I don't think the flowrate is criticial.  What you need is a positive pressure.  A slight positive nitrogen pressure will prevent the ingress of O2.  Your flowrate will be driven by the total leakage from your system.  Usually controlled by a regulator or self regulating PCV.

If you had a tank with no flow, O2 would diffuse back through any vent lines into the tank but this is pretty slow.  If I remember correctly, the molecular seal vendor on our flare line used to claim that the purge rate only had to be about 0.1 ft/sec to keep 02 out of our flare system due to diffusion.

As suggested, if the tank is designed for a slight pressure and has pressure/vacuum vents on it, put a few inches of water column pressure of N2 on it, that will keep the O2 out.  If the tank is open to the atmosphere and you can't/don't want to put vent valves on it, try sizing an orifice on the purge gas to the tank to keep 0.1 ft/sec out through the open vent line.  Check the O2 concentration in the tank after operation for a while and determine if you need more or less purging.  I doubt this will work unless you have a fairly inexpensive bulk source of N2, bottles will be prohibitively expensive and you'll use a lot.

Thanks for the responses.

jonralph: In order to create a driving force for a continuous purge to atmosphere there must be a positive pressure of nitrogen in the vessel. Do you have a guidline to say quantitatively how much a slight positive pressure should be?

TD2K & jonralph - Here is some more detail

I'm looking at removing a PSV from an vessel that is currently blanketed under nitrogen at 2kPag and replacing this with a pipe to atmosphere with a continuous N2 purge. The vessel is cirtified for 5kPag.

The vessel is a Seal Pot and recieves high rates of nitrogen during line purges. To ensure the back-pressure created from venting these nitrogen purges to atms does not exceed 5kPag, I need to size the vent piping at a particular diameter. However while not venting at these high rates I am concerned that there will not be a sufficient back-pressure created to ensure no oxygen ingress. I could increase the nitrogen purge rate until a sufficient back-pressure is created however I have no idea what sufficient is and I would like to minimise the nitrogen rate to save costs.

If the seal pot is open to atmosphere, you won't be maintaining effectively any pressure in the vessel when you are just N2 purging it unless you use a LOT of N2 which is not the object here.  Yes, I conceed there will be a small amount of pressure to flow the N2 out the vent but it would be very, very small under any reasonable amount of purge gas.

I still suggest sizing your N2 purge with a restriction orifice to maintain a certain velocity in the new vent line, check your O2 levels and resize as necessary.  As a starting point, start with 0.1 ft/sec in the vent line.  Depending on your N2 pressure, you may have to drop the pressure through a regulator to avoid having an extrememely small orifice that is susceptible to plugging.

The difference between a pad and purge is essentially this.  With a N2 pad, you are preventing O2 from entering the tank as it is at higher pressure.  If you have a continuous purge, you just ensure that any 02 that gets in is swept out.

I am not a ChemE and would never try to argue any engineering with a well rounded ChemE, but having worked in the Chemical Process Industry as a Safety Engineer for the past 10 years I want to pose a question/statement on this topic.  
I have not seen anyone mention using a pressure regulator on the N2 purge system that Mark is inquiring about.  On many of the flammable processes that I have worked with, we have a N2 purge system that uses a pressure regualtor that maintains a very slight pressure on the vessel at all times. Of course when the vessel is being filled less N2 is being added (due to the pressure being generated inside the vessel).  If we rely on the regulator to maintain the purge during this time we may not maintain the O2 level at or below the MOC.  On the other hand when we are emptying the vessel, the N2 flow is at a higher flow rate, due to the vaccum being created as the vessel empties.

One thing that I would recommend is to determine the Minimum Oxygen Concentration (MOC) for the flammable(s) that you are trying to control with the purge.  Most plants try to achieve a randomly selected O2% for the vessel head space.  Not knowing the MOC, could not olnly put your process at major risks, but you could be wasting a tremendous amount of N2.  I did a Six Sigma Project on N2 usage at a plant that I did some consulting work for and saved them over $200,000/year in N2 usage, while greatly increasing their safe gaurds on their processes.  They had just decided that 12% O2 was a safe level of O2 in the vessels and had designed fairly sophisticated interlock logic to maintain this O2 through the use of O2 monitors.

