![]() |
CiteULike | ![]() |
RhysU's CiteULike | ![]() |
![]() |
|
![]() |
Register | ![]() |
Log in | ![]() |
Direct numerical simulation of turbulence in a model solid rocket motorby: Prem Venugopal
|
Reviews
[Write a review of this article]
Find related articles from these CiteULike users
Find related articles with these CiteULike tags
Posting History
AbstractCore flow in a solid propellant rocket motor is often modeled as the injection driven flow in a chamber closed at the head end. However, it becomes expensive to conduct direct numerical simulations in this configuration at moderately high Reynolds numbers. In the current study, a simplified model for the core flow in a solid rocket motor has been devised that allows direct numerical simulations to be conducted at a reasonable cost. In the simplified model problem, the streamwise direction has been asymptotically homogenized by assuming that at large distances from the head end of the rocket motor, both the mean and RMS of turbulent fluctuations evolve slowly in the streamwise direction. The Navier-Stokes equations are then modified to account for this slow growth. Using the simplified model, a separate direct numerical simulation can now be conducted at each streamwise location in the rocket motor. A direct numerical simulation of compressible turbulent flow in the model rocket has been conducted at a streamwise location situated 40 channel half-widths from the head end of the rocket motor and at an injection Reynolds number of approximately 190. The simulation reveals that turbulence results in increased near wall gradients for streamwise velocity, density and temperature due to the presence of the Reynolds stress term. The turbulence in the model rocket is found to be only weakly compressible, although significant compressibility existed in the mean flow. Streaks, streamwise vortices and inclined shear layers are observed in the near wall region. The vorticity vector shows a preference to be inclined at (39°-49°) to the wall, similar to observations made in non-transpired channel flows. The direct numerical simulation database is also used to make a priori estimates of the relative magnitudes of subgrid terms in the filtered evolution equations. It is found that in the equations for mean momentum and resolved scale Reynolds stress, the dominant subgrid terms are the subgrid stress terms. The contribution of subgrid stress to wall normal transport and dissipation in the resolved scale turbulent kinetic energy equation are similar to those observed in turbulent channel flows.
BibTeX record
RIS record