Structural stability of unidirectional CFRP thin-walled open-section columns
Theoretical and experimental investigations into the compressive buckling behaviour of unidirectional CFRP thin- walled channel sections subject to built-in end conditions are described. Local and overall modes of instability are considered and the effects of transverse shear on both modes are discussed. Particular emphasis is given to the development of local instability theory for orthotropic materials and the basis for design charts for a range of thin-walled orthotropic sections is included. These analytical developments are accompanied by an investigation of numerical methods in which a finite difference technique is applied to single orthotropic plates and a finite element programme is used with multi-plate sections. Good correlation is observed between analytically and numerically derived buckling loads. Buckling analyses are confined to classical linearised theories and the sensitivities to eccentric loading, applied end moments, and imperfect end restraints are demonstrated. The pultrusion process for manufacture of continuous unidirectional CFRP thin-walled sections is described and suggestions for its development to multidirectional composites are given. Test, methods for the measurement of the principal mechanical properties of unidirectional CFRP from, in some cases, small specimens are detailed. Measured properties are shown to correlate with fibre volume fractions obtained from areal analyses of polished sections. The design of a strain, gauge bridge amplifier and data logging system utilised during column testing is included. The Southwell method is shown to be applicable to flexural and torsional-flexural buckling modes and in general measured buckling loads fall short of theory by 50%. Local buckling loads are indistinct although buckled forms correspond to theoretical predictions and little post buckling strength is observed. A theoretically derived buckling chart for unidirectional CFRP channel sections is presented and a minimum design safety factor of 2 is recommended.