Self-Centring seismic-RESilient sTEEL structures
ERIES-SC-RESTEEL
Dataset Description
The ERIES-SC-RESTEEL project investigates the seismic performance, repairability, and effective placement of low-damage self-centring (SC) joints for steel MRFs. The tested solution is based on dissipative and self-centring components to be included at column bases (CBs) and beam-to-column joints (BCJs) to dissipate the earthquake input energy and prevent residual drifts.
The dissipative connections consist of two parts connected by a combination of: (1) friction devices (FDs), which dissipate the seismic input energy through the alternate slippage of the surfaces in contact; and (2) PT bars with disk springs that control the self-centring behaviour of the connection.
Tests were conducted at the National Laboratory for Civil Engineering (LNEC), Lisbon, Portugal, on the ST3D shake table on a reduced-scale structure with 3 storeys, 1 bay in the x-direction, and 2 bays in the y-direction. The model scaling is based on the material and acceleration scaling identity, and a scaling factor of 0.6 was selected to respect the Lab constraints while allowing the correct reproduction of the seismic response of the joints. The interstorey heights are equal to 1.92 m at the first storey and 1.80 m at the intermediate storeys, while the longitudinal and transversal bays have span lengths equal to 3.50 m and 4.00 m, respectively.
This dataset includes the output files associated with the shake table tests, which comprised 150 steps in between seismic tests and dynamic identification. Four configurations were tested, depending on the MRFs connections: (1) MRF ALL: low-damage and self-centring BCJs and CBs (including PT bars); (2) MRF CBs: low-damage and self-centring CBs (including PT bars) and low-damage BCJs (without PT bars); (3) MRF LD: low-damage BCJs and CBs (with untightened PT bars); (4) MRF LD2: same concept of MRF LD but removing PT bars from BCJs, loosening bolts at the base of gravity columns, removing Teflon from splice, tightening splice connection (now working as continuity connection) and controlled retightening of friction plates.
For the four configurations mentioned above, separate folders were created inside the “Output files” folder. Each contains the output processed data (“.xlsx” files) for seismic tests (GMs) and dynamic identification tests (CAT), according to the test sequence. The files are organized as follows: file “_ACC”: acceleration records; file “_DISP”: global and local response displacement channels and load cells (please refer to the instrumentation description file for more information); file “_SG”: strain gauges records.
A separate folder was created with representative videos of the seismic tests. Videos of the highest intensity tests were included. For CAT tests only accelerations were recorded, while for GM tests all instrumentation was active. All the data, sampled at 200Hz, is presented in separate “*.xlsx” files, aggregated in CAT and GMs. In each “.xlsx” file the sheet name corresponds to the test number and unique identifier indicated in the “Test_Sequence.xlsx” file.
For each test stage, more than one trial may have been executed to achieve the target motion. A unique identifier is associated with each trial. All the data, sampled at 200Hz, is presented in separate “*.xlsx” files, aggregated in CAT, GMs and Sine-Sweep stages. In each “.xlsx” file the sheet name corresponds to the test stage unique identifier indicated in the “Test_Sequence.xlsx” file.
The raw data was processed differently depending on the type of measurement: (i) acceleration (in units of mg) – band pass Fourier filter between 0.5Hz and 40Hz (filter order 2); (ii) displacement (in units of mm) – low pass Fourier filter at 15Hz (filter order 2). Each displacement sensor forms a time continuum (in pseudo time) from the first test to the last test (initial offset of the first test was removed from all tests); and (iii) all channels were cropped in 30 points at the beginning and 20 points at the end, with offset removal at the beginning of each stage but with cumulative displacements for all displacement channels.
Instrumentation included accelerometers at storey levels, displacement transducers at storey levels and at beam-column joints, optical transducers and wire potentiometers for global floor displacements, strain gauges on the MRFs; and load cells on PT bars.
