Microzonation Study For Area With Liquefaction Potential Using Artificial Neural Networks Available to Purchase

The growth of urban population and prosperity of local industries had great effect on land use. Environmental sensitive areas were developed without regard to their carrying capacity due to the highly demand of land. Then, the intensive and improper land use caused disasters. Therefore, it's important to probe into land development suitability by delimiting of microzonation of environmental sensitive areas. Extensive researches about the basic causes of liquefaction and the evaluation associated with corresponding mapping of liquefaction potential have been carried out. In general, if the shear resistance of soil is considered alone in the model, the prediction compares unfavorably to in situ liquefaction areas. Therefore, data sets from 1999 Chi-Chi Earthquake, viz. liquefaction occurances, and Yuan-Lin regional geological and intensity distribution data, were used to train and test the proposed model of back propagation neural networks. The input of the model included soil properties, environmental factors, and earthquake characteristics. An attempt was made to establish microzonation maps based on liquefaction potential for Yuan-Lin using discretization and IDW method in analysis module of geographical information system and the function of spatial analysis of ArcGIS. The field data were used to perform the verification of the model. The maps might be useful for planning of land use in that local area. MODEL CONSTRUCTION AND ANALYSIS The soil property data included in the model are from field test report for Chi-Chi earthquake (MAA Consultant Engineers, 2000), and they are earthquake magnitude (M), peak ground acceleration (a_max), soil fines content (FC), and standard-penetration-test blow count (N).