Research On Settlement Characteristics And Prediction Of Collapsible Loess Subgrade

In the Ningxia area,the collapsible loess is widely distributed due to the special engineering properties of the loess,that is,the loess is under the action of the self-stress of the overburden,or under the action of the self-stress and the additional stress,There is a significant additional deformation,resulting in the majority of road subgrade in the region have uneven settlement,the stability of the roadbed in the region caused a huge impact,which seriously affected the safety and comfort of the road,while greatly increasing the cost of road maintenance and Difficulty.Based on the research on the key technology of automatic monitoring and treatment of the collapsible loess roadbed in Ningxia,the paper analyzes the settlement characteristics of the collapsible loess roadbed with the G309 line Guyuan-Xigi section expressway project as the research object research.The main research work and conclusions are as follows:(1)Experimental study on collapsibility of collapsible loess in Guyuan,Ningxia.The water content,density,porosity ratio,saturation,liquid limit and plastic limit,compression coefficient,compressive modulus and collapsibility coefficient of the collapsible loess in Guyuan area were studied.The results show that the water content is more than 23%,the density is more than 1.7,the pore ratio is less than 1,the saturation distribution is more than 60%,the compressive coefficient is less than 0.2,the loess is not collapsible;15% ~ 23%,the density distribution is between 1.4 and 1.7,the pore ratio is between 1 and 1.2,the saturation distribution is between 35 and 60%,the compressive coefficient is between 0.2 and 0.4,Collapsibility;moisture content is less than 15%,density distribution is less than 1.4,pore ratio distribution is greater than 1.2,saturation distribution is less than 35%,compression coefficient distribution is greater than 0.4,loess has strong collapsibility.(2)Numerical Simulation of Treatment Effect of Compacted Pile in Collapsible Loess Foundation and Experimental Study of Engineering ExampleBased on the FLAC-3D platform,the numerical model of the hole expansion of the collapsible loess foundation was established,and the compaction process of 30 cm,40cm,50 cm and 60 cm with different pile diameters was simulated.The displacement and stress cloud,And the field test of the compaction pile is carried out.The results show that the effect of the single pile of the compacted piles is different,and the squeeze level of the pile is different.The range of the crowded zone is 0 ~ 0.8d,the effective tightening zone is 0.8 ~ 2d,The range of the range is 2 ~ 3.5d and the noncompacted zone is more than 3.5d.When the compacted piles are compacted,the soil will rise upwards and the vertical displacement will increase first and then decrease,1.0d,the vertical displacement at the depth of 0.4 times the pile is almost zero,that is,the soil has no vertical displacement at this time,and the vertical displacement reaches the maximum value in the pile bottom section.(3)On-site monitoring of collapsible loess foundation of Guxi Expressway in NingxiaIn this paper,on-site monitoring of collapsible loess subgrade with profile subsidence pipe,water meter and stratified sedimentation meter is carried out.The data of subsidence monitoring show that the four sections are basically from September 26,2015 to October 8,2015,10 From December 27 to December 20,the roadbed settlement rate is larger(embankment filling construction stage),from October 8,2015 to October 27,2015,December 20 to February 10,2016,roadbed settlement rate(The embankment construction is completed),the embankment filling thickness is the main settlement rate of the embankment,and the thickness of the embankment is lower than that of the embankment.Influencing factors,subgrade construction is completed under natural conditions,the subgrade rate is small and stabilized.(4)Study on Settlement Prediction Model of Collapsible Loess Foundation in Qixi Expressway in Ningxia.Based on a large number of road surface subsidence monitoring data,four kinds of curve fitting method and BP neural network model are used to predict the settlement of four sections of subgrade.Based on the comparison with the measured values,the prediction model of collapsible loess roadbed settlement Adaptability to study.The results show that the four curve fitting methods can predict the settlement of the roadbed in a certain period of time based on the settlement monitoring data.The BP neural network can forecast the settlement of the collapsible loess roadbed with high precision based on the large amount of monitoring data,Which is an accurate and reliable method for predicting curve fitting.