Three GMs were selected, namely L’Aquila (pulse-like), Northridge (non-pulse-like) and Chi-Chi (non-pulse-like), to represent a broad range of seismic characteristics relevant for structural response, capturing both forward-directivity effects (pulse-like motions) and conventional far-field ground motions (non-pulse-like). The selected records were scaled in time (by λ1/2) and at four increasing intensity levels, i.e., 15%, 40%, 66% (ULS), and 100% (CLS), for each configuration. Multiple tests were conducted at different intensities, resulting in approximately 150 tests in total. Each test was preceded by low-amplitude characterisation tests (i.e., white-noise excitations) to identify the structure’s dynamic properties, including natural periods and damping ratios. For selected configurations, some tests were carried out both with and without the vertical component of the ground motions, and additional repetitions were performed at the two reference intensity levels (ULS and CLS) to assess the repeatability of the structural response. Post-test inspections were conducted after the ULS and CLS runs to confirm the absence of significant structural damage.
The research outputs demonstrated the structure's low damage and self-centring capability throughout the full test sequence.
Specimens
1. Steel MRF with low-damage self-centring joints
4
Shake table tests were conducted on a large-scale, three-story moment-resisting steel frame. The tested specimen is extracted from a three-story, three-bay prototype structure, with a scaling factor λ = 0.6. Accordingly, length units were scaled by λ, areas and forces by λ^2, and time by λ^1/2. Apart from the shake table tests, carried out at LNEC, two subassembly quasi-static tests were carried out at the University of Salerno.
Regarding the shake table test specimen, the internal MRF (represented in red) is intended to provide the lateral capacity, while the external frames are designed for gravity loads only. The inter-story heights are equal to 1.92 m at the first story and 1.80 m at higher stories. The longitudinal and transverse bays have span lengths of 3.50 m and 2.00 m, respectively. The design is carried out in accordance with Eurocode 8. The design earthquake at the Ultimate Limit State (ULS) is defined considering the Type 1 elastic response spectrum with a PGA = 0.35g and soil type D (i.e. the amplification factors are defined by a soil factor of S = 1.35). The Collapse Limit State (CLS) is assumed to have an intensity equal to 150% of the ULS. The behavior factor based on Eurocode 8 for MRFs in Ductility Class High is assumed as q = 6.5. The inter-story drift limit for the Damage Limit State (DLS) is assumed to be 1% for non-structural elements in accordance with the Eurocode 8 provisions. Beams’ and columns’ cross-sections of the MRF are IPE240 and HEB180, while the profiles of the gravity frames are IPE180 and HEA160. All the profiles are of steel grade S355.
The floor system was a steel-concrete composite floor with a total height of 130 mm. The floor masses of the specimen to carry out the test are equal to 11.76 and 10.31 ton at intermediate floors and the roof, respectively. In addition, a set of additional masses were added and fixed to the slabs to simulate service load conditions. The total mass of the specimen, including the additional masses, is 35.68 tons (measured on the shake table).
The MRF is equipped with BCJs and CBs in two different configurations, i.e., with and without PT bars. For the sake of clarity, hereinafter, the joints without PT bars are referred to as Low-Damage (LD) joints (LD-BCJs and LD-CBs), while joints with PT bars are referred to as Self-Centring (SC) joints (SC-BCJs and SC-CBs).
The connections are equipped with FDs (Friction Devices), which dissipate the seismic input energy through the alternate slippage of the surfaces in contact. The FDs consist of properly coated steel friction shims and steel cover plates clamped with pre-loadable bolts.
The specimen was produced on a specialized manufacturer and assembled in LNEC's seismic testing room, outside the shake table, having been moved using the overhead crane with a load capacity of 40 tons. It should be noted that the X direction (West-East direction) is usually referred to as the transverse direction (T) of the LNEC's 3D shake table, while the Y direction (North-South) is usually referred to as the longitudinal direction (L). The Z direction is associated with the vertical direction (V). The ERIES-SC-RESTEEL shake table tests was executed in both X and Z directions, which correspond to the direction of the MRF and the vertical direction, respectively.
1. Full self-centring low-damage system
This configuration is named MRF ALL. The self-centring system is composed of PT bars symmetrically placed with respect to the column’s depth and arranged in series with a system of disc springs, ensuring an adaptable stiffness-resistance combination. Slotted holes are included at the beam’s flanges to accommodate the target rotation (i.e., 0.04 rads). In SC-CB, similarly to the SC-BCJ, a combination of FDs and an SC system is used. In both the SC-CBs and SC-BCJs, the moment-rotation behavior is the typical flag-shape curve, where M1 is the moment at the gap opening.
