Improving the Prediction of the Behaviour of Masonry Wall Panels Using Model Updating and Artificial Intelligence Techniques

Due to the highly anisotropic properties of mason/}' panels, it is very difficult to predict accurately their behaviour. In the past finite element analysis (FEA) micro and macro models have been used to improve the quality of the FEA to be closer to their experimental results, but still there is no established method(s) that swtably address this problem. Research m the University of Plymouth (UoP) introduced a methodology of model updating that uses the concept of Correctors and Cellular Automata combined with a nonlinear FEA, which more accurately predict the behaviour of masonry panels. This paper presents further refinements to the findings of the previous research by the team at UoP and uses evolutional}' computation and regression analysis methods. New correc/or values obtained by this method further improve the results of the non-linear FEA, which shows better agreement with the expenmental results. In this study fal1ure load and the load deflection relationships are studied for comparison.