Experimental investigation of the member buckling of cold-rolled aluminium alloy 5052 channel columns

In recent years, BlueScope Permalite in Australia has used the existing rolling system for coldformed steel sections to successfully rollform aluminium alloy sections. In comparison with the conventional extrusion method, the rollforming process is faster, more costeffective and may enhance the strength of the aluminium alloy. This paper presents a series of compression tests on coldrolled aluminium alloy 5052 lipped channels subjected to member buckling. A total of 28 tests of three commercially available sections including C10030, C25025 and C40030 were conducted. Two end boundary condition configurations were designed for test rigs. In the first configuration, the end sections of the specimens rotate freely about the major axis whereas rotation about the minor axis is restrained. In contrast, for the second configuration, restrained and free rotation conditions are used for major and minor axes respectively. Both flexuraltorsional buckling and localglobal interaction buckling failure modes were observed during the tests. The experimental results are compared with the predictions in Specification for Aluminum Structures 2015 by The Aluminum Association.

Ninth International Conference on Advances in Steel Structures (ICASS’2018) 5-7 December 2018 - Hong Kong, China  EXPERIMENTAL INVESTIGATION OF THE MEMBER BUCKLING OF COLD-ROLLED ALUMINIUM ALLOY 5052 CHANNEL COLUMNS Ngoc Hieu Pham 1*, Cao Hung Pham and Kim J.R Rasmussen 1 School of Civil Engineering, The University of Sydney, New South Wales, Australia E-mails: ngochieu.pham@sydney.edu.au, caohung.pham@sydney.edu.au, kim.rasmussen@sydney.edu.au Abstract: In recent years, BlueScope Permalite in Australia has used the existing rolling system for cold-formed steel sections to successfully roll-form aluminium alloy sections In comparison with the conventional extrusion method, the roll-forming process is faster, more cost-effective and may enhance the strength of the aluminium alloy This paper presents a series of compression tests on cold-rolled aluminium alloy 5052 lipped channels subjected to member buckling A total of 28 tests of three commercially available sections including C10030, C25025 and C40030 were conducted Two end boundary condition configurations were designed for test rigs In the first configuration, the end sections of the specimens rotate freely about the major axis whereas rotation about the minor axis is restrained In contrast, for the second configuration, restrained and free rotation conditions are used for major and minor axes respectively Both flexural-torsional buckling and local-global interaction buckling failure modes were observed during the tests The experimental results are compared with the predictions in Specification for Aluminum Structures 2015 by The Aluminum Association Keywords: Cold-rolled Aluminium Alloy; Experimental Investigation; Member buckling; Channel columns DOI: 10.18057/ICASS2018.xxx INTRODUCTION Aluminium alloy structures have increasingly been used in structures such as bridges, building facades, roof and wall systems [1] due to their excellent corrosion resistance, high strength to weight ratio and the convenience of transportation and assembly Two methods are normally used to produce aluminium alloy sections The first and conventional method is extrusion commonly used around the world whereas roll-forming is the second and new method recently introduced by BlueScope Permalite in Australia [2] to produce roll-formed aluminium alloy sections In recent years, BlueScope Permalite has successfully roll-formed two trial sections, including Channel and Zed sections, within acceptable dimensional tolerances, and demonstrated that it is possible to roll-form other aluminium alloy sections by utilizing the existing rolling system of cold-formed steel sections It is found that the roll-forming process is faster and more cost-effective than the extrusion process, and may enhance the strength of the material Hence, the competitive advantages of roll-formed aluminium sections should be commercialized in the worldwide market