ESTIMATION OF MAXIMUM PEAK GROUND ACCELERATION VIA THE ANFIS AND RBF NEURAL NETWORKS

Abstract eng:
Peak ground acceleration is one of the most important factors which need to be investigated in order to predict the devastation potential resulting from earthquakes in reconstruction sites. Also, the maximum level of shaking control and the maximum load to which a structure is subjected to are criteria that can be worth considering. In this research, a training algorithm based on gradient descent (Traingd) and Levenberg-Marquart (Train LM) were developed and employed by using strong ground motion records. The ANN algorithm indicated that the fitting between the predicted PGA values by the networks and the observed PGA values were able to yield high correlation coefficients of 0.78 for PGA. We attempt to provide a suitable prediction of the large acceleration peak from ground gravity acceleration (1g) in different areas. Methods are defined by using fuzzy inference systems based on adaptive networks, feed-forward neural networks by four basic parameters as input variables which influence an earthquake in an area. The affected indices of an earthquake include the moment magnitude, rupture distance, fault mechanism and site class. The ANFIS network –– with an average error of 0.012 –– is a more precise network than FFBP neural networks. The FFBP network has a mean square error of 0.017 accordingly. Nonetheless, these two networks can have a suitable estimation of probable acceleration peaks (PGA) with levels higher than 1g in the area.

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