SOUSA, HIPÓLITO1; SOUSA, RUI2;

1) Professor, University of Porto – Faculty of Engineering, Department of Civil Engineering, hipolito@fe.up.pt

2) Researcher, University of Porto – Faculty of Engineering, Department of Civil Engineering, ruysousa@fe.up.pt

The characterization of the thermal behaviour of masonry can be performed using two methods. One method is through experimental testing in laboratory conditions, in which a relatively large masonry sample is mounted in a calibrated hot box. The other method is through numerical simulations based on the 2D or 3D finite element models.

The EN 1745 defines both methods to determine the thermal transmittance of masonry walls (or U value) for design purposes. Moreover, tabulated values can be used to determine the thermal resistance of masonry made from the most common types of materials and shapes. However, the units must have a similar geometry, otherwise laboratory tests or numerical simulations must be performed.

The experimental method is more accurate but is more expensive than the numerical method, since it requires a complex test set-up and the sampling of masonry is difficult. The consideration of these aspects is important in parametric studies based only on experimental method, because it may require the use of a high number of test samples. In these cases the numerical simulations are preferable, but must be validated since differences between experimental and numerical results are always present.

In this paper a numerical sensitivity analysis of the thermal resistance of a masonry system is presented. This analysis was applied to a concrete masonry system tested in a laboratory in order to evaluate the influence of the geometrical properties of the constituent materials. The numerical simulations are based on 3D FE model of laboratory tests and available test data to validate the numerical model. The main results are presented and discussed.

Keywords: Sensitivity analysis, thermal behaviour, masonry