Nonlinear Behavior Identification of HDR-S Bearing Using Neural Network for Seismic Structural Design

The initial parameter selection is the bottleneck of optimization method in determining the nonlinear parameter of seismic isolators during seismic isolation design. Bilinear model is easy to understand physically but more complicated nonlinear models are hard to explain due to hidden characteristics. Therefore, this study used a machine learning based approach which aims to classify the suitable nonlinear model and predict the nonlinear parameters of an HDR-S experiment data at both low and room temperature. The trained neural network model A shows that at low amplitude, Bilinear Model was classified, however at higher amplitude, Modified Park-Wen model governs. On the other hand, neural network model B successfully predicted the five parameters of Modified Park-Wen model and solve the initial parameter assumption problem of KH Method. The proposed inverse approach can be used to train an ANN model using more complicated nonlinear models.

Die Urheberrechte (Copyright) für dieses Werk sind rechtlich geschützt. Es darf nicht ohne die Zustimmung des Autors/der Autorin oder Rechteinhabers/-in weiter benutzt werden.