Currently, there are no design guidelines for cold-rolled aluminium alloy members Past research has focused on extruded aluminium alloy and has produced a huge amount of experimental data However, experimental studies of cold-rolled aluminium alloy members remain limited Roll-formed aluminium sections considered as thin-walled sections are therefore prone to local, distortional and global buckling instabilities Hence, the behaviours of Proceedings of the ninth international conference on Advances in Steel Structures Edited by Siu-Lai Chan, Tak-Ming Chan and Songye Zhu Copyright © 2018 by The Hong Kong Institute of Steel Construction Ngoc Hieu Pham et al buckling modes and the possible interactions between them need to be intensively investigated to provide design guidelines for those types of cold-rolled aluminium sections This paper presents a series of compression tests on cold-rolled aluminium alloy lipped channels subjected to member buckling The test results are compared with the predictions determined by the Specification for Aluminum Structures 2015 [3] MATERIAL PROPERTIES The tensile coupon tests for three types of sections studied in this paper were reported in Huynh et al.[4] Both flat and corner coupons taken from the flat and corner portions of the sections were tested The dimensions of tensile flat coupons conformed to the Australian Standard AS 1391-2007 [5] with a width of 12.5 mm and a gauge length of 50 mm The actual thickness of each coupon was measured using a micrometer The tests were performed in a 50 kN capacity MTS Model 43 machine using displacement control Strain gauges attached to both sides of the coupons and an extensometer were used to record the deformation processes Based on a total of 218 coupon tests conducted in Huynh et al [4], the average material properties are summarized in Table Table 1: Tensile material properties [4] Coupon Position C10030 C25025 C40030 C10030 C25025 C40030 flat flat flat corner corner corner σ0.2 MPa 220 215 220 244 248 240 σu MPa 267 262 268 290 292 282 E0 GPa 69.3 69.7 69.9 70.1 71.2 70.5 εf n 4.4 4.8 5.2 7.3 8.6 7.0 12.0 12.8 13.5 15.8 16.4 18.1 Note: σ0.2 is the 0.2% proof strength; σu is the ultimate tensile strength; E0 is the Young’s modulus; εf is the uniform elongation corresponding the total elongation after fracture; and ‘n’ is the Ramberg-Osgood index SPECIMEN DIMENSIONS AND GEOMETRIC IMPERFECTIONS The cross section dimensions of the specimens are commercially available in the catalogue “Roll-Formed Aluminium Purlins Solutions” from BlueScope Permalite [2] The nominal dimensions of the channel sections are given in Figure 1, where: t is the thickness of the cross-section; D, B, L are the overall web depth, the flange widths and the lip lengths of the section, respectively The representative sections are chosen on the basis of the nominal slenderness of the sections which are categorised as low, intermediate and high The nominal slenderness of a section is defined as follows: l  fy (1) f cr where fy is the proof yield stress of the aluminium alloy computed from the result of coupon tests; and fcr is the elastic local or distortional buckling stress determined from the Thin-Wall2 Ngoc Hieu Pham et al software [6] The selected sections of C10030, C25025 and C40030 correspond to low, intermediate and high slenderness, respectively, as given in Table Figure 1: The nominal dimensions of the channel section Table 2: Thin-Wall results for channel sections in compression Sections C10025 C10030 C15025 C15030 C20025 C20030 C25025 C25030 C30025 C30030 C35030 C40030 t mm 2.5 3.0 2.5 3.0 2.5 3.0 2.5 3.0 2.5 3.0 3.0 3.0 D mm 105 105 155 155 205 205 255 255 300 300 350 400 B mm 59 60.5 62 63 76 77.5 76 77 109 110.5 125.5 125.5 L mm 16 16 23 23.5 25 25 25.5 26 30 30 30 30 fol MPa 210.4 302.8 97.64 140.8 55.33 79.72 36.21 52.11 25.46 36.69 26.87 20.74 fod MPa 196.9 240.6 149.1 185.4 97.94 121.5 66.03 82.35 55.67 68.63 50.98 39.43 λl λd 0.999 0.8328 1.4665 1.2213 1.9482 1.623 2.4082 2.0075 2.872 2.3924 2.7956 3.182 1.0327 0.9342 1.1868 1.0643 1.4643 1.3147 1.7834 1.5969 1.9422 1.7493 2.0296 2.3078 The lengths of the specimens were selected to allow the global buckling