Abstract: Draining is one of most effective seepage control measures for geomaterials and engineering structures, of which the essence is through forming the potential spillover boundary or low water level boundary in flow region to realize the seepage control. Induced by the seepage boundary of drains, the seepage flow may exhibit a strong nonlinear effect. The variational inequality formulation of Signorini condition has been theoretically proved to be an effective approach for steady and non-steady seepage problems with drains, but needs further experimental validations. For this purpose, a laboratory test is performed on the seepage behavior in a sand flume with five horizontal drainage galleries, then the effectiveness and correctness of the variational inequality method of Signorini condition is validated by experimental data. The experimental results show that: in steady seepage flow condition, the upstream three drainage galleries in the sand flume have a dominant effect on controlling the seepage flow, while the other two drainage galleries in the downstream side lose the draining effect; and the numerical results agree well with experimental observations. In non-steady seepage flow condition, the numerical results somewhat deviate from the experimental results due to low accuracy of piezometer tubes, inhomogeneity of the sand sample and negligence of capillary effect. The non-steady seepage flow process in the sand flume with complex drains, however, is still satisfactorily demonstrated. The effectiveness and correctness of the variational inequality method of Signorini condition is validated by the seepage flow test; and it provides an effective approach for optimization design of seepage control structure in geotechnical engineering under complex drains.

Key words: sand flume model, drainage test, non-steady seepage, variational inequality of Signorini condition