Research On The Settlement Prediction Of Soft Foundation Based On Combination Model In Desert Region

With the rapid development of our economic construction, the speed of the highway construction is also increasing in the northwest China. Since the size of our country land is vast, the areas of the topography are complex along the road route. The distribution of soft soil is very broad, so some swamp weak foundation areas are often encountered in the process of the highway construction in Inner Mongolia desert region. The strength of the soft soil is low and the speed of the consolidation is slow in the swamp areas. It may cause the settlement easily because of foundation instability under the load. So the post-construction settlement has a bad effect on the stability of the roadbed structure and the life of the road pavement.In the process of the highway foundation embankment construction or after the completed filling, the soft soil foundation may cause much larger settlement and residual settlement. In particular, the enormous flow of vehicles is very sensitive to the settlement in the high-grade highways, and it needs post-construction settlement not exceed the canonical value. So it has raised a higher technical requirement to the settlement prediction of the soft soil roadbed in the highway construction. The stability of the foundation settlement is a complex problem. Up to now, many common methods are used in the settlement prediction of soft soil foundation also have some limitations and forecasting errors. It leads to a lower predictive accuracy. There has been little research on the theory of the settlement of soft soil foundation in the special project geological conditions of the desert swamp regions. Therefore, it has an important significance in the theory and engineering practice to solve the issue of the soft soil foundation settlement as well as to forecast precisely in this region.This thesis has stated the project characteristics of the soft soil and the settlement distortion mechanism of the foundation as well as its influencing factors at first. It sets the soft soil foundation settlement of the highway from Wushenzhao Eco-industrial Park (Wu) to Galutu (Ga) as an engineering background in Inner Mongolia desert region, and analyzes the particularity of the engineering geological conditions in the desert swamp area. It also discusses the comprehensive reinforcement treatments which can reduce the settlement of the soft soil foundation as well as its monitoring technology. The soft soil settlement is associated with soil properties, subgrade stress history of soil, roadbed filling height, groundwater level changes, lateral displacement, permeable foundation, the relevant factors of load and the loading rate. There are mainly with silty-fine sand, silty clay, silty soil and humus soil or peat, and fine grained soil in the soft soil desert swamp area. The rate of consolidation settlement of soft soil subgrade condition of the special geological condition is very slow and the bearing capacity is low. It can effectively reduce the settlement of the soft soil subgrade by setting the gravel pile, placing the gravel cushion and geogrid, and increasing the number of back pressure protects comprehensive strengthening treatment technology.Secondly, it summarizes the predictive power of the single prediction models, as well as its advantages and disadvantages, according to the settlement monitoring data of the typical observation section K35+270in the soft soil foundation sections of Wu Ga highway which locates in Inner Mongolia desert region, by using three single prediction model which are Grey Model (1,1), Hyperbolic fitting and Poisson curve fitting for the prediction of the soft soil foundation settlement. There are three forecast methods which can be used to predict the settlement of the soft soil foundation. For the effects of the settlement prediction over time, Grey Model (1,1) is the best, followed by the method of Poisson curve, but the prediction accuracy of the two methods need to be improved.Ultimately, the author discusses the theory of the combination forecasting and the method to establish model. It chooses Grey Model (1,1) and Poisson curve model are chosen to compose the combination forecasting models which assign different weights to the single forecasting methods. The author analyzes the model prediction effects of several different optimal and non-optimal combination forecasting models whose weights are determined by different methods comparing with submodels. The author researches the combination forecasting model which is suitable for the settlement prediction of the soft soil foundation in desert swamp region. The research has shown that the combination forecasting model reduces the systematic error of the individual prediction method and its general effect is superior to the single prediction methods. The prediction accuracy of this mode land’s data reliability is higher than the two single forecasting models.