For outlet if I use pressure boundary default "zero". Does Phonenics fix this pressure in the domain(the outlet region surely) or just use "cm(vm-Pexp)T " to make this region close to "zero pressure" during the iteration process if I get converged results finally?

Anyone can help me? Thanks

PHOENICS will use the "mdot=cm(vm-Pexp)T " expression to keep the pressure in the outlet region close to vm. vm is the external pressure, cm is the pressure coefficient and Pexp is the calculated pressure in a cell.

In PHOENICS-VR the external pressure and pressure coefficient are set in the Attributes dialog for an Outlet object. You can see the PATCH and COVAL created by looking in Group 13 of the RESULT file.

The bigger the coefficient, the closer the in-cell pressure will be to the set external pressure.

Some more information on pressure and mass boundaries can be found in the POLIS Encyclopaedia at http://www.cham.co.uk/phoenics/d_pol...nc/coval.htm#e.

Thanks,John, Let us talk this in some more detail which is my real concern. If mdot=cm(vm-Pexp)T is used I know this term must go into P correction equation then flux correction on boundary must equal to cm*T(-P') if cm is big enough then P' will close to zero so this will get that P is close to vm. Then another problem is if P is close to or equal to vm then how do you determine the inflow or outflow by compare P(incell pressure) and vm(external pressure)since P is forced not to be corrected during iteration and the local continuity is not satisfied.

Problem 2:Sometimes when P is large than vm(incell large than outside phoenics may say it is outflow) if you apply local continuity you will find the flow flux on the boundary may be inflow. So I guess the local continuity is not used to get flow flux on boundaries Phoenics only use cmT(vm-p) but how this term is corrected for each iteration especially for outflows you never specify any outflow rate there.

I am confused by phoenics method for long time. Thanks,

The source term in the P' equation is the continuity error for the cell. This is calculated from the velocity field and any mass-source contributions. The resulting pressure corrections are then used to adjust the pressure and the velocities to get a continuity-satisfying velocity field. The velocities are then solved again using the new pressure field and the procedure repeats. The basic pressure-correction method is described in this book:

S.V. Patankar 'Numerical heat transfer and fluid flow', Hemisphere,1980.

PHOENICS does not use this algorithm exactly, but the basic principle is there.

The mass source reported at mass-source patches is purely based on sum(Cm.T.(Vm-Pp)) over the cells in the patch. It is quite common for some cells in the patch to be outflow (Pp>Vm), and some to be inflow (Pp<Vm).

The outflow rate at a fixed pressure boundary is never specified - it comes from the solutions of the momentum and continuity equations. If you want to specify a fixed mass-outflow rate, you have to specify a negative mass flow rate (COVAL(name, P1, FIXFLU,-mdot)).

The pressure/mass flow boundaries are like voltage and current. Fixed pressure is like a fixed voltage - the mass flow (current) will adjust according to the resistance. Fixed mass flow is like setting the current - the pressure (voltage) will adjust to push that much mass through the system.

You can try running library case 240, which is a very simple duct flow. Add the settings STORE(IMB1,PCOR) and NPRINT=1 at the end of the Q1. This will make the continuity errors (IMB1) and pressure corrections (PCOR) be printed in the RESULT file, and the flow fields will be printed every sweep. You can then compare how the fields change with sweep for different values of the pressure coefficient at the exit. The case is set up with a value of 1.0, which is too small to keep the pressure near to zero.

You can also try replacing the inlfow boundary with another pressure boundary, but with a higher external pressure. The flow-rate through the system will be determined by the balance between the pressure difference across the ends and the wall-friction from the duct wall.

Good suggestions!Thank you. Leon