LEL measurement inside piping - is it needed? 2

[mjpetrag]

Currently I am in the process of determining whether or not our current LEL detection system is necessary from a process safety and regulatory standpoint (located in US)

The way this system is set up is several vacuum distillation column vents, on the vapor outlets of the jet condensers, are tied together and routed to a knockout pot where several other vapor streams from other processes combined. The outlet of this knockout pot is sent to a catalytic oxidizer, which Is equipped with redundant LEL probes that shut down the unit and vent the upstream flows to atmosphere.

On the distillation column vent system, upstream of the knockout pot, are another set of LEL probes that upon high LEL detected, open a vent to atmosphere and close the valve from the distillation column vents to the KO pot. There is also dilution air that bleeds into the system between the vent to atm/flow to KO pot control valves and LEL probes for dilution air.

The question I have is - are these LEL probes necessary if I have probes at the catalytic oxidizer? The only thing I can think is that there is a potentially explosive mixture in the piping from the distillation column vents and knockout pot, however there would still need to be an ignition source.

We cannot find any documentation saying why these were installed so I am looking for a regulation/best practice in this instance

-Mike

 

[Compositepro]

It is up to you to prevent an explosion under any reasonable scenario. I have seen an oxidizer explode because solvent was dispensed into a hot mixer, and the lel sensors did not respond fast enough to bypass the incinerator. The fix in that case was to install condensers on all of the mixer vents. There was a couple million dollars in damage and a lawsuit.

 

[georgeverghese]

The sensors at the catalytic oxidiser may not help much in the event of of a flammable - explosive gas air mix that is approaching the oxidiser, even when you have fast closing isolation valves at the oxidiser. Flashback / detonation velocities are very high. Sensors should be located as far upstream as possible.
Electrical grounding of all gathering lines is also required. Another protection measure is continuous purge gas flow into the burners at the oxidiser, with purge gas velocity exceeding the min forward flame velocity. Purge gas can be fuel gas or inert gas.

 

[TiCl4]

Mike,

The best practice in this case (in my opinion only!) would be to include explosion isolation valves sufficiently far upstream to react to flashback or to include blowout panels somewhere near the oxidizer.

You didn't mention this, but another good practice would be to ensure the stream containing LEL has a velocity above the flame propagation speed of whatever your VoC is.

Vents and dilution air systems do NOT respond quickly enough to spikes in LEL that can cause flashback. Explosion isolation valves are actuated via small explosives blow the valve closed very quickly when flashback is detected. Consult with companies (such as Fike) that specialize in such designs.

You also need to define your system better, and state whether your mixture will detonate or deflagrate. Such information will be required for safe design.

 

[shvet]

Detonation arrestor would be cheaper and more simple/reliable. We used detonation arrestors for incineration of a potentially explosive vent gas.

https://www.dag-ts.cz/products/arresters-in-line-bidirectional-detonation-deflagration/?lang=en#

 

[georgeverghese]

@shvet,
Detonation arrestors are typically used in addition to LEL sensors, continuous purge and other protection measures. Many codes require the installation of detonation (or deflagration) arrestors in these applications, but these devices only work well when internal flame arresting surfaces are very clean. In many cases, they will be plugged with dirt and will most likely not be able to extinguish a high velocity flame front travelling in the reverse direction.

 

[mjpetrag]

All great information. For now we are keeping the LEL meters in place and going to engage experts as recommended such as Fike to determine more reliable means of preventing flashback.

-Mike

 

[shvet]

@georgeverghese
for info
We have a process unit designed by a proven licensor - process Merox by UOP. There are no other devices than detonation arrestors and an in-line oxygen analyzers.