Study on Bolted and Blind-Bolted Single Shear Connections of Cold-Formed Steel Members

Modern industry increasingly tends towards the so-called industry 4.0, which consists among others on the integration of systems used for design and production. This also applies to steel structures made of cold-formed steel (CFS), where LGS / HGS (Light / Heavy Gauge Steel) systems are more and more popular. These are comprehensive solutions combining the most important elements of the production process: from BIM design, through prefabrication to assembly. Individual structural members can be produced on one compact device, which automatically shapes, cuts, makes holes for fasteners and marks them accordingly. Open sections with various dimensions (most often types C and U) cold-formed from zinc-coated high strength steel sheets with a thicknesses in the range of 1.2 to 2 mm (LGS) and 3 to 6 mm (HGS) can be made on the same machine. These systems are characterized by very high precision and production speed reaching over a dozen running meters of profile per minute.

Cold-formed structural systems are most often designed as frame or lattice structures, taking into account their cooperation with sheeting [1]. Individual members are joined with various types of self-drilling screws, standard bolts or other types of mechanical fasteners. This allows for designing structures with a span of over a dozen meters and a self-weight not exceeding 0.2 kN/m², that are easy to prefabricate, transport and install.

For structures with larger spans, e.g. twenty-four meters (Fig. 1a), it is rational to use Rectangular Hollow Section (RHS) members with wall thicknesses from 3 to 6 mm [2]. To connect hollow section with open section members, the so-called blind fasteners HUCK BOM (Blind, Oversize, Mechanically locked) [3] (see Fig. 1d, e and f) can be sued. The introduction of eccentricity at nodes allows the design of the lattice structure without gusset plates (Fig. 1b and c).

Figure 1.

The main advantages of the BOM fasteners are the possibility of connecting elements with only one side access and tight filling of the hole by the sleeve of the connector during installation. After installation with special hydraulic or pneumatic tool the design diameter of the fastener d is equal to the diameter of the hole d ₀ (cf. Fig. 1d and f), which significantly limits the slip in the connection. According to the producer’s data [3], the BOM R16-4 fastener with a diameter d = 13.6 mm can be installed in a hole with a diameter d ₀ = 13.8÷14.8 mm; thus the permissible tolerance is greater than in the case of fitted holes, where the clearance between the bolt shank and the wall in the hole should not exceed 0.2 mm. The technology of these connections is described in more detail in [4] and [5].

Due to the higher cost of BOM fasteners compared to conventional bolted connections, blind bolts should be used only where it is necessary for technological reasons (cf. Fig. 1b) and, above all, due to the required stiffness of the connection (slip limitation). In other cases, when both sides access is possible (cf. Fig. 1c), ordinary bolted connections can be used. The use of M16 bolts i.a. in the field splices of frame can also significantly facilitate the installation of the structure on site, since the standard bolts holes are enlarged by 2 mm in relation to the bolt diameter.

In the single-shear bolted connections, after application of the load, permanent mutual displacement of connected plates can be observed. It is caused by slip which is related with bolt-hole clearance, plastic deformations of the material in the place of bearing (hole elongation) and bolt tilting. The magnitude of these deformations undertakes to treat these connections definitely as semi-rigid, and the basis for including this phenomenon in static calculations of CFS frames is to determine the physical relationship in the form of load-displacement characteristics [6 to 9].

The behaviour of bolted shear connections is the subject of many studies. Zadanfarrokh and Bryan [6] analysed both experimentally and theoretically the flexibility of connections with M16 bolts in cold-formed steel sections with a thickness of 1.5 to 3.0 mm, and gave a formula for the flexibility of a single lap bolted joint. Lim and Nethercot [7] have performed experimental and numerical tests of similar connections to analyse the behaviour of CFS portal frames. Zaharia and Dubina [8] tested connections with M8 to M16 bolts to extend the formula purposed in [6] for various bolt diameters in order to predict the structural behaviour of CFS trusses. In [9] Dubina summarises i.a. the results of experimental and analytical study on full scale portal frame including the effect of stiffness of bolted-shear joints on the basis of component method in EC3 [10]. Wallace and Shuster [11] developed an investigation to study the behaviour of bearing failure on bolted connections without washers and recommended 25% reduction of bearing strength in the case of single shear joints. In [12] He and Wang focused on developing the method of calculating the deformation capacity of thin-walled plates with a single bolt shear connections. The extensive parametric study on bearing strength were made by Konkong and Phuvoravan [13] to modify the bearing factor for thin G550 and G300 steel sheets. Aziz and Lip [14] studied the effect of bolt thread on stiffness of double-shear connections finding that the threads reduce the initial stiffness but increase the final stiffness. In [12, 15, 16] attention was drawn to the possibility of using the component method in EC3 [10] to assess the initial translational stiffness of bolted shear connections. Some other recent research in the field of bolted shear connections in CFS members are referred in [17], while structural response of CFS frames with such joints is discussed in [18].

