1 Introduction

The CFS sections have a good strength-to-weight ratio and can be produced economically using cold forming techniques. As a result, the use of cold-formed steel (CFS) in buildings has grown in popularity [1,2,3,4,5,6,7,8]. However, they lack the ductility required for good seismic resistance [9]. As bending members, back-to-back built-up CFS section beams are becoming increasingly popular. Intermediate and primary beams in modular buildings, lower and upper caps in struts in steel trusses, and space frames are all examples of applications. The angle sections are prevented from buckling independently in this design by supplying intermediate screws/bolts at separate positions along the length. Screw connections have recently been employed in cold-formed steel elements because they open their holes first, eliminating the necessity for pre-holes. Then their teeth are removed so that they might not be damaged, and the two members are tightly joined. The screw's length is determined by the thickness of the material [10, 11]. Furthermore, screw connections outperform traditional ones in terms of moment capacity and rigidity [12, 13]. While the behavior of CFS bolted beam-to-column connections is well-documented, the behavior of CFS screwed beam-to-column connections is still required to be explored in order to improve the practical application of cold-formed steel structures. The behavior of built-up cold-formed steel studs was investigated in an experimental study by Stone and LaBoube [14] to compare with the current design code of the North American Specification for Design of Cold-Formed Steel Structural Members [15]. In this investigation, the built-up studs were made up of two C-sections oriented back-to-back to form an I-shaped cross-section. As a consequence of the analysis, the present design standards were determined to be conservative in estimating the ultimate capacity of built-up studs. Zhang and Young [16,17,18] investigated the behavior of cold-formed steel built-up open section columns with edge and web stiffeners. These works used a non-linear finite element analysis and experimental investigations with various parameters. The purpose of the parametric study was to see how edge and web stiffeners affected the built-up open sections. The design predictions generated from the existing design rules in the North American Specification [15] and the Australian/New Zealand Standard [19] were compared to the finite element and test findings. The modified direct strength approach predicts design strengths that are generally in good agreement with the ultimate loads of built-up open section columns. In addition, reliability analysis was used to assess the current design guidelines and the improved direct strength approach.

Fang et al [20] provides a deep-learning architecture, namely a Deep Belief Network (DBN) to evaluate the web crippling performance of cold-formed stainless steel channel sections with centered and offset web holes under the end-two-flange loading condition. Comparing the DBN predictions to an additional 53 experimental outcomes reported in the scientific literature revealed that they were conservative by around 10%. Dai et al [21] proposed a unique machine learning algorithm, eXtreme Gradient Boosting (XGBoost) to evaluate the moment capacity of cold-formed steel (CFS) channel beams with edge-stiffened web holes due to bending. Fang et al [22] introduces a unique machine-learning model utilizing the eXtreme Gradient Boosting (XGBoost) tool for evaluating the web crippling behavior of perforated roll-formed aluminum alloy (RFA) unlipped channels (both fastened and loosened) under interior-two-flange loading. In this study, it is demonstrated that the existing design guidelines cannot accurately estimate the web-crushing strength of perforated RFA unlipped channels. As a result of the parametric study, new formulas for web crippling strength and a web crippling strength reduction factor were proposed for perforated RFA unlipped channels. Fang et al [23] investigated the end-two-flange (ETF) web crippling behavior of RFA unlipped channels with web holes utilizing experimental tests, numerical simulations, and a deep learning model known as a deep belief network (DBN). They showed that the present design standards cannot accurately estimate the web crippling strength of RFA unlipped channels with web holes, as demonstrated. As a result of the parametric study, new web crippling strength and web crippling strength reduction factor equations were developed for perforated RFA unlipped channels. In another study of Fang et al [24], they provide a deep-learning model known as a "deep belief network (DBN)" to examine the end-two-flange (ETF) web crippling behavior of roll-formed aluminum alloy lipped channels (both loosened and fastened to the supports) with center and offset web circular holes. For training the DBN, a total of 1080 data points were created using an elastoplastic finite element model, which was confirmed by 30 new experimental results provided in this study. Comparing the DBN forecasts to the experimental data revealed that they were, on average, 5% more conservative. Chi et al [25] presented 73 new results, including 29 compression tests and 44 finite element analyses (FEA) on the axial capacity of such back-to-back CFS channels. The results indicate that the axial capacity of back-to-back channels with seven edge-stiffened holes is 19.2% higher than that of plain channels without web holes. Roy et al [26] give a numerical investigation of the behavior of back-to-back built-up cold-formed stainless steel channel section columns without lipped ends. Consideration was given to three distinct types of stainless steel: duplex EN1.4462, ferritic EN1.4003, and austenitic EN1.4404. The influence of screw spacing on the axial strength of such built-up, unlipped channels was examined. The majority of short and intermediate columns failed due to local-global or local-distortion buckling interactions, while the majority of long columns failed due to global buckling.

