Learning Forum

[Wu PeiNan]

[Rob]

Yes, that will work. 

[Wu PeiNan]

Thanks for you replied, Rob.

When I choose a semi-closed pipe to simulate a methane explosion, I always fail to ignite in this model, why is that? 

I used the simple pipe model, and I think  if that there are any requirements for the mesh?  Or which setting I ignore?

I would appreciate it.

[Ren]

Could you provide more details of what you are doing, for example,

geometry dimensions

combustion model

initial conditions

chemical mechanism

solver settings

 

[Wu PeiNan]

The model is a rectangle pipe with length of 5 meters and width of 0.3 meters. The combustion model is the finite rate/eddy dissipation model， and its initial condition is shown as following figure， with the CH4 combustion reaction.
For the solver， I choose pressure solver and simple algorithm with step of 0.005 seconds.

[Ren]

So what is your actual question?

 

[Wu PeiNan]

I can't light this model a lot of time, and even if I do, there will be irrational situations like this. I don't understand what is wrong with this. Is it the grid or a certain setting? Like this one, it lit up but the flame spread along the wall.

Thanks for your response， this question has been bothering

 me for a long time.

[Wu PeiNan]

Or do I need any additional setttings for ignition via patch to ensure that my model can simulate CH4 ignition and explosion？

[Ren]

Hello,

1. Is the pipe open-ended on both sides, or closed at one end and which end?
2. If detonation is expected, I suggest you use the density-based solver (DBNS). The pressure-based solver (PBNS) should also work if the Mach numebr does not go very high.
3. I suggest you use the "Finite-Rate/No TCI" option.
4. Are you using a 1-step reaction scheme or multiple reaction scheme. For 1-step reaction, you need to patch a temperature and the composition that correspond to the burnt state of the mixture.
5. The mesh needs to be sufficient fine around the flame front (and also shock wave). You can use automatic mesh refinement/coarsening
6. The time step size also needs to be much smaller because everything goes very fast. I think 0.005 s is too large. It needs to be around 1E-5 s, or even smaller.