ABSTRACT
ABSTRACT This paper analyzes the precision of the Artificial Neural Networks ANN prediction for a set of 3D building models subjected to various seismic actions. The ANN was trained based on the results of the Nonlinear Time-History Analysis performed with SAP2000 v14.1.0 for 243 reinforced concrete frame structures. The finite element models were obtained by varying six structural parameters. The database of seismic actions used in the ANN training includes 14 seismic recordings with different Peak Ground Accelerations PGA, each characterized by four parameters. The analysis presented in this paper implies modifying some input parameters by selecting values different than the ones used in the ANN training, in order to test the ANN accuracy. The accuracy is tested by comparing the results of the Finite Element Analysis FEA, representing the desired results, with the output parameters given by the ANN.
