Piping potential of unfiltered soils in existing levees and dams.

摘要

The research presented here covers the topic of soil piping, a phenomena that is responsible for approximately half of all dam failures. Current methods for evaluation of piping have not changed in close to 100-years, even though work during the past 15-years has indicated the methods may be deficient. Laboratory methods to study piping have also not been well developed. The pinhole piping test was developed for a specific type of piping related to dispersive soils. Recent advances have developed the hole erosion test; however, this test is limited to cohesive soils, as is the pinhole test. A new apparatus was developed here to complete some basic research into piping of non-cohesive as well as cohesive soils. The apparatus is capable of applying confining stresses in three orthogonal directions while inducing piping by controlling the flow through the load cell. Reproducibility tests indicate there is good reproducibility in tests, with a standard deviation of 0.10 cm/sec in the critical velocity required to induce piping in laboratory sands. Natural soils and mixed soils varying from clean uniform sands to silts and clays were tested to assess various parameters that influence piping. The results of the tests indicate that there are a number of parameters that influence piping potential; (1) the nature of the fines in the soil, (2) seepage angle, (3) maximum principal stress, (4) the rate of change to the seepage flow rate. Current methods to estimate the critical hydraulic gradient overestimated the required gradient for piping. New methods are proposed to evaluate piping potential, which are based on the dimensionless friction head and shear stress at pipe initiation.