Font Size: a A A

Evaluation Of Loess Collapsibility For The Foundation Of A Simulation Dam

Posted on:2013-07-02Degree:MasterType:Thesis
Country:ChinaCandidate:X D FengFull Text:PDF
GTID:2232330395461163Subject:Geological Engineering
Abstract/Summary:PDF Full Text Request
Loess is a Quaternary sediments that widely distributed on the Loess Plateau. It is very thick and has specific engineering properties, of which the most significant is the collapsibility of loess that brought a great impact to the project construction with uneven settlement and dumping of the building, the roadbed damage. In the process of large-scale development of the western region, study on the collapsibility of loess, especially correctly evaluate the collapsibility of loess have some engineering and theoretical significance.In this paper, on the basis of studying the regional geological conditions of the Heping town of Lanzhou city, we find the topography of the region is relatively simple, the formation structure is simple, new tectonic movement and seismic activity is strongly active. Based on the study of the basic conditions of the site, get physical and mechanical parameters and coefficient of collapsibility through a lot of test, on the basis of statistical rock-soil parameter, evaluated collapsibility of loess and collapsibility of site.Through comprehensive analysis of impact factor of the collapsibility of loess, we find the impact factor is numerous, mainly involving density, void ratio, moisture content, external loads, plasticity index. Based on data mining technology, identified four main factors of the collapsibility of loess by using factor analysis method, the factors is dry density, moisture content, void ratio and plasticity index, and then evaluated the impact degree of each impact factor. By using gradually regression analysis method, the paper determined the regression equation of coefficient of collapsibility and individual factor, multiple factors. Last based on the regression equation and BP neural network establish prediction model for coefficient of collapsibility, according to the test results, the predicted error is small and precision is high, it can give some opinions for engineering construction.
Keywords/Search Tags:loess collapsibility, data mining, factor analysis, gradually regression analysismethod, BP neural network
PDF Full Text Request
Related items