If transient dynamic simulations do not require any assumption of laminar, transitional or turbulent flow, why is not used for all CFD simulations? Is there no reason other than computational expense?
For "transient simulation" you should properly address the Direct Numerical Simulation (DNS).
In principle the answer is yes, you could use DNS for any problem. In practice, not only DNS is very computationally expensive but it can become unrealizable with present computational power. Complex geometry, high Reynolds number flow, shock wave, reaction front, would require too many computational nodes.
For "transient simulation" you should properly address the Direct Numerical Simulation (DNS).
In principle the answer is yes, you could use DNS for any problem. In practice, not only DNS is very computationally expensive but it can become unrealizable with present computational power. Complex geometry, high Reynolds number flow, shock wave, reaction front, would require too many computational nodes.
You can use it for any problem but it is a general solution and require much more computation power as compared if you use a steady state compressible laminar flow analysis. Why will you use a complex technique if you have the simple one?
Thanks
Hi,
in this generality your statement is just not correct. If you use for your transient simulation URANS equations, than still a turbulence model is involved. If you apply a LES, than a subgrid-scale model is involved. I.e. you still need to make underlying assumptions and considerations whether the flow of interest is laminar, laminar-turbulent transitional or fully turbulent.
Regards,
Dr. Th. Frank.
- Badges
- Science topic