How choose the artificial sound of speed to use an artificial bulk modulus for water?

Altair Forum User

Good evening to every body

 

i'm working on sph modelling of water; to obtain reasonable time step

i have to use an artificial bulk modulus in the MAT LAW 6.

I have found in the Monagham teory that it should be chosen using an artificial (lower of the real one) sound of speed that

should be at least 10 times the maximum phenomena speed.

 

I'm testing a water entry of a wedge, and the water surface imapact speed is 3 m/s. The water tank is in complete rest situatuion.

 

Is this speed that i should use as 'maximum speed of phenomena'?

 

Thanks 

Altair Forum User

Hi Damir,

 

Please refer to help.

 

For imcompressible fluids C1 = Rho * c2

 

 

where C is the sound velocity. For instance, C for sea water is ~1500mm/ms and Rho is 1e-6 kg/mm3

 

this makes C1 as 2.25

 

<?xml version="1.0" encoding="UTF-8"?>

Altair Forum User

Hello Prakash

 

yes, i found this in the help...

 

but if i use a real bulk modulus for water, i obtain  a time step too small....

 

In the letterature, Monagham, says to use as sound speed a speed that is ten times the maximum velocity of 

 

phenomena, so very far from real sound speed in water, and it leads to a C1 coefficient lower than real value.

 

Do you know something about this way of thinking?

Altair Forum User

I think this is the it should be.

 

from one of our experts 'You have to be very careful that velocity in the model is not close to the artificial celerity of the sound in water because in this case your flow will become 'unphysically' transonic, resulting in compression and even shocks in the calculation, which are clearly unwanted and often create numerical problems'

 

and here are few recommendations:

 

1. Run your reference with normal bulk modulus (the reference should have the highest speed obviously)
   
2. Run your reference with reduced bulk modulus
3. If both results are close enough : run the variants wth reduced bulk modulus

Altair Forum User

Thanks Prakash

 

i'll try with the normal bulk modulus; in the letterature in fact, i found that the artificial sound speed must be at least 10 times the velocity in the model,

so, far enought from the velocity in the model.

 

 Prakash i would need an other help from you or from anyone else

I don't know how to estimate the quadratic bulk viscosity and the linear bulk viscosity coefficients Qa and Qb in the /PROP/SPH; if i'm not wrong

the quadratic bulk viscosity should be set at zero, because in my case i have a newtonaian fluid with a mach number < 0.1 

But the linear coefficient Qb?

 

I'm not able to understand the role of ALPHAcs coefficient!! i know that this is a coefficient that identifies a kind of approximation of the particle's speed;

it should smooth the velocity of a certain particle adding the contribution of the close psrticle's velocitied weighed with ALPHAcs coefficient.

How to estimate this?

 

What kind of order to use in SPH property card? 0 or -1? i don't know what coefficient use in my simulation, water entry of a wedge.

 

waiting for a help 

 

thank for all

Altair Forum User

Hi Prakash

 

Have you had time to check the .hm file and the videos of simulations results i sent you?

 

(About the wedge water entry by SPH method)

 

thank you 

Altair Forum User

Hi Damir88,

 

yes, 

 

 regarding oder in SPH, please use 0 as it was recommended.

 

Regarding Alphacs, since you are running a shock simulation with sph, it may not be an ideal case to use this because it may produce wave propagation.

 

This is recommended to solve for stability in traction problems.

 

FYI,

 

http://www.ihed.ras.ru/rusbank/expsim/

 

this website may interest you /emoticons/default_smile.png' alt=':)' srcset='/emoticons/smile@2x.png 2x' width='20' height='20'>

Altair Forum User

Hi Prakash

 

the website you suggest me, make me calculate the value of the particle velocity in a target material (water)

knowing the velocity of the striker: by the water particle velocity i can determinate the maximum speed in water

and so the water bulk modulus...was this the aim of the link you send me?

 

Have you any advice about the values Qa and Qb sph parameters? i don't know what to do, because my sph behaves too

viscous and i can't obtain the same behaviour of my experience of sphere entry in water .

 

thank you Prakash

Altair Forum User

Hi DAMIR88.

 

For shock and wave propagation cases Qa=2 and Qb=1.

 

when in-compressible is set then drop the Qa and Qb values to much lower range.

Altair Forum User

ok, but i have a simulation with Qa and Qb set to 2e-3 and 1e-30 , but the water continue to behave

too viscous; and the sphere remains trapped in water, seems it behaves like an oil.

 

The material used is MAT/LAW6;

 

in this simulation i use the real bulk modulus.

 

Could it be too low values of parameters?

Altair Forum User

Damir88,

 

Can you input physical viscosity in the material law and set Qa and Qb to zero and check the behavior.

Altair Forum User

Prakash

 

Now i'm trying with these Properties, do you think they are right?

 

Mat/LAW6  Water: Kinematic Viscosity=1e-6 Pa and C1=2,2 GPa and Pmin -1e-30Pa  and all others parameters set to zero

Sph Property: Qa=2 Qb=1 Alpha_cs=none Order=0

Altair Forum User

Damir88,

 

Yes I think you should get some decent result now.

Altair Forum User

Hi Prakash

 

The units in the model i sent you yesterday in dropbox  are:  meters, kilograms and seconds

Altair Forum User

okay thanks, I will check and update to you /emoticons/default_smile.png' alt=':)' srcset='/emoticons/smile@2x.png 2x' width='20' height='20'>

Altair Forum User

okay thanks, I will check and update to you /emoticons/default_smile.png' alt=':)' srcset='/emoticons/smile@2x.png 2x' width='20' height='20'>

Altair Forum User

Hi Damir,

 

I used your model (for a run time 1) with Qa=2 and Qb=1 and Order= 0(default values) and compared it to your previous results.

 

I can see a bit of splashing in your previous results compared to the one which I had. But the it is time expensive with the values I used.

 

Other than this I did not see any much difference in the model. 

Altair Forum User

Hi Prakash

I have run my model with Qa=1e-20 and Qb=1e-20 and it behaves very naturally

I have run, also the model with Qa=2 and Qa=1 and it behaves strongly viscous

The modes are the sames and i used an ORDER=-1 and Alpha_cs=0.001 in both of them.

They were the sames, a part of Qa and Qb.

i send you two small video in dropbox

thanks

Altair Forum User

So the question is:

 

is there a relation to take in account to understand whitch values of Qa and Qb to use,

to get a natural behaviour of water?

 referring to the previous two models, for me, the real behaviour of water was that described by Qa and Qb both 1e-20,

and not with the default values 2 and 1

 

what do you think?

Altair Forum User

Hi Damir,

 

from one of the text books:

 

Qa is the Quadratic bulk viscosity helps to smear the shock (example oil) and Qb reduces the oscillations. 

 

So, it is obvious that when you have higher Qa and Qb values, you see a sticky and viscous behavior. I hope this is clear.

Altair Forum User

Hi Prakash

 

yes, now is clear.

 

i was occuped for an exam, i'm going to continue the work with SPH.

 

Damir

Altair Forum User

Good Luck Damir /emoticons/default_smile.png' alt=':)' srcset='/emoticons/smile@2x.png 2x' width='20' height='20'>

 

Let us know if you need any help from our side.

 

Altair Support forum is always open /emoticons/default_smile.png' alt=':)' srcset='/emoticons/smile@2x.png 2x' width='20' height='20'>