I strongly would recommend using the pressure regualtors, while using some other form of engineering controls while the vessel is filling, and also determine the MOC so that you know how much N2 is needed to safely operate the purge.  Keep in mind that this entire system should fall under your Management of Change system (regardless if the process is under PSM/RMP regualtions), due to the fact that any make-up change in the process stream could greatly impact the MOC, thuse requiring an engineering study to revise the purge system.

I hope this make since.  I would also love to discuss with anyone willing to, the topic of proper purging (i.e. purge gas, purge inlet(s) v.s. outlet(s), measurement points, etc.)

Bryan Haywood
www.SAFTENG.net

My Jan 16th post mentioned using an N2 regulator upstream of the proposed RO to avoid otherwise getting a very small RO if the N2 source pressure is high.  The regulator isn't a necessity but frequently is needed from a practical point of view.

Your other points are all quite valid.

Use API 2000 (Venting Atmospheric and Low Pressure Tanks) to determine and calculate required venting and makeup volumes.  This well followed industry standard for venting and determing purge volume requirements

I don't recall that API 2000 addresses the purging requirements of a tank to prevent oxygen infiltration.  

It does address the outbreathing and inbreathing requirements when filling and draining the tank respectively to prevent damage and emergency venting requirements during a fire.

Dear MarkkraM and all,

Anyone involved in the design of a purging system might benefit from looking at the US Chemical Safety Board report into an overpressurisation incident at a new plant at Pitkin, Louisiana in 1998:-

http://www.csb.gov/reports/2000/Sonat/

The photo on the cover sheet of this report shows how badly a purging operation can go wrong. The tall tower at right had a twin on the adjacent corner. That one was obliterated.

There had been no hazard analysis, no engineering controls such as pressure relief devices, no written procedure and inadequate training.

The purging gas was natural gas, but the deaths were due to trauma, so nitrogen could do the same damage.

Also, it musn't be forgotten that nitrogen itself is a killer if it displaces our breathing oxygen (which is the whole point of the exercise!). An LPG tank was been commissioned at a service station. It was purged with nitrogen. A man entered the tank to retrieve a tool and instantly collapsed and died. So simple. You can't see nitrogen and you can't smell it.

Any design must be accompanied with written procedure.

Sorry, MarkkraM, if this doesn't really relate to your case.

Cheers,
John.

(a precis of the Sonat report is at:-
http://au.geocities.com/jom3205/oil_gas_petrochem/sonat_review.htm
)

Hi,
Thanks all for the comments so far, they have been useful for me to get a feel for the situation.

The vessel in question is a Seal Pot filled with mineral oil. The seal is to dilute a pyrophoric substance to well below its pyrophoric concentration. The nitrogen environment in the head space of the Seal Pot is essentially a secondary level of protection in case the mineral oil becomes too concentrated. Obviously, if this situation arose the reactivity would be directly related to the oxygen concentration in the vessel and the extent of dilution in mineral oil. With this in mind a small amount of oxygen (say 1wt%) would not be a hazard.

The vessel currently has a blanket system with a nitrogen regulator on the inlet set at 2kPag and a pressure relief valve set at 5kPag relieving into a fire pit.. This device regularly lifts as we need to purge process lines with mineral oil and nitrogen into the Seal Pot every week or so. Due to this frequent lifting, the reliability of the relief valve is reduced and can drift above 5kPag between overhauls. To improve the integrity of pressure relief in the system an option I started investigating is simply removing the PRV and replacing this with an open pipe to atmosphere and a nitrogen purge. The thread I have started is specifically to get an idea of the effectiveness of a nitrogen sweep in maintaining a moisture and oxygen free environment, and what design guidelines are useful in achieving this.

Regards,
Mark

Why not replace the PRV with a PCV (downstrem regulator)?

Regards, Bill Birch

Bill,
The issue here is integrity of pressure relief.

I am considering an option to install a PCV and retain the PSV for safety purposes. Obviously the PCV will be set below the PSV. This option is very expensive and has issues relating to maintainability of the PCV due to the nature of the service.