On the other hand, Ting et al [27] presented the findings of 30 experimental tests on back-to-back built-up cold-formed steel channel sections ranging from stub columns to slender columns. After that, a FEM is described, which agrees well with the experimental test results. The FEM is then employed in a parametric analysis that includes 144 models. While the modified slenderness technique is generally conservative, it is found to be unconservative by roughly 10% for stub columns. Other works include that of Biggs et al [28], 71 built-up members were tested experimentally. Multiple lengths, intermediate welds, orientations, and thicknesses are among the test specimen variations. Reyes and Guzmán [29] discovered that the modified slenderness ratio could be utilized in place of the actual slenderness ratio for calculating the ultimate load capacity of built-up box sections with steel gauges of 1.5 mm and 2.0 mm when the seam weld spacing is less than or equal to 600 mm. Chen et al [30,31,32] reported a number of studies on the buckling capacity of back-to-back built-up CFS channels with edge-stiffened holes in perforated CFS channels. It should be noted, however, that all of these sections are distinct from the built-up section explored in this work. Wong and Chung [33] also tested the bolted beam-to-column connections with gusset plates under lateral loads, demonstrating that the suggested connection could effectively transfer the moment between the jointed sections. Anwer et al [34] investigated the strength and stiffness of bolted beam-to-column connections in various configurations. Through experiments, they discovered two types of failure in sections: bearing failure (a ductile mode) and flexural failure (a rapid collapse mode). In addition, several works in the literature have investigated the rotation capacity of CFS beams both experimentally [35] and numerically [35, 36] under the assumption of rigid support conditions for the column. Bagheri Sabbagh et al [37] examined the connections between various beam sections and different out-of-plane stiffener arrangements. Due to higher moment resistance, ductility, and energy dissipation, stiffeners have improved connection behavior once web buckling has been identified as the primary cause of failure. Due to non-existing of failure of bolts in tests, using screws instead of bolts in cold-formed steel structures can be cost-effective. This paper investigates the effect of beam/gusset plate thicknesses on the moment–rotation capacity of cold-formed steel screwed back-to-back sections without stiffeners in the beams. The ratio of beam thickness to gusset thickness was chosen to be between 0.25 and 1.25. The screwed beam-to-column connections in cold-formed steel structures with varied element dimensions were examined and compared for moment rotation \(\left(M-\theta \right)\) behavior, moment–rotation characteristic values, and failure modes. To make a relevant comparison, different beam thicknesses and gusset plate thicknesses were used. A total of 12 connections have been tested based on these parameters. Two of them have already been completed previously, and the results have been published in a paper [38].

2 Experimental investigation

2.1 Test specimens

Figure 1 illustrates the details of back-to-back built-up CFS sections used in the experimental program, which began with varied beam thicknesses and gusset plate stiffener thicknesses and progressed to screwed connections. The details of back-to-back built-up CFS sections were designed by the chapter J of the (AS/NZS) [19] specification. This study looks at 12 CFS examples with various beam and gusset plate thicknesses, as well as a screw diameter of 8 mm and a grade of 8.8. A total of 13 screws were used in the test specimens, which included nine screws at the connection point and four screws in the beam. All experiments were carried out on a one-to-one scale. The ratio of beam thickness to gusset was chosen to be between 0.25 and 1.25. Table 1 lists the specimens and their characteristics.

