C. SUI (1), M.Y. RAFIQ (2), G.ZHOU (3)
(1) Corresponding author, email: firstname.lastname@example.org. School of Civil Engineering, Queen’s University, Belfast, Northern Ireland, BT9 5AG
(2) School of Engineering, University of Plymouth, UK, PL4 8AA
(3) Harbin University of Technology, China, 150006
This paper introduces a method to get correctors by model updating, to further improve FEA of laterally loaded masonry wall panels. Static deflection data of the panels is used for model updating and the genetic algorithm (GA) is employed for the model updating process. Firstly the parameter of the existing FEA model is studied; secondly, a special objective function is carefully developed and then constraint functions are proposed to obtain convergence. Thirdly, the uniqueness of model updating is studied. The derived correctors by model updating indicated that the actual built?in edge of the masonry panel is not absolutely fixed as assumed in the FEA model, but appears to display some rotation; on the other hand, the simply supported edges are not in free rotation, as assumed, but are slightly constrained. The uniqueness study shows that the physical meaning of model updating is reliable. In the end the derived correctors are used to analyse other masonry panels and improved analyses are achieved.
Model updating, FEA model, Genetic algorithm, uniqueness, masonry panel, lateral loading