Settlement Prediction Of Embankment With CFG Pile Composite Foundation In Peking-Shanghai High-speed Railway

With the rapid development of high-speed railway, the quality of railway subgrade stability has been demand more stringent. However, the settlement of high-speed railway subgrade is an important aspect for stability evaluation, especially the factory after the settlement control is in very strict control. This paper research the CFG pile composite foundation settlement of subgrade about the Peking-Shanghai high speed railway, analyzing the theory of soft soil foundation deformation and the concept of post-construction settlement railway roadbed; Focus on subsidence prediction, using the former engineering measured data to predict the settlement rules; Because subgrade settlement scale of the high-speed railway track is small and fluctuations, in order to improve the traditional hyperbola method, exponential curve method, combining with three point method to solve two kinds of curve model. This paper is based on engineering settlement observation data of the Peking-Shanghai high-speed railway track DK511+787～DK513+050sections. The subsidence prediction results of DK512+625section show that the improved hyperbolic method selection at three times interval is larger, and the accuracy of subsidence prediction is better. Improving the initial time point selection in the exponential curve method before the subgrade filling is completed, the accuracy of subsidence prediction is better by selecting a larger time interval. The original data in the GM (1,1) grey prediction method should be chosen before the roadbed filling is completed, and the original data sequence selecting is in a larger time interval (the proposed interval is bigger than28d).The original sequence number only five can achieve good prediction results; The all conclusions above were applied to the other section of the segment, and the results showed that the three kinds of prediction methods have good applicability in the segment; The final settlement data using the GM (1,1) grey prediction method is between using the improved hyperbolic method and using the exponential curve method. The prediction accuracy using the GM (1,1) grey prediction model is higher than using the segment modified hyperbola method and using the exponential curve method, and there is a certain reference value about subsidence prediction for future similar projects.