or local global interaction buckling to occur According to the Australian standard AS/NZS 4600:2005 [7], the global buckling stress is the lesser of the flexural or flexural-torsional buckling stresses which are determined as follows: The elastic flexural buckling stress (foy): f oy   2E  ley / r  (2) The elastic flexural torsional buckling stress (foxz): f oxz   ( f ox  f oz )  ( f ox  f oz )   f ox f oz    2 (3) Ngoc Hieu Pham et al where: ley is the effective length for buckling about the y-axis; r is the radius of gyration of the full, unreduced cross-section; fox is the elastic buckling stress in an axially loaded compression member for flexural buckling about the x-axis; foz is the elastic buckling stress in an axially loaded compression member for torsional buckling; β is the coefficient specified in Clause 3.4.3 of the Australian standard AS/NZS 4600:2005[7] To capture both the flexural bucking and the flexural-torsional buckling modes, two boundary configurations were designed: i) the specimens can rotate freely about the strong axis (x-axis) for the first configuration to capture the flexural-torsional buckling mode; and (ii) the specimens can rotate freely about the weak axis (y-axis) for the second configuration to capture the flexural buckling mode Warping displacements are restrained at two ends for both two configurations The effective lengths in each axis are given in Table 3, where lex, ley and lez are the effective lengths for buckling about the x-, y- axes and for twisting, respectively; and l is the actual length of a specimen The lengths of the compression specimens chosen on the basis of elastic global buckling analyses are listed in Table Table 3: Compression specimen lengths and the effective lengths Configuration Configuration Configuration Sections C10030 C25025 C25025 C40030 Specimen lengths (m) 2.0 2.5 3.0 3.0 5.0 7.0 2.0 3.0 3.0 5.0 7.0 lex ley lez l 0.5l 0.5l l 0.5l 0.5l 0.5l l 0.5l 0.5l l 0.5l Cold-rolled aluminium sections as thin-walled sections are sensitive to buckling, which is significantly impacted by geometric imperfections It is, therefore, crucial to capture the magnitude and pattern of the imperfections of each specimen prior to testing The geometric imperfections of specimens were measured using a laser scanner method as reported in Pham et al [8] TEST SET-UP With the excessive specimen length of meters, column specimens were impossible for testing in a vertical position; instead, they were tested in a horizontal position Figure shows the test rig where two reaction I-beams were clamped into a rail system that was anchored into the strong floor This set-up prevents the reaction I-beams from pulling out of the ground during the testing process The two brace angle sections of L100mm×100mm×10mm were bolted through the web of each reaction I-beam by three M20 high strength 8.8 bolts to prevent two reaction I-beams from sliding while testing Although the specimens were tested horizontally, Ngoc Hieu Pham et al the self-weight of the aluminium alloy specimens were small and neglectable A 500 kN MOOG jack was bolted on top of the reaction I-beam to apply a load to the specimens as shown in Figure (a) The compression test rig (b) Front view (c) Section A-A Figure 2: Compression test set-up Ngoc Hieu Pham et al (a) The first configuration (b) The second configuration Figure 3: End boundary conditions Two configurations of end boundary conditions were designed for the test rigs In the first configuration, the end sections of the specimens freely rotate about the major axis whereas their rotation is prevented in the minor axis In contrast, for the second configuration, restrained and free rotation conditions are designed for major and minor axes, respectively The details of two configurations are shown in Figure The distances between the ends of the specimen and the centre of the hinges are 75 mm The specimen displacements were monitored using 13 transducers as shown in Figure Four transducers including LVDT (L1, L2) on the left hand side and (R1, R2) in the right hand side were attached at the ends of specimens to measure the end rotations and the axial shortening Five transducers