A Neural Network-Based Methodology for Generating Spectrum -Compatible Earthquake Accelerograms

A new neural network based methodology for generating artificial earthquake spectrum compatible accelerograms is proposed using the learning capabilities of neural networks to develop the knowledge of the inverse mapping directly from the response spectra to earthquake accelerograms. A two-stage approach is used. In the first stage, a replicator neural network is used as a data compression tool. The replicator neural network compresses the vector of the discrete Fourier spectra of the accelerograms to vectors of much smaller dimension. In the second stage, a multi-layer feed-forward neural network learns to relate the response spectrum to the compressed Fourier spectrum. A simple example, in which thirty accelerograms are used to train the two-stage neural network, is presented to demonstrate how the method works. This two-stage neural network methodology is further extended and enhanced. Two types of new stochastic neural networks that are capable of generating multiple earthquake accelerograms from a single response spectrum are presented. The stochastic neural networks have been combined with a new data compression scheme using replicator neural networks. A benefit of this extended neural network methodology is gaining efficiency in compressing the earthquake accelerograms and extracting their characteristics. The proposed method produces a stochastic ensemble of earthquake accelerograms, from any single response spectrum or design spectrum. An example is presented that uses one hundred recorded accelerograms for training the neural network and several design/response spectra for testing. The methodology presented in this research is promising and full of potential for generating multiple spectrum compatible earthquake accelerograms and solving one-to-many inverse mapping problems.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Civil Engineering

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2015

Author dc:creator

• Lin, Chu-Chieh Jay

Contributors dc:contributor

• Ghaboussi, Jamshid

• Artificial Intelligence

Language dc:language

eng

Identifier

(MiAaPQ)AAI9955643

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/83497