Design of single-angle compression members according to the Canadian standards: Discussion

Abstract

In this note, the author has examined the design requirements according to the Canadian design standards (CSA 1989, 1994) and American standards (AISC 1986, 1994). The note is a direct follow up of the paper by Temple and Sakla (1996). It considers the design capacity of single-angle compression members attached by one leg as compared to the values obtained by experimental work and (or) by finite element analysis. The author has summarized the applicable code clauses in the form of a flowchart that should serve as a useful guide to a structural designer. However, in discussing the code clauses the author has drawn conclusions and made some statements which could be misleading. Although it is not stated clearly in the note or in Temple and Sakla (1996), the author is discussing only one type of application of single-angle equal-leg section with the end connection detail as shown in Fig. 1 of the note and Fig. 5 of Temple and Sakla (1996). The moments that are considered as affecting the design capacity of the single-angle compression brace are a result of eccentricity between the supporting gusset plate and the angle compression member. In practice, moments on single-angle compression members can also result from frame action when single angles are often used to form structural frames to support light equipment. This application is not included in the discussion. The following comments are therefore limited to the application discussed in the note or in Temple and Sakla (1996). 1. The author states that there are two design approaches, Canadian and American, and these two approaches yield different results. To summarize, the approaches stated by the author are (a) to ignore the end eccentricities and treat the members as concentrically loaded members (the simple column approach or Canadian approach), and (b) to consider end eccentricities by treating the angle as a beam-column (American or AISC approach). The author states that “The CISC Handbook of Steel Construction (1991) provides no guidance as to preferred design approach. It seems that the practice generally used in Canada is to ignore the end eccentricities and to design the angle as if it were concentrically loaded and as if it buckles about the z-axis, the minor principal axis of the angle cross section.” In fact, the Canadian Standard CSA S16.1-M89, Clause 21.1 (or CSA S16.1-M94, Clause 21.1) states that “axially loaded members that meet at a joint shall have their gravity axes intersect at a common point if practicable; otherwise, the results of bending due to the joint eccentricity shall be provided for.” While CSA S16.1 does not tell the designer how to account for the effects of eccentricity at joints, it nevertheless does not ask the designer to ignore it. Since there is no specific guidance given in the CISC handbook, the author interprets that it is a common design practice to ignore end eccentricities and design as a concentrically loaded member. In this regard, there is actually very little difference between the CSA Standard S16.1 and AISC specification. In the recent revision of the CSA S16.1 (CSA 1994), the beam-column formula may appear to be different from the AISC formula, owing to differences in the curve fitting expressions used by the AISC and CSA (see Fig. D1 of this discussion). However, the predicted failure loads by the CSA S16.1 differ very little from the calculated values using the AISC standard. The difference in the results becomes less significant when differences in the load factors and the resistance factors are taken into account. 2. The author states that “it is concluded that the simple column approach results in better estimates of the failure loads than does the AISC approach.” It is assumed this statement is based on the results presented in Fig. 8 of Temple and Sakla (1996). The author arrives at the above conclusion, since as per this figure, the failure loads predicted using the expressions for concentrically loaded column lie near or in between the two curves which show the experimental/finite element results, one corresponding to an angle section with pinned end condition and the other curve with fixed end condition. However, Fig. 8 of Temple and Sakla (1996) can be misleading. The figure shows the failure load versus slenderness parameter, λ, which the author appears to have defined without the effective length factor, K. As per the CSA S16.1 or the AISC specifications, the definition of slenderness parameter includes K (i.e., K = 1.0 for pinned end condition and 0.5 for fixed end condition). Therefore, to be consistent with the CSA S16.1 definition of the slenderness parameter, there should be Received October 25, 1996. Revised manuscript accepted January 15, 1997.