Estimation of the axial behaviour of masonry walls based on Artificial Neural Networks

•Masonry walls made of cement mortar and clay bricks are investigated.•Artificial Neural Networks are used.•The network is trained with around 1950 data items.•An estimation for the bearing capacity of brick masonry is obtained.

Estimating the load-bearing capacity of brick masonry walls is a fundamental aspect of the design or retrofitting of this type of structures. This paper presents a new ANN-based proposal as an alternative to the different existing methods. The proposal takes into account load eccentricity, wall slenderness ratio and stiffness and masonry tensile strength, and is validated by a comparison with the Eurocode 6 and other formulations as well as three other experimental studies. The proposal closely agrees with the experimental results and is less conservative than Eurocode 6 and therefore more likely to provide the optimum design for masonry walls.