[TheBoyce]

I am trying to model stratified spin-up from rest. My pressure update equation is implicit and therefore the result is large set of linear equations to be solved. The structure of the coefficient matrix is tridiagonal with fringes which is a very sparse matrix. To solve this system of equations I am using the iterative biconjugate gradient method. My problem is that in order to get a solution an arbitrary point needs to be set to an arbitrary value, however, the point at which I define the pressure is acting as a sink. The value of the pressure I set is simply added to the entire pressure field and the pressure gradients are not changed. The location of where I set the pressure does make a difference to the pressure gradients and therefore alters the strength of the sink as it is the pressure gradients that are used in the velocity update equations. Any ideas as to why this sink is occurring and any ideas as to how I can remove it would be greatly appreciated.

Thanks in advance.

[TheBoyce]

Any ideas? Please? I'm desperate.

[CFDtoy]

If I were you, I will look closely at my pressure correction eqn and see how the af terms (mass flux term) is modeled. Be careful implementing your rhie-chow interpolation and since you have body force, make sure you are absorbing rho*g within p, pstatic should be treated as effectively pcell-rho*g...body force involved.

I think your treatment of the pressure for evaluating face vel will be crucial..take a look at that.

good luk.

CFDtoy

Originally Posted by TheBoyce

Any ideas? Please? I'm desperate.