Behaviour of Headed Anchor Blind Bolts Embedded in Concrete Filled Circular Hollow Section Column

Transcription

1 Behaviour of Headed Anchor Blind Bolts Ebedded in Concrete Filled Circular Hollow Section Colun Yusak Oktavianus 1, Helen M. Goldsworthy 2, Ead F. Gad 3 1. Corresponding Author. PhD Candidate, Departent of Infrastructure Engineering, The University of Melbourne, Parkville, Victoria 31. Eail: yoktavianus@student.unielb.edu.au 2. Associate Professor, Departent of Infrastructure Engineering, The University of Melbourne, Parkville, Victoria 31. Eail: heleng@unielb.edu.au 3. Professor, Faculty of Science, Engineering & Technology, Swinburne University of Technology, Hawthorn, Victoria Eail: egad@swin.edu.au Abstract Blind bolts allow technicians to connect the steel bea and Concrete Filled Circular Hollow Sections (CFCHS) colun fro the outside of CHS. In addition, using blind bolts as the connection ay avoid the brittle behaviour generated by the welded connections. However, there are currently no guidelines and clear understanding of the behaviour of this connection. This paper presents the results of finite eleent studies on pull-out behaviour of various diensions of groups of headed anchor blind bolts (HABBs) ebedded in CFCHS with various D/t ratios in both 9 and 18 T-stub. Several paraeters, such as θ value, additional blind bolt in shear, additional through bolt, D/t ratio, and ebedent depth, were investigated to gain a better understanding of the behaviour of this connection. Keywords: pull out, blind bolt, headed anchor blind bolt, concrete filled circular hollow section 1. Introduction Concrete Filled Circular Hollow Sections (CFCHS) have a pronounced advantage relative to steel coluns in ters of structural strength and stiffness, and aesthetic appearance. Also, they are inherently ductile because of the high level of confineent of concrete provided by the steel tube and are hence favoured for use in oent-resisting fraes in areas of high seisicity such as Japan. Learning fro the 1994 Northridge earthquake, a lot of brittle fractures were initiated at coplete joint penetration (CJP) weld between the botto flange of the bea and colun flange at very low levels of plastic deand (FEMA-351, 2). A suary of the typical failure odes in the welded joint is presented by Treblay et al. (1995). Moreover, the lack of access to the inside of circular hollow sections challenges the wider use of this section in the industry. The use of a new type of bolts in which the nut can be tightened fro only one side (so called blind bolts ) should overcoe the two probles discussed above. Blind bolts allow technicians to connect the steel bea and Concrete Filled Circular Hollow Sections (CFCHS) colun fro the outside of CHS; also using blind bolts to connect T-stubs to the CFCHS

