Study On The Collapsible Zoning Of Loess Along The Second Phase Of The Han-Ji-Yi River Diversion Project

With the development of the western region's economy and the continuous advancement of urbanization,the urban population continues to increase,and the demand for domestic and industrial water is increasing.The problem of water shortage in the Guanzhong region is becoming increasingly serious,which severely restricts the development of Guanzhong region led by Xi'an.The project of introducing the Han River into the Wei River is to solve this problem.The second phase of the project of introducing the Han River into the Wei River is mainly to construct various structures such as tunnels,pipes and culverts.The Guanzhong area belongs to the loess area.So,each line of the second phase of the project of introducing the Han River into the Wei River will inevitably cross the loess area.Because loess is formed in a special natural environment,it has some special properties,especially its sensitivity to water,which is the collapsibility of loess.Many engineering problems are caused directly or indirectly by collapsing.Therefore,the research on the collapsibility of loess has some significance.In this paper,the water transfer structures of the second phase of the project of introducing the Han River into the Wei River are used to study the loess area along the line,and the collapsing of the loess along the line is researched in order to provide a certain reference for similar projects.The research has the following contents:Fully collected the geological survey data of the second phase of the project of introducing the Han River into the Wei River,we can sort out the loess texture and landform profile of the main trunk line of the project,and selected the physical properties of the loess:dry density,porosity ratio,natural moisture content,saturation,and liquid limit,Plastic limit,liquidity index,plasticity index.These eight indicators are used as research objects to study the relationship between physical property indicators and collapse coefficients,the eight indicators are divided into three types of indicators based on related relationships and cluster analysis.Four typical geomorphic units along the line of the project:Piedmont alluvial fan,third-order terrace of the Ba River,third-order terrace of the Wei river,and undisturbed loess taken from the loess loess were tested indoor,and physical property index values were obtained.And the collapsibility coefficient of the loess,and the relationship between each physical property index and the collapsibility coefficient was studied separately,and the mathematical expression and correlation between the two were obtained,which provided Indicator reference for a predictive model of the collapsibility coefficient.From each of the three types of physical indicators,we could select an index that has the most correlation with the collapse coefficient,and establish an input layer node number of 3 layers,an implicit layer node number of 8,and an output layer node number of 1.The BP neural network model is used to predict the collapsing coefficient of the loess.The model not only considers the influence of physical properties on the collapsing coefficient of the loess.It is also fully considered that the differences of the geomorphic units may affect the collapsibility coefficient of the loess.The loess collapsibility coefficient calculated by testing the model meets the accuracy required for engineering applications.Finally,by using the established prediction model of collapsing coefficient and combining common collapsing zoning rules,a collapsing zoning method suitable for the project under study is proposed.The collapsing process of the loess is nonlinear.Neural networks have been proven to simulate nonlinear processes well in many fields.Based on a large amount of preliminary data collection,field surveys,and laboratory tests,this paper studies several types of projects along the line.The loess physical index and collapsing coefficient of typical geomorphic units,on the basis of which a BP neural network prediction model of the loess collapsing coefficient is established,and the model is used to partition the collapsing of some of the lines studied.The collapsible prediction method of the collapsible prediction model can provide a certain reference for the study of loess collapsibility and the construction of similar projects.