentiometers’ configuration for the measure of deformations of the second storey, out-of-plane loaded URM walls on the north face of the building; all measures are in mm

4.3 Omega gages

One goal of the shake-table test was to investigate the contributing width of the RC slab spanning between vertical walls. For this reason, 24 omega gages were installed on the east side of the first-storey slab to monitor the strains of the top and bottom face of the slab. The omega gages had a base length of 120 mm and a measurement range of ±2 mm. The layout of the omega gages as well as their labels is presented in Fig. 39. Figure 40 shows a photo of the gages of “Location B”. Positive readings corresponded to an elongation of the omega gages.

Fig. 39

Configuration of omega gages at the first-storey slab on the east side of the test unit; all measures are in mm

Fig. 40

Omega gages of “Location B”: top face of the slab (a) and bottom face (b)

4.4 Optical measurement system

The displacement response of the structure during the tests was recorded using an optical measurement system. This is a 2D measurement system, i.e., only displacements in the direction of motion and vertical displacements were recorded. This system was developed by the TREES-laboratory (Lunghi et al. 2012) and uses high definition cameras to compute the position of reflecting markers (Fig. 41). Ten cameras were used, each one was equipped with a near-infrared filter and an infrared lamp projecting light against the reflecting markers. The cameras then recorded the light reflected from the markers with a sampling rate of 60 Hz. Each camera was able to cover a rectangular surface area of approximately 1.5 m × 2 m and could record the position of about 70 markers. During the test, the ten cameras were not synchronised, but the data was synchronised when it was post-processed (Appendix 4). Due to the limited number of cameras, only the motion of the URM walls and spandrels of the west façade could be recorded. Figure 42 shows the marker layout and the surface covered by each camera. The average distance between gridlines was approximately 0.2 m. The markers, with a square label in Fig. 42, are glued directly onto the URM surface, while the markers with a round label are glued onto small “L”-shaped steel plates, which were fixed to the RC slabs or to the foundation. In total, 492 markers were recorded of which 444 were glued onto the masonry and 48 were fixed to slabs and the foundation.

Fig. 41

Optical measurement system: high definition cameras (a), marker distribution on the west façade of the building (b)

Fig. 42

Optical measurement system: layout of markers and cameras

The camera layout was chosen in such a way that several markers were recorded by two cameras. This configuration allows for reconstruction of the displacement profile of the entire URM façade by synchronising the data measured by one camera with the data measured by an adjacent camera. The label of a marker consists of the camera number followed by a number identifying the marker within the subset of markers measured by one camera. The markers recorded by one camera are numbered from left to right, top to bottom. Note that the markers belonging to two subsets therefore have two names. The coordinate histories of each marker are computed in a global coordinate system whose origin is located at the initial position of the bottom, left corner of the structure (Lunghi et al. 2012). The x-axis is the horizontal axis and points towards the north, and the y-axis is the vertical axis and points upwards (Fig. 42).

Appendix 4: Test data

The data can be downloaded as one zip file from www.zenodo.org using the doi: 10.5281/zenodo.11578 (10 files of 0.26–0.41 GB). The platform ZENODO (www.zenodo.org) was developed under the European FP7 project (http://www.openaire.eu/) and is hosted by the research facility CERN, which operates a Large Hadron Collider. The database is organised into twelve subfolders: “Specimen Level”, “Experiment Level”, “Signal Level” and nine folders named after each test as follows: “Test<test number>”. Upon unzipping, the folder structure as shown in Fig. 43 unfolds.