2 colun ay avoid the brittle behaviour generated by the welded connections. There are several blind bolts that are available in the arket. Those are the Huck high-strength blind bolt (Huck International Inc., 199), the Lindapter Hollo-bolt (Lindapter International Ltd., 1995), Flow-drilling (France et al., 1999a), and the AJAX ONESIDE blind bolt (AJAX Engineered Fasteners, 22). AJAX ONESIDE blind bolts have been chosen and will be utilized in this paper to describe the echanical behaviour of the blind bolts ebedded in CFCHS. A feasibility study for cogged anchor blind bolt (CABB) connections to CFCHS was carried out by Goldsworthy & Gardner (26). Yao et al. (28; 211) used single and groups of cogged anchor blind bolts (CABBs) and single headed anchor blind bolts (HABBs) in CFCHS. Speciens with D/t ratio fro 32.4 to 54 were used in pull-out tests of single HABBs in CFCHS. Since there are currently no guidelines and clear understanding of the behaviour of this connection, this paper presents finite eleent studies on pull out tests of various diensions of groups of headed anchor blind bolts (HABBs) ebedded in CFCHS with various D/t ratios and in both 9 and 18 T-stubs. The D/t ratios are chosen on the basis that local buckling should be prevented. AISC (21) liits the D/t ratio for copression steel eleents in coposite ebers subject to axial copression (.15E/ f y ) and flexure (.9E/f y ). Both of those liitations are for copact sections. Hence, if 35L steel tubes are used, the axiu D/t ratio that can be used is approxiately equal to 51. Therefore, D/t ratios between 3 to 51 will be used in this paper. The endplate of the curved T-stub used in this study has been designed to have a thickness of 25 so that it will reain elastic. Several paraeters, such as θ value, additional blind bolt in shear, additional through bolt (TB), D/t ratio, and ebedent depth, were investigated to gain a better understanding of the behaviour of this connection. 2. Details of the speciens Details of the speciens are shown in Table 1 and Figure 1. BB and TB represent ordinary blind bolts (without extensions) and through bolt, respectively. Table 1. Details of the speciens Specien Tube size D/t Top of curved Botto of degree of T-stub curved T-stub T-stub ( ) Note H9-1 CHS 457x M2 HABB 1M2 HABB 9 H9-2 CHS 457x M2 HABB 2M2 HABB 9 H9-3 CHS 457x M24 HABB 2M24 HABB 9 H18 CHS 457x M24 HABB 4M24 HABB 18 HB18 CHS 457x M24 HABB + 2M24 HABB M24 BB 2M24 BB and BB HT9-1 CHS 457x M2 HABB + 2M24 HABB 9 1M24 TB and TB HT9-2 CHS 457x M2 HABB + 2M24 9 1M24 TB HABB_12 and TB HT9-3 CHS 2M2 HABB M24 HABB 9 457x12.7 1M24 TB and TB HT9-4 CHS 457x M24 HABB + 2M3 HABB 9 1M3 TB and TB

3 All of the HABBs have effective ebedent depth, h eff, of 1 plus the thickness of the washer, except the HABB in the botto of curved T-stub in specien HT9-2 which has h eff equal to 12 plus the thickness of the washer. The thicknesses of the washer for M2, M24 and M3 are 6, 7 and 8.5, respectively. e e H ep t s t ep e e H ep t s t ep t s=15 t ep =25 H ep =29 e= 6 =85 =45 =25 r ep =228.5 R ep =253.5 r ep R ep t ep Elastic NA y c2 y c1 t s=15 t ep =25 H ep =29 e= 6 =85 1 =25 =4 r ep =228.5 R ep =253.5 (a) 9 T-stub (H9-2 and H9-3) (b) 18 T-stub (H18 and HB18) Figure 1. Configuration of the specien and the FEA odelling Rep t ep r ep HABB or BB HABB or BB Effective ebedent depth 3. Finite eleent (FE) odelling Sleeve Figure 2. Headed anchor blind bolt (HABB) ABAQUS (212) was used for executing the finite eleent analyses. Figure 1 shows the quarter odel for the FE odelling of speciens H9-2, H9-3, H18, and HB18. A displaceent boundary condition was applied at the tip of the T-stub ste to represent the pull-out action fro the hydraulic jack. The bolt and nut were odelled as one solid eleent.

