Gobi Desert Soil Foundation Settlement Method And Prediction Research

In pace with the development of economic of our country, the process of engineering construction in west area where Gobi was distributed widely has been accelerating markedly. So far there are short of researches about the physical-mechanical properties of Gobi desert soil how to strengthen it and predict its sedimentation exactly after course of construction but focus on how to develop and use Gobi desert soil. Therefore it is important for the construction of highway and railway in these Gobi areas and also roadbed problems and remedial measures and treatments to research the problems as above mentioned.By taking the second tow-lane roadbed section of Lan-xin Railway as test section, in-situ test and numerical analysis method are applied respectively to research the property of Gobi desert soil foundation stabilization and prediction of subgrade. The achievements are as follows:(1)Experimental methods such as in-situ test were used to explore the physical-mechanical properties of Gobi desert soil.(2)Based on load test and dynamic sounding test heavy-tamping method, the best strengthen method for Gobi foundation, were singled out from several layouts of reinforcement.(3)Studied four settlement prediction models which are hyperbolic method, Hoshino method, Asaoka, GM(1,1)model. And analysis the effects of this four models in different time starting point to forecast results’relative error and correlation coefficient.(4)Relying on the Lanxin railway embankment test section of the second double road, using different forecasting methods, study the Gobi roadbed settlement of prediction methods. Predict post-construction of the Gobi roadbed, and analyse the impact about different time point on the predicted results. The results show that hyperbolic method is not suitable for settlement prediction in the Gobi roadbed. Because the Gobi roadbed’s subside instability and the hyperbolic method is sensitive to abnormal data. Asaoka method have a poor result for the initial irregular settlement data. The results of Hushino method and GM(1,1) model are very well and the relative prediction error is small. So they are appropriate to the Gobi roadbed.