O.ZARRIN(1), Y.Z. TOTOEV(2) and M. RAMEZANSHIRAZI(3)
(1)PhD Candidate, Centre for Infrastructure Performance and Reliability, The University of Newcastle, Callaghan,NSW 2308, Australia, firstname.lastname@example.org
(2)Senior Lecturer, Centre for Infrastructure Performance and Reliability, The University of Newcastle, Callaghan,NSW 2308, Australia, Yuri.Totoev@newcastle.edu.au
(3)PhD Candidate, Sapienza University of Rome, Italy, email@example.com
ABSTRACT The Semi-Interlocking Masonry (SIM) system has been developed by the Masonry Research Group at the University of Newcastle, Australia. The main purpose of this system is to enhance the seismic resistance of framed structures with masonry panels. In this system, SIM panels dissipate energy through the sliding friction between rows of SIM units during earthquake excitation. This paper aimed to find the applicability of artificial neural networks (ANN) to predict the displacement behaviour of the SIM panel under out-of-plane loading. The general concept of ANN requires them to be trained by related force-displacement data of SIM panels. The overall data for the training and testing of the network was 70 increments of force-displacement from three tests, which comprise of nine input nodes, these being height and length of panels, height, length and width of the brick, friction, geometry angle of brick, compressive strength of the brick with the lateral load. The aim of this study is to predict the displacement of the SIM panel using the Radial Basis Function network (RBF). The mean square error (MSE) of the network was 0.00032 and the coefficient of determination (R2) values showed as 0.96. The result revealed that ANN has significant potential to predict the behaviour of SIM panels.
KEYWORDS: Semi Interlocking Masonry (SIM), artificial neural network (ANN), Radial Basis Function (RBF), Neuron, Displacement