from LC1 to LC5 were attached to the aluminium transducer frame which was held at two corners of the web and the flanges as shown in Figures 2(c) This set-up allows LVDTs attached to the transducer frame to locally measure the distortion of the crosssection at the mid-length when the global buckling occurs For measurement of global deformations, four transducers from G1 to G4 were attached to two stable columns and were connected with two corners of the transducer frame through four steel strings as shown in Figures 2(c) Based on the recorded data from those four transducers, the global deformations of the specimens including the lateral deformation and the global twist can be determined The first step is to determine the displacements of two corners of the transducer frame in the xdirection and y-direction Subsequently, based on the dimensions of the transducer frame and the centroid of the specimen, the displacements of the centroid and the global twist of the specimen at the mid-length are calculated The specimens were tested both concentrically and eccentrically The concentric tests are used to validate against the theoretical buckling load whereas eccentric tests are used to propose design curves with a nominal eccentricity of le/1500 at both ends of each specimen, where le is the actual length of the specimen The test specimens are labelled and listed in Table For example, a typical test “C10030-2.0m-2e” is defined as follows: (i) “C10030” indicates a channel section with the depth of 100 mm and the thickness of 3.0 mm; “2.0 m” is the length of the specimen; (ii) (iii) “2e” indicates the eccentric test number Without letter “e” at the end, the test is a concentric test Ngoc Hieu Pham et al Configuration Configuration Table 4: Compression specimens for both two configurations Specimens Concentricity Eccentricity Configuration Sections C10030-2.0m-1e C10030-2.0m-1 C10030-2.0m-2e C10030-2.5m-2e C10030-2.5m-1 C10030 C10030-2.5m-3e C10030-3.0m-1e C10030-3.0m-2 C10030-3.0m-3e C25025-3.0m-3 C25025-3.0m-1e C25025-5.0m-3 C25025-5.0m-2e C25025 C25025-7.0m-1 C25025-7.0m-2e C25025-2.0m-1 C25025-2.0m-2e C25025 C25025-3.0m-1 C25025-3.0m-2e C40030-3.0m-2e C40030-3.0m-1 C40030-3.0m-3e C40030-5.0m-2e C40030-5.0m-1 C40030 C40030-5.0m-3e C40030-7.0m-2e C40030-7.0m-1 C40030-7.0m-3e TEST BEHAVIOR 5.1 The first configuration For C10030 section of all studied lengths, the failure modes were observed in flexuraltorsional buckling mode as shown in Figure 4a The local buckling mode was not observed because the ultimate load is much lower than the local buckling load Figure shows the eccentric test results of C10030-2.0m-1e including the relationship curves between the axial load versus the end-shortening and the end-rotation at the end section, and global twist at the mid-length of the specimen The effect of nominal eccentricity and initial imperfections lead to the gradual overall buckling process as shown in Figure 4(b) and the early development of the global twist and end rotation as shown in Figures 4(c) and 4(d) 100 Axial Load (kN) 80 60 40 20 0 (a) The flexural-torsional buckling mode End-shortening (mm) (b) Axial load versus End-shortening Ngoc Hieu Pham et al 100 80 80 Axial Load (kN) Axial Load (kN) 100 60 40 20 60 40 20 0 0.2 0.4 Global twist (rad) 0.6 (c) Axial load versus Global twist 0.01 0.02 0.03 End-rotation (rad) 0.04 (d) Axial load versus End-rotation Figure 4: The test results of C10030-2.0m-1e in the first configuration For C25025 section, the interactions of local-global interaction buckling modes were seen in various lengths as shown in Figure The local buckling occurred and was recorded at small load of nearly 40 kN This load is in good agreement with the elastic local buckling load from Thin-Wall-2 software [6] Figure presents the test results of the specimen C25025-3.0m-3 Local buckling modes were observed on the web of the specimen as shown in Figure 6(d); this leads to a shift of the effective centroid of the cross-section The curve of axial load versus end-shortening shows a change in the slope at the local buckling point as shown in Figure 6(a) After the local buckling point, the vertical deformation at