One of the most important processes in the groundwater flow is a macroscopic dispersion process in geological media. In this study, the laboratory experiment and numerical simulation are carried out in order to understand the transport of the non-reactive tracer in non-uniform porous medium. The experimental apparatus consists of six different glass beads of diameters of 0.1,0.15,0.2,0.4,0.6 and 0.8mm to simulate the fluid flow and mass transport in a synthetic non-uniform medium. Both dye and NaCl solution were used as tracers in the experiments. The method of characteristics and the implicit finite difference method are applied for the numerical computations of the convective-dispersion and the groundwater flow equations, respectively. The results of the computation agreed with the measurements and reveal that the applied numerical method is valid for the mass transport in the non-uniform porous medium.

View full abstract

For the practical analysis combined with infiltration from surfaces, we have established the simulation code composed of multi-layered and heterogeneous properties system in transient three dimensions with two-phase treatment. Field experiments are accompanied and in accordance with the simulation (Yokoyama et al.,1997a). Further practical consideration, however, should be stressed on well functions, which locally characterize flow and propagation-field around wells. Though well capability could be analyzed in finer grid system, it requires tremendous nodal points. Insteadly, in this paper sub-division system is laid on the normal-main grid system in order to save the number of nodal points.
According to the two-phase treatment (Yokoyama et al.,1997b), soil properties such as porosity, particle-sizes, etc. are important to forecast the propagation of heat/solute materials. We have dug the additional row of wells in the shallow aquifer and located within several meters with expectation of isotropy. Heat propagation is mainly tested repeatedly then salt-solute propagation is added. Break-through curves not only in depth but also by wells are observed satisfactorily with attenuation and time-lag. Then these experimental date got agreement with this numerical values. Potentiality of this simulation code is just used in this experiment on injection wells. Influences by observation wells as to flow- and heat-field are resulted in little significance experimentally and numerically. Variations in penetration among layers from wells, however, is analyzed clearly.

View full abstract

Hauser Lake Dam is located on the Missouri River about 29.0km from the city of Helena, Montana. The subsoil consisted of 20.1m thick gravel. The water was retained by a skin of steel plates supported by a steel framework that rested on large footings. The dam failed in 1908, one year after the beginning of the first filling. It was concluded that the presence of the footings produced a local concentration of flow, and spring erosion caused the failure of the dam judging from the fact that it did not fail immediately.
The FEM seepage flow analysis and stability analysis against seepage failure revealed that:
(1) In case that the installed filter does work, the local concentration of flow does not occur in the part between two footings.
(2) In case that the filter does not work, e. g. the fine soil particles are washed in the interstices of the filter, accumulate and gradually obstruct the flow after the water filling, the concentration of the flow occurs enough to cause seepage failure of the soil. In this case the part between the two footings becomes a critical state against Prismatic failure. The soil mass in this part is possible to move upward and collapse i. e., seepage failure of bulk-heave-type will occur. The plugging of the filter will be also caused by luxuriant growth of algae, deterioration and disintegration of soil particles due to drying and wetting, etc.

View full abstract