Experimental and Finite Element Data of T-joints with Circular-Hollow-Section chord members

Abstract

This paper presents a dataset of cyclic tests and FE simulations of T-joints with Circular-Hollow-Section (CHS) chords and passing-through plates. These specimens were designed to be representative of components of beam-to-column joints between CHS columns and passing-through IPE beams, subjected to the cyclic load protocol proposed by AISC 341. Specifically, cyclic displacement histories were applied at the ends of the plates using a hydraulic actuator. The provided dataset includes the displacement histories recorded through a potentiometric transducer during the tests and the reaction forces recorded using the actuator's load cell.

Subsequently, Finite Element (FE) models of the specimens were developed and validated against the experimental results. The FE models accurately replicate the geometrical and mechanical properties of the tested specimens, and the displacement histories experienced by the specimens were applied.

Starting from the validated FE models, a parametric analysis explored the behaviour of a more extensive dataset comprising 44 geometric configurations of the analysed connection. Key geometric parameters influencing the connection's response were varied, including the ratio between plate width and tube diameter (β, ranging between 0.44 and 0.74), the ratio between tube diameter and twice its thickness (γ, ranging between 15.28 and 27.39), and the ratio between plate and tube thicknesses (τ, ranging between 2 and 8.75). For each of the 44 cases, cyclic simulations were performed, adopting the same protocol applied to the tested specimens and elaborating the force-displacement response.

The significance of this dataset lies in its derivation from numerical simulations based on FE models validated against experimental results, making it a reliable resource for researchers aiming to develop mathematical and mechanical models for predicting the cyclic response of T-joints between CHS chord members and passing-through plates.

Specifications Table

Subject	Civil and Structural Engineering
Specific subject area	Experimental testing and numerical simulations investigating the cyclic response of T-joints with Circular-Hollow-Section chord and passing-through plate.
Data format	Raw
Type of data	Table
Data collection	The experimental data were gathered during a testing campaign at the University of Salerno. Specifically, three cyclic tests were executed on T-joints with CHS chord and passing-through plate. The displacements and forces were recorded using a transducer and a load cell attached to a vertical actuator with a maximum load capacity of 2000 kN in tension and 3000 kN in compression, featuring a stroke of ±75 mm. Finite Element models of the tested specimens were developed thanks to the software Abaqus and validated against the experimental responses. The validated models allowed to perform cyclic simulations on 44 distinct geometrical configurations of CHS-to-plate connections.
Data source location	Materials and Structures Laboratory. Department of Civil Engineering, University of Salerno, via Giovanni Paolo II, 132, Fisciano, 84084, Italy (40°46′13.3"N 14°47′22.9"E)
Data accessibility	Repository name: Mendeley Data Data identification number: 10.17632/bg7gcgy9wk.1 Direct URL to data: https://data.mendeley.com/datasets/bg7gcgy9wk/1

