Application of stiffness/strength corrector and cellular automata in predicting response of laterally loaded masonry panels
This research has introduced a new concept, 'stiffness/strength corrector', which more accurately models variation in masonry properties at various locations (zones) within a masonry wall panel. Derivation of these correctors was based on a closer mapping of the laboratory experimental results to those obtained from a non-linear finite element analysis of full-scale masonry panels subjected to a uniformly distributed lateral load. In this research only one panel, which was tested in a previous research, was used as the "base panel" and correctors for new panels with and without openings with various boundary conditions were derived by matching similar regions and zones between the new panel and the base panel. The research has also derived the concept of zone similarity between the base panel and any new panel. It was discovered that the types of panel boundaries surrounding specific regions within the two panels govern zone similarity. At first, a manual method for matching zone similarity was proposed based on careful visual inspection to identify similar regions within the two panels. It was found that this method is difficult to implement as the user needs to have a deep knowledge of the behaviour of the panel to be able to accurately locate similar regions/zones. As it was established that the zone similarity was mainly related to the panel boundaries, this knowledge was used to derive appropriate rules for matching zone similarity. These rules were implemented in a cellular automata model which was able to automatically locate similar zones between the base panel and a new panel and assign appropriate corrector values to zones within the new panel. The stiffness/strength corrector values were used to, modify global material properties of the panel. A specialised non-linear FEA program for masonry panels was used to analyse a number of panels provided by CERAM with modified rigidities or tensile strength values. Comparison of results with laboratory experimental values shows that with this new method an average 18% improvement in the prediction of failure load, in comparison with the non-linear FEA results with smeared masonry properties, was possible. The failure patterns for the majority of panels with or without openings, having various sizes and boundary conditions, were much closer to the experimental results. The results of case studies using the new method clearly show that the proposed method is a much better representation of the true behaviour of the masonry panels which models variation in masonry properties and the boundary effects more accurately. The corrector values for any type of new panel are derived from a single base panel in which there was not sufficient data available at different locations on the panel, particularly near the panel boundaries. Thus, in some cases it uses a crude approximation of the boundary types to establish corrector values for a new panel. If sufficient data points were available more accurate results would have been possible to achieve.