The Settlement Influencing Factors Study And Gray Prediction On Soft Soil Ground Of Expressway

In recent years, with the great development of coastal building, it is necessary to build more expressways. There often exists large area of deep soft soils in Pearl River Delta. Due to the character of soft soil, such as high compressibility,bulky water content, low strength and bad hydraulic permeability, the first problem is the stability and deformation. Especially the strictly limited pavement deformation, control of excessive settlement of soft soil ground becomes the key engineering technique problem. As a result, the precision of the embankment on soft soil ground settlement predication is exactly higher. Based on the influencing factors of expressway soft soil roadbed and the settlement predication problem, the paper makes some research work.In this paper, it has recommended the engineering properties of soft soil roadbed at first. According to the multianalysis of expressway in-site observation information and geologic information in Pearl River Delta area, summarizing the characteristics, influencing factors and decreasing law of settlement. It has directly proportional relation between landscape settlement and loading. The consolidation developing of soft soil ground is from surface layer to deep layer, and the settlement of each depth layer become decrease with time. Based on the study of soft soil structure, the sand layer in the soft soil could accelerate the soil consolidation. According to classifying the sand layer and comparing the settlement of different section, it is proved that the sand layer could decrease the back-work settlement of soft soil roadbed at a certain degree when the soft soil is not very deep.Explore the application of gray system theory on the prediction of settlement in this paper. Due to the character of unequal time-interval of settlement observation data in engineering practice, making the data interpolated with spline-interpolate and interpl, then transforming the data series to equal time-interval. Building gray model and predicting settlement with the unequal time-interval data. In order to increase the precision of prediction, it has presented three methods to improve the gray prediction; they are original data series rebuilding, model improving and prediction method improving. By means of calculating these improved models and checking the precision of prediction models by completing the coding with MATLAB, searching for the best fit prediction model with the settlement character of soft soil ground in study area.At last, several curve fitting methods are used to deal with the in-site observation data and to predict the roadbed; then calculating the final settlement of roadbed section. Making the adaptability analyses of different kinds of curve prediction method with the totality mean variance method and partial mean variance method; then select out the best curve model of settlement prediction for study area.