The joints with blind fasteners of the type BOM are not included in the provisions of the design standards, so it was required to carry out extensive research, which aimed to determine their load capacity and stiffness. In the paper [19] Wuwer presents a method of calculating any arbitrary lap joint loaded simultaneously with shear force and bending moment, which is based on a non-linear force-displacement relationship. In [20] Swierczyna and Wuwer provide formulas for determining the bearing resistance of a single fastener. Paper [5] by Słowiński and Wuwer describes the method of strengthening the RHS columns. The structural response of a CFS frame with semi-rigid blind-bolt joints were discussed by Swierczyna and Wuwer in [2]. Paper [21] presents the results of tests of connections in alternating load. A summary of previous activities in this area is included in [4].

The load transfer mechanism in both blind fasteners BOM and standard bolts is the same, but their structure varies significantly. The effective use of these connectors require a good understanding of their structural behaviour, thus the research was developed to analyse the behaviour and compare these connectors. The paper presents the experimental study on single-shear bolted connections of cold-formed steel members with a wall thickness of 4 mm using M16 bolts. The results are illustrated in the form of static equilibrium paths on the graphs of the relationship between the load F and the relative displacement δ of the connected walls. The appropriate relationship determined on the basis of the EC3 [10] standard and the proposed formulas enabling the determination of non-linear load-displacement characteristics of the tested joints were illustrated on the graphs. Test results are referred to the previous study [4, 20] of analogous connections with BOM R16-4 fasteners to compare the bearing strength and stiffness of these joints. Furthermore the effect of joints on structural response of exemplary CFS frame were shown.

The results of tests for specimens with M16 bolts are presented on the graph of the relationship between the loading F and the relative displacement δ (Fig. 3), which was calculated as the mean value of the measurements made by IDT1 and IDT2 sensors. On the graph, dashed lines indicates the values of characteristic 2F _b,Rk =2k ₁α _b f _u dt and design 2F _b,Rd =2F _b,Rk /γ _{M 2} bearing resistance calculated according to Table 3.4 in Eurocode 3 [10] for the actual measured connection parameters listed in Table 1, where the bearing factor k ₁α _b = 2.5·0.926 = 2.31 and γ _{M 2} = 1.25.

Figure 3.

The load-displacement characteristics of the tested connections proved to be non-linear over the entire load range. The deformations in individual specimens are characterized by a significant dispersion which is caused, among others, by various values of slip at the initial stage of loading. This phenomena is related mostly with the clearance in standard holes with a diameter of d ₀ = 18 mm. For example, in the M16.4/2-5 specimen, that was deliberately installed in such a way that the connected walls were offset from each other in the opposite direction to the loading direction, the slip was 3.8 mm (cf. Fig. 3). Similar behaviour was also observed in many studies [6, 8, 11].

The behaviour of all specimens was similar. The beginning of clearly visible plastic deformations, which manifested in the tilting of the bolts and local deformation of the connected walls out of the connection plane, was observed at a load approximately equal to 2/3F _b,Rd . In the final stage of the test, when the tilt of the connectors was near 45°, the bolt heads were pulled through strongly deformed holes with simultaneous punching shear of the walls material under the bolt heads (Fig. 4).

Figure 4.

The exception were the specimens M16.4/4-1 and M16.4/4-2 in which washers were used both under the nut and the bolt head (w = 4), where the test was stopped at a load F = 120 kN due to the limited capacity of anchors. In the specimen M16.4/4-3 the shear failure occurred in anchor at load F = 142.5 kN. The shape of the paths of static equilibrium without characteristic plateau indicates the possibility of a different form of failure in those specimens. The increased strength of these connections was associated with better stabilization of the bolts by washers fitted on both sides, which limited the tilting and pulling of the fasteners.

Tables 2 to 4 list the values of the maximum load F _max and displacement δ (F _max ) recorded during the test. The bearing factors calculated as a ratio of F _max /2f _u dt can be compared there with the value of bearing factor k ₁α _b = 2.31 according to [10].

In all connections where two washers were used (w = 2) the value F _max slightly exceeded the characteristic bearing resistance 2F _b,Rk , determined according to EC3 [10]. The maximum loading in the connections where w = 4 proved to be significantly higher than in the other specimens, although the failure loading was not reached during the test. In the connections where no washers were used (w = 0), the bearing capacity was lower by up to 10% compared to the theoretical value (cf. Table 4, specimen M16.4/0-2).

