 At NASA Ames Research Center, our basic product is information. Information is produced in the form of reports, technical reports, publications, and also computer software. We export hundreds of computer programs each year to the aerospace industry to be used by them to help them in their design process of various vehicle components or configurations. The technology developed here at NASA Ames Research Center in the form of computational simulation software can benefit the aerospace industry by helping them produce safer and more fuel efficient aircraft and also producing those aircraft at a reduced cost. There are basically three aerodynamic simulation sciences that exist. There's flight testing, there's experimental testing, and then there's computational fluid dynamics. Each has its advantages and disadvantages. Used together, that's what we call computation to flight, which is one of the visions at NASA Ames Research Center. Computational fluid dynamics is used to complement both experimental and flight testing. We can sometimes use computational fluid dynamics to do smart experimental testing. In other words, eliminate the need for doing a lot of repetitive testing and only test at very important or very critical test conditions. We can do the same thing with flight testing. As you know, flight testing is a very expensive process. We can use computational fluid dynamics to help augment that testing program and the data that it's produced. The computers that are used to generate some of the results that you see today are contained in our NAS building, the Miracle Aerodynamic Simulation Facility. They include a Cray II computer and a Cray YMP computer. These are two of the fastest computers in the world today. The process by which we perform CFD simulations involves, first of all, defining a geometry or getting that geometry into the computer. The next step involves discretizing the flow field about that configuration. Once that flow field is discretized, then we can apply a flow solver in order to compute the flow about that configuration. That generates an enormous amount of data and we use graphics workstations to reduce that data and produce the results. Once that technology has been validated against experimental data, then we disseminate that technology for industry's use. During launch of the space shuttle, provisions have been made to employ an abort mode in case of emergency. One of the elements of this abort mode is known as FASTSEP. FASTSEP is the FAST separation of shuttle orbiter from the rest of the launch vehicle. As the shuttle goes through the launch sequence, we will look at the results of two different numerical simulations. The first occurs during ascent. As the shuttle passes through the transonic range, the external tanks of the rocket motors experience a hysteresis effect due to shock movement. Once the Mach number is changing from 0.8 to 1.02 over 6 seconds, as shown here, the unsteady flow field that evolves around the vehicle lags behind what it would be for steady flow or a fixed Mach number. Notice the shock development shown by the pressure contours on the leading surface of the tank. On the base of the tank, we see a highly unstable flow regime. The second numerical simulation occurs at about 2 minutes into the flight at an elevation of about 50,000 meters. At this time, the solid rocket boosters or SRBs fall away from the tank and the orbiter. These bodies moving relative to each other make analysis using other methods much more difficult. Notice the interaction of the pressure contours as separation occurs. The shuttle under the body of the shuttle are the pressure contours created by shock wave interaction between the orbiter, the external tank, and the SRBs. Robert Meakin, research scientist and member of the NASA Ames Space Shuttle Flow Simulation Team, explains the work done with the solid rocket booster, or SRB, as a stepping stone to simulating fast set. The work that the Space Shuttle group at Ames, they interact with the Johnson Space Center, and we're doing numerical simulations of various conditions of the shuttle during the ascent. Of course, there's a great deal of wind tunnel data that's available and also flight data that's been accumulated over the flight history of the shuttle. The shuttle group here at Ames is augmenting the data that there is there filling in missing data and carrying out calculations that really aren't possible to model in any other way. Fast set being one of those cases, but really there's no other way to get that sort of information. A turbo pump is the main element in a rocket engine that supplies fuel from the fuel tank to the combustion chamber. An important component in the turbo pump is the inducer. A massive flow separation or cavitation in the inducer can block the fuel supply and result in total engine failure. Two, one, we have ignition, we have liftoff, liftoff of Columbia, ignition 61C. Seekhwan Yoon, a senior research scientist of MCAT Institute at Ames Research Center, explains. A computational study of this kind of problem was impractical even on supercomputers since the existing computer programs were not fast enough. The objective of our project is to develop a very efficient computer program which can give a direct impact on the design of future rocket