Strong Ground Motion Duration and Response Spectra Using Artificial Neural Networks

Abstract eng:
Artificial neural networks (ANN) were used to estimate strong ground motion duration and response spectra using accelerograms recorded in and around the Mexican cities of Puebla and Oaxaca. These networks were developed using a back propagation algorithm and multi-layer feed-forward architecture in the training stage. For strong ground motion duration, we integrate data considering that the phenomenon is characterized by seismic magnitude, epicenter distance, site period and azimuth. Results were compared with those obtained from the Arias method and Reinoso&Ordaz equation. Regarding to response spectra, besides the previous parameters we also considered a vector of spectral amplitudes. In order to evaluate the forecasting capacity of the ANN strong ground motion duration and response spectra were estimated from earthquakes whose data were not included in the training phase. An acceptable concordance is observed between them and those provided by the ANN. Overall, the results presented show that ANN provide good and reasonable estimates of strong ground motion duration and response spectra in each one of the three orthogonal components of the accelerograms recorded in the cities of Puebla and Oaxaca. Furthermore, the networks have a good predictive capacity to estimate duration and response spectra.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket