Entropy calculation using CFD
 

Hello there,

I am looking for a way to calculate sum of entropy rises due to viscous effects, shoks, temp. gradient etc. throughout the computational domain and compare it with entropy difference of outlet and inlet.

I wonder if anyone can help or introduce a reference.

Appreciated

MJK

Re: Entropy calculation using CFD
 

try these links ...

http://www.memagazine.org/backissues...y/entropy.html

http://www.memagazine.org/backissues...py/entgen.html

Re: Entropy calculation using CFD
 

Thanks Walter,

In fact that was the only source I had found on this matter- of course other than what Bejan has explained in his text.

I wonder if anyone has tried such an approach and been able to match entropy generation due to viscous dissipation and temp gradient WITH the overall entropy rise throughout the domain.

Regards

MJK

Re: Entropy calculation using CFD
 

Entropy is a state variable, as such it may be computed from two other state variables (like pressure and temperature). Why can you just not compute it at each of your points in the domain.

Re: Entropy calculation using CFD
 

Bob,

That is right- entropy at any point can be obtained in terms of two other independent variables.

I was asking how is it possible to obatin and compare entropy generation within the domain due to (1) viscous effects, (2) thermal effects and (3) shock waves.

Some good explanation is given in text by A Bejan- but it does say no wrod about the shock effect.

Regards

Mohammad

Re: Entropy calculation using CFD
 

Since the CFD solution gives you just the values of the variables (including entropy if you've computed that), it is likely to assess how much a change at any point came from what physics. However, you may be able to get an idea by figuring out where you are in the domain. For instance, if you are near a shock wave, compare the entropy rise with the theoretical. The difference will be the sum of the other physics and the CFD error associated with any nonconxservation of entropy. Away from shock waves, there will be no entropy generation due to that effect. Similarly, within a boundary layer, you may assume the entropy rise is due viscous and thermal effects. To separate out the thermal effects from the viscous, I suppose you could do an inviscid computation for the same conditions (except for using a slip BC on solid surfaces) and check the difference.