Research On Seepage Erosion Of Filter System Based On Pore Structure

The seepage erosion phenomenon poses a certain threat to the safety and stability of earth-rock dams,and filter layers is one of the effective measures to prevent seepage damage of earth-rock dams,which has been widely used in engineering examples at home and abroad.In view of the seepage failure of some existing reservoirs in different degrees,scholars have made intense research on the effective setting of the filter layer and the loss process of protected soil particles under the action of seepage,but they mainly focus on the proposal of the filter criteria,the critical conditions of seepage erosion failure,and the changing laws of soil parameters,and seldom consider the influence of soil pore structure on seepage erosion phenomenon.In this thesis,referring to the pore model proposed in a few literature and its limitations,based on the theory of pore channel size distribution,the formula for calculating the infiltration depth of protected soil particles in filter media is derived by using probability method,and a tetrahedral pore network model which is more in line with the dense state of artificial fill is proposed.The rationality of the new model is verified by comparing the infiltration depth values predicted by different methods.The PFC particle flow numerical model of infiltration erosion of cubic and tetrahedral filtration systems is established,and the feasibility of the numerical model is analyzed.The difference of filtration performance between the two structures and the similarities and differences of maximum infiltration depth predicted by different methods are compared,and the movement process of particles is explored,revealing the infiltration erosion mechanism of the filtration system.Finally,combined with the design case of Qiaojiagou Reservoir,the effective setting scheme of filter layer is evaluated,and the application of pore model in filter design is analyzed.The main research findings are as follows:(1)The verification results of the new model show that the reduction of porosity caused by the close arrangement of particles makes the infiltration depth predicted by the new model smaller than that predicted by Locke and Indraratna model;With the increase of the probability of protected soil particles passing through the pore channel of filter,the infiltration depth predicted by the new model gradually changes from slightly smaller than that predicted by Silveira model to larger than that predicted by Silveira model;The infiltration depth of particles increases with the increase of characteristic particle size ratio;When the characteristic particle size ratio D15/d85<4,the maximum infiltration depth predicted by the new model is similar to that predicted by Zou numerical simulation.When the characteristic particle size ratio D15/d85>4,the predicted results are quite different due to the short model size and infiltration time.(2)The PFC numerical simulation results of cubic pore structure show that the numerical model can well reproduce the variation of filter material gradation curve and the corresponding value of particle loss at different depths from the interface predicted by Locke and Indraratna test and model,which affirms the applicability of PFC numerical simulation method;The particle loss,flow velocity,porosity and permeability coefficient of protected soil increase with the increase of characteristic particle size ratio and calculation steps at 1 cm away from the interface The porosity and permeability coefficient of filter media decrease with the increase of calculation steps,and decrease with the increase of characteristic particle size ratio The parameters finally tend to be stable,and the filter material effectively blocks the loss of protected soil particles.(3)PFC numerical simulation of tetrahedral pore structure shows that the smaller pore size of tetrahedral filter material reduces the particle passing rate,which makes the self-filtration process more stable.Therefore,the parameters obtained by tetrahedral structure simulation are smaller than those obtained by cubic structure simulation,but their changing trends are the same,so tetrahedral structure has stronger filter performance.The maximum infiltration depth of particles predicted by numerical simulation method and new model method increases with the increase of characteristic particle size ratio,but due to the reduction of model size,the truncation of filter material gradation and the failure to fully reach the complex structure of natural soil,the predicted values of the two methods are quite different;Numerical simulation results further verify the feasibility and applicability of the new model.(4)By analyzing the numerical simulation results of two kinds of pore structures,it is concluded that the infiltration erosion process of the filtration system includes three stages:initial stage,development stage and stability stage.In the initial stage,the protected soil particles move slightly under the action of seepage to realize local rearrangement,but they do not enter the filter material;In the development stage,fine particles begin to move greatly and enter the filter layer,and stop moving after encountering pore channels smaller than their own size,thus preventing the loss of finer particles.On the contrary,they continue to move forward and eventually form a self-filtering layer;In the stable stage,the particles do not move further,the loss tends to saturation,and the parameters reach stable values.(5)By using several representative filter criteria and pore model methods,the filter design of Qiaojiagou Reservoir in the loess hilly region of southern Ningxia was applied.The results show that the filter design of silty clay loess is more suitable for the Design Code for Rolled Earth-Rock Fill Dams(SL 274—2020).Different filter criteria have their own scope of application and limitations,and provide reference for filter design of different types of soil.The pore model method provides a new way for filter design by predicting and analyzing the infiltration depth of protected soil particles.