Prediction Of Cracking Mode Of Masonry Wall Panel With Opening Based On Exprimental Record And Intelligence Technology

For a long time, much research on the performance of masonry wall panels subjected to in-plane loads has been done from either experimental techniques or theoretical approaches. However, due to the complexity of the working behavior of masonry structure and the limitation of the current traditional analysis technology, it is difficult to accurately analyze the masonry wall panels with door and window openings, especially under laterally loads. As a result, little relevant research has been performed so far. At present, the research methods concerning masonry wall panels are mainly mechanistic study on basis of many a presumption. Nevertheless, resulting from the heterogeneous and anisotropy property of masonry materials, constitutive relation and failure criterion of the masonry materials have always been the difficult parts in the investigation of cracking mode of masonry wall panels. And the presumptions mentioned above lead to inaccurately evaluation of the relationship between the variability of masonry materials and performance of masonry wall panels and the high invalidity of traditional numerical analysis methods. Therefore, the study carries out the following tasks:First of all, the cellular automata (CA) model of masonry wall panels with opening is established. Based on the existing CA model of the real masonry wall panels, the newly proposed CA model is to reflect the influence of opening on the structural performance. Thus, it broadens application of the existing CA model in predicting the cracking mode of masonry wall panels.Next, a criterion for matching zone similarity is proposed based on maximum correlation coefficient. Since that existing criterion for matching zone similarity is all based on minimum error and the resolution of cellular state value decreases as transition coefficient value increases, which results in the failure of minimum error method. Hence, the criterion for matching zone similarity on basis of maximum correlation coefficient is provided. Numerical examples show that the criterion based on maximum correlation coefficient performs much better than that based on minimum error, when predicting the cracking mode of the solid masonry walls.Finally, by adopting the artificial intelligence structural analysis approach developed by Zhou et al, a new method based on the finite element (FEA) analysis and CA is proposed to numerically describe the state at local zones within the panel in sense of both configuration and loading case. This new CA numerical mode of masonry wall panels is based on the normalized displacement of each of the corresponding elements given by the simplified FEA of masonry wall panels with walls but also accounts for the influence of load cases. Therefore, the numerical mode is more suitable to predict the cracking mode of masonry wall panel with opening than the existing CA numerical mode.