engines. Shown here is the actual hardware of the space shuttle main engine. The turbo pump consists of an inducer with a stationary casing and a shrouded impeller with partial blades. In this computer generated image, we remove the shroud and view the inducer and partial blades of the impeller. We can see a pressure gradient across the blades as well as along the hub due to the action of the centrifugal force. The particle traces over the suction side of the blade and through the tip clearance are seen here. Traces over the pressure side of the blade, shown in purple, are sucked into the tip clearance and become the leakage flow. The stream-wise particle flow is shown in green. The interaction of the leakage and stream-wise flows results in a region of concentrated vorticity. Do Chan Kwok, research scientist with the Applied Computational Fluids Branch of Ames Research Center and group leader for this project, continues the explanation. The turbo pump is especially one key area we can improve the performance. Aircraft engines normally perform in the deficiencies in the 90 percentile range. Typically the turbo pump operates in 80 percentile efficiency range. So there certainly we can see lots of room to improve. The shuttle can be benefited by this computational simulation and by improving the performance and improving the reliability. And eventually we will meet the high launch capability in the future. Recent advances in heart surgery have led to the development of the artificial heart and the use of artificial heart valves. At the same time, technology developed to compute the flow in components of the space shuttle main engine is being applied to simulate the unsteady flow in the Penn State artificial heart. Do Chan Kwok explains NASA's involvement in this spin-off technology. In general we are interested in reapplying NASA developed technology, especially the CFD technology can be reapplied in many different instances. An artificial heart is particularly interesting because it will help national health problems and the demand for this type of mechanical device can really contribute to human health and also animal health in the future. Chetan Kiris and Stuart Rogers, research scientists, were responsible for the flow code on this project. Stuart Rogers further explains. The data we started with in this case was basically taken straight off of blueprints which were used to build models of this heart which were tested by Penn State. Given the blueprints from that model, we then generated a series of codes which would then describe that shape to the computer as a series of discrete points. Once we had those discrete points then our flow solver could take them and compute the flow inside the heart. Here we see the main chamber of the heart and the particle traces which indicate the flow. The color of the traces indicates the release height at the inflow valve opening. This is the computer generated image of the tilting disc heart valve. This valve can be used in conjunction with an artificial heart or used as a separate device. The inflow conditions are specified at the entrance for the valve opening and they are specified at the exit for the valve closing. The tilting disc reacts from the forces applied to it by the blood flow. The valve motion is made possible by using the chimera grid embedding technique. We can view valve operation from different rotational views. Red particles are released from the vertical plane at the entrance and magenta particles are released from the sinus region of the aorta which is located just beyond the tilting disc. The flow between the disc and the aortic wall is highly accelerated. These kinds of changes in the local blood flow conditions can greatly affect the blood structure. Technology developments that result from solving this problem will yield spin back applications for other flow problems associated with the space shuttle main engine. An important contribution made by NASA to medicine results in a contribution made by medicine back to NASA. Science aiding science through interdisciplinary cooperation. The F-18 is a jet fighter currently used by the United States Navy. It is used in air to air and air to ground fighter and attack roles by the fleet. It has a great deal of maneuverability and performs at high Gs and at high angles of attack. National Cummings, National Research Council Research Associate explains why the F-18 was chosen as a research vehicle. The F-18 because it's capable of pulling such high maneuvers and high Gs gets into regimes of aerodynamics that other aircraft don't even experience. Because of that we're using it as a test bed and a computation basis for producing predictions for flows over high angle of attack aircraft. The CFD process begins when the aircraft manufacturer supplies the surface geometry of the aircraft to be studied. From this information a surface grid and a flow field grid is created. Then flow solving begins using the three dimensional partially flux split time marching F3D Navier Stokes code. Conditions are made for turbulent flow by using the Baldwin-Lomax turbulence model. Here we compare the flow visualization around the one thirty second scale model of the F-18 in the idetics international water tunnel with the computational results. Clearly vortices from the