NEURAL NETWORKS FOR QUICK EARTHQUAKE DAMAGE ESTIMATION

This paper proposes the use of neural networks to predict damage due t o earthquakes from the indices of recorded ground motion. Since the re lationship between ground motion indices and resulting damage is diffi cult to express in mathematical form, neural networks are conveniently applied for this problem. Simulated earthquake ground motions are use d to have a well-distributed data set and the ductility factor from no n-linear analysis of two single-degree-of-freedom structural models is used to represent the damage. A sensitivity analysis procedure is des cribed to identify qualitatively the input parameters that have a grea ter influence on the damage. The result of the trained neural network is then verified by using several recorded earthquake ground motions. It is found that some instability in the prediction can occur. Instabi lity occurs when input values exceed the range of the training data. T he neural network model using PGA and SI as input give the best perfor mance in the recall tests using actual earthquake ground motion, demon strating the usefulness of neural network models for the quick estimat ion of damage through earthquake intensity monitoring.