Figure 1
figure 1

The configuration of beam-to-column connection (geometrical property of the models: (a) beams models, (b) end-plate, (c) gusset plate and (d) beams details).

Full size image
Table 1 Properties of the test.
Full size table

As shown in table 1, the label No-tpA-tB can be explained as follows:

  • The No is labelled as test number.

  • The gusset plate thickness is labelled as "tp".

  • The A is the thickness of the gusset plate.

  • The beam thickness is labelled as "t".

  • The column thickness is labelled as "tc".

  • The B is the thickness of the beam.

2.2 Mechanical properties

The coupon tension test was conducted to determine the test specimens' material properties. A total of three coupons were set in accordance with UNE-EN 10002-1 [9]. Table 2 shows the typical characteristic values for cold-formed steel sheets and screws (8.8). According to ISO 898-1999 (E) [10], each screw (8.8) was tested under tension to determine the mechanical properties of the screw material.

Table 2 Average characteristic values for cold formed steel sheets and screw (8.8).
Full size table

The stress-strain curves for each thickness of the sections have been presented in figure 2.

Figure 2
figure 2

Stress–strain relationship for each thickness of the sections considered.

Full size image

2.3 Test set-up and loading procedure

The moment–rotation curve of cold-formed steel screwed back-to-back sections of beam-to-column connections was studied using experimental models under static loading. The test setup is shown in figure 3. The descriptions in the test setup section are separated into two sections:

Figure 3
figure 3

Test set-up.

Full size image

2.3.1 Experimental tools

The experimental instruments utilized in this study are similar [39,40,41]. The maximum load machine had a maximum stroke of 300 mm and a continuous speed of 0.016 mm/s until the specimens collapsed with 250-kN loading. The pump's slowest speed is 0.016 mm/s. Deformations under stress are measured using displacement transducers. To measure the relevant displacement, three linear variable displacement transducers (DT) with a maximum displacement of 300 mm are utilized. The data is gathered via a data-logging device, which records all measurements from the load cells at one-second intervals.

2.3.2 Experimental procedure

To conduct realistic scale experiments, the frame dimensions in this investigation were determined on a 1/1 scale. The beam and column lengths were chosen to be 1500 mm. A two-column guidance device was supplied for the beam to prevent lateral-torsional buckling while loading (figure 3). However, in the experiments, lateral-torsional buckling of the beam did not occur in general. As a result, the goal of this investigation was to obtain a realistic stress pattern at the connection, and the fracture of the various specimens, i.e., ultimate loads, was achieved using a particular testing machine.

2.4 Conversion of experimental data to desired data by formulas

This experimental investigation examines the moment–rotation behavior. In the study, cantilever tests are used to design the connections. In the cantilever test, the constant stiffness of the connection is determined by the relationship between the moment and rotation of a beam can be written as follows [42,43,44,45,46,47]:

$$k=\frac{M}{\theta }$$
(1)

where moment \(\left(M\right)\) is:

$$M=P{L}_{load}$$
(2)

\(k\) is also given by deflection of the free end of the cantilever, \(\delta \) in the following expression [42]:

$$k=\frac{1}{\frac{\delta }{P{{L}_{b}}^{2}}-\frac{{L}_{c}}{16E{I}_{c}}-\frac{{L}_{b}}{3E{I}_{b}}}$$
(3)

where \({L}_{b}\) and \({L}_{c}\) are the length of the beam and column, \(E\) is the modulus of elasticity, \({I}_{b}\) ve \({I}_{c}\) are the moments of inertia of the beam and column, respectively.

