|
Résumé :
|
Forty-nine experimental tests were conducted on rectangular and square tubular columns filled with high-strength concrete (f'c=90 MPa) subjected to axial load and a nonconstant bending moment distribution. The test parameters were the length (2, 3, and 4 m), the cross-section aspect ratio (square or rectangular), the wall thickness (4 or 5 mm), and the ratio of the top and bottom first-order eccentricities r=etop/ebottom (1, 0.5, 0, and −0.5). The effect of slenderness combined with the influence of variable curvature is compared with the design loads from AISC 360-10 and Eurocode 4. The results show that, for the slender elements of this experimental campaign, both codes present a similar error close to 4%. The small error margin indicates that both standards are applicable to high-strength concrete, although the concrete tested is outside the upper bond of Eurocode 4 and AISC’s specifications. However, there are particular cases that present excessive unsafe errors. These errors correspond to the cases where the second-order effects are more important and are attributable to an overestimation of the flexural stiffness E⋅I, which indicates that it needs correction.
|
