Modelling of RC walls with ductile detailing subjected to high axial loads

Abstract

Observations following the 2010/2011 Canterbury Earthquakes revealed unexpected damage to reinforced concrete walls, characterized by undesirable failure modes such as crushing of concrete and buckling of longitudinal reinforcement in the web and end regions. In an effort to address these failures, a number of changes have been made to the New Zealand Concrete Structures Standard (NZS3101:2006) that will be published in the next amendment. Major changes include an introduction of an axial load limit (P ≤ 0.3Agf’c) and changes to confinement detailing of the wall web and end regions. Four large-scale walls (C10-A30) are currently being tested to assess the effects that variation of these parameters will have on wall response. Prior to commencing experimental testing, blind predictions have been made using nonlinear finite element models developed in VecTor2. The models were calibrated against test results for five walls with similar detailing and loading conditions to C10-A30. The calibrated models successfully matched strength capacity, failure mode (flexural crushing). Drift at failure observed in experiments was simulated with an average accuracy of 5-30%. Using the modelling approach, C10-A30 were predicted to exhibit generally poor performance characterized by low ductility. Specifically, C10 and A10, with axial load ratios of 0.1Agf’c, were predicted to reach 1.6% and 1.8% drift, respectively, before failure with an average ductility of 3. A longer confinement length in A10 compared to C10 had only a minor effect on the deformation capacity. Specimen A20 (axial load of 0.2Agf’c) and A30 (axial load of 0.3Agf’c) were predicted to reach 1.2% drift and 1.05% drift, respectively, with no ductility due to compression-controlled flexural failure resulting from the large axial loads. Failure was characterized by a rapid loss of strength in a brittle compression failure.