Difference between sum(Wall Heat Flow)@Wall1 and areaInt(Wall Heat Flux)@Wall1

danr · October 11, 2009, 13:25

I'm doing a 2D steady state simulation of a hot pipe in a cold pipe filled with air, with CFX12.
I've prescribed the inner pipe and outer pipe walls at fixed temperatures, and want to find out how much heat is transferred between them by natural convection (no radiation, SST turbulence model). The grid is a one cell thick slice (1mm), with symmetry boundaries either side.

When converged, the heat flowing through the inner wall and outer wall are equal and opposite (sum = 0, as you would expect with a steady state solution).

Previously I've relied on the expression "areaInt(Wall Heat Flux)@Wall1" to give me the heat transfer in [W], but in cfx12 have also opted to save "Wall Heat Flow" to the results files. The values for "sum(Wall Heat Flow)@Wall1" are exactly half that of the flux derived values - dimensions also in [W].

I've done this with many simulations with additional pipes etc and the values always differ by a factor of 2.

Does anyone know why they differ and which is the correct value?

Thanks,
Dan

ghorrocks · October 11, 2009, 18:37

Very strange. You should report this to CFX support as it may be a bug.

I suspect the areaInt(Wall Heat Flux) is the correct number as it should take the integration points into account properly. sum(Wall Heat Flow) would not take into account the integration points so would probably be an approximation. But it is strange the difference is a factor of 2 exactly each time.

I would check the results against the heat flows reported in the output file. I am pretty sure areaInt will be the correct one.

October 11, 2009, 13:25	Difference between sum(Wall Heat Flow)@Wall1 and areaInt(Wall Heat Flux)@Wall1	#1
danr New Member Dan Rhys Join Date: Oct 2009 Posts: 1 Rep Power: 0	I'm doing a 2D steady state simulation of a hot pipe in a cold pipe filled with air, with CFX12. I've prescribed the inner pipe and outer pipe walls at fixed temperatures, and want to find out how much heat is transferred between them by natural convection (no radiation, SST turbulence model). The grid is a one cell thick slice (1mm), with symmetry boundaries either side. When converged, the heat flowing through the inner wall and outer wall are equal and opposite (sum = 0, as you would expect with a steady state solution). Previously I've relied on the expression "areaInt(Wall Heat Flux)@Wall1" to give me the heat transfer in [W], but in cfx12 have also opted to save "Wall Heat Flow" to the results files. The values for "sum(Wall Heat Flow)@Wall1" are exactly half that of the flux derived values - dimensions also in [W]. I've done this with many simulations with additional pipes etc and the values always differ by a factor of 2. Does anyone know why they differ and which is the correct value? Thanks, Dan

October 11, 2009, 18:37		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	Very strange. You should report this to CFX support as it may be a bug. I suspect the areaInt(Wall Heat Flux) is the correct number as it should take the integration points into account properly. sum(Wall Heat Flow) would not take into account the integration points so would probably be an approximation. But it is strange the difference is a factor of 2 exactly each time. I would check the results against the heat flows reported in the output file. I am pretty sure areaInt will be the correct one.