A Finite Element and Experimental Study of Reinforced Concrete in Torsion

[摘要] This thesis describes a study into the torsional behaviour of reinforced concrete members, in particular solid Lsections under pure torsion. It consists of three distinct but linked phases: (1) development, assessment and application of a three dimensional nonlinear finite element model for short-term behaviour of reinforced concrete, (2) an experimental programme and (3) a numerical parametric study. Nonlinear behaviour takes into account concrete cracking, nonlinear triaxial stress-strain relations of concrete, concrete crushing and yielding of steel reinforcement. Cracking behaviour is modelled by a fixed orthotropic smeared crack approach, allowing up to three cracks to occur at any sampling point. Modelling of post-cracking behaviour allows for shear transfer and tension stiffening effects. Concrete behaviour under all multiaxial stress states is governed by a short-term constitutive law and a peak stress failure criterion. A bilinear uniaxial stress-strain law allowing for isotropic strain hardening is used for steel reinforcement. 20-noded isoparametric brick element is used to represent concrete, with single bars embedded within the concrete elements to simulate reinforcement. A modified Newton-Raphson approach was used for solving the nonlinear problem, based on the evaluation of a secantial elasticity matrix. The program was assessed by studying the behaviour of deep beams, shallow beams simulating beam-column behaviour, rectangular beams subject to pure and combined torsion and L-sections under pure torsion. Through systematic study, the influence of some of the major nonlinear material and solution parameters was established for these applications and limits on their values were set. The experimental programme involved testing a series of solid reinforced concrete models of L-shaped cross sections under pure torsion in a specially designed and built test-rig. The tests were devised to: (1) assess the current British Code design procedure for torsion of solid L-sections, (2) obtain an insight into the torsional behaviour of these types of sections and (3) provide detailed results to assess the reliability of the finite element model in the analysis of torsion of fully reinforced flanged sections. A dual approach of complementing the experimental results with a numerical parametric study, using the developed finite element model, is adopted where more variables not included in the experimental programme were investigated. The current British Code design procedure for torsion was found to be too conservative for solid flanged sections, and it is concluded that the code's rules can be less stringent. Some recommendations that might help bring the code's torsion design procedure in line with other major codes of practice are given.

[发布机构] University:University of Glasgow

[关键词] Civil engineering