1. Value of the Data

• •Recent research efforts to employ hollow section members in steel structures have been carried out considering the significant advantages these profiles offer compared to traditional double-tee solutions [1], [2], [3]. These advantages include: (i) enhanced aesthetic appeal; (ii) high values of the radius of gyration and the absence of a weak axis; (iii) reduced costs for paintings, fire and corrosion protection; (iv) lower drag wind coefficients. However, the primary limitation hindering their use is the complexity of manufacturing beam-to-column connections. Nevertheless, the recent application of 3D Laser Cutting Technology (3D-LCT) in Civil Engineering [4], [5], [6], [7] has allowed the fabrication of welded connections by precisely cutting the tubular profile with the imprint of the cross-section shape of the double-tee beam or plate.
• •The Component Method (CM) is an advanced strategy to predict the stiffness and strength of steel beam-to-column joints [8], [9], [10], [11]. This approach requires identifying and characterising individual nodal components in terms of strength and stiffness to model the connection's monotonic moment-rotation response. The practical application of the CM is ruled by Eurocode 3 Part 1.8 provision [12], which provides formulas to estimate the strength and stiffness of key nodal components of typical welded and bolted steel beam-to-column joints. However, joints may experience loading and unloading cycles during seismic events, fostering nodal components to exhibit hysteresis loops. This introduces a limitation to the applicability of the CM, restricting its use to specific loading conditions and posing significant research challenges for extending its application to other scenarios, such as those encountered in seismic or robustness assessments.
• •Recent studies have been conducted to extend the range of application of the CM also to seismic loading conditions. In this framework, Oliveira et al. [13,14] emphasised that, although various components influence joint behaviour, only some are essential for capturing the connection's dissipative characteristics. Consequently, examining joint hysteretic behaviour can be streamlined by concentrating on these critical components employing numerical modelling through appropriate multi-parameter mathematical expressions. In this framework, Sica et al. [15] have focused attention on the cyclic behaviour of CHS to passing-through plate connections, which can be regarded as a standalone connection or a component of CHS to passing-through IPE beam-to-column joints [7]. Specifically, experimental tests and numerical simulations have been conducted on CHS to axially loaded passing-through plate connections under cyclic conditions. The results of this research are collected in [16].
• •The Finite Element (FE) analyses evaluated the force-displacement response of simulated T-joints, specifically focusing on assessing their initial stiffness, yield strength and damage. As highlighted in Sica et al. [15], these data were employed to formulate mathematical expressions for predicting the cyclic response of the analysed joint.
• •The data discussed in [16] can be valuable for researchers and practitioners in the field of steel structures and beam-to-column joints, particularly for those focused on enhancing existing guidelines for predicting the flexural behaviour of connections involving tubular columns. In fact, given the inclusion of results from experimental tests and Finite Element simulations, researchers can verify the provided data and potentially expand the scope of the investigated configurations in the parametric analysis. Furthermore, relying on the knowledge about the cyclic responses of different cases, researchers can develop mathematical expressions for implementation in design guidelines and commonly used structural software.

2. Background

The motivation behind creating this dataset stems from the observation that current research efforts predominantly focus on nodal components using the component method approach, aiming to provide only a monotonic characterisation of individual nodal components. Conversely, there has been a notable lack of studies addressing the cyclic investigation of these components. In this context, the present study seeks to present a robust dataset of simulations focusing on connections between Circular-Hollow-Section (CHS) and passing-through plates. These connections can be viewed as standalone components or integral parts of CHS to passing-through IPE beam-to-column joints. By doing so, researchers interested in this field can conduct cyclic analyses of such components, offering mechanical models and mathematical expressions to predict their cyclic responses. This approach aligns with the work by Sica et al. [15], wherein the same dataset was exploited to characterise the behaviour of the examined component mathematically, according to the hysteretic uniaxial material property of the OpenSees library [17].

3. Data Description

The dataset is collected in a Microsoft Excel file in .xlsx format, featuring four worksheets (Table 1).

Table 1.

Organisation of the Excel file.

Worksheet	Title	Content
1	Geometry_tested_specimens	Geometrical properties of the tested specimens
2	Material_properties_specimens	Mechanical properties of the tested specimens
3	Experiments_and_FE_validation	Validation of the implemented FE model
4	Parametric_analysis	FE simulations on additional 44 joints

The first worksheet details the geometry of the three CHS-to-plate specimens that experienced cyclic experimental tests at the laboratory of the University of Salerno. In particular, the worksheet includes a table with the dimensions of the tubular profiles and plates composing the specimens, supplemented by an image that provides a clearer understanding of the parameters to which the given values pertain (Fig. 1).

Fig. 1.

Worksheet 1: geometry of the tested specimens.

The second worksheet presents the outcomes of the tensile coupon tests conducted to characterise the material properties of the three specimens under examination according to UNI EN ISO 6892-1:2020 recommendations [18]. To be precise, three coupon tests were performed on dog-bones extracted from each tube belonging to the three specimens. This worksheet details the engineering stress-engineering strain and true stress-true strain data for each of the nine tests, along with the geometry specifications of the tested coupons (Fig. 2).

Fig. 2.

Worksheet 2: material properties evaluated through tensile coupon tests according to [18] (left) and Schenck Hydropuls S56 Universal Machine (right).

The “Experiments_and_FE_validation” worksheet collects the results of the cyclic tests on the abovementioned specimens and the numerical outcomes of the implemented FE models in Abaqus [19]. Specifically, this worksheet provides the displacements and forces for each test/simulation (Fig. 3).

Fig. 3.

Worksheet 3: results of the experimental tests and numerical simulations of the three specimens.

The last worksheet includes a table containing information about the 44 cases examined in the parametric analysis. Additionally, it presents the force-displacement curves for each case obtained by the FE simulations (Fig. 4). For clarity, it is crucial to highlight that the force-displacement curves presented have been properly modified to eliminate the elastic deformations of the plates and tubes.

