Pipeflow Experiments to Quantify Pore-Water Pressure Buildup due to Pipe Clogging

Abstract. Clogging of soil pipes can be detrimental to the stability of a hillslope leading to landslide, streambank, dam, and gully failures. A soil pipe may become clogged through internal erosion or pipe collapse; therefore, it is important to obtain more information surrounding these clogging occurrences in order to better predict their effects. When a pipe becomes clogged, an instant pressure buildup occurs upstream from the clog. This pressure may be enough to remove the clog, or the pressure may continue to build in the surrounding soil matrix due to the clog which could lead to landscape failures. Field observations indicate occurrences of both, and this study will investigate characteristics for which the clog is removed or remains intact. Baseline laboratory experiments were conducted with a 100 cm long, artificial (clear PVC) soil pipe. A pipe clog was established at approximately 90 cm along the pipe length. Triplicate experiments were conducted with two pipe diameters (20 mm and 30 mm), two soil types (sand and a more cohesive sandy clay loam with 5% clay content), two clog lengths (3 cm and 6 cm), three pipe roughness, various packing densities, and both dynamic and constant head experiments to determine their importance in clog removal. Digital pressure gauges capable of recording pressures every 0.1 s were installed along the second half of the pipe to monitor pressures both before and after the clog. These transducers recorded the pressure response due to clogging, the pressures that built up behind the clog, and the length of time that the plug withstood the applied pressure before removal. Pipeflow rates were monitored by continuous weighing of the outflow. Regardless of pressurized time (time until clog removal), all clogs were removed as plugs. Sand clogs were all removed in under 100 s, regardless of length, head, and pipe roughness. Adding roughness to the pipe increased the clog removal time for the sandy clay loam soil by more than 50%, but had no effect on the sand clogs. The relationship between applied head and pressurized time was a negative exponential relationship. The bulk density had an exponential relationship to the amount of time the plug withstood the pressure. In the dynamic head experiments, the plug was removed in less time at lower pressures than in the constant head experiments. The data obtained through the experiments outlined above will assist model developers in creating improved models for soil piping and internal erosion. This will allow researchers to better understand and predict internal erosion, eventually leading to the ability to prevent major landscape failures.