Chimney effect in the staged flare system

[enguest]

Gents,

I am calculating the emergency purge rate for the HP flare system which has two stacks - main for the small reliefs and emergency for the large reliefs. Calculation is done using dynamic simulation for the case when the emergency stack is open and main is closed. The flow which we assumed based on manual calculation (gas contraction after hot relief followed by rainstorm) is adequate for this case. However when I tryied to simulate the case when both stacks are opened large backflow occurs at the main stack. Stacks are of same height (more that 100 m) and emergency stack diameter is two times bigger than the main stack diameter. I assume that this can be a chimney effect between two stacks and the larger stack sucks vapour through the main stack. My assumption is based on the fact that this phenomena is not taking place when i disable the static head contribution in my dynamic model.
There is no liquid seal in this flare system.

Do you think that this is a real case or just maybe the wrong setup of dynamic model.

Thanks in advance.

 

[enguest]

And also if this is a realistic scenario, how to solve this problem?

Thanks

 

[sshep]

Hey Enquest,

Your work sounds very interesting. I would not ignore your result without checking it out further. It could indicate an indaquate system of molecular seals, liquid seals, purge gas flow, etc.

best wishes,
sshep

 

[byronius]

I have seen the chimney effect pull a vacuum on flare lines after a flare was shut down and no flow remained. This vacuum existed on every flare line with an open bleeder or air gap. The plant was being decommissioned so we were taking the flare system apart. It was a bit surprising how noticeable the effect was.

That said, it was on a flare header that had no flow in it except for the pull created by the chimney effect down at the stack end. Something like a 48" flare stack 250 ft in the air.

Is the main flare staying lit with some flow throughout all this? I would be somewhat surprised to see your main flare stack reverse direction in the event of an emergency relief. This would seem odd especially if there was some constant flow maintained on the main header (purge flow, continuous vent, etc.) and it was lit. I would expect both flare stacks to be at similar outlet pressures because they are both burning gas. From there, I don't see why the main flare would reverse to get to the emergency, even though it is a bigger outlet.

I agree that there is some chimney effect, but I would think the burning gas would be enough on the main flare to overcome any chimney going the other direction. And, again, the burning main flare stack is already at a low pressure. Stands to reason the gas would continue out that way, maybe at a lower flow rate due to the emergency flare providing another sink for the flows.

Your model is predicting it though, so you definitely need to figure out why. I'm not knowledgeable enough to just dismiss it. Does the model provide justification for the pressures it is calculating/using to judge the flows?

 

[enguest]

Gents, Thank you very much for yoyr answers.

The model showing back flow in the main flare stack after hot relief is stopped and system being cooled down due to a rainstorm.