Learning Forum

[J.LoongHee]

Hi all, I have 4 questions to ask (they are related to each other):

• Q1. When there is no radiation switched on (just combustion species transport + turbulence model), the net results (W) for total sensible heat transfer is 1% of the heat of reaction source and we could 'judge' that the simulation is converged to a final state (of course we do check the temperature outlet and velocity outlet monitoring probes). However, this general rule doesn't apply when wall radiation and wall conductivity model is switch on. WHY? (see point 2 & 3)
• Q2 As you can see in the below details (2 tubes cooling case), when radiation model is switch on along with wall thermal conduction (temp boundary conditions). The net results (W) for total sensible heat transfer can only be balanced if we take the outlet value VS the heat of reaction source value. 
• Q3 In another simulation (10 tubes cooling), same settings for all tubes and submodels. The net results (W) for total sensible heat transfer is even further from heat of reaction source. Again why? (see screenshot Q3). From the physics perspective, the 10 tubes seems to reduce the heat transfer exiting the outlet domain, that make sense but what is the new convergence criteria in this scenario since by comparing the outlet value VS the heat of reaction source, it is further away from each other. 
• Q4 During the simulation processes, the solver will occasionally prompt (temperature limited to 1.0+0) in 12 cells on zone 10 in domain 1 (see image attached), by trying to identify the cells which has high temperature using ISO-surface, it is impossible to find these cells. How do we completely eliminate these temperature limited cells? Does it affects the overall convergence?
  
  Q2 (Details):
  I have a furnace with a burner (Mesh size 5.5m cells) it has 10 circular tubes (2 tubes set as wall with prescribed lower wall temperature i.e., XX^C double digit Celcius, 8 tubes set as interior).
  Settings:
• Steady-state simulation
• Turbulence KW GEKO
• Energy On
• Radiation On 
• Species Transport Volumetric Methane-Air (appropriate Air:fuel flow rate 10:1 Ratio) 
• 8 tubes set as INTERIOR (no wall)
• 2 tubes set as wall with constant lower temperature (cooling - double digit celcius)
  Domain: Rectangular with burner at one end, several tubes at the sides of the furnace walls and a converging exit at one end
  
  
  Q3: Heat of Reaction Source: 241.9kW vs  outlet -213.4 kW. 
  
  
  

 

Q4: 



Thank you
Jee 

[Essence]

Hello,

Do not use GEKO model. Try SST model. How is the mesh quality? What is the maximum skewness? Which Turbulence-Chemistry model are you using? Did you input correct mixture properties in the Materials?
It would be better if you would upload an image of the model.

[J.LoongHee]

Hello Essence,

Do not use GEKO model: Okay. I will change the model to SST. The mesh comprised of both mainly hexahedral + very small tetrahedral zone (bridging).

Which Turbulence-Chemistry model are you using?: I am using species-transport (Eddy Dissipation) - Fluent Data Base Methane-Air (see Image below), the methane mass fraction is set as 1 at one fuel inlet & 0.23 for air at another swirling inlet:



How is the mesh quality?: Skewness max 0.99856 only comprised of very small amount of cells. Please see chart below from workbench. I think the mesh is quite good (could I send the images in Customer Service Platform? Prefer it that way).






Did you input correct mixture properties in the Materials? It is correct one (I checked), Fluent default (methane-air) as basic test case.

New Update: Another mesh without 10 internal fluids (i.e., just walls tubes, no shadows) + SST KW models steady-state seems to achieve better convergence (the continuity at least one order magnitude lower, I haven't put wall thermal conductivity for the side walls but will test it soon) and it doesn't have that "temperature limit" prompt in the console:

 

[Essence]

Hello,

Please generate a mesh with minimum orthogonal quality of 0.15. You don't know. Even because of some bad elements, you might be facing the issues.

And related to the materials, I referred to the properties such as viscosity, density, thermal conductivity which need to be set according to the models (sutherland, ideal gas etc.) and not the constant values.