Fig. 43

Organization of the data

The folder “Specimen Level” contains the subfolder “Construction Drawings” with the construction drawings of the test unit as pdf files. The folder “Experiment Level” is organized into two subfolders: “Photos” and “Videos”. The folder “Videos” contains the videos recorded during the nine different tests; the files are labelled as follows: CoMaWallS_Test_<test number>.mov. Videos are available for all tests except for Test 1. The folder “Photos” contains pictures taken during the experimental campaign, documenting the damage to the structure. All photos are JPG files, and the file names comprise three components: first, the test number after which the photo was taken; second, the type of damage the picture is documenting; and third, which element of the test specimen is shown on the photo. Figure 44 presents the labelling of the structural elements for the first storey of the structure. Hence, the file Test_2_HorizontalCracks_ERC1.jpg presents the formation of horizontal cracks after Test 2 in the first storey of the east RC wall. The subfolder “Signal Level” contains the subfolder “Metadata” which is organised into the two subfolders “Conventional” and “Optical”. The subfolder “Conventional” contains the files Conventional_Instruments.xls and Omega_Gages_Conversion_Parameters.xls. The first file is organised into six columns, which indicate for each instrument the following information: instrument label, channel number of the instrument in the conventional data files, type of instrument, producing company, model of the instrument and measurement range. The second file contains the parameters for the conversion of the recordings of the omega gages from voltage to deformations. The file comprises three columns. In the first column, the names of all 24 omega gages are listed (see Appendix 3), while the second and third columns report the parameters “a” and “b”. These two parameters are necessary for the definition of the linear relation between the variation of voltage recorded by the gage and its deformation in millimetres (Disp = a*Voltage + b). The strain is obtained by dividing the displacement by the base length of the instrument (see Appendix 3); positive values of deformation correspond to an elongation of the instrument. The subfolder “optical” contains ten JPG images named after each camera of the optical measurement system. The name of the files are Cam_<camera number>.jpg. Each image represents the layout of the markers recorded by the camera and their labels (see Appendix 3).

Fig. 44

Labelling of the structural elements of the first storey

The data recorded during the nine tests are organised into nine folders, one for each test. Each folder “Test<test number>” is subdivided into two subfolders: “Unprocessed Data” and “Post-processed Data”. These folders contain two different types of data: the recordings from conventional instruments (load cell of actuator, accelerometers, potentiometers and omega gages) and the recordings from the optical measurement system. The folder “Unprocessed Data” also contains the white-noise data recorded by the conventional measurement system during the tuning of the shake-table between the tests. These recordings can be used for dynamic identification analysis of the test specimen in order to investigate the variation of the dynamic properties during the test campaign.

5.1 Unprocessed data set

The folder “Unprocessed Data” contains the original and unmodified recorded data from the nine tests and the white-noise measurements recorded between the different tests. The data are organised into two subfolders: “Test” and “White Noise”. The folder “Test” is subdivided into the two subfolders “Conventional” and “Optical”, which contain the original files recorded during the shake-table testing by the conventional and optical measurement systems. The folder “White Noise” contains only the subfolder “Conventional”; optical measurements were not carried out during the white-noise excitation. The files are column-oriented files, and the different columns are separated by a tabulator (\t). The numbers are in scientific format with six decimals, and the decimal separator is the dot (.).

5.1.1 Conventional measurement data

The file contained in this folder reports the data recorded by the 96 conventional instruments, i.e., accelerometers, potentiometers and omega gages as well as the accelerometer, the displacement transducer and the load cell of the hydraulic actuator of the shake-table. Readings of the latter were exported from the shake-table controller and read into the data-acquisition system of the conventional instruments. The folder contains one file named after the test number as follows: Conventional_Meas_Unprocessed_Test_<test number>.txt. In total, nine files are available, one for each test.

The first line of each file contains the name of the instrument, the second line contains the unit, and from the third line onwards, the recorded data are reported. Each file comprises 97 columns. The first column includes an index which starts for each test with one. It is recalled that the sampling frequency of the conventional instruments was 1024 Hz. Columns 2–4 contain acceleration and displacement of the shake-table and the actuator force. For all three quantities, readings were positive when the table moved towards the north. Columns 5–24 contain the readings of the accelerometers on the shake-table, foundation and test unit; for the labelling of the accelerometers, the reader is referred to Appendix 3. For the “unprocessed” data set, readings are positive for accelerations towards the north (longitudinal direction) and east (transverse direction, Fig. 26). Columns 25–73 and columns 74–97 contain readings from potentiometers and omega gages, respectively. For potentiometers and omega gages, the sign convention of the unprocessed data set was provided in Appendix 3.

