Alireza Mortezaei - Visiting Scholar, School of Civil Engineering, the University of Queensland, Brisbane, Australia
Hamid Reza Ronagh - Senior Lecturer, School of Civil Engineering, the University of Queensland, Brisbane, Australia

The experiences from past earthquakes have shown that the regularity either in plan or in elevation is of crucial importance on earthquake performance of structures. Besides buildings that are irregular in plan,buildings with vertical irregularity have a paramount presence in the built infrastructure of Iran. Sudden changes in stiffness and strength between adjacent stories are associated with changes in structural system along the height, changes in story height, setbacks and changes in materials. The objective of this study is to investigate the adequacy of Artificial Neural Networks (ANN) to determine the dynamic response of vertically irregular reinforced concrete buildings in three dimensions subjected to ground motions. For this purpose, an ANN model is proposed to estimate the base shear forces, base bending moments and roof displacement of buildings in two directions. In the ANN model, a multilayer perceptron (MLP) with a back-propagation (BP) algorithm is employed using a scaled conjugate gradient. ANN model is developed, trained and tested in a MATLAB based program. A training set of 84 and a validation set of28 buildings are produced from dynamic response of vertically irregular RC buildings under the seismic forces. Finite Element Analysis (FEA) is used to generate training and testing set of ANN model. It is demonstrated that the neural network based approach is highly successful to determine the response of RC buildings subjected to ground motions.

Dynamic analysis; RC buildings; Artificial neural networks; Earthquake