Identification of Brick-Masonry Material Properties Through Inverse Analysis and Genetic Algorithms

Unreinforced brick/block masonry (URM) has been used for centuries as an effective building material. However, the response of URM is very complex because of its inherent heterogeneity and nonlinear behaviour, which is governed by the interaction between units and mortar. Mesoscale modelling, though computationally demanding, can provide a very good representation of the actual structural response when using adequate mechanical parameters for URM component materials. These can be obtained using numerical calibration based on the results of experimental tests. In this paper, inverse analysis techniques utilising genetic algorithms are employed to calibrate material parameters of an advanced nonlinear mesoscale description, which uses zero-thickness interfaces for representing mortar joints. In particular, the elastic material parameters of mortar interfaces are derived from measurements at the macroscale. In order to apply this procedure to in-situ nondestructive tests, a non-conventional flat-jack test setup has been investigated. The potential and limitations of the proposed method are assessed using computergenerated pseudo-experimental data, where modelling errors are ruled out. Sensitivity and random noise analysis are performed to evaluate the influence of the precision of the measurement equipment employed in the tests.