CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   CONVERGENCE PROBLEM - oil boiler (http://www.cfd-online.com/Forums/fluent/43748-convergence-problem-oil-boiler.html)

MM February 14, 2007 06:48

CONVERGENCE PROBLEM - oil boiler
 
I have convergence problem and I am asking kindly for help! I am using Fluent to simulate the combustion in oil boiler (25 kW). My domain consists of two separated models (when using non-premixed combustion it is not possible to define any new fluid in my case water) Model 1 â€" non-premixed combustion (model-1.cas, model-1.dat) Model 2 â€" water (model-2.cas, model-2.dat) that I have connected writing commands in Solve®Execute commands…

Each model has around 2 millions cells. First I run model-1 at the moment 800 iterations In model-1 under Solve®Execute commands… I wrote following three commands:

1. file/write-profile flux.prof yes surface, inv-flux

2. !fluent 3d â€"t2 â€"path/usr/local/Fluent.Inc â€"g â€"i /home/marina/marina-02/water.jou

3. file/read-profile temp.prof

that are executed every 800 iterations.

So model-1 runs 800 iterations, than the profile flux.prof is written on the connected areas between two models (command 1). With the second command I am starting to run journal file water.jou that opens model-2, reads the flux.prof, runs 200 iterations, saves case/date file and at the end exports the temperature distribution temp.flux on the connected areas between two models. Than the model-1 import the temp.flux and so on…. im Kreis.

My problem is that I can't obtain convergence.I have first obtained a non-reacting flow solution... My model 1 is defined as follow:

k-e model,Realizable, Enhanced Wall Treatment

Number of continuous phase iterations per DPM iteration: 100

Tracking parameters Max. number of steps: 25000

Step length factor: 20

Injection types:pressure-swirl-atomizer

Number of particle streams: 10

Particle type: Droplet

Material: fuel-oil-liquid

Evaporating species: CH4

Stohastic tracking - Discrete random walk model

Number of tries: 8

Pressure-velocity coupling: SIMPLE

Discretization

Pressure: PRESTO!

All other at the moment First order upwind

Solution controls under relaxation factor:

Pressure: 0.5

Density: 1

Body forces: 1

Momentum: 0.5

Turbulence kinetic energy: 0.8

Turbulence dissipation rate: 0.8

Turbulent viscosity: 1

Energy: 1

Temperature: 1

Discrete ordinates: 1 0.975

Mean mixture fraction: 1

Mixture fraction variance: 0.9

Discrete phase sources: 0.25

Please, can someone who is dealing with the non-premixed combustion suggest me what to do in order to obtain convergence ?? In 14000 iterations i wasn't able to obtain the convergence. I observe the temperature and velocity at the exit and those value are still varing a lot (temperature lies between 700 and 200°C and should be around 200°C). also the total heat transfer rate in kW exchanged between two models vary between 30 and 45 kW (it should be 25 kW) There is no problem with the water side...

I hope someone can help me.


nicoslaw February 15, 2007 06:24

Re: CONVERGENCE PROBLEM - oil boiler
 
Please try to change:

Define->models->solver->gradient option to Green-Gauss node based Or Least Squares cell based maybe it would work, regards, nicoslaw


All times are GMT -4. The time now is 18:39.