Study on Fast Prediction Model of Seismic Economic Loss Based on GA-ANN Macroscopic Vulnerability Method

Abstract:

Article Preview

It is well-known that seismic disaster will cause serious damage, so the prediction and evaluation of seismic loss before earthquake event happened can provide foundation of disaster reduction program. And after seismic event happened, it is very important to fast evaluate the seismic disaster loss. Traditional seismic loss prediction method is vulnerability list method which need detailed information about all kinds of structures and facilities of the evaluated areas, and for most areas, it is very difficult to get such detailed information. This paper brings forward a new simple fast prediction method for seismic economic loss. Firstly, a macroscopic vulnerability model was discussed. secondly, a three-layer BP network model for seismic economic loss prediction is built up, BP network model is one of the most widely used artificial neural network (ANN) model, but due to its deficiencies such as easy to get into local extreme minimum value, genetic algorithm (GA) is introduced to overcome the deficiencies. In this model, seismic intensity, average GDP per capita and per area, population density are selected to be the input layer index, and GDP loss ratio as the output layer, 80 earthquake events which happened recent years are regarded as training and check up samples. Finally, the economic loss caused by Wenchuan earthquake is evaluated by the proposed model, and the feasibility and practicability are validated by the numerical example.

Info:

Periodical:

Advanced Materials Research (Volumes 243-249)

Edited by:

Chaohe Chen, Yong Huang and Guangfan Li

Pages:

5106-5110

DOI:

10.4028/www.scientific.net/AMR.243-249.5106

Citation:

J. Li et al., "Study on Fast Prediction Model of Seismic Economic Loss Based on GA-ANN Macroscopic Vulnerability Method", Advanced Materials Research, Vols. 243-249, pp. 5106-5110, 2011

Online since:

May 2011

Export:

Price:

$35.00

In order to see related information, you need to Login.

In order to see related information, you need to Login.