CFD prediction and model experiment on suction vortices in pump sump

The sump size is being reduced in order to lower the construction costs of urban drainage pump stations in Japan. As a result of such size reductions, undesirable vortices such as air-entrained and submerged vortices are apt to appear in sumps because of the higher flow velocities. The Turbomachinery Society of Japan (TSJ) Standard S002:2005 states that the appearance of such visible vortices is not permissible for conventional sumps, and experiments with scale models usually have been done to assess the performance of sumps. Such tests, however, are expensive and time-consuming, and therefore, alternative computational fluid dynamics (CFD) methods for evaluating sump performance are desirable. The Research Committee on Pump Sump Model Testing, which is an organization in the TSJ, carried out a benchmark for flows in model sumps. They contributed commercial CFD codes such as “Virtual Fluid System 3D”, “Star-CD 3.22”, “Star-CD 3.26”, and “ANSYS CFX 10.0”. Some of the benchmark results were reported by Matsui, J. at the 23 rd IAHR Symposium in Yokohama, Oct 2006. The remaining results comprise this second paper. The calculated results were compared with experimental ones for flow patterns, locations of vortices, and their vorticity. In the experiments, the critical submergences for flow rates were minutely examined through visual observation with a video camera. The locations of the vortices were obtained by using the laser light sheet visualization method. The velocity and vorticity distribution in the sump were measured by using a PIV method. The following results were obtained. 1) The critical submergence for the air-entrained vortex is almost proportional to the flow rate in the sump. The vortex behavior is unsteady and the duration of the vortex varies greatly. 2) The submerged vortex appears accompanying the air-entrained vortex in the region of low submergences and high flow rates. The critical submergence for the submerged vortex is also proportional to the flow rate. 3) Some CFD codes can predict the visible vortex occurrence and its location for submergence and flow rate conditions with enough accuracy for industrial use. 4) The calculated velocity distribution at the bell entrance qualitatively agrees with the experimental results. However, the agreement is poor in terms of the magnitude and distribution patterns of the vorticity. This difference is caused by the lack of accuracy of the experiment and CFD computation. 5) Predicting the critical submergence for the visible vortices was not imposed in the benchmark. The calculated stream lines and vortex core lines are not able to be used to predict the visible vortices with much accuracy. An additional post-processing such as obtaining the vortex core static pressure and comparing it with ambient pressure for an air-entrained vortex or with the saturated vapor pressure of the water for a submerged vortex would be necessary to predict the visible vortices.