Apart from the specimens in which washers were installed under the nut and bolt head, relative displacements at F _max were on average 16 mm, while the maximum displacement at failure reached about 28 mm, which indicates a high deformation capacity of these connections.

The results of the previous studies on BOM fasteners were summarized in [4] and [20]. A total of fifty-five axially loaded specimens (cf. Fig. 2b) with connections of cold formed sections made of steel with f _u = (340÷540) MPa and thickness t = (3÷5) mm made with BOM R10 and R16 fasteners installed in the holes with a diameter d ₀ = (9÷14.5) mm were included. The behaviour and form of failure of connections with BOM fasteners (Fig. 5) is similar to that observed in the specimens M16.4/2 and M16.4/0 (cf. Fig. 4). The results of selected tests will be discussed in more detail in the next section.

Figure 5.

Methods for predicting the load-displacement characteristic of single-shear bolted connections in the form of a linear relationship were proposed, among others, in works [6] and [8]. Currently, however, the most widely known way to determine the stiffness of joints is the so-called component method proposed in EC3 [10]. A formulas to determine the elastic stiffness coefficients for bolts in shear k ₁₁ and bolts in bearing k ₁₂ have been provided there. They are associated with a specific case of beam-to-column joint with a bolted angle flange-cleats, however, as mentioned in the standard, they are for general application. In works [15] and [16] those formulas are presented in a form suitable to determine the stiffness of a single fastener:

(1a, b) k 11 = 8 d 2 f u b d M 16 and k 12 = 12 k b k 1 d f u ,

where:

k b = min { 0 , 25 e b / d + 0 , 5 0 , 25 p b / d + 0 , 375 1 , 25 } , k t = min { 1 , 5 t j / d M 16 2 , 5 } ,

d – diameter of the bolt,

d _{M 16} – nominal diameter of an M16 bolt,

f _ub – ultimate tensile strength of the bolts,

f _u – ultimate tensile strength of the steel on which the bolt are pressed,

e _b – distance from the bolt-row to the free edge of the plate in the direction of load transfer,

p _b – spacing of the bolt-rows in the direction of load transfer,

t _j – thickness of the elements against which the bolt is pressed; t _j = t’ or t _j = t”, (Fig. 6).

Figure 6.

In the case of single shear connection (Fig. 6), the stiffness is a sum of flexibilities of bolts in shear 1/(m·k ₁₁) and bolts in bearing to wall with a thickness t’: 1/(mk’₁₂) and t”: 1/(mk”₁₂).

The translational stiffness of a joint is directly proportional to the number m of fasteners:

(2) S δ , m = 1 1 m k 12 ′ ′ + 1 m k 11 + 1 m k 12 ′ .

Straight lines corresponding to the stiffness values calculated for M16.4 specimens according to formula (2) and formulas proposed in [6] and [8] are plotted on the graph in Fig. 3. Referring those lines to the test results, it can be stated, that the stiffness determined on the basis of EC3 [10] and [8] represent the initial stiffness, whereas the value calculated according to [6] corresponds to the secant stiffness (cf. Fig. 3). However, taking into account the characteristic dispersion of the test results, it can be assumed that these differences are not significant.

As the tests has shown, at loading about 70% of the design bearing resistance, the cumulative plastic deformations cause the load-deformation characteristic become clearly non-linear. These observations are consistent with the provisions of EC3 [10], where the elastic initial stiffness determined on the basis of the component method should be used for loads not exceeding 2/3 of the design load capacity of the joint. Above this value, secant stiffness or nonlinear relation should be adopted on the basis of appropriate coefficients, but the standard [10] does not provide values suitable for the tested joints.

In [23] it was proposed that the non-linear load-displacement characteristic of bolted shear connections can be described by means of an exponential function. Assuming that the translational stiffness of the connection is directly proportional to the number m of fasteners, the function takes the form:

(3) F = m a F ( 1 − e − b F δ ) .

The important advantages of the above formula are simplicity and the possibility of physical interpretation of its parameters. The value of a _F can be associated with the load causing plasticizing of the connection, while the product of a _F and b _F determines the initial stiffness of the connection.

In the case of bolted connections, the values of the parameters of the exponential function are proposed to be determined as follows:

(4a, b) a F , b = F b , R k = k 1 α b f u d t , b F , b = S δ , 1 / a F , b ,

where S _δ,1 according to eq. (2) for m = 1.

