Design of nitrogen dumping system at the exhaust of relief valves.

[simoncb]

I would like to receive advice on designing a nitrogen dumping system at the exhaust of full flow relief valves, located downstream of a pressure regulating valve, in order to dilute the discharge gas below the flammable limit.

Our Client's specification has only the following statement about this requirement.

"At each relief valve exhaust pipe an automatic Nitrogen dumping system shall be provided to dilute the discharge gas below the flammable limit. Sensing the rise in pressure in the exhaust pipe Nitrogen stored in bottles will be automatically discharged into the exhaust pipe. Adequate Nitrogen cylinders shall be stored such that it can dilute the full flow of safety valves at least for ten minutes".

The gas has 97% methane, which has LFL of 5% & UFL of 15%. The full relief valve capacity will depend on the orifice size (likely to be P) and is around 82000 kg/hr . Diluting this large flowrate of gas to the desired flammability limit will certainly require large amount of nitrogen ( approx 700,000 Nm3).

I have come across this requirement for the first time and therefore would be grateful for assistance in confirming the nitrogen requirement.

 

[pmureiko]

I believe this specification is for gases with a composition already in the flammable region. Because of the risk of this material igniting and possibly flashing back to the vessel, you want to bring it out of the flammable region by dilution. In this case there is no value to dilute methane, unless local ordinances (?)require it.

A flare should be able to handle a PSV discharging methane without any problems. If the PSV goes directly to atmosphere, would a flame arrestor provide passive protection better than depending on nitrogen dilution.

 

[EGT01]

I agree with Pmureiko and can't believe the client spec is intended for diluting a methane rich stream such as described though you don't indicate what the other 3% of the gas is.

Just as a quick check, assuming the stream is all methane then
82000 kg/hr CH4 * 1 kgmol/16 kg CH4 = 5125 kgmols/hr CH4

In order to add N2 to bring it to the lower flammable limit
LFL = 5 vol% CH4 = 5 mol% CH4

5125 / (x + 5125) = 0.05
x = 97375 kgmols N2

97375 kgmols N2 * 22.4 Nm3/kgmol (0°C, 1 atm) = 2181200 Nm3/hr N2

For 10 min storage
2181200 Nm3/hr * 1 hr/60 min * 10 min = 363533 Nm3

If you want to get below LFL, say to 2.5 vol% CH4, then the required dump rate and storage amount will increase by slightly more than 2x indicated above. Is that where you get your 700000 Nm3 from? Anyway, that's a whale of amount of nitrogen to have to store and deal with storage issues besides being able to deliver at that kind of rate. And don't forget what that will do to sizing of your relief valve exhaust line so you don't exceed the allowable back pressure at the relief valve.

I would suggest clarifying the requirement with your client.

 

[CHD01]

Find a polite way to inform the client that they do fully understand what they are specifying. It might work for a small pilot plant, but not a large facility.

First, you are getting good advice in that the amount of N2 required is ridiculous; and that a flare is one possible way to address the situation.

But lets back up a second. The objective is NOT normally to reduce the discharge below its flammability limit AT THE DISCHARGE VENT! Much of the time you WANT the vent to light off (of course this requires a calculation to determine the radiant heat effects of this - vent must be a minimum distance from equipment and personnel for safety). The layout of the plant needs to be determined and you need to calculate the PPM of the release at ground level or at the highest point of human occupancy and the radius of the working area from the vent discharge. This depends on the release rate, the vent tip diameter and the vent height along with wind properties.

If you really want to avoid a flame at the vent, there is another way - use high pressure steam and inject it a reduced pressure into the vent to snuff flame. Again, however, check this against reality as others did for N2 in this thread to see if steam is even realistic; if not, fall back to the flare as a solution.

The more you learn, the less you are certain of.