5.1.2 White noise conventional measurement data

Before a test, the shake-table was excited with a white-noise signal (frequency range: 0.2–40 Hz; RMS amplitude: 0.05 m/s²). The white-noise excitation served to tune the shake-table, but the results can also be used to determine the dynamic characteristics of the structure. For this reason, they are included in the unprocessed data set. Note that only the conventional and not the optical measurement system was recording during the white-noise excitation.

The files used for recording the response during the excitation with white noise have the same structure as the ones presented in the previous section. The files are named using the following convention: WhiteNoise_Unprocessed_Pre_Test_<test number>.txt. Since this recording was performed between two tests, the suffix “pre-” in the name of the files indicates that the data are recorded prior to a certain test. For example, the file WhiteNoise_Unprocessed_Pre_Test_2.txt reports the white noise data recorded between Tests 1 and 2. In total, eight files are available, and no white noise recordings were carried out prior to Test 7.

5.1.3 Optical measurement data

This folder contains one file with the data recorded by the optical measurement system. The file is named Optical_Meas_Unprocessed_Test_<test number>.txt. The sampling frequency of the optical measurement system is 60 Hz. In total, nine files, one for each test, are available.

Each file contains the time histories of the x and y coordinates of each marker in the afore-defined global reference system (Fig. 26). The first line of the file reports the name of the marker and the coordinate. The column label is composed of a number identifying the camera, a number identifying the marker within the subset of markers measured by this camera and the coordinate (x or y). For example, the column named “5.10.x” reports the x-coordinate of the 10th marker recorded by Camera 5. The second line of the file reports a variable called “good signal” that indicates if the recorded data for the marker is reliable or not. During the test, a camera may lose contact with the signal coming from one marker; in this case, the software keeps reading a constant coordinate value equal to the last recorded value before the signal was lost. At the end of the test, the software checks if there were markers whose coordinates remained constant during the test for a certain time interval or if there were markers with residual deformations exceeding a certain pre-defined limit value. If the recording of one marker fulfils one of the two conditions, it is considered unreliable. The variable “good marker” is binary, i.e., it takes a value of “one” when the measurements are reliable over the entire duration of the test and “zero” when the measurements are considered to be not fully reliable. The third line reports a variable which takes a value of “one” if the marker was attached to the masonry surface and “zero” if the marker was attached to the foundation or one of the slabs by means of steel angles (Appendix 3). From the fourth line onwards, the measured coordinates are reported.

Each file is composed of 1219 columns; the first column contains an index, and the remaining 1218 columns are the time histories of the horizontal and vertical coordinates of 492 markers; 111 of these 492 markers are measured by two cameras and 3 markers from three cameras.

5.2 Post-processed data set

This final data set contains the post-processed and synchronised conventional and optical data. In addition to the recorded channels, the post-processed data set contains a series of variables that were computed from the original channels; these variables are defined in the following and intended to facilitate the use of the data by allowing the user to easily plot fundamental graphs like global hysteretic response curves. The folder “Post-processed Data” contains the subfolder “Test” which is organised into two subfolders: “Conventional” and “Optical”.

The goals of the post-processing were as follows: (1) to obtain a data set in which conventional and optical recordings refer to the same global reference system; (2) to clean the recordings from bias and noise that, for different reasons, had been introduced during the tests; (3) to remove data that are considered unreliable; (4) and to synchronise the measurements coming from the two different systems (conventional and optical).

The synchronisation procedure was based on three steps: first, a re-sampling was made of the optical data from the original sampling frequency of 60 Hz to the sampling frequency of the conventional measurements of 1024 Hz. Second, the data from the different optical cameras had to be synchronised. Third, the conventional and optical data were synchronised. The synchronisation procedure is presented in the following sections.

The global reference system, which is shown in Fig. 26, follows the reference system of the optical measurement system. The z-axis, which points to the east, was defined by the TREES laboratory as a positive axis for the transverse direction. This sign convention was maintained in the post-processed data, thus leading to a left-handed reference system. Note that the z-axis is only used for the accelerometers that recorded east–west accelerations, which were very small due to the symmetry of the test unit with regard to the north–south axis. Note that the data of the white-noise measurements between tests are not included in the post-processed data set but only in the unprocessed data set.

