Learning Forum

[Ohiro]

Hi, I'm simulating gasflow into the cylinder, but mass flow rate does not match for each cross section.(residuals less than 1.0E-15 so convergence is good enough)

What do you think about the reasons? Is the difference acceptable?

[Rob]

Are those planes/iso surfaces that you're monitoring?  How does the mass flux look on inlet & outlet?

[Ohiro]

Hi Rob, inlet is 1.9755497e-05, and outlet is -1.9755497e-05.

Inlet should be the same surface with delta_y=0.06 and outlet should be with delta_y=0, but the values are not equivalent to those in the second image.

[Rob]

OK, and you're using a plane or isosurface for those other faces? 

[Ohiro]

Right, all 7 faces normal to the y-axis are isosurfaces. They are named delta_y=0, delta_y=0.01,... from the right side of the picture.

[Rob]

OK. Iso surfaces & planes use interpolated data, ie they interpolate from the nearest cell centres and then integrate/average/whatever. As a result they'll always be marginally less accurate than a defined surface from the mesh (interior, inlet, wall etc) which stores data on the facets. Assuming y=0 & y=0.06 are on the flow boundaries the interpolation may be a little worse than surfaces inside the domain as there's less data available. 

[Ohiro]

I understand.

So, if there is a plane I want to see the flow rate (or other physical quantities) on the inside, should I make a defined surface by cutting the body so that the face is boundary?

[Rob]

If you need exact values yes, split in geometry or (whole cells only) you can split the mesh in Fluent. In most cases the difference is negligable, so you may need to check mesh resolution. 

[Ohiro]

OK. Thank you so much for your answer, Rob!