four body and wing leading edge extension or LEX can be seen. We visualize our numerical predictions using a variety of methods including simulated surface oil flows which help us to see the primary and secondary cross flow separation lines on both the fuselage and the leading edge extension. We also use helicity density contours which enable us to see both positive and negative senses of rotation of vortices. We can further visualize the vortices by passing particle traces back through the helicity density contours which help us see the vortices as they pass back over the fuselage. The goal of the research is to be able to predict the flow of a full aircraft such as the F-18 so we can see the interaction of things such as the LEX vortex as it comes up over and top of the LEX runs down the body and pinches on the vertical tail and possibly causes structural damage. In actual flight test shown here on the F-A-18 high alpha research vehicle or harv note the effect of the LEX vortices on the vertical stabilizer as Russell Cummings continued. There are flight tests being currently conducted down at Dryden and we're comparing our CFD predictions concurrently with them taking their data and it's very exciting since I don't believe that very many people have been able to do that before to actually have their CFD predictions hand in hand with flight test data and as we've compared the two side by side we've seen that the CFD has been able to very well predict the type of aerodynamics both for surface pressures and off-surface flow visualizations. A multi-stage compressor is used on jet aircraft engines to compress the air before it goes into the combustion phase. A multi-stage compressor consists of many rotor stator pairs. Rotors are rotating airfoils and stators are stationary airfoils. In a multi-stage compressor you may have as many as 17 to 20 of these rotor stator pairs. Karen Gundy-Berlett, research scientist at Ames Research Center, explains the difficulty of doing research in this arena. The goals of my project are to compute the three-dimensional flow within a multi-stage compressor and hopefully by doing this we can understand the fluid physics of the flow within the compressor see if we can design compressors that are much more efficient and much more reliable while reducing the weight and the size of the compressor. Here we see the results of this research. These are the pressure contours within the aircraft engine compressor. The flow is moving from left to right. Low pressure is indicated by blue whereas high pressure is indicated by red. The pressure contours show the inviscid part of the flow field. By seeing the pressure difference across each of the airfoils you can see what forces are occurring on the airfoil. Notice that the pressure within the system is quite unsteady as the pressure rises from the first stage to the second stage. These are the entropy contours within the two-and-a-half stage compressor. The entropy shows the viscous part of the flow field. It points out the slow fluid that sticks to the surface of the airfoils. Notice that the slow fluid is convected back through the system for three or four chord lengths. In a multi-stage compressor the flow within the latter stages is much more complex than the flow to the initial stages easily seen here. The entropy plot does a good job showing the wakes due to the viscous dissipation of the air next to the airfoil. As the wakes progress along the surfaces of the airfoils note that there are varying forces applied that tend to twist and rotate the airfoils. In the latter stages where there are many wakes being convected through the compressor the forces are even more severe. The flow fields are very complicated and the unsteady forces appear to be varying quite rapidly. There is experimental data available for this compressor that's why we chose to simulate the flow within this compressor. So far the comparisons have been very good. Time average pressures on the surface are very close to the experimental values. Wake average data is in good comparison for computations where I have an extremely fine grid in the second stage of the compressor. One of the directions I see for computational fluid dynamics in the future is in an area we call multidisciplinary physics. In that area we combine not only the fluid equations but the equations governing electromagnetics or propulsion or controls into one software simulation tool. In order to solve problems like that the problems of multidisciplinary fluid physics is going to require computers a thousand times faster than the computers we have today. Computers on the speed of one teraflop that's one trillion floating point operations per second. In order to obtain the one teraflop capability that we'll need for performing multidisciplinary simulations it's going to require massively parallel computers. Computers that have thousands and thousands of processors as compared to the YMP which has eight processors. The internationalization of the aerospace business is going to cause CFD to play an even greater role in the simulation sciences area. Our American aerospace manufacturers are going to rely more heavily on computational fluid dynamics to produce better and more efficient aircraft in order to be competitive with our overseas competitors.