[CFD-POST] How are gradients calculated?

ghobold · February 27, 2014, 09:00

I'm having trouble understanding how gradients are obtained by CFD-Post. I am running an adimensional problem in Fluent and I want to calculate the Nusselt number in CFD-Post. However, whenever I probe the temperature gradient on the pipe wall, Fourier's law is never obeyed. In other words:

probe(Wall Heat Flux)@wallPoint ≃ 1 (which is what is set up in Fluent, so it's OK), but

-probe(Thermal Conductivity)@wallPoint*probe(Temperature.Gradient )@wallPoint ≃ 0.719, which contradicts Fourier's law.

It's an axisymmetric problem (simple circular duct), so Temperature.Gradient Y ≃ Temperature.Gradient, but Fourier's law is not obeyed.

brunoc · March 26, 2014, 08:24

CFD-Post uses shape functions to calculate gradients. They are detailed in the CFX documentation.

Fluent stores the results in the mesh elements/cells, whereas CFX stores them in the mesh nodes. So when CFD-Post read a Fluent result, it interpolates the cell values to nodes, and all calculations are then node-based (including the gradients based on shape functions). That is why you see a difference. Fluent, on the other hand, calculates them internally, but the temperature gradient calculated by Fluent is not available to CFD-Post by default.

You should always trust the solver computed values (Wall Heat Flux, Wall Shear, etc), not the one you manually computed.

Cheers.

February 27, 2014, 09:00	[CFD-POST] How are gradients calculated?	#1
ghobold New Member Gustavo Join Date: Jun 2013 Posts: 26 Rep Power: 13	I'm having trouble understanding how gradients are obtained by CFD-Post. I am running an adimensional problem in Fluent and I want to calculate the Nusselt number in CFD-Post. However, whenever I probe the temperature gradient on the pipe wall, Fourier's law is never obeyed. In other words: probe(Wall Heat Flux)@wallPoint ≃ 1 (which is what is set up in Fluent, so it's OK), but -probe(Thermal Conductivity)@wallPoint*probe(Temperature.Gradient )@wallPoint ≃ 0.719, which contradicts Fourier's law. It's an axisymmetric problem (simple circular duct), so Temperature.Gradient Y ≃ Temperature.Gradient, but Fourier's law is not obeyed.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
dsmcFoam setup	hherbol	OpenFOAM Pre-Processing	1	November 19, 2021 02:52
[CFD - Post] Discrepancy between areaInt of Pressure	ziggo	Visualization & Post-Processing	0	September 4, 2013 12:25
[CFD Post] Rescale the Viewer	max3.2	ANSYS	1	September 5, 2011 15:13
[CFD Post] Surface streamlines not located where needed	siw	CFX	1	July 7, 2011 20:13
[CFD Post] Displaying boundary layer profiles	siw	CFX	0	July 4, 2011 13:15

March 26, 2014, 08:24		#2
brunoc Senior Member Bruno Join Date: Mar 2009 Location: Brazil Posts: 277 Rep Power: 21	CFD-Post uses shape functions to calculate gradients. They are detailed in the CFX documentation. Fluent stores the results in the mesh elements/cells, whereas CFX stores them in the mesh nodes. So when CFD-Post read a Fluent result, it interpolates the cell values to nodes, and all calculations are then node-based (including the gradients based on shape functions). That is why you see a difference. Fluent, on the other hand, calculates them internally, but the temperature gradient calculated by Fluent is not available to CFD-Post by default. You should always trust the solver computed values (Wall Heat Flux, Wall Shear, etc), not the one you manually computed. Cheers.