Simplified micro-modeling of partially-grouted reinforced masonry shear walls with bed-joint reinforcement: Implementation and validation

Abstract

Partially grouted reinforced masonry (PG-RM) shear walls of hollow concrete blocks (HCB) have been an object of study during the last years. The non-constant cross-section of this type of structural element and the presence of reinforcement set a challenging scenario when assessing their lateral resistance. This scenario makes simple approaches (e.g., design expressions) lacking accuracy. Besides, the most accurate existent analysis methodologies rely on user-defined sub-routines that are not available for commercial use. Therefore, proper analysis methodologies are still a need. In this regard, this research aims at reproducing the behavior of PG-RM shear walls with bed-joint reinforcement with a simple but also accurate approach. In this line, the in-plane behavior of PG-RM shear walls was reproduced by implementing 2D micro-models in a multi-purpose commercial FE code without requiring excessive work, advanced programming skills, and unaffordable hardware. The model approach was validated by reproducing two identical full-scale PG-RM shear walls. Although the model was not able to reproduce cyclic loading as in the tests, the model captured the experimental failure mode and lateral resistance with an acceptable degree of accuracy. Moreover, the distribution of cracks and deformations in horizontal reinforcement elements were appropriately reproduced at the lateral resistance, indicating the most demanded reinforcement portions. Additionally, the proposed modeling approach was compared with two alternative approaches: a 2D model that reproduced tensile failure employing interfaces and a smeared crack model and a 3D model that reproduced tensile failure utilizing a smeared crack model. The benchmark results pointed out the advantages of the reference model over the alternative modeling approaches. The first alternative model reproduced an excessive displacement capacity, and the second alternative model simulated an inaccurate crack pattern and was associated with a heavy computational burden.