CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   FMG initialization query (http://www.cfd-online.com/Forums/fluent/79006-fmg-initialization-query.html)

Mohsin August 10, 2010 00:25

FMG initialization query
 
Hello

I have following querries regarding FMG initialization. I will be grateful for your kind look at these:

1. When I run Fmg initialization without starting any iterations then it gives me the following error:

/solve/initialize> fmg-initialization
invalid command [fmg-initialization]

Please specify. Why it is so. As its prime function is to give the initial solution so isn't it should be specified before iterations?

2. When FMG initialization is complete I get the following message:

FMG: Initialize flow for Segregated solution.. . end

It means that FMG initialization has been done and after this I dont need to change the initial values from solve...> initialize. And I only have to iterate now?

3. I am using multiphase flow DPM. But first i think i should run FMG initialization without DPM and then introduce DPM later(making it single phase in order to for FMG initialization to run). Is it possible.

Please help me. I will be grateful

Mohsin

xrs333 August 15, 2010 02:56

Hi Mohsin,
1 the flow need to be initialized, but iterations is not pre-requisite, before fmg-initialization, because fmg-initialization is also a iterative algorithm.
2 Yes, just click on 'iter'.
3 yes, it is preferable to get a converged flow before the particles are introduced into the compute domain if you are solving a steady case.

best wishes

Mohsin August 15, 2010 03:57

Quote:

Originally Posted by xrs333 (Post 271519)
1 the flow need to be initialized, but iterations is not pre-requisite, before fmg-initialization, because fmg-initialization is also a iterative algorithm.
2 Yes, just click on 'iter'.
3 yes, it is preferable to get a converged flow before the particles are introduced into the compute domain if you are solving a steady case.

Thank you. I have some queries regarding your answers.

1. Do i need to initialize it with any value. I mean for instance can I initialize it with "all zones" or from "inlet"? it wont matter coz FMG initialization will give initial values by itself?

2. But after running FMG initialziation the values in solve..>initialize panel have not been updated to FMG initial values. Aren't they should be updated and how can i know the initial vaues for my pressure and velocity?

3. I m using steady state two way coupled DPM with "update discrete phase iteration with every flow iteration" feature turned on. Is it still possible to calculate velocity field prior to calculating DPM.

Thank you for your kind help.
Mohsin

xrs333 August 15, 2010 04:32

1 yes, just like with no fmg-init. As to the initial value, my suggestion is that you initialize the flow field with a set of values of the variables for a physical state, say, a state of zero velocity everywhere in the domain is a real world state.

2 the initialize panel need not to be cared for after iteration starts or fmg-init runs. Actually, fmg-initialization solve the Euler eq.'s and you can evaluate the results using conventional post-processing tools after fmg-initialization, just like you do whenever the iteration is interrupted.

3 yes. You may start the iteration before any injection is defined. If the injection has been created already, you can uncheck the interact with continuous phase option temporarily, and then turn it on any time suitable.

Mohsin August 16, 2010 01:21

Thank you very much.

I have 2 last questions to clarify my query

1. It is written that FMG initialization is not valid for multiphase flow. But when DPM injections are activated. Still i am able to run FMG initialization. isn't it contradict the manual's statement?

Quote:

Originally Posted by xrs333 (Post 271524)
3 yes. You may start the iteration before any injection is defined. If the injection has been created already, you can uncheck the interact with continuous phase option temporarily, and then turn it on any time suitable.

2. If "continous phase iteration with every DPM iteration" is to be calculated in the flow field. Isn't it necessary to calculate DPM iterations with flow iterations?? or its fine to calculate it seperately by first calculating flow field then from particle tracks DPM.

(if "continous phase iteration with every DPM iteration" is checked then the calcutaions for conitnious phase are updated as source terms are created in the continous phase from discrete phase after specified iterations) so can we still calculate DPM seperately?

Just a bit confused:confused:


All times are GMT -4. The time now is 22:07.