5.2.1 Conventional measurement data

The data contained in the folder “Conventional” are organised into TXT files with the name Conventional_Meas_Postprocessed_Test_<test number>.txt. Each file comprises 101 columns as follows: The first column is for the index, columns 2–97 are for the 96 instruments, and columns 98–101 are for derived channels which will be defined later in this section.

The following post-processing procedure was applied to the original data set:

• The sign of the acceleration recordings along the longitudinal direction as well as the read-out from the shake-table controller (acceleration, force and displacement), and the displacement recordings from the potentiometer measuring the displacement of the shake-table (“Pot_49”) were switched to match the global reference system.

• Readings of all accelerometers were filtered to remove the noise introduced by the recording instruments. For the filtering, a sixth-order, low-pass Butterworth filter with a cut-off frequency of 40 Hz was used.

• Due to saturation, some acceleration time histories were showing anomalies such as drifting. These recordings were considered unreliable and were therefore replaced with “NaN” elements (“Not a Number”). If the recordings of an instrument were considered unreliable during a test, the entire channel was replaced with NaN elements during this test.

• For all tests, the readings of all accelerometers and the read-out from the shake-table controller (acceleration, force and displacement) were shifted to remove the initial offset. The record of each instrument was shifted by subtracting the mean value of the first 500 values of the record.

• For the first test, the readings of potentiometers and omega gages were shifted to remove the initial offset using the procedure described in the previous point. For the other eight tests, the records were shifted so that the initial displacement of a specific test corresponded to the residual displacement of the previous test. The initial value was computed as the mean value of the first 500 readings of the record, and the residual value as the mean value of the last 500 readings. The objective was to eliminate offsets in readings introduced between tests, such as due to accidental movements of the instruments when cracks were marked on the test unit.

• The data recorded by the 24 omega gages were converted from voltage to strain using factors “a” and “b” from the file “Omega_Gage_Conversion_Parameters.xls” previously presented.

5.2.2 Optical measurement data

The folder “Optical” contains TXT files, one for each test over the different test folders. The file is named Optical_Meas_Postprocessed_Test_<test number>.txt. The file has the same format and is organised in the same way as the optical measurement data of the unprocessed data set.

Two modifications were applied to this data set with respect to the unprocessed data set, as follows: first, the data set was re-sampled from the original frequency of 60–1024 Hz, to match the sampling frequency of the conventional data. Second, the data from different cameras were synchronised among each other. The optical measurement system was composed of ten cameras recording different subsets of markers. These cameras, despite having the same sampling frequency, did not take pictures at exactly the same instant and therefore the measurements of the different cameras are slightly offset in time. To eliminate this offset, the cameras were synchronised by means of the markers that were recorded by two cameras. To minimize the time lag due to the mixing of signals from different cameras, the signals were synchronised to maximize the cross-correlation coefficient of marker signals that were recorded by more than one camera. In addition, the order of the cameras was chosen to maximize the sum of these computed cross-correlation coefficients.

The adopted synchronisation procedure was the following: first, the cameras of the north-west URM wall were synchronised starting with Cameras 5 and 9 (Fig. 45). To do so, the records of Marker 2A by the two cameras were shifted in time so that the cross-correlation was maximized. The signals were then padded with NaN-elements at the beginning and end. The synchronisation procedure continued subsequently with Cameras 10, 6, 4 and 3 using the Markers 3A–6A, respectively. The same procedure was then applied to synchronise the cameras recording the south-west URM walls. Finally, the two sets of synchronised data (north-west and south-west) were aligned by synchronising the recordings of Markers 1A and 1B, assuming that the building foundation did not undergo axial deformations during the tests. In Table 11, the correspondence between the point used in the synchronisation procedure, presented in Fig. 45, and the corresponding markers’ labelling is presented.

