Learning Forum

[cptblackburn]

Hello,

I am doing a 2D, supersonic analysis on an intake. I extended the domain on top and bottom of my intake geometry for setting up the boundary conditions(pressure- outlet). When I run simulation at design Mach number (2.2) I am getting good results. By good result i mean mass imbalace is less than 1% but residuals are oscillating in the 10^-3 range.

But when i reduce the mach number I am getting divergence. I made markers at maximum aspect ratio, max skewness points. Max AR point is outside the intake at the end of external domain. There after some iterations temparature is getting limited and sometimes time step is getting reduced due to large temperature change at the external points. Can I solve this by changing the dimension of external domains or through mesh refinements?. I was also thinking about using a pressure based solver instead of density based.

[Rob]

I tend to use pressure based for most models, density based would be used for very stiff combustion models and flows over about Mach 3 or so. Then again, I also avoid most aero models... 

Assuming the flow enters at the left you may want to extend the domain - it's very close to the area of interest. I can't tell if the problem is flow, mesh or physics related, if you look at the solution before it divereges or with a more "sensible" value range how does it look?

Residuals/convergence like you describe could mean there's a transient effect in the flow - use monitor points in areas of interest to see what's going on. 

[cptblackburn]

hello, thanks for the reply.

I monitored temperature and pressure outside the intake, near the top and bottom right corner of the extended domain. Temperature and pressure increase as iterations progress and when the temperature reaches 5000 K issues arise. I know the upper limit for temperature in fluent is 5000K. So I did pressure-based study in the domain and found that a bow shock is formed from the inlet and it is extending beyond the domain. After increasing the domain size upwards I am getting closer to the values obtained from pressure-based analysis on the density-based solver. 

[Rob]

Yes, you do need to ensure everything happens inside the modelled domian! 

The 5000K limit can be altered, but that's only done if you really expect things to get that hot. So, some combustion cases will alter the limits.