The rotational deformation of the connection (\(\theta \)) was obtained via the following equation which is approximately given by [48]:

$$\theta =arctan\frac{(\delta {DT}_{2}-\delta {DT}_{1}-{\delta b}_{el}\left({DT}_{2}\right)}{{L}_{1}}.$$
(4)

3 Eurocode 3

Eurocode 3 Part 1.8 [49] was used to calculate moment resistance, stiffness, and rotation, as well as to produce graphics [38]. Moment–rotation characteristics are shown in figure 4. Additionally, the characteristic values are the plastic flexural resistance (\({M}_{j.Rd}\)), rotation (\({\theta }_{M.Rd}\)) corresponding to \({M}_{j.Rd}\), the initial stiffness (\({S}_{j.ini}\)), the post-limit stiffness (\({S}_{j.{p}^{-1}}\)), the maximum bending moment (\({M}_{j.max}\)), rotation (\({\theta }_{M.j.max}\)) corresponding to \({M}_{j.max}\), the bending moment capacity (\({M}_{\theta .Cd}\)), rotation (\({\theta }_{Cd}\)) relating to \({M}_{\theta .Cd}\), the knee-range of the \(M-\theta \) curve, which is expressed as the transition zone between the post-limit and initial stiffness.

Figure 4
figure 4

Moment–rotation curve characteristics.

Full size image

The ductility of a joint \(\left({\Psi }_{j}\right)\) is a feature that represents the length of the moment–rotation response yield plateau. The difference between the rotation value corresponding to the joint plastic resistance \(\left({\theta }_{M.Rd}\right)\) and the total rotation capacity \(\left({\theta }_{Cd}\right)\) is the proposed definition of joint ductility. Thus, a joint's ductility is defined as the relationship between the maximum rotation of the joint and the rotation value corresponding to the joint's plastic flexural resistance \(\left({\theta }_{M.Rd}\right)\) [44]:

$${\Psi }_{j}=\frac{{\theta }_{Cd}}{{\theta }_{M.Rd}}$$
(5)

Furthermore, the rotation values at the maximum load and corresponding ductility levels, \({\Psi }_{j.maxload}\), can be derived as follows:

$${\Psi }_{j.maxload}=\frac{{\theta }_{M.j.max}}{{\theta }_{M.Rd}}$$
(6)

The conclusions about joint ductility and rotation values at maximum load and corresponding ductility levels will be explained in the Results section. Eurocode 3 [49] specifies quantitative principles for predicting the joint flexural plastic resistance and initial rotational stiffness of end-plate joints. These structural parameters are analyzed below using geometric and mechanical nominal properties mentioned in Eurocode 3.

4 Experimental results

The moment–rotation characteristic determined from the 12 Cold-Formed Steel Screwed column-beam connections experiments are drawn and reported in figures 5, 7 and 10 and table 3.

Figure 5
figure 5

Moment–rotation curves for the comparison of specimens with 2 mm of gusset plate thickness.

Full size image
Table 3 Moment–rotation characteristic properties for all specimens.
Full size table

4.1 Comparison of beam thickness

Numerical data from the moment–rotation curves in figures 5, 7 and 10 and table 3 are examined. For the specimens with 2 mm gusset plate thickness, increasing the beam thicknesses from 1 to 2.5 mm raised the \({M}_{j.Rd}\), \({M}_{j.max}\) and \({M}_{\theta .Cd}\) values by 57.03%, 55%, and 48.49%, respectively, while those values increased by 59.83%, 68.21%, and 71.54% for the specimens with 3 mm gusset plate thickness. Moreover, \({M}_{j.Rd}\), \({M}_{j.max}\) and \({M}_{\theta .Cd}\) values rose by 39.39%, 48.28%, and 74.01% for the specimens with 4 mm gusset plate thickness when the beam thickness increased from 1 to 2.5 mm. Thus, beam thickness directly increases moment resistance values.