4 The collapsible washer was also odelled as a solid washer. Miniu pretension force was applied using teperature pressure to the iddle of the shank of the blind bolt to clap the curved T-stub flange and steel tube together. The coefficient of friction used between the concrete and the HABB is equal to.3 as recoended by Guezouli and Lachal (212), whereas the coefficient of friction used between steel and steel is assued to be equal to.3. Noinal aterial properties were used for steel eleents as shown in Table 2, where f y, f p, and f u represent yield stress for steel, proof stress for bolt, and ultiate strength, respectively. A concrete copressive strength of 48 MPa was used. Concrete daage plasticity was chosen for representing the concrete plastic behaviour. The fracture energy stated in the ABAQUS docuentation (ABAQUS, 212) was used for odelling the concrete behaviour in tension. Fracture energy which is equal to 15 N/ was used for the concrete strength of 48 MPa. The tensile strength of the concrete was assued to be equal to.56 f c (ACI 318M, 211). Table 2. Material properties for steel tube, T-stub, blind bolt, nut, and washer Note Steel tube Curved T-stub Blind bolt, nut, Endplate Flared flange and washer f y or f p (MPa) f u (MPa) Finite eleent analysis (FEA) results and discussion The results fro FEA were separated into different sub-sections to understand the effect of different paraeters to the pull-out force vs displaceent relationship. The paraeters are θ value, additional blind bolt in shear, additional through bolt, D/t ratio, and ebedent depth Effect of inclination Speciens with 9 curved T-stubs, H9-1 and H9-2 in particular, were copared. Specien H9-1 which has 2M2 blind bolts needed to be ultiplied by 2 to be copatible with H9-2 which has 4M2 blind bolts. The difference is that H9-1 has no θ value, whereas H9-2 has θ value equal to 25 (see Figure 1 for clear presentation of θ). Figure 3 gives a plot of the force vs displaceent fro the FEA. The force represents the pull-out force applied to the tip of the curved T-stub, whereas the displaceent represents the outward displaceent at the iddle of blind bolt s head (Figure 3(b)) and the outward displaceent at the iddle of the curved endplate of T-stub (Figure 3(a) and other Figures). Figure 3 shows that both 2x H9-1 and H9-2 will produce siilar ultiate strengths. However, the value of ultiate strength of H9-2 is 2% less than the ultiate tensile capacity of 4M2 blind bolts. Figure 4 also shows that the ultiate strength of HB18 with 4 HABBs at θ = 25 and 4 HABBs at θ = 65 is less than the ultiate tensile strength of 8M2 blind bolts. This is because in the case of H9-2 (θ equal to 25 ) and H18 (θ equal to 25 and 65 ), the pullout force fro the T-stub will be distributed as tensile and shear forces to the bolt, and the bolt will fail in cobined shear and tension instead of in pure tension. Table 3 gives the capacity reduction ratio for blind bolts at different value of θ. These have been derived using Eq. (1) below fro AS 41 (AS 41, 1998). ( V f 2 ) + ( N tf 2 ) 1. V f N tf (1) V f V f = 1.62 k r f uf (n n A c + n x A o ) (2)

5 N tf N tf = 1 A s f uf (3) Using several assuption as follows: = 1; thread area excluded when calculating the shear area; and using M2 grade 8.8 as an exaple, Table 3 shows that the pull-out capacity is expected to reduce as the θ value is increased. AS41 provided ore conservative result copared with FEA at θ = 25. Using a capacity reduction factor equal to.95 for blind bolts at θ = 25, the capacity reduction factor for blind bolts at θ = 65 in HB18 was calculated using FEA and it resulted the sae value as obtained in AS41. θ ( ) Pull-out capacity (kn) Table 3. Pull-out capacity at different value of θ V f N tf cobination pull out capacity (kn) (kn) ratio pull out capacity at θ = AS41 FEA result (H9-2) (HB18) M2.6x4M2 H9-1 2xH9-1 H Displaceent () (a) Including the effect of elastic curved T-stub = bolts slip 4M2.6x4M2 H9-1 2xH9-1 H Displaceent () (b) Without the effect of elastic curved T-stub Figure 3. Effect of inclination

