Learning Forum

[tosuccess]

I'm experimenting about a negative pressure device.

I used the pressure inlet and exhaust fan outlet, and assiged a gauge pressure value of 0 [Pa] and -2.5 [Pa] to those values, respectively.

In addition, I used default values for other values, but the residuals do not converge and diverge.

Can you point out the problem in my project?

[Rob]

You may. 

[NickFL]

If you want help, we would need a little bit more details. The fact of the matter is, the inlet pressure & outlet pressure boundary conditions are inherently unstable because we have no condition that connects the velocity and the pressure. A better approach would be to specify a mass flow and a pressure and iterate the mass flow until you achieve the pressure drop you want.

[tosuccess]

I have a question.

Should I use the 'mass-flow-outlet' condition at the boundary conditions if I know the air volume (Q) for the initial condition?

[NickFL]

- Yes, if you have the volumetric flow rate, and assuming you know the density, then the mass flow rate can be computed. This can be set at the boundary condition, or you could convert it to a velocity condition.

- You mentioned that you have an exhaust fan outlet, have you validated that the polynomial coefficients are correct? Perhaps it is best if you switch this to the mass-flow-outlet and use the value you computed from above.

[tosuccess]

thank you.

Then, do I have to turn the simulation while changing the inlet condition by fixing the outlet setting?

[NickFL]

I am a bit unsure on your question. If you specify the mass flow rate (from the volumetric) you would compute the corresponding pressure drop through the domain. If you are certain of this flow rate, then we would not need to iterate because you know it.

It really depends on what you are trying to model and what known boundary conditions you have. Earlier you stated that you knew the pressure difference, and later added you knew the volumetric flow rate. Mathematically we cannot specify all three (look at the NS equations and BCs closely--it is not possible). A general approach would be take it to model it with two (mass flow & a pressure are the best) and use the third (other pressure) as a validation point. But best approach we cannot exactly say without knowing all the details.