A study was conducted to determine experimentally and theoretically the losses in radial inflow turbine nozzles. Extensive experimental data were obtained to investigate the flow behavior in a full-scale radial turbine stator annulus. A theoretical model to predict the losses in both the vaned and vaneless regions of the nozzle was developed. In this analysis, the interaction effects between the stator and the rotor are not considered. It was found that the losses incurred due to the end wall boundary layers can be significant, especially if they are characterized by a strong crossflow. The losses estimated using the analytical study are compared with the experimentally determined values.

She is a social scientist, holding a PhD in cultural anthropology from the University of California at Irvine and an MA in Social Science from the University of Chicago. She is Co-Chair of the Committee for the Anthropology of Science, Technology and Computing in the American Anthropological Association. She studies engineers and their work in relation to environments, technologies, and human lives.

The structure of developing flows inside curved channels has been investigated numerically using the time-averaged Navier Stokes equations in three dimensions. The equations are solved in primitive variables using finite difference techniques. The solution procedure involves a combination of repeated space-marching integration of the governing equations and correction for elliptic effects between two marching sweeps. Type-dependent differencing is used to permit downstream marching even in the reverse-flow regions. The procedure is shown to allow efficient calculations of turbulent flow inside strongly curved channels as well as laminar flow inside a moderately curved passage. Results obtained in both cases indicate that the flow structure is strongly controlled by local imbalance between centrifugal forces and pressure gradients. Furthermore, distortion of primary flow due to migration of low momentum fluid caused by secondary flow is found to be largely dependent on the Reynolds number and Dean number. Comparison with experimental data is also included.

Laboratory-scale experiments simulating the injection of fresh water into brine in a Strategic Petroleum Reserve (SPR) cavern were performed at Sandia National Laboratories for various conditions of injection rate and small and large injection tube diameters. The computational fluid dynamic (CFD) code FLUENT was used to simulate these experiments to evaluate the predictive capability of FLUENT for brine-water mixing in an SPR cavern.The data-model comparisons show that FLUENT simulations predict the mixing plume depth reasonably well. Predictions of the near-wall brine concentrations compare very well with the experimental data. The simulated time for the mixing plume to reach the vessel wall was underpredicted for the small injection tubes but reasonable for the large injection tubes. The difference in the time to reach the wall is probably due to the three-dimensional nature of the mixing plume as it spreads out at the air-brine or oil-brine interface. The depth of the mixing plume as it spreads out along the interface was within a factor of 2 of the experimental data.

No abstract