Font Size: a A A

Predicting Permanent Settlement Of Concrete Faced Rockfill Dams Using RBF Network

Posted on:2016-01-24Degree:MasterType:Thesis
Country:ChinaCandidate:S X HanFull Text:PDF
GTID:2272330461478933Subject:Structure engineering
Abstract/Summary:PDF Full Text Request
Concrete faced rockfill dams(CFRD) are an applicable type of the earth-rock dams. As water resources to undertake the strategy of large-scale development in our country, the seismic stability, security and other issues on high CFRD will be particularly important and prominent. Earthquakes belong to natural disasters, which caused serious economic loss, environmental pollution and casualties of life, etc. The prediction of the seismic damage, therefore, is a critical issue for us. The permanent settlement is one of the main seismic damage types of CFRD. We perform research on methods for predicting the crest permanent settlement, the main work and conclusions are as follows:1. The RBF network model is used for predicting the static settlement of CFRD. The RBF network model is established according to the static settlement including completion period and storage period of actual engineered CFRD. Samples are divided into training samples and testing samples. Finally, Shuibuya design data as a forecast sample for forecasting analysis.2. We propose a method based on RBF network for predicting crest permanent settlement of CFRD. Using three dimensional finite element methods,240 kinds of working conditions are calculated as samples. The dam highlights considered are 100 m, 150 m,200 m and 250 m. The adopted four seismic waves are El Centro seismic wave, Taft seismic wave, Kobe seismic wave and artificial wave, and three magnitudes considered are 7,8 and 9. Three aspect ratios considered are 1:1,2:1 and 3:1. These samples are divided as training samples and testing samples respectively to establish a RBF network model. Experimental results verified the feasibility of the RBF network model to predict the permanent deformation of CFRD. Take Zipingpu CFRD as an example, seismic permanent deformation predicted by the proposed model is compared with the measured deformation. Results show that the RBF network model is accurate, convenient and effective in predicting static and crest permanent settlement of CFRD.3. By analyzing a variety of ideal CFRD, the effects of dam height, aspect ratio, water depth ratio, average density of rockfill materials, peak acceleration and different seismic waves spectrum characteristics were focused on crest permanent settlement of CFRD.
Keywords/Search Tags:CFRD, radial basis function networks, permanent settlement, predicting
PDF Full Text Request
Related items