| At present,over 95% of the tailings dams in China belong to the upstream tailings dam.And the saturation line of the dam is relatively high.In addition,earthquake activities have become more and more intense in recent years in China.Therefore,once the tailings reservoir is liquefied,there is always the risk of dam break.The consequences are also difficult to estimate.Based on these,the tailings sand of a tailings dam is taken as the research subject.By developing a small rigid model box which can significantly reduce the boundary effect.The liquefaction test of tailings sand under seismic action was carried out.Based on the introduction and improvement of the particle swarm optimization(PSO)-support vector machine(SVM)algorithm,the threshold model of liquefaction was established.Accroading to the concept of dynamic apparent viscosity in fluid mechanics and the one-dimensional shear beam model,the flow equation after the tailings sand liquefaction and the flow discriminating factor is proposed.The following results are obtained:(1)Experimental study on tailings liquefaction under earthquake action shows that under the same peak ground motion acceleration(PGA),the tailings sand reacts more under the action of El Centro wave,which makes it easier to liquefy the tailings sand.The earth pressure of tailings sand and The ground motion energy is closely related and presents a typical spatial effect.Under the action of three peak ground motion accelerations at a depth of 45 cm from the surface,the ground motions have obvious reductions similar to the ground motion shear zone.The pore pressure development of tailings sand can be divided into three typical stages: the rapid increase of pore pressure at the beginning of vibration,the high pore pressure state after initial liquefaction,and the rapid dissipation phase after full liquefaction.When the vibration is carried out to a certain extent,the stress-strain hysteresis curve indicates that the saturated tailings sand has an alternating dilatancy and shearing in one cycle.And the development of pore pressure ratio has hysteresis with respect to the development of shear stress and shear strain underground motion.(2)Based on the liquefaction characterization factors extracted from the experiment and the factor-dimensional analysis based on SPSS software,the liquefaction influencing factors suitable for factor analysis are obtained.On base of the particle swarm-support vector machine(PSO-SVM)algorithm,the liquefaction threshold model is established.And compared with the existing genetic algorithm(GA)and grid node search method(GS),it shows that the PSO-SVM algorithm not only has faster calculation speed but also has higher prediction accuracy and higher accuracy 92.7% which reflects its better applicability.(3)Applying the theory of fluid mechanics,the tailings sand exhibits the characteristics of non-time-varying non-newtonian fluid underground motion.After the liquefaction,the apparent viscosity in the tailings decreases instantaneously.And the apparent viscosity increases almost as the strain rate increases.At the same time,the flow factor of the tailings sand is less than 1 after liquefaction underground motion.Under the action of the same seismic wave,the larger the PGA,the more obvious the shear thinning effect of the tailings sand,and the more obvious the flow trend when the duration of ground motion is the same.Under the same PGA,El Centro showed stronger flow characteristics in the tailings sand.(4)The application of the proposed liquefaction tailings flow model is compared with the existing research results and found to have good applicability.The tailings sand has similar fluid properties to ordinary saturated sand after liquefaction,which also indicates that it is feasible to study the flow characteristics of tailings sand after liquefaction by changing the apparent viscosity and shear strain rate.The paper has 94 pictures,30 tables,and 108 references. |
PDF Full Text Request