Nitrogen Compressor Leak Survey

[RJB32482]

Hello,
I am looking to justify/reject a project to add some O2 analyzers in a room where we have some Nitrogen Compressors. I believe the worst case would be a leak from a flange. The PSV's on the compressor discharge outside, so this should not be an issue. I have source models from Crowl and Loveur's "Chemical Process Safety" Book to determine the N2 leak rate. Questions that arise are:

1. Should I assume the leak from a flange will be a full diameter leak? The book has source model on "leak from a hole". I was just going to assume the leak would be the full diameter and calculate it that way.

2. What do I need to look at to see if we need the O2 analyzers (room size, ventilation, proximity to doors and windows)? Anyone perform an analysis like this before? Should there be air turnover in the room of some ratio? Or does another "rule of thumb" exist?

Please give me a path to move forward to make this determination. Thanks.

 

[dcasto]

API RP500 has some change per minute recommendations for natural gas. If you think about it, at 20% LEL you'd have 99.8% air and the 6 changes per minute would keep the amount of air above 99.8%. Now the 6 changes per minute is a starting point and should cover most leaks. A complete flange failure or line failure would be a whole lot more. But, you'll have O2 analyzers that will keep people out or shutdown the system.

 

[zdas04]

dcasto,
I think you have a typo, 6 air changes per minute would be a huricane. At 6 air changes per hour I have a real hard time keeping processes from freezing this winter.

RJB32482,
Oxygen defeciency is a real problem in spaces where nitrogen is processed. It is a pretty common practice to either have whole-room oxygen detectors or have proceedures that require persons entering the space to have portable detectors. With regard to your leak size, API RP 500 has a proceedure for estimating leak potential from flanges and fittings (oddly enough called the "Fugitive Emissions Method") that has a factor for flages that is far smaller than a full-flange leak.

David

 

[dcasto]

Yeah, 6 per hour. I just looked at the API 500 to be a model knowing that you need at least 19.5% O2 or about a 10% reduction to have a low O2 situation. Whereas 10% LEL (or 10% of 5% HC gas) is .5% reduction of air by methane gas. So if the 6 changes is recommended to circulate and keep most methane leaks to a .5% concentration, it should do well for when you can a 10% concentration (10% N2 + 90% air).

 

[mbeychok]

RJB32482:

This is not a direct answer to your question, but I think it is still useful to understand the seriousness of possible nitrogen leakage into a room or any other closed space.

In the book "Forensic Pathology", Second Edition, Vincent J. M. DiMaio and Dominick J. DiMaio, ISBN: 0-8493-0072-X, at page 231, it is stated that: At oxygen concentrations (in air) of 4 to 6%, there is loss of consciousness in 40 seconds and death within a few minutes”.

You can read that at

http://www.charlydmiller.com/LIB/fo...thasphyxia.html

and scrolling down to the section entitled "Asphyxia" on page 231.

The cost of installing oxygen analyzers in the room is probably insignificant compared to the possible serious consequences not only of a flange leak but also of a catastrophic pipe failure for one reason or another. Ventilation systems can also fail.


Milton Beychok
(Visit me at www.air-dispersion.com)
.

 

[schnipp]

O2 Meters are very cheap compared to the costs of catastrophic failure. Nitrogen asphyxiation fools the bodies breathing mechanism. The signal to the brain that starts a new breath is high CO2 levels. With N2 enriched atmospheres, the CO2 levels in the blood never increases and you go to sleep. In those that are untrained, the hole watch usually goes down, as well.

I know someone that had a nitrogen exposure and quick response by the fellow workers lowering him into the tank saved his life by raising him back up. He just slumped over and went to sleep. It can happen that fast.

The oxygen analyzers we use here are less than $1500. They put out a 4-20 ma signal to the DCS. Calibration is required once a month. People are comfortable working around it. Contact your local MSA rep for more info.

 

[dcasto]

I don't think RJB was questioning the installation of O2 sensors and safety, I thought he was looking for what would happen if the sensor was activated. What would be a threshold level to evacuate and how could you clear the area to get back into it.

What say you RJB32482?

 

[RJB32482]

Yes I agree. I am not questioning putting in the sensors. But others in my group are. They think the room is too large that if there was a leak, it would dissipate quickly and no danger would exist. So I wanted to give them a theoretical scenario (maximum leak rate) and how fast it would fill the room without any ventilation.

So if I gave them the data that the room could fill with nitrogen in seconds or even minutes, that the O2 sensors would be justified. So that's why I asked about leak rates and ventilation.

Please reply if any other scenarios like this occurred in your plants and how you justified it to the other workers.

Thanks.

 

[dcasto]

I kind of pointed out the dilution factors earlier using natural gas. If the air in the room is still, then a pinhole leak could kill a person that walks into the area around the leak. The N2, unlike natural gas, will not rise nor dispurse quickly. OSHA regs require O2 monitors for labs that have bottled N2 in them.

A 1" nipple breaking would make the building unsafe in minutes. I can't remember the rule of thumb now, is pressure psig * hole diameter inches = escape rate in MSCFD I believe.