Hello !

I am simulating a laminar flame in a Rijke tube with a honeycomb structure over which the flame is supposed to reside. I used a 3000 K temperature patch just in front of the honeycomb. The case is adiabatic so the flame actually flashes back into the honeycomb. The problem lies with the temperature patch: I'm unable to get rid of it easily. The under-relaxation I can go upto without making the solution go haywire is 0.3 for energy. For the species I'm using a 0.9 relaxation. The species profiles develop pretty fast but the temperature patch refuses to dissapear. I'm using 4 processors in an SGI Origin machines and it has already taken me 10000 iterations to bring the max temp in the system down to 2400 K. The patch seems to be moving downstream but at a very slow rate. I have a 2000 K patch after the 3000 K patch downstream till the exit.

Can I get any suggestions as to what I can do ?

Thanks

Prateep

That looks quite similar to my problem and question. Did you had a look to the 'laminar combustion" question which can be found in the 'archive of old contributions".

Advises were:1. Using the coupled solver, start with a coarse grid...

I also had problems in getting the temperature low. It always hit the upper limit. I think it has s.th. to do with the kinetic constants. Are your kinetic coefficients suitable for a laminar problem. I think the build in coeff. of Fluent are not. They seem to be valid only in turbulent cases where the reaction rate is calculated based on the mean temperature values. This would lead to rates which are too low (factors of about 1e4). Therefore the coeff. are adapted to be suitable for turbulent conditions, i.e. to give high reaction rates which may lead to high temperatures in a laminar case. This is still a supposition of mine, since I yet did not arrive in comparing coeff. from literature (e.g of Westbrook and dryer) which are expresed in kmol with the ones in Fluent (expressed in kg).