Study On Engineering Property And Unfrozen Water Content Prediction Of Saline Soil In Nong’an

Nong’an county is located in western Jilin Province, southern Songnen Plain, typical seasonal frozen soil in northeast region.Different types of saline soil are distributed in the erea. This area belongs to the temperate continental monsoon climate.It is obvious that four seasons change here yearly and the erea is sufferringfreezing and thawing cycles.The average annual evaporation exceeds the amount of rainfall.Soil salinization is very serious, such as salt expansion, dissolved depression, corrosion and frost heaving and undesirable properties. Seasonal frozen soil salinization is not only bad for local agricultural,but also production has brought huge problems.It brings a lot of problems to engineering construction here at the same time, such as the reduction of arable land, grassland degradation, building corrosion based, muddy road. Therefore it is necessary to study engineering properties of Nong’an saline soil. Soil water, salt material morphology and migration were studied, and accordingly, establishingrational and effective prediction model of soil freezing temperature and unfrozen water content.In this paper, Nong’an upland soil samples were used for the research, firstly, the experimental analysis of material composition and physicochemical properties were conducted, and the distribution of physicochemical indexes of the soil profile were discussed. The influence factors ofsoil salinization process were studied; Second, due to the freezing temperature of soil is an important factor affecting soil water migration, ice formation and frost heave. Therefor, through indoor test, the change of freezing temperature of saline soil in the area are analyzed in detail.And the effects of dry density,salt content, initial water on the rate of salt freezing temperature of frozen soil in Nongan were studied; Thirdly, unfrozen water content is not only an important evaluation the index of moisture migration in soil characteristics of permafrost, it is also commonly used in thermal calculation parameters. Thus, the unfrozen water content of saline soil is calculated by temperature measurement method, in addition, the unfreezing water content of non saline soil in Changchun area was measured by nuclear magnetic resonance method.Finally, because soil freezing temperature and unfrozen water content are related with soil water content, salt type and concentration, dry density and other factors which have a complex nonlinear relationship. Therefore, in this paper, we introduced the support vector machine method.Freezing temperature and unfrozen water content data of the saline soil in Nong’an, Mo Ling clay and soil in Changchun area, using grid search, genetic algorithm, particle swarm optimization algorithm for model parameter optimization, were predicted. the freezing temperature and unfrozen water content prediction model of three types of soil wereestablished through the measured data.