Douglas L. Sondak and Erik Brisson – Boston University
Daniel J. Dorney – NASA Marshall Space Flight Center
The availability of large, fast computers has greatly increased the scale and complexity of scientific computation. The amount of data generated by these simulations has also grown, and deducing physical structures from the results can be a daunting task. Individuals in different physical locations often collaborate on large scale simulations, and the results are frequently of interest to other individuals who were not involved with the original computations. A distributed computing environment is an ideal venue to allow collaborators to participate in the interrogation of the results.
Interactive Particle Tracking
One technique for visualizing velocity fields is to render the path, or “streakline” of a particle released at a specified point. Important flow features such as separated flow regions and vortices are then readily apparent. The streakline demo allows a user to pick points anywhere within the simulation domain and see the resulting streaklines. This consists of a visualization component running on a graphics workstation and driving the Boston University Deep Vision Display Wall, a navigation/picking client running on a Windows PC, and a streakline generator program running on a high performance parallel system.
The streakline generator initially reads in a precomputed solution completely defining the flow within the inducer. The user picks points within the computational domain displayed on the Wall, using a wireless Gyromouse. These points are sent to the streakline generation program, which integrates them over a sequence of time steps. For each time step, the current positions for all streaklines are sent back to the visualization application. This demo is built on top of SCV’s Distributed Applications Framework For Immersive Environments (DAFFIE) system.
Application to a Rocket Engine Turbopump Inducer
Previous applications of the interactive streakline rendering system have been to small subsets of full turbomachinery geometries, typically encompassing several blade passages. In the current application, the system has been applied to the complete annulus of a rocket engine turbopump inducer.
Inducers are often used in fuel and oxidizer pumps in rocket engines. The inducer is something of a “pre-pump,” increasing the pressure of the fluid to minimize cavitation and improve performance of the main pump. Inducer flow fields typically exhibit unsteadiness, including a backflow region upstream of the blades. It is important to predict the unsteady flow as accurately as possible, since it can cause vibration of structural components upstream of the inducer, lowering their life cycles. Interactive streakline visualization is quite useful to examine the details of these complex flow fields.
Figure 1 above shows the computational surface grids on the inducer blades and the hub flowpath. The passages are discretized using overset O- and H-grids. In the figure the hub H-grids are shown in blue and the hub O-grids are shown in black. Figure 2 shows an example of the streaklines introduced into a snapshot of the unsteady solution. Surfaces are colored by static pressure, with blue representing low pressure and red representing high pressure. The streaklines have been introduced along a radial line to simulate a rake upstream of the blades. The tip vortex can be seen as the streaklines exit the inducer blades.