6 With regards to the effect of the angle of inclination, θ, on the stiffness, coparison have been ade between specien 2x H9-1 and H9-2, and between 2x H9-3 and H18. With the elastic deforation of the curved T-stub included in the deterination of deforation (Figure 3(a)), the secant stiffness at 6% of ultiate capacity was reduced fro 519 kn/ to 34 kn/ (decreased about 35%). In Figure 4, the secant stiffness at 6% of ultiate capacity of 2x H9-3 also decreased about 35% copare with that of H18. Figure 3(b) shows clearly when the slip occurred in the case of H9-2. The slip force, S f, can be calculated as follows: S f = n μ s F p sin θ (4) where n is the nuber of blind bolts that have sae θ value, F p is the pretensioned force applied to the blind bolts (approxiately equal to 7% of ultiate capacity of the blind bolt); μ s is static coefficient of friction (assued to be equal to.3). In the case of H9-2, S f = sin 25 kn. = 412 kn. This value is siilar to the value obtained in FEA which is equal to Effect of additional blind bolt in shear The specien with a 9 curved T-stub (H9-3) was copared with speciens with a 18 curved T-stub (H18 and HB18) to coprehend the effect of the additional HABBs or BBs in shear. As shown in Figure 4, additional HABBs or BBs in shear increased the ultiate capacity by less than twofold. This is because the effect of the inclination that has been explained above. For the stiffness at 6% of ultiate capacity, the stiffness of H18 and HB18 was about 31% and 21% larger than that of H9-3. Hence, HABBs in shear in H18 were not very effective in increasing the stiffness, and there was only a sall increase in stiffness induced by using headed anchor blind bolts (HABBs) rather than ones without anchors (BBs) M24.6x8M24 H9-3 HB18 H18 2x H Displaceent () Figure 4. Effect of HABBs or BBs in shear

7 4.3. Effect of additional through bolt (TB) Figure 5 shows that though the ultiate strength of H18 is uch larger than that of HT9-4, the secant stiffness at 6% of ultiate capacity of H18 is lower than that of HT9-4. This eans an additional through bolt will be ore effective than 4 additional HABBs in shear in increasing the stiffness. The TB will prevent the bolts in the sae elevation of the TB fro slipping. Using the value given in Table 3, the contribution of the TB was calculated. It was shown that only 86% capacity of the TB is able to be added to the syste since the ultiate tensile capacity of the groups of HABBs at the botto of the T-stub was reached at this stage Displaceent () Figure 5. Effect of through bolt 4.4. Effect of D/t ratio and ebedent depth K at.6 Fu of H9-3 K at.6 Fu of HT9-4 K at.6 Fu of HB18 H9-3 HT9-4 H18 Concrete cones fored by the bearing of the headed anchor on the concrete could overlap horizontally, vertically, or both. In general, the overlap between the cones occurred horizontally prior to occurring vertically. This is siply because of the saller distance between the heads of the adjacent HABBs in the horizontal direction than in the vertical direction. Figure 6 presents the effect of increasing the ebedent depth and decreasing the D/t ratio. As the effective ebedent depth of HABBs at the botto of the curved T-stub was increased fro 17 in the case HT9-1 to 127 in the case of HT9-2, both stiffness and strength increased and the concrete cone was delayed until.7 of the ultiate capacity of (3M24+2M2). This eans that in the case of D=457 and t=9.5 (D/t = 48.1), effective ebedent depth approxiately equal to 5 ties the diaeter of the blind bolt should be provided to ensure that the concrete has sufficient cone area to take the pullout force transferred by the HABBs. The strength and stiffness increased as the D/t ratio was reduced fro 48.1 in the case HT9-1 to 36 in the case of HT9-3. The increased participation of bearing of the washer on the steel tube wall helped to delay the foration of the cone in this case. Increasing the diaeter of the blind bolt without increasing the ebedent depth or reducing the D/t ratio will not effectively increase the stiffness and the capacity of the syste (copare case HT9-4 with HT9-1). The ultiate capacity of HT9-4 was only about 75% of ultiate capacity of 2M24+3M3 and concrete cone developed early (below.6x(2m24+3m3)).

