FE analysis of pultruded FRP built-up columns

Presently, very few research papers concerning built-up members made of Pultruded FRP (PFRP) material are available and no technical standard reports recommendations and design rules specifically suited for these members. Dicuonzo et al. (4th Int. Conf. on FRP Comp. in Civil Eng., CICE 2008) presented a temporary, modular all-GFRP structure designed for an expo fair stand. In this structure, the 3 m-high columns are comprised of two standard PFRP C-shaped profiles with dimensions 203 x 56 x 9.5 mm, connected with one another through 500 mm-spaced bonded battens. The modular, 6 m-span roof girders allow for covering 12 m-span lengths without requiring intermediate columns. Top and bottom chords of the roof girders are comprised of two standard PFRP equal-leg angles with dimensions 102 x 102 x 9.5 mm, once again connected with one another using bonded battens. Bai et al. (Mater Struct 2013; 46: 1143-54) performed four-point bending tests on built-up members used in the five-storey GFRP Eyecatcher building in Basel, Switzerland. That structure, built in 1999, still remains the tallest FRP building in the world. Its structural skeleton is comprised of three parallel trapezoidal GFRP frames connected by wooden decks. The various members of the frames were designed by assembling individual standard PFRP shapes by continuous bonding. Boscato et al. (Compos Struct 2015; 121: 46-63) presented experimental results on PFRP built-up columns comprised of four standard C-shaped profiles mutually connected at discrete points either by bolting or by bonding. In a first attempt to interpret the experimental results with simple closed-form equations, the authors compared the experimentally evaluated buckling loads with those estimated using (1) Euler's and (2) Engesser's equations (applied to the built-up columns considered as members with uniform cross-section) and (3) an expression typically adopted for steel built-up members. Unfortunately, none of these expressions was capable to predict the experimental buckling loads and the authors concluded that geometrically nonlinear FE analyses are necessarily to be performed to capture the actual column behavior. The previously quoted papers do not raise the issue of a systematic approach to the design of PFRP built-up members, which should consider: various basic profiles (C- and L-shaped, etcetera); different types of connection between basic profiles (battens and or lacings, connected by bolting and/or bonding, etcetera); changes in geometric characteristics (member total length, spacing of the battens, etcetera); and various types of loading. In this context, the present paper deals with a series of parametric analyses of PFRP built-up columns comprised of two closely spaced C-shaped profiles. Numerical models of the columns are developed using bidimensional orthotropic finite elements. In order to define the maximum distance between the battens that ensures an overall buckling behavior, linear buckling analyses are carried out for different spacings of the battens. Moreover, to investigate the influence of the batten stiffness, both rigid and deformable battens are used in the models. Finally, geometrically nonlinear analyses of the coumns with initial imperfection are carried out to capture the ultimate resistance. In these analyses, the ultimate conditions are located either by a peak in the load-lateral deflection column response or by the attainment of the limiting material strenght at some point of the column. Since several battens spacings are considered, the numerical results obtained could be used, in analogy to the case of PFRP I-section columns (see National Research Council of Italy, CNR-DT 205/2007), to define a buckling design curve for built-up columns