Ideal value for the \({M}_{j.max}{/M}_{j.Rd}\) ratio is between 1 and 1.3. Table 3 demonstrates that the cold-formed steel screwed column-beam connections are not cost-effective since the \({M}_{j.max}{/M}_{j.Rd}\) ratio increased with increasing gusset thicknesses. Additionally, the cold-formed steel screwed column-beam connections design is the most economical when the \(t/{t}_{p}\) ratio is greater than 1. With increased beam thicknesses of 1–2.5 mm in the test specimens with 2 mm gusset plate thicknesses, the stiffness values went up by about 89.97%, the stiffness values dropped by 89.06% and 834.98%, respectively for the specimens with the gusset plate thickness of 3 and 4 mm. Hence, the stiffness values fell as the thickness of the gusset plates grew from 2 to 4 mm. Alternatively, the stiffness value is calculated by the \({M}_{j.max}{/M}_{j.Rd}\) ratio if it is less than 1.3. As the beam's thickness increased, so did its stiffness value. When the thickness of the gusset plates increased, the rotational capacity decreased. Furthermore, when \(t/{t}_{p}\) exceeds 1, the rotation capacity is greater than when \(t/{t}_{p}\) is less than 1. The rotation capacity is determined by the \(t/{t}_{p}\) ratio. Moreover, the rotational capacity demonstrates that all models exhibit semirigid behavior. The \({\Psi }_{j}\) values for the gusset plate thicknesses of 2 and 3 mm fell by 144.64% and 156%, respectively, with increased beam thicknesses of 1 to 2.5 mm, whereas the \({\Psi }_{j}\) values for the test specimens with 4 mm gusset plate thicknesses increased by 8.79% with the same increase in beam thicknesses. Hence, the \({\Psi }_{j}\) values depend on the gusset plate thicknesses. The \({\Psi }_{j.maxload}\) values for the gusset plate thickness of 2 and 3 mm increased about 49.02%, and 4.94%, respectively, with increased beam thicknesses of 1–2.5 mm whereas the \({\Psi }_{j.maxload}\) values decreased about 38% with the same increase in beam thicknesses in the test’s specimens with 4 mm gusset plate thicknesses. Therefore, the \({\Psi }_{j.maxload}\) values depend on the gusset plate thicknesses. The energy dissipation values for the gusset plate thickness of 2, 3 and 4 mm increased about 53.20%, 23.97% and 65.75%, respectively, with increased beam thicknesses of 1–2.5 mm. As a result, the energy dissipation value rises in direct proportion to the thickness of the beam and gusset plates. Furthermore, when the \(t/{t}_{p}\) ratio is greater than 1, the energy dissipation capacity of the specimen is greater than when the \(t/{t}_{p}\) ratio is less than 1.

The failure modes of the specimens depicted in figures 6, 8 and 11 are analyzed based on their photographs. In experiments 3 and 10, the beam undergoes torsion. In this circumstance, the bending moment and rotational behavior of these experiments were difficult to observe. However, it was still analyzed due to failure modes that occurred at the connection's endpoint.

Figure 6
figure 6

Failure modes for the comparison of 1-2-3-4 tests.

Full size image

The moment–rotation curve behavior and failure modes for tests 1, 2, 3, and 4, where the gusset plate thickness is 2 mm and the beam thickness is 1, 1.5, 2, and 2.5 mm, were compared in figures 5 and 6. It was determined that only when the beam thickness was 1.5 mm or below did stiffness and moment capacity rise as the beam thickness grew. Experiments on test specimens with a thickness of 2 mm for the gusset plate have shown that increasing the thickness of the beam profile increases its strength values. It was discovered that stiffness values increase proportionally with profile thickness. When comparing the first and fourth experiments, the increase in profile thickness from 1 to 2.5 mm produced higher stiffness values.

Figures 7 and 8 show the moment–rotation curve behavior and failure modes for tests 5, 6, 7, and 8 with 3 mm gusset plate thickness and 1, 1.5, 2, and 2.5 mm beam thicknesses. The beam with a thickness of 1 mm deformed at the point where the gusset plate of the beam body ends, while the beam with a thickness greater than 1 mm broke into screw components.

Figure 7
figure 7

Moment–rotation curves for the comparison of specimens with 3 mm of gusset plate thickness.

Full size image
Figure 8
figure 8

Failure modes for the comparison of 5-6-7-8 tests.