8 M24+2M2.6x(3M24+2M2) HT9-1 HT9-2 HT9-3 HT Displaceent () Figure 6. Effect of ebedent depth and D/t ratio 5. Conclusions Extensive finite eleent analyses (FEA) on the group behaviour of headed anchor blind bolts (HABBs) ebedded in concrete filled circular hollow section (CFCHS) coluns were accoplished. Several paraeters, such as θ value, additional blind bolt in shear, additional through bolt, D/t ratio, and ebedent depth, were investigated to gain a better understanding of the behaviour of this connection. The HABBs were pretensioned to ensure that the curved T-stub was well claped to the CHS. Curved T-stub was designed to reain in the elastic range of the aterial behaviour. Several conclusions were ade and can be suarised as follows: 1. Pull-out capacity of the connection decreased as the θ value increased. This is because there is a cobined force in shear and tension. Reduction of the pull-out capacity of the connection can be calculated by aintaining the cobination ratio to be equal to 1. Moreover, reduction in stiffness by 35% each tie should be considered as θ changed fro to 25 and fro 25 to 65 when the elastic deforation of the curved endplate of T-stub is taken into account. 2. HABBs in shear provided siilar stiffness copared with BBs in shear. This is because the shear is doinating and hence the headed anchor will not contribute uch in increasing both the stiffness and strength. 3. It was shown that only 86% capacity of the TB can be added to the syste and it depended on the total ultiate tensile capacity of the groups of HABBs in the botto of the T-stub (which is always lower than that of in the top). It was also shown that the secant stiffness at 6% of ultiate capacity of the syste using TB is larger than that of the syste using 4 additional BBs in shear. 4. As the effective ebedent depth of HABBs at the botto of the curved T-stub was increased fro 17 in the case HT9-1 to 127 (roughly equal to 5xHABB diaeter) in the case of HT9-2, both the stiffness and strength increased and the concrete cone was delayed until.7 of the ultiate capacity of (3M24+2M2). 5. The strength and stiffness also increased as the D/t ratio was reduced fro 48.1 in the case of HT9-1 to 36 in the case of HT9-3. Increasing the diaeter of the blind bolt without increasing the ebedent depth or reducing the D/t ratio will not effectively increase the stiffness and the capacity of the syste.

9 References ABAQUS. (212). ABAQUS 6.12 docuentation. Providence, USA: Dassault Systèes Siulia Corp. ACI 318M. (211). Building code requireents for structural concrete (ACI 318M-11) and coentary. USA: Aerican Concrete Institute. AISC. (21). Specification for Structural Steel Buildings. United States of Aerica: Aerican Institute of Steel Construction. AJAX Engineered Fasteners. (22). ONESIDE brochure. B-N12 data sheet. Victoria, Australia. AS 41. (1998). Australian Standard: Steel Structures. Sydney, Australia: Standards Australia Liited. FEMA-351. (2) Recoended seisic evaluation and upgrade criteria for existing welded steel oent-frae buildings (pp. 2-9). France, J. E. Davison, J. B. and Kirby, P. A. (1999a) Strength and rotational stiffness of siple connections to tubular coluns using flowdrill connectors. Journal of Constructional Steel Research Vol 5, pp Goldsworthy, H. M. and Gardner, A. P. (26) Feasibility study for blind-bolted connections to concrete-filled circular steel tubular coluns. Structural Engineering and Mechanics Vol 24, pp Guezouli, S. and Lachal, A. (212) Nuerical analysis of frictional contact effects in pushout tests. Engineering Structures Vol 4, pp Huck International Inc. (199). Industrial fastening systes. Arizona, USA. Lindapter International Ltd. (1995). Type HB hollo-bolt for blind connection to structural steel and structural tubes. Bradford, England. Treblay, R. Tiler, P. Bruneau, M. and Filiatrault, A. (1995) Perforance of steel structures during the 1994 Northridge earthquake. Canadian Journal of Civil Engineering Vol 22, pp Yao, H. Goldsworthy, H. M. and Gad, E. F. (28) Experiental and Nuerical investigation of the tensile behaviour of blind-bolted T-stub connections to concretefilled circular colun. Journal of Structural Engineering, Aerican Society of Civil Engineers Vol 134, pp Yao, H. Goldsworthy, H. M. Gad, E. F. and Fernando, S. (211) Experiental Study on Modified Blind Bolts Anchored in Concrete-Filled Steel Tubular Coluns, Australian Earthquake Engineering Society Conference, Novotel Barossa Valley Resort, South Australia.