A Physical Investigation of Air/Water Interactions Leading to Geyser Events in Rapid Filling Pipelines.

Abstract

Geyser events have been observed during the filling of large diameter conduits such as CSO storage tunnels which may experience an explosive release of water and air mixtures from a vertical shaft. Previous studies have not focused sufficient attention on the role of air in the process. This dissertation is devoted primarily to the study of large air pockets trapped in a nearly horizontal conduit and their release through a vertical shaft. Field data collected by others was reviewed and it is concluded these geysers cannot be explained by single phase water flow alone. An alternative explanation involving the release of large air pockets is offered. The propagation and reflection of hydraulic bores is one mechanism by which large air volumes can become trapped in an otherwise water-filled conduit. Experiments demonstrate the interaction of multiple bores resulting in the formation of several air pockets which interact in succession at a vertical shaft to lift water significant vertical distances. Additional experiments involved either the continuous injection of air or the release of a single large air pocket in a horizontal conduit to investigate the air dynamics at an attached vertical shaft. The arrival of the leading edge of the air pocket at the shaft pushes water upwards ahead of it and small diameter shafts are particularly susceptible to large vertical rises, especially with air pockets larger than the liquid volume in the shaft. Numerical modeling is presented to confirm this observation. The vertical rise following the expulsion of an air pocket is demonstrated to primarily be an inertial surge process.. The situation is more complex when multiple air pockets are present in the pipeline since the release of one pocket can influence the behavior of subsequent ones. A modest expansion of the diameter a short distance above the connection with the tunnel was found to provide reductions in water rise. Extending conclusions to large scale systems is difficult due to the inability to produce a process known as flooding in air/water exchange flow. This dissertation demonstrates the need to consider both vertical air/water interactions and horizontal dynamics during system design.