Loess is a kind of quaternary sediment formed in arid and semi-arid conditions. The distribution area, burial depth and formation integrity of loess in our country are the highest in the world. Loess in native state has high strength and low compressibility,but strength fell sharply after immersion.Collapsible loess’ s original structure will be destroyed under water and additional load, causing collapse deformation. Collapsible loess caused a lot of geological and engineering disasters, which has brought huge losses to loess area economic construction.Therefore, the study of loess collapse deformation law is particularly important.In thepast, collapsible loess research was focused on fully immersed saturated collapse deformation, but payed less attention to humidification deformation of humidified to a moisture content case. In the article, based on the analysis of reading a lot of literature,it summarizes the main causes of collapsible loess.Based on Xinjiang Ili region remolded indoor compression testby using the double line method, analyses the influence of the moisture content, dry density and overburden pressure to compression deformation characteristics and collapse deformation characteristics in the humidifying process of remodeling loess, summarizes the variation of remolded loess collapsibility and compressibility in humidifying process, and gets the following conclusions:(1) Ili remolded loess compression decreased with dry density increasing and enhanced with moisture content increasing. Loess compression after flooding has little relation with moisture content before immersion.(2) Ili remolded loess collapsible weakened with humidification, enhanced with dehumidification. With different overburden pressure, even given the same degree of humidification to remolded loess, moistening deformation and increasing moistening collapse coefficient are quite different. There’s lower sensitivity to humidifying under low pressure and higher sensitivity to humidifying under high pressure.(3) With the increase of degree of saturation, increasing moistening collapse coefficient increases. It shows "micro-wet effect" under high pressure and a small amount of humidification can cause greater depression. When humidifying moisture content is lower, increasing moistening collapse coefficient is more sensitive to changes of humidifying, and when humidifying moisture content is higher, the sensitivity declines.Finally, in thisarticle, the influence factors of the collapsibility loess is analyzed, for this experiment soil, using principal component analysis to determinethe factors affectingcoefficient of collapsibilityaremoisture content, dry density and the external load,the factors affectinginitial collapse pressurearemoisture content and dry density. Finally, it uses the BP neural network method to establish forecasting network model of coefficient of collapsibility and initial collapse pressure. The prediction results show that the prediction models have small error and high precision. And the generality of network model was validated by the data in the literature. It illustrates that with artificial neural network method to calculate the coefficient of collapsibilityand initial collapse pressure of loess is accurate and reliable and provides a new method for predicting the loess collapsibility. |