Fig. 4.

Worksheet 4: parametric analysis.

The Finite Element (FE) analyses evaluated the force-displacement response of simulated T-joints, with a specific focus on assessing their initial stiffness, yield strength and cyclic damage. As highlighted in Sica et al. [15], these data were employed to formulate mathematical expressions for predicting the cyclic response of the analysed joint.

4. Experimental Design, Materials and Methods

The experimental tests on CHS-to-plate connections were conducted on the three specimens by applying cyclic displacement histories at the free ends of the plates. The objective was to replicate the displacements experienced by the flanges of CHS-to-passing-through IPE beam joints when following the protocol outlined in AISC 341-16 [20]. Specifically, a transducer and a load cell were employed to record the applied displacements and the corresponding reaction forces. In Fig. 5, the experimental setup is depicted, wherein the CHS tube is horizontally oriented and welded to end plates bolted to rigid steel supports. These supports are securely fixed to the strong floor. A vertical actuator is employed to load the passing-through plate, and a hinge is strategically positioned between the specimen and the actuator. This configuration ensures that no bending forces are transmitted to the specimen. The vertical actuator has a maximum load capacity of 2000 kN in tension and 3000 kN in compression, featuring a stroke of ± 75 mm. Details of specimen 3 and the supports are shown in Figs. 5 and 6.

Fig. 5.

Details of specimen 3 (dimensions in mm).

Fig. 6.

Details of the supports (dimensions in mm).

The experiments consisted in applying cyclic displacement histories to the passing-through plates. Specifically, the selected cyclic displacement histories align with the testing protocol outlined in AISC 341-16 [20]. It is worth highlighting that no predefined cyclic load protocol is specifically tailored for T-joints. Nevertheless, since the examined T-joints are assumed to be a component of more complex beam-to-column joints (specifically, CHS-to-IPE connections), the displacement history was derived through analogy by transforming the rotation ($\phi$) of the beam-to-column joint into the corresponding displacement ($\delta$) at the tube-to-flange attachment. By employing this analogy, the displacement can be equivalently expressed as $\delta =\phi ·h_1/2$. Considering that, for standard European IPE profiles, the ratio between the height ($h_1$) and the flange width ($b_1$) is 2:1, it is possible to express $\delta$ equivalently as $\phi ·b_1$. With these assumptions, the adopted cyclic displacement history is summarised in Table 2.

Table 2.

Cyclic loading history.

Number of cycles	Amplitude δ (mm)	Velocity (mm/s)
6	0.0075· b1	0.05
6	0.01· b1	0.07
6	0.015· b1	0.10
4	0.02· b1	0.13
2	0.03· b1	0.20
2	0.04· b1	0.27
2	0.06· b1	0.40
2	0.08· b1	0.53
2	0.10· b1	0.67

The tensile coupon tests aimed at characterising the material properties of the tested specimens were carried out using the Schenck Hydropuls S56 Universal Machine (Fig. 2). This machine has a load capacity of ±630 kN and a piston stroke of ±125 mm.

The Schenck Hydropuls machine records the force-displacement curves exhibited by the tested dog-bones. The stress-strain constitutive laws have been collected in the dataset available in [16]. The transition from force-displacement to stress-strain curves was achieved through Eqs. (1)–(4). Specifically, in Eqs. (1)–(4), $\delta$, $L$, $b$, and $t$ denote the elongation, length, width, and thickness of the dog-bone, respectively. $F$ represents the applied force, ${\epsilon }_E$ and ${\epsilon }_T$ denote engineering and true strain while ${\sigma }_E$ and ${\sigma }_T$ represent engineering and true stress, respectively.

$${\epsilon }_E=\frac{\delta }{L}$$ (1)

$${\sigma }_E=\frac{F}{bt}$$ (2)

$${\epsilon }_T=ln\left(1+{\epsilon }_E\right)$$ (3)

$${\sigma }_T={\sigma }_E\left(1+{\epsilon }_E\right)$$ (4)

Finite Element (FE) models of the tested specimens were created in the Abaqus software [19] by faithfully reproducing their geometries and mechanical properties. As illustrated in Fig. 7, the FE models were implemented during the validation phase, including the rigid base and supports. The connection between the plate and the tubular profile was established without explicitly modelling the welds; instead, a tie interaction between the holes and the plates was employed to simulate the welding. All degrees of freedom of the rigid base were constrained. A “Hard” normal response and “Penalty” tangential behaviour with a friction coefficient of 0.3 were specified and applied to all contact surfaces, including those between the basement and the base of the supports and between heads/nuts/shafts and plate surfaces/holes. The materials were modelled using stress-strain laws obtained from the coupon tests. Young's modulus and Poisson's ratio were 210 GPa and 0.3, respectively. The members constituting the investigated specimens were meshed with a size of 5 mm, utilising C3D8-type (8-node linear brick) elements, ensuring that at least two elements were included within the thickness of all components.

