Affiliations: Center of Excellence for Fundamental Studies in Structural Engineering, School of Civil Engineering, Iran University of Science & Technology, Narmak, Tehran, Iran | School of Civil Engineering, Iran University of Science & Technology, Narmak, Tehran, Iran | College of Civil Engineering, Babol Noshirvani University of Technology, Babol, Iran
Note: [] Corresponding author. G. Ghodrati Amiri, Center of Excellence for Fundamental Studies in Structural Engineering, School of Civil Engineering, Iran University of Science & Technology, Narmak, P. O. Box 6765-163, Tehran, Iran. E-mails: ghodrati@iust.ac.ir (G.G. Amiri), khorasani@iust.ac.ir (M. Khorasani), s_aghajari@iust.ac.ir (S. Aghajari), zahra_tabrizian@stu.nit.ac.ir (Z. Tabrizian).
Abstract: Based on Adaptive Neural Network Fuzzy Inference System (ANFIS) networks, this paper presents a novel approach to generate artificial earthquake accelerograms from available data, which are compatible with specified design or response spectra. The proposed procedure uses the learning abilities of ANFIS networks as a powerful tool to develop the knowledge of the inverse mapping from response spectrum to earthquake records. Furthermore, to obtain better simulation results, Wavelet Packet Transform (WPT) and Principle Component Analysis (PCA) are used to convert records and response spectra from real to transformed spaces. Then, ANFISs are trained to relate response spectrum of records to their wavelet packet coefficients. In this process, the same results of different training levels of ANFIS method are obtained. In order to clarify the efficiency and accuracy of the proposed method, the results have been compared with the outcomes of previous artificial earthquake accelerograms generation methods. Finally, several interpretive examples are provided to demonstrate success of the suggested method.
