Activating gas radiation leads to energy imbalance
 

Hi folks!

I am currently doing the official ANSYS tutorial with the can combustor (see picture, I think it's available in Fluent 2019 R1 only, Chapter 14), however with radiation included. I am using the k-epsilon standard, eddy dissipation for species transport and P1 model for radiation. When I keep the Absorption Coefficient of the mixture (methane-air) at 0 and check the Flux Reports of the (converged) solution, there is almost no energy imbalance in all three heat flux reports (Net result = 0-1 W).

However, when I change the Absorption Coefficient of the mixture to a non-zero value (either putting a constant or using the WSGGM), there is always an energy imbalance for the Radiation Heat Transfer Rate Flux Report, the other flux reports, however, are again 0-1 W (Mass Flow Rate is also at -1e-9). The residuals look pretty good and converged (Continuity at 1e-4, Energy at 1e-6, P1 below 1e-08, Species between 1e-4 and 1e-6, staying at constant level). I decided to observe the net radiation heat transfer rate during simulation and it also seems to converge but it stays at level that is way too high with a few thousands Watt (see pictures)). The mass-averaged temperature of the domain is also stable and does not change anymore. Same goes for the incident radiation for the walls.

The under-relaxation factors are all set to default. I am using the coupled scheme with pretty much everything second order (except Gradient). Steady. No pseudo transient. I've tried other cases including a simple cylinder and seeing always the same phenomenon: Absorption Coefficient set to zero works fine, but non-zero cases always lead to an energy imbalance for the Radiation Heat Transfer Rate.

You can find the boundary conditions below. Everything's the same as in the tutorial except for the wall whose temperature is set to 1300K (max material temperature, may not be a proper BC in this case but that's a different story) and the inner emissivity factor to 0.85. The walls of the swirl vanes are assumed to be adiabatic but they're rather small compared to the rest. The temperature contour (see picture) looks fine and pretty much the same as in the case of no radiation at all.

Hence the question is: why is there an energy imbalance in the Radiation Heat Transfer Rate flux report?

Hi Seph


is maybe a while, but did you get any information on that?

I am having a similar problem when using a DO model in a mesh that uses symmetry and has steam as participating media. The energy balance is totally wrong when using radiation..