Fig. 7.

FE model.

The validated finite element models were used to explore a wider range of geometrical configurations. In this context, 44 combinations of tubular profiles and plates were examined. These configurations were selected based on identifying various beam-to-column joints with passing-through beams. The beams in these joints ranged from IPE240 to IPE500, while the CHS columns had diameters varying between 193.7 mm and 406.4 mm and thickness ranging from 4 to 10 mm. For each case, the following non-dimensional parameters were systematically varied: i) the ratio between the width of the plate and the tube diameter ($\beta =b_1/d_0$); ii) the ratio between the tube diameter and twice its thickness ($\gamma =d_0/\left(2t_0\right)$); iii) the ratio between the thicknesses of the plate and tube ($\tau =t_1/t_0$);. Specifically, $\beta$ ranged from 0.44 to 0.74, $\gamma$ varied between 15.28 and 27.39, and $\tau$ ranged from 2 to 8.75.

All tube-to-plate connection configurations in the examined cases were simulated using the equivalent quadrilinear true-stress true-strain curve suggested by Faella et al. [11], considering a widely used steel grade, specifically S275. The 44 “simulated experiments” analyses were conducted using a static solver, applying cyclic displacement amplitudes as specified in the protocol outlined in Table 2.

It is noteworthy that the force-displacement curves, reflecting the cyclic behaviour of the simulated 44 cases (belonging to the fourth worksheet of the dataset [16]), have been appropriately adjusted by subtracting the elastic contributions associated with the deformability of both plates and tubes from the recorded displacements. This adjustment enabled the analysis to focus only on the plastic behaviour of the connection and was achieved by applying the expression indicated in Eq. (5):

$${\delta }_p={\delta }_{p+e}-\frac{F{L}_1}{E{A}_1}-\frac{F{L}_0^3}{192E{I}_0}-\frac{\chi F}{G{A}_0}\frac{L_0}{2}$$ (5)

In Eq. (5), ${\delta }_p$ and ${\delta }_{p+e}$ represent the displacements with and without adjusting for the elastic deformability of the tube and the plate. $F{L}_1/\left(E{A}_1\right)$ denotes the axial deformability of the plate, where $F$ is the applied force, $L_1$ is the free side of the passing plate, $E$ is the Young's modulus, $A_1$ is the area of the plate. Additionally, $F{L}_0^3/\left(192E{I}_0\right)$ and $\chi F{L}_0/\left(2G{A}_0\right)$ correspond to the flexural and shear deformability of the tube, respectively; $\chi$ is the shear factor for a CHS profile, while $A_0$, $I_0$, and $L_0$ represent the area, inertia, and length of the tube, respectively.

In experimental tests and numerical simulations, the axially loaded passing-through plates transfer the loads to the hollow sections. Nevertheless, as the plates have been designed to prevent buckling, damages are concentrated only on the Circular-Hollow-Section (CHS) profiles. Specifically, in all cases, the collapse of the tested or simulated specimens is attributed to the transverse crushing of the tubes induced by the passing-through plates.

Limitations

None.

Ethics Statement

The authors declare that they have followed the general ethics rules of scientific research performance and publishing. This work did not involve human subjects, animal experiments, or data collected from social media platforms.

CRediT authorship contribution statement

Sabatino Di Benedetto: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Data curation, Writing – original draft, Visualization. Massimo Latour: Conceptualization, Methodology, Formal analysis, Investigation, Data curation, Resources, Writing – review & editing, Supervision. Gianvittorio Rizzano: Conceptualization, Methodology, Resources, Supervision, Project administration, Funding acquisition.

Data Availability

• Dataset of cyclic tests and simulations of T-joints with CHS chord and passing-through plate (Original data) (Mendeley Data)