Flash arrestor on H2 applications. 1

[Prometheus21]

Hi everyone,

This is a bit outside my usual purview, so I'm turning to you.

My company sells pressure regulators for (mostly) industrial gases, like oxygen, nitrogen and hydrogen, as well as flexible hoses that gets attached to the pressure regulator. The assembly is performed by the customer.

Now the pressure regulator gets attached to the 200 bar H2 cylinder. The flexible hose (typically 1/4'' inner diameter tubing) then gets attached to regulator. For applications involving welding, cutting ect. you would install a flashback arrestors (also known as flash arrestor) between the hose and the regulator; the idea being it would stop the reverse flow of gas as well as flash back into the regulator and the cylinder itself.

Now the pressure regulator is rated for a cylinder pressure of 200 bar (P1), where the working pressure (P2) can either be 1-10 bar, or 10-200 bar. The highest rated flashback arrestor for this application (to my knowledge - and after asking multiple distributors in my region (Europe)) is for a working pressure of 17 bar, and the cost is sky-high.

I now have a customer (plant manager) that wants a flashback arrestor rated for 200 bar H2 fitted between the flexible hose and the pressure regulator.

Now the pressure regulator is of course equipped with a pressure relief device in case of failure. But I have never before seen or heard anything about installing a 200 bar flashback arrestor between a flexible hose and the pressure regulator mounted on the H2 cylinder. It just hasnt been done before to my knowledge. In the words of one of the distributors I reached out to: a flashback arrestor with that spec has never been, and will never be, manufactured.

Now my question is this: Has any of you ever heard of such a thing? Are there any rules and regulations in place (Europe) I'm not aware of? In my experience so far every customer that uses H2 for cutting and welding uses a 200 bar cylinder, with a mounted pressure regulator with a working pressure between 1-10 bar, with a flashback arrestor if needed. For higher working pressures than 17 bar, no such equipment exists...

Thoughts?

 

[LittleInch]

I'm not familiar with the details but these just look like a spring loaded non return valve to me.

If you ask for a 200 bar spring loaded non return valve as opposed to a flame arrestor you'll probably find one.

Industrial gases are a world of their own. I assume you've asked or looked in EIGA and similar industry bodies for information?

https://www.eiga.eu/

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Prometheus21]

Hi LittleInch,

A non-return valve has no particular flame stopping properties, as opposed to a flashback arrestor. It can of course reduce the probability of a flashback, but is not regarded as a proper failsafe on the same level a flash arrestor. The arrestor contains a pressure or temperature actuated cut-off valve (usually pressure).

Now for acetylene cylinders it is always recommended to install a flashback arrestor to the pressure regulator outlet, which is not a problem since the working pressure of acetylene cylinders is very limited (approx. 1 bar). I have yet to see it on high pressure H2 cylinders...

"I assume you've asked or looked in EIGA..." - I have. I have also reached out repeatedly and not gotten an answer (yet). As a general note; I know the number of hydrogen applications are increasing by the day (at least from what I can tell based on customer feedback), but I also see that the industrial gas industry is lagging behind when it comes to desigining proper equipment. This is especially true for 300 bar hydrogen cylinders where the application demands a working pressure of 0.2-0.5 bar... most two-stage pressure regulators on the market is struggling at that point.

 

[LittleInch]

Maybe the other way of approaching this is to say is there any possibility of this actually occurring, i.e. the flashback thing presumably requires an excess pressure at the flame end going back down the hose. Can this actually happen at say >50 bar?

Also looking at the OP again, the plant manager seems to be ignoring the relief systems and similar on the regulator which prevent the hose ever seeing > 17 bar. This also should be followed up as it sounds like someone just freebasing his own thoughts and coming up with scenarios which are not credible or very low risk. If really needed, just add another relief valve u/s the flashback arrestor?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Prometheus21]

"Maybe the other way of approaching this is to say is there any possibility of this actually occurring, i.e. the flashback thing presumably requires an excess pressure at the flame end going back down the hose. Can this actually happen at say >50 bar?"

As to my knowledge the data is lacking in this area. At lower working pressures ( up to 17 bar I suppose) the burn rate is quicker than the exit velocity (to put it simply). Air gets sucked in with the hydrogen and you have a sustained retrogression of the flame upstream at very high speeds. Now at very high working pressures >50 bar I don't know if the exit velocity would prevent the fire from flowing upstream back into the system. From a practical standpoint hydrogen has a very broad range for combustion in air; LEL: 4% H2 in air, UEL: 96 % H2 in air (some sources state 75%). Combined with a very low ignition energy I can imagine that an initially small flame zone could quickly go from a "slow burning" speed (sub-sonic) to detonation (supersonic reaction front). This is what happens at lower pressures and I would be hard pressed to imagine anything different would occur at higher working pressures. (I can very well be in the wrong here).