The midspan splice connection, consisting of bolted plates and low-friction Teflon sheets, is intended to mitigate the frame expansion effect.
The shake table testing campaign was designed to evaluate the seismic performance of the specimen under progressively increasing earthquake demands while continuously monitoring changes in its dynamic characteristics. The test sequence began with a series of dynamic identification (CAT) tests in the X and Z directions, which were used to identify the initial dynamic properties of the structure, including its natural frequencies, damping ratios, and mode shapes. These dynamic identification tests were repeated throughout the campaign to track possible stiffness degradation and assess the accumulation of damage after each major seismic excitation.
Following the initial characterization, the specimen was subjected to three earthquake records: L’Aquila (GM1), Northridge (GM2), and Chi-Chi (GM3). Each record was applied at increasing intensity levels, typically 15%, 40%, 66%, and 100% of the target ground motion. The excitations were performed either in a single horizontal direction (X or Z) or simultaneously in both directions (XZ), allowing the evaluation of the structure under realistic multi-directional seismic loading.
Instrumentation
The shake table motion was monitored by 3 accelerometers (ACC) and 3 linear variable displacement transducers (LVDTs). Regarding the specimen, it was extensively instrumented with a range of sensors, including displacement transducers (wire potentiometers and LVDTs), optical sensors, accelerometers, load cells and strain gauges.
To monitor the accelerations at each floor of the specimen, accelerometers were employed. Each device is capable of measuring acceleration at a single point along one direction only; therefore, more than one accelerometer was combined to record multi-axial accelerations on several points at each floor. To capture potential in-plane rotations of the foundation plane and the two upper diaphragms, at least two points were monitored at each floor.
The following naming convention was adopted to identify each device: ACC_ID_Floor_Position_Direction. For example, the accelerometer placed on the third floor, at the South-East corner, oriented along the Y direction, is labelled: ACC11_3F_SE_Y.
To measure the displacements of the structure’s floors, three external reference frames rigidly connected to the laboratory floor were installed, outside the shaking table, aligned with it along the two horizontal directions, X and Y (two on the east side and one on the north side). The displacements of the prototype measured with respect to these frames are referred to as global displacements. Wire potentiometers and optical transducers, which are well suited for measuring displacements over distances up to 250mm, were used for this purpose. Structural response parameters (e.g., interstory drifts) can be derived based on these absolute measurements.
The following naming convention was adopted to identify each device: DISP_ID_SLAB_Floor_Position_Direction. For example, the displacement at the first floor in the X direction - NE corner is labelled: DISP02_1F_NE_X.
Relative displacements, rotations, and strains, were monitored through LVDTs and strain gauges installed at key structural components. In particular, LVDTs were positioned at the beam-column joints and at column bases to evaluate the rotational demands of the joints, and at the splice of the second story to assess the displacement at the gap opening. Load cells and instrumented bolts were used to monitor the force variation at the PT bars and at the friction devices at the BCJs and CBs. Strain gauges were used to monitor tension/compression or bending response at key structural sections. The Figure below shows the different instrumentation used at column bases, namely LVDTs (green highlight), load cells (yellow), strain gauges (at the back of the disk springs) and instrumented bolts (purple highlight). The instrumented bolts are not part of this dataset.
The following naming convention was adopted to identify each device:
• Load cells: LC_ID_Element_Position. In what concerns the columns, load cells are place the column bases, so that the Position refers to the side of the column in which the load cells are placed. Regarding the beams, the load cells are placed on the outside of the column, so that the Position refers to the beam end that is being monitored. For example, the load cell placed on the West end of the first story beam is named as LC6_BEAM1_W.
• LVDTs: LVDT_ID_Element_Position_Side. For example, the LVDT placed at the top flange, East side of the first story beam is labelled: LVDT6_BEAM1_top_E.
• Straing gauges: please refer to the Figure and Table below for the strain gauges description.