Regards,
Mark

 this is one good discussion ...few notes first :

 is this really a vessel ? its not if design pressure is < 15 psig as per ASME code so check API relief protection for atmospheric tanks...

 is that a psv ? or breather valve ? guess u r talking about a breather and there is no isolation on outlet

 if ur breather valve is not reliable how about a pcv on N2 supply and a small oil pot in which outlet is imersed deep enough to maint your blanketting pressure !

 note :

  for whatever solution you select please see how your system will work in case of loss of N2 supply or inlet /outlet devices fail close /open ..

 hope this helps..

 if this discussion is still a live ..for my comment above ..for the outlet secondary seal pot ..consider make up /over flow ..in case oil level is blown ,,

 regards

MOC's intended for flammability in gas-air mixtures wouldn't do any good when oxygen must be kept out for other "chemical" reasons. Sometimes the purging gas (nitrogen, etc.) itself contains traces of oxygen that must be taken into consideration. Purging estimations usually assume that there is perfect mixing within the containing vessel, constant temperatures and pressures.
An frequent (sometimes uninterrupted) check on effluent gas quality is always recommended. As always, when having decided what is the permissible oxygen level, the best purging technique is selected based on cost and performance.

The purge rate must exceed the rate at which atmospheric pressure can increase to cause inflow to, or the rate at which you can drain, the sealpot AND any attached vessel.
The latter is usually the limiting factor, unless you suffer from tornadoes.
There are several suppliers of 'transzero' regulators designed for this purpose, but usually they are used in conjunction with a pressure relief set c. 0.25kPa to provide a positive backpressure for the regulator to work against.

MarkkraM ,

Simply put a nitrogen bubbler in the seal leg. You can adjust your regulator (low pressure) for the liquid head you will see. The sizing flow of the N2 purge (when it kicks in) should match the maximum draw that your process would have on the seal in the event it is blown.

You can also size on the basis for so many gas changes per minute basically a dilution rate in the seal piping.

With this arrangement you will not consume N2 until the seal blows.


Be sure to use lock-out tag-out proceedures on N2 purged equipment, and posted notices in the area.

Hi,
   Following two suggestions may be useful:
 1. Keep continuous flow ( through 1" control valve on 1" line)into the tank. Put split range control valve on the 1" vent line  . The set point may be 0.2- 0.25kg/cm2g. At
0-50% range  inlet control valve will open and try to maintain the tank pressure. There should be a dead band from 50--75% where make up will be stopped. If pressure  is still increasing, vent control valve will open at 75%-100% controller output to purge the tank.
  I this system purge is intermittent only.
2. Another option is: put PCV on make up line only and keep the vent line open. This system will have continuous purge.

 Hope the suggestions are useful.
Best of luck!
satyajit

MakkraM,

The sulphuric acid tank that collapsed at the Motiva plant did not have adequate inert gas blanketing. You might like to look at the report on the US Chemical Sfatey Board's site (www.csb.gov). It doesn't tell you how to design an inerting system, but it shows the potential consequences of inadequate inert blanketing.

John.

Cheers,
John.

There is a paper by Husa which basically says that the oxygen profile which establishes in a vertical open pipe to atmosphere varies as an exponential relationship of the pipe diameter, purge rate and the distance down the pipe.

This formula can be expressed as

O2% = 21 * exp{- U * L /0.0036/Fb/(D^1.46)}
U = purge velociy - fps
L = distance from top - ft
Fb = Gas buoyancy factor
D = diameter - inches

If the purge is a single component, then
Fb = exp{0.065*(28.96 - MW)}
If the gas is multicomponent, Husa's formula becomes a little top heavy and I recommend using
Fb = 6.25*[ 1 - 0.75*((MW/28.96)^1.5)]
MW = Mol wt of purge gas (mixture)


David

Flareman,
Thanks for the equation.
You mention that it applies to vertical open pipes.
In the vent to atmosphere, I have a horizontal run approximately 4m long. Can I claim any benefit from this?

As far as the effectiveness of a nitrogen sweep in maintaining an oxygen free environment and design guidelines you may also want to look at what API RP-521 says about flare headers, here is an excerpt
"Alternatively, the continuous introduction of purge gas can be used to prevent flashback. Studies [13] have shown that a safe condition exists in situations that involve hydrocarbon-air mixtures if a positive flow of oxygen-free gas is maintained, allowing the oxygen concentration to be no greater than 6 percent at a point 25 feet (7.6 meters) from the flare tip."  

