CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   need help with natural convection problem (https://www.cfd-online.com/Forums/main/194574-need-help-natural-convection-problem.html)

yhs0365 October 18, 2017 21:37

need help with natural convection problem
 
4 Attachment(s)
hello everyone

I am having trouble with solving simple natural convection problem.

The problem is that momentum residual is so high and do not go down...:confused:

I attached my design image... the red one is heat source so maybe there has to be movement of inner gas, circulation around heat source

Attachment 59081

When I set laminar model and set initial velocity as 0m/s momentum residual was as below Attachment 59082

and velocity profile was very dirty...

for example region that has to be have downstream had upstream and so on...:confused:

so i ask for colleague how to reduce momentum residual he said that I have to change turbulence model or set initial velocity as twice of my goal velocity

so I set k-e model (because it's residual was lower than sst) and set initial velocity as 0.1m/s.

then, momentum and turbulence residual were as below

Attachment 59083
Attachment 59084

velocity profile was clearer than before but, still momentum residual has to be lower and velocity profile was still incorrect.

is there somebody know well about natural convection?
and is there other factors to lower momentum residual and make clear velocity profile??:(:(

thank u!

lcarasik October 18, 2017 22:54

Quote:

Originally Posted by yhs0365 (Post 668387)
hello everyone

I am having trouble with solving simple natural convection problem.

The problem is that momentum residual is so high and do not go down...:confused:

I attached my design image... the red one is heat source so maybe there has to be movement of inner gas, circulation around heat source

Attachment 59081

When I set laminar model and set initial velocity as 0m/s momentum residual was as below Attachment 59082

and velocity profile was very dirty...

for example region that has to be have downstream had upstream and so on...:confused:

so i ask for colleague how to reduce momentum residual he said that I have to change turbulence model or set initial velocity as twice of my goal velocity

so I set k-e model (because it's residual was lower than sst) and set initial velocity as 0.1m/s.

then, momentum and turbulence residual were as below

Attachment 59083
Attachment 59084

velocity profile was clearer than before but, still momentum residual has to be lower and velocity profile was still incorrect.

is there somebody know well about natural convection?
and is there other factors to lower momentum residual and make clear velocity profile??:(:(

thank u!

Is your flow in a turbulent regime for natural convection? If not, it probably isn't a correct assumption to be using a RANS formulation.

How are you accounting for density changes? Is gravity set in the correct direction? Are you properly resolving the highest gradients in your flow?

yhs0365 October 20, 2017 02:14

Quote:

Originally Posted by lcarasik (Post 668395)
Is your flow in a turbulent regime for natural convection? If not, it probably isn't a correct assumption to be using a RANS formulation.

How are you accounting for density changes? Is gravity set in the correct direction? Are you properly resolving the highest gradients in your flow?


:o:o:o

thanks for reply.:D


It is natural convection so, I couldn't know whether the flow is turbulent or laminar. So, I assumed the flow as laminar flow but, its reynold's number was far larger than 2300 so, I set the flow as turbulent flow.

i made fluid properties (conductivity,viscosity, density, heat capacity) as function of temperature.

also, I set gravity as -9.8m/s^2

and i can't understand ur last sentence... resolve highest gradient of flow?? what's that mean?

now I'm trying various trials to make residual lower. such as increasing heat transfer coefficient, make pressure loss 0 in porous media...etc...

but, I can't lower momentum residual more below than 1e-2...

lcarasik October 20, 2017 02:16

Quote:

Originally Posted by yhs0365 (Post 668553)
:o:o:o

thanks for reply.:D


It is natural convection so, I couldn't know whether the flow is turbulent or laminar. So, I assumed the flow as laminar flow but, its reynold's number was far larger than 2300 so, I set the flow as turbulent flow.

i made fluid properties (conductivity,viscosity, density, heat capacity) as function of temperature.

also, I set gravity as -9.8m/s^2

and i can't understand ur last sentence... resolve highest gradient of flow?? what's that mean?

now I'm trying various trials to make residual lower. such as increasing heat transfer coefficient, make pressure loss 0 in porous media...etc...

but, I can't lower momentum residual more below than 1e-2...

How are you calculating your Reynolds number? If it is natural convection, you should be using a Raleigh or Grashof number.

yhs0365 October 20, 2017 03:01

Quote:

Originally Posted by lcarasik (Post 668554)
How are you calculating your Reynolds number? If it is natural convection, you should be using a Raleigh or Grashof number.

oh sorry... I mean rayleigh number... I wrote it by mistake

it is far larger than 1e9


All times are GMT -4. The time now is 16:30.