2. Self-centring low-damage column-bases
In this configuration, named as MRF CBs, column bases are kept as self-centring low-damage connections (SC-CBs), while beam-column connections are now low-damage connections (LD-BCJs), i.e. PT bars were untightened. It is worth noting that LD-BCJs are equipped with FDs (only PT bars were untightened).
In the case of the SC-BCJs, the moment-rotation behavior is still the typical flag-shape curve, where M1 is the moment at the gap opening. Conversely, in the case of LD-BCJs, the moment-rotation behavior is a hysteretic rectangular curve where Mslip is the slippage moment of the FDs.
The second configuration of the shake table testing campaign focused on the seismic response of the MRF-CBs configuration under progressively increasing earthquake intensities. As in the previous test series, the program began with dynamic identification (CAT) tests in the X and Z directions to establish the specimen's dynamic properties before seismic loading. These calibration runs were repeated at key points throughout the campaign to monitor changes in stiffness, damping, and overall structural condition resulting from accumulated damage.
The specimen was then subjected to the L’Aquila (GM1) and Northridge (GM2) earthquake records at multiple intensity levels. The seismic inputs were applied primarily as biaxial excitations (XZ), beginning at 15% intensity and increasing through 40%, 66%, and 100% of the target ground motion. Between major loading stages, additional dynamic identification tests were performed to evaluate the evolution of the structural response and quantify any degradation in dynamic characteristics.
Instrumentation
The instrumentation is the same as described for the Full self-centring low-damage system.
3. Low-damage system
This configuration is named MRF LD. In this configuration both column bases and beam-column connections are low-damage but not self-centring, i.e. the PT bars were untightened.
The MRF-LD shake table testing campaign was conducted to investigate the seismic performance and damage progression of the specimen under increasing earthquake intensities. The test sequence incorporated multiple Dynamic Identification (CAT) stages, performed in the X and Z directions, to monitor the evolution of the structure's dynamic properties throughout the experiment.
The seismic loading program consisted of the L’Aquila (GM1) and Northridge (GM2) earthquake records applied at increasing intensity levels. For the L’Aquila record, the specimen was first tested under 15%, 40%, 66%, and 100% intensity levels in the X direction, with Dynamic Identification tests performed between major loading stages. Following the attainment of the full-intensity motion, additional 100% runs were conducted to evaluate the stability of the response and the accumulation of damage under repeated severe seismic excitation. A similar approach was adopted for the Northridge record, where the structure was subjected to biaxial (XZ) excitations at 15%, 40%, 66%, and 100% intensity levels, interspersed with Dynamic Identification tests.
Instrumentation
The instrumentation is the same as described for the Full self-centring low-damage system.
4. Low-damage system - version 2
This configuration is named MRF LD2. Building upon the outcomes of the MRF LD configuration some structural modifications were introduced: (i) removal of PT bars from BCJs (for the MRF LD they had been untightened); (ii) loosening of bolts at the base of gravity columns (ensuring they work as pinned connections); (iii) removal of Teflon from splice and tightening of splice bolts (now working as continuity connection); (iv) controlled tightening of friction plates.
The program began with Dynamic Identification (CAT) tests in the X and Z directions to establish the condition of the specimen prior to the application of strong earthquake motions. The specimen was then subjected to the Northridge (GM2) earthquake record at 100% intensity under biaxial (XZ) excitation, followed by a subsequent Dynamic Identification stage. The testing then continued with the L’Aquila (GM1) record, where two additional runs at 100% intensity were performed to assess the repeatability of the structural response and the progression of damage under repeated severe loading. After another Dynamic Identification stage, the seismic demand was increased beyond the original target level through the application of a 120% intensity L’Aquila motion. This test was intended to evaluate the residual capacity and robustness under extreme seismic conditions. The final Dynamic Identification tests provided a measure of the remaining dynamic characteristics of the specimen after exposure to the maximum loading demands of the experimental program.
Instrumentation
The instrumentation is the same as described for the Full self-centring low-damage system, except for the load cells that were removed with the PT bars.
Project Metadata
Rights
Creative Commons Attribution 4.0 International.
CC BY 4.0
71 sessions
42 downloads
71 views
1 metadata
22 file previews
Feedback
We are always looking to improve the quality of our data and metadata. If you have any feedback or suggestions, please let us know.
Send Feedback