| Expansive soil,due to its high expansibility,low permeability and strong adsorption of nuclide,is often selected as the antiseepage material for disposal of nuclear waste and landfill.It expands when it meets water to close the gap,and its special low permeability can isolate nuclear waste from outside facilities.However,the general underground environment is complicated,and the expansive characteristics of expansive soil will be affected by many factors.This paper mainly considers the influence factors of pore water solution concentration,load,dry density and stress history.One-dimensional consolidation instrument is used to carry out a series of expansion tests on expansive soil,and a prediction model of expansion and shrinkage of expansive soil under the influence of multiple factors is established.The main research contents of this paper are as follows:1.The effect of pore water solution on expansion deformation characteristics is studied by using no-load expansion rate test.The results show that the time history curve of expansive soil can be divided into three stages,namely,initial rapid expansion stage,slow expansion stage and stable expansion stage.When the concentration of internal and external solution is the same,the concentration of internal solution is unchanged and the concentration of external solution is changed,the maximum expansion rate is positively correlated with the concentration of solution,and also positively correlated with the expansion limit.The expression between expansion rate difference and expansion limit difference is derived.When the concentration of external solution remains unchanged and the concentration of internal solution changes,the maximum expansion rate increases first and then decreases with the increase of concentration.2.The effects of load and dry density on the characteristics of expansive soil are studied.The results show that the overlying load is negatively correlated with the expansive strain,while the dry density is positively correlated with the expansive strain,and the expansion rate expression is obtained considering the influence of the initial dry density and the overlying load.The larger the load,the smaller the expansion limit,and the larger the dry density,the smaller the expansion limit.The pore ratio showed the same trend with overburden load and dry density,and both showed a downward trend.3.The relationship between stress history and expansion characteristics is studied and analyzed.The results show that the stress history affects the expansion deformation of expansive soil.When the soil is under the load,the stress history inhibition effect is more significant,and the influence of stress history decreases gradually with the increase of the soil force.The pore ratio decreases greatly under the stress history.When the overburden load is small,its influence on the pore ratio is more significant.There is a negative correlation between the load and the expansion limit,and the expansion limit decreases obviously under the historical action of stress.There is a positive correlation between the expansion rate and the expansion limit,and the stress history results in a low expansion rate and expansion limit.4.Establish a prediction model of swelling and shrinking of expansive soil coupled with firefly optimization algorithm and artificial neural network.Five indexes affecting expansion deformation,including external solution concentration,internal solution concentration,overlying load,dry density and stress history,were selected.The above test data were used as the input of firefly optimization BP neural network model to predict the model,and the model in this paper was compared with the BP neural network model.The results show that the model can accurately predict the expansion and shrinkage of expansive soil,and it is feasible to be applied in this field.By comparing the simulation results of the proposed method and BP neural network method,and establishing the evaluation index system,it can be concluded that the proposed method is better than the latter,and the prediction results are more accurate and more extensive. |
