Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Expansion - Deflection Nozzle

Hello everybody,

I'm a student in the Von Karman Institute in Belgium.
I've a project about an expansion - deflection nozzle. I would like to realize a Comsol model of it. You can see on the pic that it's pretty simple. Just a nozlle with a chamber pressure at the top (102 bar and 2400 K) and a atmosperic pressure for the outlet. I've choose a k-eps model but my calculations never converge !! I don't know what to do. If you can help me.

Just to know, it's not my first model with Comsol. But I cannot understand what happend with this one.
If you need more details, tell me.

Thanks a lot for your answers !


3 Replies Last Post May 25, 2010, 12:10 PM EDT
Jean-Francois Leon
Posted: 7 years ago May 25, 2010, 6:53 AM EDT
well you dont provide a lot of info and the attached pic is useless... is it compressible fluid or incompressible for example air water other...
But anyway,
I dont think Comsol can solve this problem. Based on the pressure and temperature you give , it is going to be fully turbulent probably supersonic expansion .. you will need state of the art specialized CFD codes and even there the guidance of an expert in these calculations will most probably be needed. also 2400K is pretty hot you will need meaningful EOS because at least initially your fluid should have some level of molecular dissociations and recombination during expansion might be significant.
And again except MAYBE if you code everything using pde modes, and this will be a major work to optimize the solver and validate it, I dont htink Comsol can deal with that ...
sorry for the bad news...
JF


well you dont provide a lot of info and the attached pic is useless... is it compressible fluid or incompressible for example air water other... But anyway, I dont think Comsol can solve this problem. Based on the pressure and temperature you give , it is going to be fully turbulent probably supersonic expansion .. you will need state of the art specialized CFD codes and even there the guidance of an expert in these calculations will most probably be needed. also 2400K is pretty hot you will need meaningful EOS because at least initially your fluid should have some level of molecular dissociations and recombination during expansion might be significant. And again except MAYBE if you code everything using pde modes, and this will be a major work to optimize the solver and validate it, I dont htink Comsol can deal with that ... sorry for the bad news... JF

Posted: 7 years ago May 25, 2010, 8:33 AM EDT
It's a compressible flow highly hypersonic.
We assume that the flow is frozen, so there are no dissociations inside the flow (yes it's a big assumption but we have no other choice for the moment).
So according to you, it's not possible to perform these calculations under Comsol ?
I'm sad to hear that !
Yeah, I also use Fluent, but I wondered that Comsol'll do the same thing !
Thanks for your answer
It's a compressible flow highly hypersonic. We assume that the flow is frozen, so there are no dissociations inside the flow (yes it's a big assumption but we have no other choice for the moment). So according to you, it's not possible to perform these calculations under Comsol ? I'm sad to hear that ! Yeah, I also use Fluent, but I wondered that Comsol'll do the same thing ! Thanks for your answer

Jean-Francois Leon
Posted: 7 years ago May 25, 2010, 12:10 PM EDT
can you explicit what you mean by "frozen flow" ...do you assume adiabatic expansion? .. why are you doing the modeling...? [ depending on the goal you can define your "accuracy need" and from there see what modeling you can do to get wha you need].

Being more a physicist than an engineer myself , I dont like "engineer model" like the k-eps stuff..even if I know they are very useful in a lot of case.
It is not my cup of tea...
So if i were you I will write the gas dynamic equations [ I guess that mass and momentum should be enough if you dont have heat exchange or real world eos..] and see what it gives... you will probably dont succeed in getting any practically usefull result [ sorry I am just expressing my opinion here for euler equations are a bitch to solve...] but you could learn a lot in the process and comsol is a good tool for that.
I know a cfd module is announced for comsol in a few months maybe it will solve your problem in a more "black boxed way" .... or not... so being patient is another option you have... :-)
good luck
JF
edit depending on what is important for you, you might find some configurations locally stable that could be approached by simpler methods... it goes back to why do you want to do the modeling?
with nozzle usually it is modelling the divergence of the gas an predict the angular distribution of the mass at the exit of the nozzle..
These problems can be solved in simpler way than full fledged modeling...to support nozzle shape optimization/design for example.
many things have been published here you should find good papers with proper reference that will tell you how do the modeling in a practically usefull way
can you explicit what you mean by "frozen flow" ...do you assume adiabatic expansion? .. why are you doing the modeling...? [ depending on the goal you can define your "accuracy need" and from there see what modeling you can do to get wha you need]. Being more a physicist than an engineer myself , I dont like "engineer model" like the k-eps stuff..even if I know they are very useful in a lot of case. It is not my cup of tea... So if i were you I will write the gas dynamic equations [ I guess that mass and momentum should be enough if you dont have heat exchange or real world eos..] and see what it gives... you will probably dont succeed in getting any practically usefull result [ sorry I am just expressing my opinion here for euler equations are a bitch to solve...] but you could learn a lot in the process and comsol is a good tool for that. I know a cfd module is announced for comsol in a few months maybe it will solve your problem in a more "black boxed way" .... or not... so being patient is another option you have... :-) good luck JF edit depending on what is important for you, you might find some configurations locally stable that could be approached by simpler methods... it goes back to why do you want to do the modeling? with nozzle usually it is modelling the divergence of the gas an predict the angular distribution of the mass at the exit of the nozzle.. These problems can be solved in simpler way than full fledged modeling...to support nozzle shape optimization/design for example. many things have been published here you should find good papers with proper reference that will tell you how do the modeling in a practically usefull way

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.