How to compute the viscous dissipation term
 

Hi, everybody. I am performing 2D unsteady CFD simulations on a cross-flow fan. I am using ANSYS FLUENT 12.1. I adopt the standard k-omega model to simulate turbulence.
I want to compute the viscous dissipation term. I cannot find this variable in FLUENT. Hence I exported the solution to Tecplot, and then write some equations to compute it.
I use the following equation to compute the viscous dissipation term: (This equation can be found in ANSYS FLUENT 12.0 Theory Guide, page 4-24 and 4-25)
viscous dissipation=
where


I use [Data]|[Alter]|[Specify Equations…] in Tecplot to implement these equations. The expressions which I type into Tecplot are listed as follows:
{duxdx}=ddx({x-velocity})
{duxdy}=ddy({x-velocity})
{duydx}=ddx({y-velocity})
{duydy}=ddy({y-velocity})
{viscous_heating2}=ddx({x-velocity}*({viscosity-eff}*({duxdx}*2)-2.0/3.0*{viscosity-eff}*({duxdx}+{duydy}))+{y-velocity}*({viscosity-eff}*({duxdy}+{duydx}))) + ddy({x-velocity}*({viscosity-eff}*({duydx}+{duxdy}))+{y-velocity}*({viscosity-eff}*({duydy}*2)-2.0/3.0*{viscosity-eff}*({duxdx}+{duydy})))
Finally, I display contour of the variable “viscous_heating2”. However, to my surprise, the minimum value of this variable is -679152, and the maximum value is 352578. This is weird – because the dissipation should be always positive. How could negative values occur?

One possible cause of your troubles is that you are using the ddx functions in tecplot to compute derivatives. That's among the worst methods you can use, they are first oder accurate (at most second), and depending on your grid, that can be way to inaccurate.
Can you compute your derivatives in your code? On the original mesh maybe? Then there's at least the chance that you are not hunting tecplot ghosts with this....

Thank you very much. I'll try it.