BP Neural Network Computation of Response Spectra Using a Subpicture Idea

The paper deals with an application of neural networks for computation of response spectra on the fourth floor of a prefabricated medium height building under seismic-type excitations. The subpicture idea from picture transmission is also adapted to this problem. Both the training and testing patterns were formulated on the basis of measurements performed on an actual structure. Neurally evaluated spectra were compared with spectra computed on the basis of experimental data. The main advantage of the neural approach with the subpicture idea is that the prediction of building response spectra can be performed using very simple neural networks and without analysis of structural motion equations.