Regarding choosing correct boundary condition

Hi,

I have a condition where at the left boundary and the right boundary I need the flux to be non-zero.

Thus, the equations

pL + qL*fL = 0 and pr + qr*fr = 0 are such that fL and fr being non zero is making my equation hard to determine the values for pl, ql as well as pr, qr.

Plz tell me how should I approach for the same.

with regards,

rc

Answer 1

[Torsten]

If you set qL = 1 and qr = 1 and you defined f as f = D*du/dx in the function where you need to specify [c,f,s], you can specify the flux in pL and pr.

[Rahul]

Hi Torsten,

Thanks a lot.

But I think the term Pi_s in the flux is a problem. How to fix this term?

rc

[Torsten]

I cannot open your bc.pdf. Thus I don't know what Pi_s is.

[Rahul]

Hi Torsten,

plz find attached.

rc

[Torsten]

What is under the d/dr ? What did you specify as f ? What do you want to set as boundary condition ?

[Rahul]

Hi,

the flux is

 f = -1*u(2)*((v0+v1*epsilon_v)*DuDx(4)+ pi_s);                            

Here,

v_phi denotes the velocity denoted here as u(4).

Also,

pi_s = k * epsilon_v (1-sigma * dp/dr)*gamma

where, k denotes coefficient of equation.

rgds

[Torsten]

What is u(2) ? Is it n ?

The flux for u(4) should not include pi_s.

Put 1/r d/dr(r*n*pi_s) in the source term s(4) and remove it in the specification of f(4).

[Rahul]

Hi, Torsten,

Yes, u(2) is n.

Ok, I'll do so. Thanks a lot.

rgds,

rc