Fig. 45

Definition of the markers for the synchronization of the optical measurement data

Table 11 Labels of markers that were used for the synchronisation of the cameras

5.2.3 Synchronisation of conventional and optical data

The conventional and optical data sets were synchronised using a procedure similar to the one presented in the previous section for the synchronisation of the different cameras. For the synchronisation of the two data sets, the displacement histories of the shake-table were used, i.e., the reading of the potentiometer “Pot_49” (see Appendix 3) and the history of the x-coordinate of Marker 2B (Fig. 45). The displacement histories were matched in time, maximising the cross-correlation. To apply this procedure, it was necessary to pad records with “NaN” elements at the beginning and at the end.

5.2.4 Derived channels

The additional channels described in this section are included in the data files for the conventional measurement system. The first derived data channel is Channel 98, which contains an estimate of the base shear V _b,ACC of the test unit. The base shear is computed as the sum of the inertia forces of the four stories. The inertia forces of each storey are computed from the two accelerometers that measure accelerations in the longitudinal direction: at Storey 1 from “Acc_5” and “Acc_6”, at Storey 2 from “Acc_9” and “Acc_10”, at Storey 3 from “Acc_13” and “Acc_14” and at Storey 4 from “Acc_17” and “Acc_18”. The unit of the accelerometer readings is g. The inertia force, in kN, was then computed multiplying the storey acceleration by the storey mass and the gravitational acceleration. The inertia force of storey 4, I ₄ , was, for example, calculated as follows:

$$ I_{4} = - M_{4} \cdot {\frac{{Acc_{17} + Acc_{18} }}{2}}\cdot {g} $$

(8)

where M ₄ is the mass of Storey 4. The storey masses were computed as the mass of the slab, the mass of the additional masses on the slab and half the mass of the walls above and below the slab; a summary of the storey masses can be found in Table 12. The base shear is the sum of the inertia forces of the four stories (Eq. 2).

Table 12 Summary of storey masses used for the computation of the base shear force

Both accelerometers of the fourth storey and one of the accelerometers of the third storey were saturated during the last two tests. Hence, for those tests, it was not possible to derive the inertia forces of the fourth storey and consequently the base shear force. For these last two tests, column 98 of the data file therefore contains “NaN” elements.

Channel 99 reports a further estimate of the base shear, V _b,ACT . This second estimate is evaluated as the actuator force of the shake-table, F _ACT , minus the sum of the inertia force of the shake-table, the inertia force of the foundation and the inertia force of half the mass of the walls of the first storey. The inertia force was obtained by multiplying the acceleration measured by the accelerometer on the shake-table, “Acc_1”, by the sum of the masses of the shake-table and the foundation and by the gravitational acceleration (Eq. 1).

A third estimate of the base shear is derived from the recordings of the optical measurement system and is reported in Channel 100. For this method, for each storey, the storey acceleration was computed from the displacement histories of three markers that were connected to the RC slab in correspondence with the central URM wall. These markers correspond to Markers 2.51, 2.54 and 2.57 for the first storey; 1.55, 1.58 and 1.61 for the second storey, 8.64, 8.67 and 8.70 for the third storey and 7.64, 7.67 and 7.70 for the fourth storey. The storey accelerations were obtained by taking the second time derivative of the storey displacements. The noise introduced by this procedure was reduced by applying, after each derivation step, a second-order Butterworth Lowpass filter with a cut-off frequency of 12 Hz. Once the storey accelerations were computed, the base shear, V _b,MARK , was evaluated by multiplying the storey acceleration by the storey mass, and finally, the base shear was derived as the sum of the inertia forces at all stories (Eq. 3).

Channel 101 reports the relative top displacement of the structure, D _top , with respect to its foundation. This top displacement was evaluated from the recordings of the Optical Measurement System as the difference of the average horizontal displacement of three markers connected to the top slab (Markers 7.64, 7.67 and 7.70) and three markers connected to the foundation (Markers 2.02, 2.05 and 2.08).

Note that all the variables described in this section were computed from the post-processed database, i.e., after the synchronisation of the conventional and optical measurement systems was performed.