Jinzhong Loess Collapsible Characteristics

With the further development of the west, the shortage of land resources in major cities in loess plateau region has increased. The gradual change of industrial and civil building construction from low to high terrace has led to the survey、evaluation and treatment has become the most important issue in in engineering design in loess areas.The construction in Jin-zhong city Shanxi province is located in loess plateau area is threated by the loess collapsibility. The existing evaluation system of loess collapsibility is based on correction of indoor test results, however decades of project experience and incidents prove that the regional correction factor in Jin-zhong city is too small, so it is vital to make a systematic research on loess collapsibility in Jin-zhong city.In order to research the loess collapsibility in Jinzhong city, in typic collapse loess area, a couple of immersion tests on self-weight collapse loess site were conducted.By monitoring the surface and deep subsidence, the lateral displacement, the emergence and development of crack, as well as the change of soil pressure and the water migration, the rules of development of the collapse deformation of ground surface and underground were researched. The results indicate the rules of development of the influence region of loess collapse, the development process of fissures around the test pit and the underground lateral deformation. The computing method of area correction factor (3o was discussed. Combining with the results of immersion tests, a series of laboratory tests on triaxial tests and uniaxial tests were conducted which aimed to research the vary of salinity and compare the influence of different consolidation stress on loess collapsibility.The influence of axial compressive force and axial compressive force on the rate of collapse also were researched.The techniques of data analysis were used to analyzed the correlation of loess’s physical mechanical index and the coefficient of self-weight collapsibility.Based on qualitative and quantitative analysis,six index which were closely tied to the coefficient of self-weight collapsibility were choosed to model a radial basis function neural network model to forecast the coefficient of self-weight collapsibility of the region.