Operating principle(s) of Detonation Arrestors

[BJFBama]

I'm designing a 2" x ~10"OAL feed manifold for feeding highly enriched nat. gas/O2 mixtures to a small scale burner. The incoming gas/air/O2 lines connect at the upstream end of the manifold. The initial FEMA analysis shows adequate venting through the burner to handle a normal deflagration-type flashback. My main concern however is the possibility of detonation and so I'm requesting input to the following questions:

(1) The flame front of course will self extinguish where the incoming gases mix but I'm mainly wanting to protect a detonation pressure wave from continuing to run up and damaging feed lines, flowmeters, etc. Do detonation arrestors also stop a shock wave from propogating through or do they funtion just to quench the flame front? For a few I've checked out, the normal flow pressure drops seem pretty low.

(2) Do rupture discs (or similar devices) respond quickly enough to mitigate a supersonic detonation wave?

(3) Considering the short 10" distance, is detonation a real concern even at say 50 - 100% O2? Most literature speaks of so many pipe diameters needed to transition to detonation but I'm assuming they're speaking mainly about air and not enriched mixtures.

Thanks in advance.

 

[Flareman]

Detonation Arrestors are really just compex flame arrestors which have a high heat transfer capability to quench the flame during a short passage time and usually a double element arrangement to break up "pressure cells" which form in the first element. The pressure build up ahead of the flame front can pass through the element.

Most quoted run up distances to detonation are unreliable because they are usually obtained in plain pipe whereas there is often a bunch of various diameters in a real world installation. The pressure build up to detonation occurs as a combination of the flame eating through the mixture, the forward pressure wave compressing (and collapsing) the upstream volume and the flue gases having to push out of the open end. If you base your estimate on a quoted run up, assume that you are using the smallest discharge diameter not the diameter in the upstream tube, although 10" run up hardly seems to be a problem in any circumstance, (just my feeling not a legal opinion) but most arrestor manufacturers would allow you 10 diameters x your 2" pipe.

There's also a pipe diameter factor because of the heat transfer to the wall of the pipe and the relative amount of perimeter to flow area (reaction surface). For this reason small pipes tend to be less of a problem that large ones.

Small pipes also tend to have a much better pressure rating than large ones so you are less likley to damage the pipe.
In a real world application which regularly mixes air and gas in a pipe of as great a length as 1/2 mile and ignites it, (and sometimes goes to detonation), a shedule 40 x 1" pipe is just fine. We protect the upstream systems with spring loaded non-return valves which slam shut when the pressure wave comes along.

Rupture discs are commonly used to relieve detonation pressures. You just have to be able to calculate how much hot gas to relieve how quickly. I'm pretty sure that one of the NFPA codes covers this. You might try a question in the Codes forum.


David

 

[BJFBama]

Thanks Flareman for the reply.

Your point about detonation arrestors allowing the pressure to pass through is key. Are you aware of any calculations or specs as to what degree the pressure of the shock wave might be attenuated through a reduced diameter? The flamefront will extinguish at the upstream end of the 10" manifold where the gas/O2 mix. And, we've estimated that the pressure of a true detonation shock wave could be as high as ~ 1,250 psi at this point. We can easily design the pipe, flanges, connections to handle this but the main objective is to contain it here and prevent the pressure wave from moving further up into rotameters, etc. where personnel could be injured. It seems intuitive to simply install as small of an ID / thick wall restriction as possible in the incoming lines. However, I don't understand what modeling equations or rules of thumb to use to help quantify just how much the 1,250 psi might be reduced.

Regarding the second issue of rupture discs, since my initial posting I've learned that the fastest acting rupture discs have a response lag of ~5 millisec. Velocities of CH4/2O2 supersonic detonation waves have been reported around 2,330 m/s (Lewis, von Elbe). At this speed the wave could travel 11 - 12 meters past the disc before any venting begins. My simplistic understanding of a detonation model is that compression is beginning only a couple of differential elements ahead of the wave and, once activated, no amount of venting behind the front is effective.

Please feel free to challenge the above as I have no practical experience here and these texts seem very academic in my opinion.

Regards,

 

[25362]

May be these links shed some light:

http://www.clarkson.edu/~wilcox/Design/flamarst.pdf

http://www.banksengineering.com/about_flame_arrestors_and_detona.htm

 

[Flareman]

(I can't believe no one else has jumped on this thread. There have to manufacturer's who read eng-tips!)

25362's link to Stan Grossels paper shows the crimped ribbon style nicely, which is the primary design used in detonation arrestors (as a double element).

On the pressure attenuation issue, it depends on the upstream volume and diameter as well as the orifice size, and whether the downstream volume is closed or open end. Downstream pressure depends on the rate at which you can add new molecules through the hole.

But do you really need to beat yourself up with a heavy theory when you can solve the upstream problem with non- return valves. (Really ! We do it!).
Also, what is the final discharge velocity at the point you think the flame will start its run up. If the discharge velocity exceeds the initial flame speed it's not going into the pipe anyway.

David

 

[BJFBama]

Thanks 25362 for the links. I had already seen the second one but the Grossel link looks great.

Flareman, thanks also for the reply.
I might not have made it clear but this is a small research burner with 3/4" incoming gas/O2 lines and with rotameters/controls only about 15ft away. Are you talking about simple check-valves or a controlled sensing-type fast-acting valve. I could understand the suitability of the latter in a larger dia pipeline that covers some distance where you have a chance to sense the flashback and signal a valve some distance upstream. But that doesn't seem so practical here? I'll check into the CMR type arrestors, they look promising and seem to be based on the same idea of creating small dia restrictions to disrupt the shock wave.

Downstream pressure is not an issue. There's a burner grid downstream that can easily vent this small volume of flu gas. We've also designed in a grid that should (theoretically) maintain velocities above the critical flame velocity. However, there are way too many possibilities for upsets to rely solely on this. High %O2 mixtures can detonate easily and the PHA team is requiring that they be addressed.

Thanks again for the feedback and I too am surprised not to have heard from any equip manufacturers.