# Problem with cfx Solver Results

 Register Blogs Members List Search Today's Posts Mark Forums Read

 August 26, 2008, 05:00 Problem with cfx Solver Results #1 Ordoumpozanis Guest   Posts: n/a Hi I am running a case of double facade with buoyancy. I have adiabatic walls except one witch have a heat flux and a boundary of p=patm at inlet and m= const at outlet. I noticed that at outlet, the velocity that I get is different from the velocity witch I expect. I tried to run the same case with boundary condition of normal velocity = const at outlet and the mass flow that I get is different from the one I except. In the problem I use a K-e model with constant density. Can anyone think a reason for this situation ??? Tnanks

 August 26, 2008, 16:25 Re: Problem with cfx Solver Results #2 CycLone Guest   Posts: n/a What velocity did you expect?

 August 26, 2008, 18:15 Re: Problem with cfx Solver Results #3 brunoc Guest   Posts: n/a So, correct massflow = wrong velocity correct velocity = wrong massflow Is the density of your material correct? Can you assume it to be constant?

 August 27, 2008, 04:07 Re: Problem with cfx Solver Results #4 Ordoumpozanis Guest   Posts: n/a Hi, thnk for the reply. I managed to get the right results by adding the normal velocity only to the calculations on the boundary, witch is logical. but if I check each cell on the boundary, the equation of Acell = mass flow (m3/s)/ velocity (m/s) is not equal to the area of the cell on the boundary. The values for the calculations are output from post process thnks again

 August 27, 2008, 11:54 Re: Problem with cfx Solver Results #5 CycLone Guest   Posts: n/a Hi Ordoumpozanis, This is actually expected. The solver calculates mass flow rates at integration points, where the velocity is not equal to the nodal velocity. If you calculate the mass flow rate as rho*A*V, you'll actually get the wrong results. Post avoids this error by using the integration point mass flows, which are written to the results file from the solver. The actual calculation of mass flow rate at the integration point is rather involved and must include the spatial variation of both velocity and pressure. You can review how this is discretized in the solver theory guide. -CycLone