Full size image

Experiments 6, 7, and 8 demonstrate the same moment–rotation trend, but experiment number 5 demonstrates a decrease in moment capacity and an increase in rotation capacity. In general, as beam thickness has increased, stiffness moment capacity and energy absorption capacity of joints have increased. However, experiment number 7 gave exceptional values since screw breakage was seen after buckling the beam (figure 9). Experimental studies for the test specimens of which gusset plate thickness is 2 mm have shown us that the increase in beam profile thickness also increases the strength values. The stiffness values were found to increase proportionally with increased profile thicknesses. As a result, when comparing the first and fourth experiments, the increase in profile thicknesses from 1 to 2.5 mm increased stiffness values.

Figure 9
figure 9

The buckling behavior of Test 7.

Full size image

Tests 9, 10, 11, and 12 are related to the specimens where the gusset plate thickness is 4 mm, and the beam thickness is 1, 1.5, 2, and 2.5, respectively. When the moment–rotation curve behavior and failure modes seen in figures 10 and 11 are compared, the beams with their thicknesses of 1.5 mm and less have been deformed from the point where the gusset plate ends at the end of the body of the beams while deformation is obtained as screws breakage in the beams with the beam thickness of more than 1.5 mm. In the experiments, as the beam thickness increased, the joint's moment capacity and energy absorption capacity increased, and the rotation capacity decreased. All experiments showed similar moment–rotation behavior. Experimental studies for the test specimens of which gusset plate thickness is 2 mm have shown us that the increase in beam profile thickness also increases the strength values. The stiffness values were found to increase proportionally with increased profile thicknesses.

Figure 10
figure 10

Moment–rotation curves for the comparison of specimens with 4 mm of gusset plate thickness.

Full size image
Figure 11
figure 11

Failure modes for the comparison of 9-10-11-12 tests.

Full size image

4.2 Comparison of gusset plate thickness

Figures 12 and 13 show the comparison of gusset plate thicknesses of the all-specimens tests. Figures 12 and 13 and table 3 show that the \({M}_{j.Rd}\), \({M}_{j.max}\) and \({M}_{\theta .Cd}\) values for the 1, 1.5, 2, and 2.5 mm beam thicknesses decreased with increased gusset plate thicknesses of 2–4 mm in the test’s specimens. Also, the energy dissipation capacity for the 1, 1.5, 2, and 2.5 mm beam thicknesses decreased with increased gusset plate thicknesses of 2–4 mm in the test’s specimens. Moreover, the moment resistance and energy dissipation capacity values increased when the \(t/{t}_{p}\) ratio more than 0.625. Therefore, the gusset plate thicknesses and \(t/{t}_{p}\) ratio affect the moment capacity and energy dissipation capacity in the all-specimens tests. If the failure modes, presented in Figs. 6, 9 and 11, are examined, the thickness of the gusset plate used in the experiments did not affect the failure shapes, as the beam thickness was kept constant at 1, 1.5, 2, 2.5 mm. When comparing the moment–rotation curves seen in Figs. 12 and 13, as the gusset plate thickness increases, screw breaks occur earlier, and their rotational and energy absorption capacities decrease. There was no significant impact on moment capacity.

Figure 12
figure 12

Moment–rotation curves for the comparison of specimens with 1 and 1.5 mm of beam thickness.

Full size image
Figure 13
figure 13

Moment–rotation curves for the comparison of specimens with 2 and 2.5 mm of beam thickness.

Full size image

4.3 Discussion

As a result, when the change of beam thickness and gusset plate was examined, the following results were obtained. In some of the experiments, there was little deformation in the gusset plate, and no failure occurred due to the rupture or excessive deformation of the gusset plate in any experiment. This shows that gusset plate thickness can be made thinner to provide a sufficient economical way. In experiments with beam thickness of 2 or 2.5 mm, since there is always a failure of screw breakage, it is seen that the thickness of the beam is excessive for this screw diameter. The beam can be made economically by using less beam thickness. In specimens where the beam thickness is low (generally 1.5 mm and less), deformation was observed in the beam. Moreover, when failure modes are examined, the failure of the beam occurs at the point where the gusset plate ends in all these experiments. Therefore, the use of stiffener at these points might provide delaying the failure of the beam and increasing the moment capacities.