the mid-length of the specimen increases significantly, and the global twist at the mid-length starts to increase and develops remarkably near the peak load as shown in Figures 6(c) and 6(b) Hence, the local-global interaction buckling mode was observed due to the local deformations on the web and flexural-torsional deformations of the specimen Figure 5: The local-global interaction buckling mode in the first configuration Ngoc Hieu Pham et al 100 100 80 Local buckling 60 Axial Load (kN) Axial Load (kN) 80 40 20 60 40 20 0 10 -0.01 End-shortening (mm) 100 100 80 80 60 40 20 60 40 12 Vertical Deformation (mm) Flange 1: LVDT-LC2 Flange 2: LVDT-LC4 Web: LVDT-LC1 20 -3 0.04 (b) Axial load versus Global twist Axial Load (kN) Axial Load (kN) (a) Axial load versus End-shortening 0.01 0.02 0.03 Global twist (rad) 15 -2 (c) Axial load versus Vertical-deformation Local Deformation(mm) (d) Axial load versus Local deformations Figure 6: The test results of C25025-3.0m-3 in the first configuration 5.2 The second configuration All specimens failed in a combination of local-flexural interaction buckling modes After the local buckling, the specimens were bent about the major axis toward the lips as shown in Figure 7(a) Figure shows the behaviour of a typical specimen, C40030-3.0m-1 The local buckling modes were observed at the load of about 40 kN as shown in Figure 7(c); this load fits with the elastic local buckling load from Thin-Wall-2 [6] The local buckling point is also seen in Figure 7(b) with the changing slope of the axial load versus the end-shortening at the load of about 40 kN After the local buckling load, the flexural buckling deformation occurred and increased quickly as shown in Figure 7(d), followed by the erosion of bearing capacity, and the drop of axial load of this specimen Hence, the local-global interaction buckling mode was observed in this test due to the local buckling deformations of the web and the significant lateral displacement of this specimen Ngoc Hieu Pham et al 120 100 Axial Load (kN) 80 Local buckling 60 40 20 0 120 120 100 100 80 80 60 60 40 Flange1: LVDT-LC4 Flange2: LVDT-LC2 Web: LVDT-LC1 20 -2 Local Deformation(mm) (c) Axial Load versus Local deformations 10 40 20 -10 10 End-shortening (mm) (b) Axial Load versus End-shortening Axial Load (kN) Axial Load (kN) (a) The local-global interaction buckling mode 10 20 30 40 50 Lateral deformation(mm) 60 (d) Axial Load versus Lateral deformation Figure 7: The test results of C40030-3.0m-1 in the second configuration COMPARISON OF THE TEST RESULTS WITH PREDICTED STRENGTHS FROM THE SPECIFICATION FOR ALUMINUM STRUCTURES 2015 The Specification [3] was used to predict the ultimate capacities of the experimental program The strength of a compression member is equal to the lowest of three available strengths: the nominal member buckling strength, the nominal local buckling strength and the nominal local-member interaction buckling strength The nominal member buckling strength is determined from Section E.2 of the Specification for Aluminium Structures 2015 [3] as follows: For λ ≤ λ1: Pnc  Fcy Ag (4)  C   For λ1 < λ < λ2: Pnc   Bc  Dc    0.85  0.15 c  Ag Cc  1   (5) For λ ≥ λ2: Pnc  0.85 E 2 (6) Ag 10 Ngoc Hieu Pham et al The nominal local-member interaction buckling strength is determined from Section E.4 of the Specification for Aluminium Structures 2015 [3] as follows: 1/  0.85 E  Pnc       Fe2 / Ag (7) where λ, λ1, λ2 are non-dimensional slenderness, and are determined from Section E.2 of the Specification [3]; Bc, Cc, Dc are determined from Tables B.4.1 or B.4.2 of the Specification [3]; and Ag is the gross cross-sectional area Table 5: Evaluation of design capacity in the first configuration Specimens Test results Ptest (kN) Aluminum Design Manual 2015  Pnc(kN) Failure mode Failure mode Ptest/Pnc C10030-2.0m-1 C10030-2.0m-1e C10030-2.0m-2e 91.25 81.00 81.38 M 81.80 M 1.12 0.99 0.99 C10030-2.5m-1 C10030-2.5m-2e C10030-2.5m-3e 70.64 60.10 61.60 M 52.34 M 1.35 1.15 1.18 C10030-3.0m-2 C10030-3.0m-1e C10030-3.0m-3e 49.45 46.76 44.65 M 37.70 M 1.31 1.24 1.18 C25025-3.0m-3 C25025-3.0m-1e 88.20 89.30 L-M 65.7 L-M 1.34 1.36 C25025-5.0m-2e C25025-5.0m-3 55.23 56.25 