If I go away from cylinders for a moment, and look at H2 pipelines (WP up to 210 bar, EIGA state the following regarding flash arrestors:

"There are varying opinions about the need for and efficacy of a flame arrestor to inhibit backward
propagation of a fire into the vent pipe. Similarly, purging of the vent line with an inert gas such as
nitrogen can help to prevent auto ignition, but there are numerous reported instances where a purge
was ineffective. Therefore, this document makes no recommendation either way at this time other
than to point out that flame arrestors and nitrogen purges do no harm" - DOC 121/14.

And for other high pressure hydrogen applications the hydrogen is often mixed with nitrogen to values below the LEL. The exception being welding applications, but here I have only seen working pressures below 17 bar, as previously stated.

"Also looking at the OP again, the plant manager seems to be ignoring the relief systems and similar on the regulator which prevent the hose ever seeing > 17 bar. This also should be followed up as it sounds like someone just freebasing his own thoughts and coming up with scenarios which are not credible or very low risk. If really needed, just add another relief valve u/s the flashback arrestor?"

Agreed. It must be stated however that a lot of applications involve using flexible hoses and pressure regulators rated for 200 bar working pressure. Meaning the hose will see a working pressure of > 50 bar hydrogen depending on the user applications. But as stated before, in those applications you would (to my knowledge) mix the hydrogen with nitrogen further downstream, and add multiple layers of relief valves, forced venting, grounding, bunker approaches, firewalls, isolation valves and so forth, as you normally would in large scale hydrogen applications. So as you say, not very credible and probably low risk if all the safety measures demanded in the different codes is being met.

 

[georgeverghese]

Flashback arrestors should be installed close to the point of flashback ie in this case, close to the welding torch. This is where the reverse flame front velocity is lowest. The further upstream the reverse flame front gets to travel, the faster the flame front velocity, and then the flashback turns into a detonation with very high velocity. Installing a flashback arrestor close to the H2 cylinder cannot stop the reverse flame front, and is bad practice.

If the flashback arrestor is good for 150psig only, install 2 regulators in series, both with setpoint of say 100psig max. The first one being a monitor regulator. Talk to regulator vendors about this.

 

[Prometheus21]

georgeverghese: and for applications with welding torches you would as you say, have a tourch-mounted flashback arrestor (check valve + arrestor) with a hose mounted arrestor AND a regulator mounted arrestor. Which is fine because the working pressure is usually no more than 10 bar (sometimes upwards to 17 bar).

Now the question is what is usually done when the working pressure of hydrogen is more than 17 bar? Let's say the hydrogen has a working pressure of 90 bar, and is mixed below the LEL; a sustained flashback would not occur, correct? But what if the mixture is above the LEL?

1. Is that a likely application? If so, based on what you are saying, a flashback arrestor at the outlet of the regulator would serve next to no purpose as it cannot stop the reverse flame front.

"The further upstream the reverse flame front gets to travel, the faster the flame front velocity, and then the flashback turns into a detonation with very high velocity" - Could you please explain how the reverse flame velocity increases the further upstream it travels? I must admit I know very little of combustion mechanics - (As a side-step: any literature pointers would be greatly appreciated as the subject is fascinating).

 

[georgeverghese]

That adds up to 3 arrestors in series - the one upstream at the regulator is of no use, since no flashback arrestor will work at this location.
If the H2 mix is well below LEL, then the flame front would extinguish itself. If the mix is above LEL, the 2 regulators in series would be a must when pressure exceeds the design pressure of of the arrestors.
I learned a few things about the behaviour of reverse flame fronts, flashbacks and detonations from some report published by DnV submitted to Shell some 25years ago.

Pls note that detonation arrestors may work further upstream of the source of ignition, but these are much more expensive, more bulky, and prone to fouling - avoid as far as possible. Its anybody's guess if these would work as intended when flame front velocities go through the roof.

See if this helps :

http://www.aiche.org/sites/default/files/docs/webinars/BanksD-FlameFlashbacksFINAL.pdf

 

[Prometheus21]

georgeverghese: correct, and there you have it, EIGAs own recommendation for torch applications...

Thank you for info, and the attached file! Interesting read and greatly appreciated!