Curtis C. OberSenior Member of the Technical Staff
Parallel Computational Science
|Curtis C. Ober
Parallel Computational Sciences, Dept. 9221
Sandia National Laboratories
Albuquerque, NM 87185-1111 U.S.A.
Phone: (505) 844-1846
Fax: (505) 845-7442
Aerodynamics of a Body Immersed in a Supersonic Wake: A Computational Study
To determine if a trailing body will remain in the wake of leading body at supersonic speeds, a computational study was performed to determine the pressure and streamlines on the surface of the trailing body. In the picture to the right, the red indicates high pressure on the upper part of the hemispherical nose, which has been exposed to near freestream conditions behind the leading body. The cyan on the middle to lower part of the hemispheical nose indicates low pressure associated with the recirculation region behind the leading body.
The streamlines on the surface of the trailing body show two rings related to stagnation and separation. The outer ring, indicated by diverging streamlines at the high pressure region, shows the locations of stagnation on the trailing-body surface. The inner ring, indicated by converging streamlines at the low pressure region, shows the locations of separation on the trailing-body surface.
High-Performance Simulations of Coastal and Basin-Scale Ocean Circulation
The major domestic oil and gas companies are extending their offshore exploration and production activities into deeper and often inhospitable waters in search of new reserves. Exploration in these untapped regions is inherently expensive and can be extremely risky from economic, environmental, and safety viewpoints. Among other factors, a significant risk occurs from ocean eddies. Although currents from persistent eddies may or may not affect a particular drilling area, conservative safety factors require rig shutdown and possible abandonment when strong eddies are in the region. Cessation of activities on rigs, even for short durations, can result in significant economic losses for typically several companies in the industry per event. Predicting current surges due to ocean eddies or current-shelf interactions can reduce the losses incurred by rig shutdown, as well as aid in deep-water rig design. Thus, accurate and timely predictive capabilities can widen an oil rig's operational window during pass-by of ocean eddies. For these reasons, development of high-preformance, high-resolution predictive capabilities for a basin-scale ocean environment is warranted.
In the example shown, overset grids are generated around islands and along the shoreline of North, Central, and South America. A background grid helps communicate fluid properties between the various grids. These grids are broken down into smaller grids which can be spread across many processors of a massively parallel computer.
Seismic Imaging on Massively Parallel Computers
A key to reducing the risks and costs associated with domestic oil and gas exploration is the ability to image complex geologies, such as thrusts in mountainous areas and sub-salt structures in the Gulf of Mexico. Current industry computational capabilities are insufficient for the widespread application of 3-D, prestack, depth, migration algorithms. A 3-D data set can be as large as several terabytes in size, and with current technology, a single image can take months to produce, and the multiple runs necessary to refine the geological model may take a year or longer. Oil and gas companies need to be able to perform complete velocity field refinements in weeks and single iterations overnight. High performance computers and state-of-the-art algorithms and software are required to meet this need.
A standard verification step of a migration algorithm is the Marmousi model shown here. The image produced by our 3-D, prestack, depth, migration algorithm, Salvo, has been superimposed over the velocity of the Marmousi model. There is very good agreement between the Salvo image and the velocity model, including the strong reflector at the lower right edge of the image.
Last modified: October 12, 1998
MPCRL Home Page || Sandia Home Page
Questions and Comments || Acknowledgment and Disclaimer