In the following figures, the results of experimental tests for the specimens with M16 bolts are plotted on the graphs of the relationship between the average load of a single fastener with a shear force F ₁ = F/2 and the mutual displacement δ, where they can be compared with the exponential curves.

In cases where there is no significant slip, the exponential function describes well the behaviour of the specimens M16.4/2 with washers fitted only under the bolt nuts. Connections M16.4/4 where washers were used both under the nut and the bolt head have a load capacity much higher than predicted by the formula (4a), thus their behaviour is also described in the safe way. In the case of joints M16.4/0 in which no washers were used, the value of the parameter a _F,b was reduced by 10% due to the lower load capacity of these joints. The value of the reduction factor was estimated on the basis of test results listed in Table 4, where the lowest ratio F _max /(2F _b,Rk ) = 0.9 for specimen M16.4/0-2. It should be noted that in [11] a 25% reduction of resistance for connection without washers was proposed which indicates that the exact prediction of the behaviour of connections without washers may require more extensive research. The lack of washers did not visibly affect the deformation capacity of these joints.

Figure 7.

Figure 8.

Figure 9.

As shown i.a. in [4] the exponential function (3) can be used to describe the behaviour of connections with BOM blind fasteners. Results of the tests were analyzed looking for the relationship between the measured values of ultimate tensile strength f _u of connected steel plates, their thickness t and diameter d = d ₀ of the fasteners and experimentally determined parameters a _F,bb and b _F,bb of exponential curves describing the behaviour of the joints in tested specimens. The dependencies adopted in [4] are illustrated in Fig. 10. The next figures illustrate the results of tests and exponential curves for selected specimens with BOM R16 fasteners. On the F ₁-δ diagrams the straight lines corresponding to stiffness S _δ,1 calculated according to eq. (2) for m = 1 and parameters of BOM fasteners were added. They can be compared with the straight lines representing the initial stiffness calculated as the product of a _F,bb and b _F,bb .

Figure 10.

A series of five BOM-S.4 specimens (cf. Table 1) were tested for displacements not exceeding 8 mm (Fig. 11). Following the ECCS Recommendations [24] the bearing capacity F _{δ ,lim} was taken for the limit displacement criterion δ _lim = 3 mm. However, the load was high enough to observe a visible plastic deformations of the joints.

Figure 11.

A five identical specimens BOM-KSW.4 (cf. Table 1) were loaded until the failure (Fig. 12), which shows that the deformation capacity of the BOM fasteners is comparable to M16 bolt connections. At the final stage of the test, at displacements exceeding 18 mm, a slight increase in load was observed, which is associated with a significant tilting of the fasteners (cf. Fig. 5). This caused that considerable part of the load were transferred by tension in the connectors.

Figure 12.

To compare the bearing capacity of the individual types of fasteners, the ratio of the a _F parameter and the product of f _u , d and t is summarized in Table 5.

It can be assumed that the value of load causing plasticizing in test specimens with M16 bolts with washers under the nut may be about 19% higher compared to the analogous value for connections with BOM fasteners, while in the case of specimens without washers the maximum load can be higher by about 9%. These differences are mainly due to the relatively small diameter of the “head” of BOM fasteners which results in low pull-trough resistance. In the case of the M16 bolt the average of the diameter inscribed and described on its head is 26 mm, while the diameter of the “head” of the BOM R16 fastener is about 21 mm (cf. Fig. 1d). It should also be noted that, according to the producer’s data [3], the sleeve of which the fastener’s “head” is formed is made of mild low carbon steel.

Table 6 summarises the initial stiffness values of connections in tested specimens. In the case of specimens BOM-S.4 the initial stiffness turned out to be about twice as high as for M16.4, wherein both types of specimens were made with steel S235 (cf. Table 1). In the case of BOM-KSW.4, where S355 were used the initial stiffness was almost three times greater.

Comparing the value of stiffness calculated for BOM fasteners as the product of a _F,bb and b _F,bb with the value of S _δ,1 according to formula (2) for m = 1 (cf. Fig. 11 and 12) also in this case the stiffness of connections with blind fasteners turned out to be about twice as high as the value predicted by the formula developed for bolts.

The relatively high initial stiffness of the connections with BOM fasteners results mainly from the lack of slip in the blind bolt joint where the fastener sleeve tightly fills the hole during installation (cf. Fig. 1e). It should also be emphasized that the connections with BOM fasteners are characterized by much smaller dispersion of results in comparison to M16 bolt connections. However, this differences could be probably significantly smaller in the case where M16 bolts with partially threaded shank would be installed in fitted holes or holes enlarged by only 1 mm in relation to the bolt diameter.