I am currently working on numerically simulating film cooling of a flat plate using compound angles holes, in which the lateral spread of the secondary fluid is a crucial study. I read in a research paper that Realizable k-epsilon turbulence model gives the worst prediction for lateral spread of the secondary fluid in numerical studies of film cooling, and yet all the literature I have referred that numerically simulate film cooling with compound angled holes use Realizable k-epsilon for their simulations. Is there any particular reason for this?
I don't think, that there is a particular reason for this besides inertia of e.g. ANSYS Fluent users from their historic tradition in this solver, where for a long time the Realizable k-eps model was the most commonly used turbulence model besides all its deficiencies.
Usually for a heat transfer application like film cooling your would like to use a k-omega based model like SST with a y+~1 mesh resolution providing you with a full boundary layer resolution.
Regards,
Dr. Th. Frank.
Dear Moon,
It is well known that 2 equation turbulence models (k-e) are under-predicting lateral spread of the secondary fluid. At the mean time over-predicts the center-line effectiveness. If one want to capture lateral spread accurately, LES is another option but definitely DNS is not feasible for the mainstream and secondary flow Reynolds number observed over turbine blades and afterburner section.
Although, my work was not on expended holes yet the following papers can help to understand why Realizable model was used widely.
The following articles might be of your interest:
1. Experimental and numerical studies on film cooling with reverse/ backward coolant injection, International Journal of Thermal Sciences 111:390-408 DOI 10.1016/j.ijthermalsci.2016.09.027
2. A Numerical Study on the 2D Film Cooling of a Flat Surface, Numerical Heat Transfer Applications 67(6):673-695, DOI: 10.1080/10407782.2014.949131
Lot of literature is available on expended holes (Experimental and Numerical LES studies), I would recommend to try the available 2 eqs models and compare the results.
Kuldeep Dagar thank you for the articles. I'll certainly go through those
Realizable and standard k-e are the most used tubulence models in film cooling, using these models doesn't require very refined mesh near the wall compared to SST model for example which can reduce solver errors and convergence instabilities due to poor mesh quality.
- Badges
- Science topic