L-M 47.8 L-M 1.16 1.18 C25025-7.0m-1 34.55 C25025-7.0m-2e 33.15 M 37.7 M 0.92 0.88 Table 6: Evaluation of design capacity in the second configuration Specimens Test results Aluminum Design Manual 2015 Ptest/Pnc Ptest (kN) Failure mode Pnc(kN) Failure mode C25025-2.0m-1 C25025-2.0m-2e 60.75 59.40 L-M 54.4 L-M 1.12 1.09 C25025-3.0m-1 C25025-3.0m-2e 36.95 34.65 L-M 42.2 L-M 0.88 0.82 C40030-7.0m-1 C40030-7.0m-2e C40030-7.0m-3e 37.45 36.55 36.10 L-M 43.8 L-M 0.86 0.83 0.82 C40030-5.0m-1 C40030-5.0m-2e C40030-5.0m-3e 60.35 57.60 58.95 L-M 54.5 L-M 1.11 1.06 1.08 C40030-3.0m-1 C40030-3.0m-2e C40030-3.0m-3e 104.25 102.25 102.65 L-M 75.6 L-M 1.38 1.35 1.36 11 Ngoc Hieu Pham et al The prediction of the ultimate capacities and failure modes are compared to the test results in Tables and 6, where M stands for Member buckling mode; and L-M stands for interaction local - member mode It can be seen that the predicted failure modes are in good agreement with those observed from the test programs The design capacities are unconservative for slender columns, and are conservative for short columns CONCLUSION The experimental investigation was conducted to determine the member capacity of coldrolled aluminium alloy channel columns The failure modes included member buckling modes and local-member buckling modes The test results were compared with the predicted strengths and failure modes determined by Specification for Aluminum Structures 2015 [3] The predicted failure modes are in good agreement with the test results In terms of design capacities, the Specification was found to give unconservative predictions for slender columns, and give slightly conservative predictions for short columns ACKNOWLEDGEMENT Funding provided by the Australian Research Council Linkage Research Grant LP140100563 between BlueScope Lysaght and the University of Sydney has been used to perform this research The authors would like to thank Permalite Aluminum Building Solutions Pty Ltd for the supply of the test specimens and financial support for the project The first author is sponsored by the scholarship provided by Australian Awards Scholarships (AAS) scheme from Australian Government REFERENCES [1] Szumigala M and Polus L., “Application of Aluminium and Concrete Composite Structures”, [2] [3] [4] [5] [6] [7] [8] FProcedia Eng., 108(61), 544-549, 2015 Blue Scope Lysaghts, Permalite – Aluminium Rollformed Purlin Solutions, Permalite Aluminium Building Solutions Pty Ltd, Eagle Farm Qld 4009, Australia, 2015 The Aluminum Association, Aluminum Design Manual 2015, 1525 Wilson Blvd, Suite 600, Arlington, VA 22209, 2015 Huynh L.A.T., Pham C.H and Rasmussen K.J.R., “Mechanical Properties of Cold-Rolled Aluminum Alloy 5052 Channel Sections”, Proceedings of 8th International Conference on Steel and Aluminium Structures, Hong Kong (7-9/12), 670-684, 2016 AS 1931:2007, Metallic materials – Tensile testing at ambient temperature, Standards Australia, Sydney, 2007 Nguyen V.V., Hancock G.J and Pham C.H., “Development of the Thin-Wall-2 Program for Buckling Analysis of Thin-Walled Sections Under Generalised Loading, Proceeding of 8th International Conference on Advances in Steel Structures, Lisbon (22-24/7), 2015 AS/NZS 4600:2005, Cold-formed Steel Structures, Standards Australila, Sydney, 2005 Pham N.H., Pham C.H and Rasmussen K.J.R., “Incorporation of Measured Geometric Imperfections into Finite Element Models for Cold-rolled Aluminium Sections”, Proceeding of 4th Congres International de Geotechnique – Ouvrages – Structures, Ho Chi Minh city (26-27/10), 161-171, 2017   12 ... The experimental investigation was conducted to determine the member capacity of coldrolled aluminium alloy channel columns The failure modes included member buckling modes and local -member buckling. .. the dimensions of the transducer frame and the centroid of the specimen, the displacements of the centroid and the global twist of the specimen at the mid-length are calculated The specimens were... program The strength of a compression member is equal to the lowest of three available strengths: the nominal member buckling strength, the nominal local buckling strength and the nominal local-member