Investigation of interlocking cracks in reinforced concrete membrane elements
Author: Nik Lüscher
Language: English
Abstract
The load-deformation behaviour of reinforced concrete membranes subjected to in-plane shear and normal forces can generally be reliably simulated using the Cracked Membrane Model with rotating, stress-free cracks (CMM-R). There are limitations for membrane elements with low reinforcement ratio in one of the two orthogonal reinforcement directions and for cyclic loading in the zero-crossing range. This Master's Thesis investigated whether the Cracked Membrane Model with fixed interlocked cracks (CMM-F) can reproduce the load-deformation behaviour observed in experiments better than the CMM-R.
Interlocking cracks are taken into account in CMM-F using established aggregate interlock models. The three most established and cited models are the Two Phase Model (TPM), the Rough Crack Model (RCM) and the Contact Density Model (CDM). These models were tested under monotonic and cyclic shear loading and compared with each other. The comparison revealed differences in the resulting stresses. In general, the CDM results in the highest stresses and the RCM in the lowest stresses. The TPM was most sensitive to changes in the ratio between crack opening and crack slip (relative displacement of the crack edges relative to each other in the direction parallel to the crack).
A comprehensive parameter study was conducted to investigate the effect of various material properties on the load-deformation behaviour according to the CMM-F for membrane elements subjected to monotonically increasing shear deformation. It was found that for small crack angles (relative to the strong reinforcement direction), the crack slip changes sign after the reinforcement in the weak reinforcement direction has exceeded the yield strength. This effect is more pronounced when compressive membrane forces are applied parallel to the strong reinforcement direction.
In order to avoid the iterative procedure for solving the CMM-F equations, it was investigated whether the concrete deformations could be neglected. When the concrete deformation is neglected, the crack opening deviates from the iterative solution by only 1%. The crack slip deviates from the iterative solution by up to 12% for small shear strains and by around 4% for large shear strains. For common problems, it seems justified to neglect concrete deformations.
For cyclic simulations, the shear stiffness in the zero-crossing range according to the CMM-F corresponds much better with the results of experiments than in simulations using CMM-R. The CMM-F assumes that there are two crack families, with the crack angle changing sign relative to the strong reinforcement direction. For simplification purposes only one crack family is assumed to be active. When transitioning from positive to negative shear strain, a change from one crack family to the other is assumed. However, this model assumption proved to be insufficient. It results in an offset of the shear force at zero crossing, which was not observed in experiments.
