Size effect of squat shear walls extrapolated by microplane model M7

Recent earthquakes revealed poor performance of very tall shear walls. This is no surprise because the design for size effect has long been hampered by the lack of large-size tests and by mingling of different concretes and different test parameters among the existing test data. Fortunately, recent progress in material modeling and computer power permits overcoming this obstacle by extending the data to large sizes computationally. Herein, a large classical set of reduced-scale shear wall tests performed (at ETH) are selected and used to verify and calibrate the finite element simulations with the crack band model, in which a powerful constitutive damage model-microplane model M7-is implemented. Then the calibrated computer code is used to predict the strength and ductility of much larger shear walls. As expected, the size effect is found to occur if the shear in concrete controls the failure load but not if the steel bar yielding does. Increasing the reinforcement ratio may compromise ductility; it may cause the shear in concrete to control the peak load and the failure to occur before the steel bars begin to yield. Increasing the height-to-width aspect ratio of the wall is shown to lead to flexural failure, which does not show size effect on the shear wall strength. Increasing the wall size may change the failure mechanism, cause the concrete to fail at a lower stress, and shorten or remove the yield plateau. Adding horizontal reinforcing bars to vertical ones tends to prevent inclined shear cracks in the web, although it has little effect on sliding shear. Computer simulations of the present type are helpful for checking the design of large shear walls.