An experimental investigation of distortional buckling of steel storage rack columns

Abstract

An experimental investigation on the behavior of steel storage rack columns subjected to compression is presented. Members of different lengths are tested, but special attention is focused on the behavior of the specimens having lengths that make them subject to distortional buckling. This mode of buckling can be observed in moderately long specimens; namely, longer than the stub columns used for the determination of the local buckling strength, but short enough to avoid the effects of global buckling. The deformation experienced by these specimens is measured, and it is observed that there is a range of member lengths where the failure mode is a combination of distortional buckling and global buckling modes. Furthermore, it is verified that, although the effect of the interaction between these modes on the member strength is not large, the accuracy of the current design procedures improves if it is considered in the calculations.

Introduction

Columns of steel storage rack structures contain a large amount of holes uniformly distributed along their length. These holes facilitate the connection between the columns and the other members of the rack structure. However, from the design point of view, the presence of holes presents problems. Although many investigations of the effect of perforations on the strength of members have been performed (see, for instance, [1], [2], [3], [4], [5], [6], [7], [8], [9], [10]), a definitive analytical solution for pallet rack members has not yet been formulated. As a consequence, current standards recommend determining the load carrying capacity of columns from experimental tests [11], [12].

The objective of the study covered in this paper, that is a part of an ongoing research at Universitat Politècnica de Catalunya, is to formulate experimental and analytical approaches to the design of industrial rack columns against distortional buckling. This paper covers the first phase of this study, to formulate test procedures for use in design. Based on the results of the tests conducted to date some tentative design formulations have also been reached. In the next phase of the study, the findings presented here will be refined and expanded.

According to the current design standards of rack structures [11], [12], distortional buckling strength tests should be carried out applying the load through rigid plates located at the end cross-sections. The specimens are bolted or welded to these plates and, consequently, the deformation and warping of these end sections are constrained. The specimens can be considered fixed with respect to the distortional buckling mode. It is difficult to test members with distortional buckling pinned end supports, because it is complicated to allow warping of the end cross-sections and to apply a compressive load at the same time.

In an actual upright frame, the connections between members are not fixed but offer only a partial distortional buckling restraint. As a consequence, the distortional buckling strength derived from the experimental tests on individual columns may be higher than the distortional buckling strength that can be developed in an upright frame. This is why distortional buckling tests should be carried out on long members, where several distortional buckling half-sine waves can develop. The favorable effect of the fixed distortional buckling ends vanishes as the member length and the number of distortional buckling half-waves increases (see [13]). However, long members should also be avoided because the participation of distortional buckling in the failure mode becomes small as the length increases and the global modes become dominant. The whole problem can be summarized as it is stated in the AISI test Standard S910-08 [14]: “the specimen length should be sufficiently short to minimize overall column effects, and sufficiently long to minimize end effects during loading”. This problem of the effect of the fixed end supports is not considered in the European code for design of rack structures EN15512 [11]. It simply prescribes that the specimen length should be chosen equal to the length of the single bracing panel closest to one meter.

It can be concluded that choosing the appropriate test length is one of the key points of the determination of the distortional buckling strength. The present investigation started with the aim of formulating a procedure to determine this length. Compression tests on different cross-sections and with different specimen lengths are carried out to find out how the appropriate test length can be selected by means of an analytical procedure. The paper demonstrates that, as it was expected, the specimen length can be approximately determined from the results of linear buckling analyses. However, the main goal of the paper is to show the behavior of the members observed in the experimental tests. In this sense, it is interesting to note that, for the reasons mentioned above concerning the effects of end supports, tests are performed on moderately long specimens. Their length is chosen at the end of the range of distortional buckling dominant failure mode, or in the transition from the range of distortional buckling dominant to the range of global dominant failure mode. For this range of specimen lengths, distortional buckling and global buckling modes are combined. Therefore, the tests provide information about the interaction phenomena between these modes. Recent investigations on different types of cold-formed cross-sections have pointed out that the distortional–global interaction may have some effects on member strength [15], [16], [17].

In this paper, Section 2 presents the cross-sections analyzed and the specimen lengths chosen for the tests. Section 3 includes an explanation on the test setup. Section 4 is devoted to describe the behavior of the compressed members. Subsequently, different methods for the calculation of the distortional buckling strength are evaluated in Sections 5 to 7. Finally, the concluding remarks of Section 8 close the paper.