|
[Sponsors] |
October 24, 2018, 07:30 |
Solution Divergence
|
#1 |
New Member
Yahia
Join Date: Aug 2018
Posts: 9
Rep Power: 8 |
Hello everyone,
I'm simulating a turbulent flow inside an S-shaped duct using k-ω SST turbulence model. BC's were specified at computational domain inlet and outlet as 'pressure inlet' and 'mass flow inlet' respectively.The mass flow rate of the outlet has to match the prescribed center-line Mach number of M = 0.6 at the inlet. Simulation required to match previous computational and experimental results. Only total and static pressures are known at the inlet center-line as shown in the attached figures. However, when initializing the solution, it indicates that pressure information is not available at the boundaries and the case will be initialized with constant pressure !. Solution process also indicates divergence when starts calculating! Any advice will be highly appreciated Imposing Inlet BC's.pdf prseesure info not available.pdf |
|
October 24, 2018, 10:15 |
|
#2 |
Senior Member
Lucky
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 5,739
Rep Power: 66 |
In order to do pressure interpolation with the hybrid initialization, you must specify the pressure at (at least) one inlet and one outlet. You only have a pressure inlet (and no pressure outlets) and that's why it posts the info message. It is not an error.
As for why your simulation diverges, good luck. |
|
October 24, 2018, 13:43 |
|
#3 | |
New Member
Yahia
Join Date: Aug 2018
Posts: 9
Rep Power: 8 |
Quote:
The first criterion used in the simulation is to achieve a steady mass flow rate at the inlet plane as prescribed by previous studies. the mass flow rate boundary condition applied at the computational domain outlet was determined by the inlet achieving a fully developed flow of Mach number M = 0.6. However, no info about mass flow rate at the outlet. Since the flow is considered compressible, I'm wondering if I can employ the equation of mass flow rate for an ideal compressible gas (m ̇=(AP_t)/√(T_t )×√(γ/R)×M(1+(γ-1)/2 M^2 )^(-(γ+1)/(2(γ-1))) in order to determine the mass flow rate at the computational domain outlet!? one more thing to note that I'm not able to set the static pressure at the outlet because the flow is considered subsonic and not supersonic. |
||
October 24, 2018, 17:46 |
|
#4 | |
Senior Member
Lucky
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 5,739
Rep Power: 66 |
We're no longer discussing initialization here. I'm not sure if initialization was causing your divergence problems or not but now you are asking about boundary conditions. So I assume initialization is no longer an issue.
Quote:
Regardless you can't directly force a Mach number anywhere. You can impose a Mach number only for supersonic conditions. A mach number of 0.6 is obviously subsonic. It may be that you have experimental data at the outlet recording the Mach number as 0.6, but this is not a boundary condition you can impose (not even in reality). Of course you want the CFD to match the experiment, but you need to think about what was done experimentally and therefore what needs to also be done numerically to obtain a Mach number of 0.6 at the outlet indirectly. |
||
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
PEMFC model with FLUENT | brahimchoice | FLUENT | 22 | April 19, 2020 15:44 |
Floating point error and divergence detected | aannjj | FLUENT | 0 | July 2, 2013 03:44 |
3d vof | Smaras | FLUENT | 2 | February 19, 2013 06:58 |
Quarter Burner mesh with periosic condition | SamCanuck | FLUENT | 2 | August 31, 2011 11:34 |
Divergence in 8 node parallel solution | hypete | FLUENT | 0 | October 14, 2010 04:01 |