Of course the 6% number should be checked versus the characteristics of the flammable material you are dealing with.

As another reference, the guidelines given in NFPA standard 69, Standard on Explosion Prevention Systems, may be helpful.  This standard covers the minimum requirements for installing systems for the prevention of explosions in enclosures that contain flammable concentrations of flammable gases, vapors, mists, dusts, or hybrid mixtures.  There web site is
http://www.nfpa.org

One other thing I might add is in the guidelines given above, it may be prudent to install an oxygen analyzer to ensure you are meeting your design guidelines.  It is not unusual to find oxygen analyzers as standard equipment for plant flare headers.  For a single equipment vent, it may be difficult to justify the cost but should at least be considered.

Your best bet is to install a conservation vent (i.e. as available from Protectoseal etc.) on your vessel, set at about 3 kPa, and sized so as not to exceed say 4.5 kPa at the maximum N2 flow from your purge lines through this tank.  This sort of unit is specifically designed for this type of duty, used in conjunction with a blanketting valve, and consists of a dead weight rather than a spring.  Keep your PSV as a safety feature only, as a back-up to prevent overpressurisation of the vessel.  Both units are effectively protecting your vessel, so both need regular checks and maintenance.
If your oil is sticky, and tends to cause the existing PSV to stick, then you might have problems with this as well.  Keeping it off the main vessel, e.g. on a branch of a tee, or with a coalescer / flame arrestor in between will help, but may not necessarily eliminate the problem.  There may be an initial cost associated with this, but it should pay for itself in next to no time compared to a constant N2 purge.  
If you really have to go for a constant N2 purge, then you need to consider the worst case scenario for this tank.  What is the maximum ingress of air that can occur at any given time during normal and abnormal operation?  Consider specifically fluctuations in level in normal operation, and inbreathing due to changes in external temperature at night, and in cases of thunderstorms etc.  Then you have to set the N2 flowrate equal to the worst case.  Just assuming it is a fixed value for all tanks / applications is suspect to say the least.  What is needed to stop a small ingress of air into a flare header is not going to be anywhere near enough to keep a storage tank free of oxygen, especially if the tank is regularly drained and re-filled.  
Incidentally, the MOC for some materials is available, but there are not that many.  There are calculation methods that have been published, but they seem a bit suspect.   Your best bet is probably to have it measured.  Typically, many substances come out at around 10%, but some substances (e.g. Hydrogen and CS2 are much lower than this).  NFPA guidelines recommend you don't exceed 80% of the MOC for cases where O2 levels are consantly monitored, or 40% of the MOC where they are not.  Hence keeping O2 below 4% is adequate for the vast majority of applications.  Your 1% you mentioned appears to be well within this.
Hope this helps.

I purged a 10000 gallon gasoline tank with nitrogen supplied from compressed nitrogen tanks used by our welding department. the equipment consisted of a pressure regulator from that deparment and some welding hoses. It took a few hours to purge the tank free of air when it was removed from site.  Flow rate was not a critical issue as long as we had a slight positive pressure keeping the air out of the tank.

Does anyone can tell me more on the issues of nitrogen purging for the flare system.How to ignite the flame for flaring and how to maintain the flame from being put off when purging with nitrogen? If the answer is to control the volume of the nitrogen purge, then how does the volume relates to the design of the N2 purging for the flare system? Are there any minimum value of purge gas volume that could prevent sudden flame-out?

PZA
There have been previous discussions on this, for example see Thread798-66873 and Thread798-72477.

Any time you purge with Nitrogen (or any other inert) you have to be concerned about extinguishing the main flame.  One way to do this is to make sure that the heating value of the total mixture (hydrocarbon plus nitrogen) never drops below a critical value.  The US EPA set this value at 200 Btu/scf, which is a good rule of thumb but doesn't actually describe the real flammability issues and, in my view, gives a misleading idea that heating value is everything.
Lots of flammability postings too such as Thread798-33615 and Thread610-33619

If you have 100% nitrogen, or you have too much nitrogen and the main flame will not burn, then you also have to worry about extinguishing the permanent pilot burners.  If this is a possibility you need to use a permanent pilot which cannot be extinguished by a blanket of inert (difficult but not impossible) and/or use a pilot monitor with automatic rapid re-ignition, see Thread124-75952.


David