CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   boundary condition for fan simulation (http://www.cfd-online.com/Forums/main/7361-boundary-condition-fan-simulation.html)

Mark March 30, 2004 09:04

boundary condition for fan simulation
 
Could someone say which boundary conditons to use if I want to simulation the flow ventilation fan. The fan should be located in open space and the sourrounding air should have zero velocity. So velocity should only be induced by the rotation of the fan.

Best regards, Mark

Li Yang March 30, 2004 15:34

Re: boundary condition for fan simulation
 
farfield boundary condition where the absolute velocity equals to zero. You may need an attached blade frame system. In an attached blade frame, the flow can be regarded as steady.

Jonas Holdeman March 31, 2004 11:46

Re: boundary condition for fan simulation
 
A problem like this has been on my "to do" list for a couple of years now. If I ever get around to it, I would first do a 2D simulation of a ceiling fan in the center of a room near the ceiling. I am interested in the incompressible flow in the room, not the details of the fan. I use divergence-free Hermite finite elements with stream function and solenoidal velocity components as degrees-of-freedom. Boundary conditions on the room wall are homogeneous: stream function and velocity components are zero. I will put two nodes at the edges of the fan and introduce a constraint that the difference in stream function between the nodes is a constant which depends on the flow rate. I got the dimensions of the fan and flow rate from a brochure or maybe a chart at Lowes or Home Depot.

If I ever get around to the 3D computation, I will constrain the tangential component of the vector potential at the nodes around the blade tips so that the integral of this component around the fan is equal to the specified flow rate. The velocity components and the tangential component of the vector potential will be set to zero at the walls.

This approach does not involve unknown quantities like pressure differences across the fan or other fan details, except that I might block off the center where the fan motor is and constrain the stream function there to be the average of the stream function at the tips in the 2D case.


All times are GMT -4. The time now is 23:44.