Table 3 shows that the maximum bending moment when the experimental has 2, 3, and 4mm gusset plates thicknesses increased about 55%, 71.15%, and 48.82%, respectively, by increased thickness beam of 1–2.5 mm. also, that the maximum bending moment decreased, by increased thickness gusset plate of 2–4 mm. Moreover, \({M}_{j.max}{/M}_{j.Rd}\) value needs to be between 1 and 1.3 and the most ideal specimens are in this region to be useful and valuable. Figure 14 shows that when the \({t}_{p}\) less than 2 mm, the specimens test is economical model, and the \({M}_{j.max}{/M}_{j.Rd}\) value directly depends on the gusset thickness and beam thickness. In other words, when the gusset thickness increases and the beam thickness does not exceed 1 mm, the specimens become economical specimens.

Figure 14
figure 14

\({M}_{j.max}{/M}_{j.Rd}\) values all specimens’ tests.

Full size image

Table 3 also presents that increasing thickness beams of 1–2.5 mm increased energy dissipation by 53.20%, 23.96%, and 65.75%, respectively, when the experimental had 2, 3, and 4 mm gusset plates thicknesses. By increasing the thickness of the gusset plate from 2 to 4 mm, the maximum bending moment was reduced.

The ductility of a joint (\({\Psi }_{j}\)) and the rotation values at the maximum load and corresponding ductility levels \({\Psi }_{j.maxload}\), behaved opposite to each other. One side showed an increase while the other showed a decrease. In general, when the \(t/{t}_{p}\) value is greater than 1, the specimen becomes an economical ideal.

5 Results

The experimental behavior of CFS back-to-back sections beam-column connections attached with self-drilling screws with different beam thicknesses and gusset plate thicknesses is reported in this work conducting 12 tests. Experimental testing with a cantilever is carried out with full-scale specimens of the connections to see how gusset plate thickness and beam thickness affect connection behavior. Moment–rotation curves are used to simulate the behavior of connections. The geometric properties of these moment–rotation curvatures change depending on the thickness of the gusset plate and the thickness of the chosen connection type's beam. Moment–rotation curves and the actual behavior of the connections were investigated using resistance, stiffness, energy dissipation and ductility. The main conclusions can be summarized as follows:

  • The moment resistance increases in direct proportion to the thickness of the beam.

  • The cold-formed steel screwed column-beam connections design is the most cost-effective when the \(t/{t}_{p}\) ratio is greater than 1.

  • Increased gusset plate thicknesses of 2–4 mm reduced stiffness values. On the other hand, the stiffness value is determined by the \({M}_{j.max}{/M}_{j.Rd}\) ratio, that is, if the \({M}_{j.max}{/M}_{j.Rd}\) ratio is smaller than 1.3. As the beam thickness was increased, the stiffness value also increased.

  • When the \(t/{t}_{p}\) ratio is greater than 1, the rotation capacity and energy dissipation capacity values are higher than when the \(t/{t}_{p}\) ratio is less than one. The \(t/{t}_{p}\) ratio determines the rotation capacity. In addition, the rotation capacity indicates that all models are semi-rigid.

  • The ductility of the joint is determined by the thickness of the gusset plates.

  • The energy dissipation value increases in direct proportion to the thickness of the beam and gusset plates.

  • There was little deformation in the gusset plate in some of the experiments, and no failure occurred owing to gusset plate rupture or excessive deformation in any of the experiments. This demonstrates that gusset plate thickness can be reduced in a cost-effective manner.

  • As there is always a failure of screw breaking in experiments with a beam thickness of 2 or 2.5 mm, it is determined that the beam thickness is excessive for this screw diameter. The beam can be created more cost-effectively by reducing the thickness of the beam.

  • Deformation in the beam was seen in specimens with a low beam thickness (usually 1.5 mm or less). Furthermore, the failure of the beam occurs at the point where the gusset plate ends in all of these experiments when failure modes are investigated. As a result, using stiffeners at these locations may help to delay beam failure while also improving moment capacities.