Learning Forum

[J.LoongHee]

Dear Experts,

I am already 1 year into combustion modelling. I think I am getting a better grasp of the concepts. 

However, it does puzzle me one thing. I did a mesh independency test of a CFD domain (3 meshes). Default numerical scheme, node gradient scheme (as recommended in one of the post). The Heat-of-Reaction Source fluctuate very minimally say around 500W or even less out of XXX,XXX watts. And the solution converged well, energy residual order e-6. the rest at least e-4 or e-5. URF around 0.7 for temperature, DO probably around 0.75. But I can revert back to full default which is 0.75 and 1.0 DO no problem.

 I took the medium mesh density and performed subsequent combustion analysis. The outcome of mesh independency is very good. I get a trend and Mesh M2 appeared to be most optimum. 

When I am monitoring the convergence, I performed industry best practices which is to monitor various variables. I monitor the fluid zone Heat-of-Reaction for a particular case. This time the Heat-of-Reaction kept fluctuate (probably 5000 Watt in between say 50 iterations). The net result at a particular instance could be 2000 Watt another instance could be -500 Watt. In order words, the net results kept fluctuate but is of course getting down to < 2% regime. 

I do view the flow field by exporting images in HPC every 250 iterations. I am quite puzzle why my technique hasn't achieve robustness! I am basically using the same input conditions though. I do aware that different type of combustor geometry could yield different flow stability due to the mixing of fuel & air? 

I have been trying to use a smaller and smaller time-scale factor (for steady-state pseudo), I do noticed for some cases the 1 default to 0.5 work well. But when I applied to some combustion cases. This cannot be replicated! I puzzled me why it can be so stable in mesh independency test but not on some actual trials. 

The second question I have is:
I do aware that if a preheating of air is used (say XXX Celcius). The above mentioned problem (which is fluctuation of Heat-of-Reaction source) is more likely to occur! Any suggestion what approach I can adopt to better stabilise the flow? 



Thanks. 
Jee 

[Vijay Narayan]

Hi Jee,

May i know your combustion model setup details.

I'm assuming it is Finite Rate model. Are you using reduced or detailed mechanism ? Are you using Stiff chemistry solver or Relax to equillibrium ?

Is it Gaseous fuel or DPM based liquid fuel ?

You could try Coupled solver w/o Pseudo transient option and also could explore combustion (w/o DO model) to begin with and then observe the HoR response.