Research Of Prediction Method For Subgrade Reinforced CFG Pile-Slab On High Speed Railway Of Deep Soft Soil Zone

The settlement of subgrade usually is one of the most important controlling factors for the high-speed railway construction in deep soft soil. That is because the excessive settlement or uneven settlement will deteriorate the road condition, reduce the comfort of passengers, and even endanger traffic safety. Therefore, it is vital to get the settlement data of the road foundation. it ensures safety during construction. On the other hand, it forecasts the post-construction settlement correctly to control it under the permit of the range of the design. Hence, forecasting the settlement deformation is of significance to the actual engineering.At present, there are many forecasting methods for subgrade built in the deep soft soil. Each method has its shortcoming and virtues and its suitable condition. This paper analyzes measured settlement data of subgrade reinforced by CFG pile raft in the deep soft soil zone for pre-pressure more than five months after filling to find that the settlement is quite small and relatively large fluctuation. Whether the curve fitting method used widely in prediction of large settlement of subgrade on soft ground is suitable for this settlement condition is still a problem needing further research, therefore, combining the feature that ettlement of subgrade is quite small and relatively large fluctuation.The applicability of the curve fitting method has been studied through settlement prediction methods based on analyzing systematically and selecting optimally.The main work in the thesis is following:according to the field observation data of high speed railway, the prediction methods used in engineering such as hyperbolic method, exponential method, Asaoka method, grey theory method and improved expone tial method are selected to evaluate experimentally and the suitability of each prediction method is analyzed, then, the magnitude of prediction error and the effects different time starting point and diferent segment to prediction accuracy are studied by comparision of field observation data and prediction data.Because the exponential curve has strict requirements monotonicity for the observation data settlement, part data are ups and downs of small magnitude may not carry into the prediction calculation. Therefore, the exponential curve method has been improved by introducing the selecting points thought of three-point into exponential curve model, Analysis of engineering examples shows that the improved exponential curve method has good stability and requirement is after the inflection point to three-point selecting. The result of optimum comparison and selecting shows that improved exponetial method can reflect the settlement trend, with higher correlation coefficient, smaller prediction error and more sensitive to abnormal data, it can timely detect the problematic section with abnormal settlement trend, so improved exponential method is recommended as the optimal method for subgrade prediction of high speed railway. Asaoka method and grey theory method is higher correlation coefficient and stronger applicability, recommended as second-class method necssarily verfy the prediction reult of improved exponential method.