This is a theoretical question regarding how dilution purging with nitrogen gas would be setup for use on a tank. The purging is for the purposes of inertion before handling flamable liquids.
Would you always need a back pressure regulator on the vent line to ensure you are getting approrpriate mixing within the tank before the nitrogen vents out? My thought is an unrestircted vent line will simply cause the nitrogen gas to “short circuit” across the interior of the vessel head and vent out and the rest of the tank atmosphere relies on diffusion to get oxygen out instead of convection.
In practice, I’ve only seen and designed systems that pressure purge vessels prior to introducing flammables. Sometimes a small sweep flow rate is used during operation, but post inertion of the vessel.
Our standard tank set up for flammables was a normal regulator with setpoint below the PVSV pressure setpoint. This was for steady state operation for conservation of N2.
Taking a flammable tank out of service and returning it to service after maintenance was handled manually with manual valves. No back pressure regulator. We did use a portable calibrated O2 meter at a vent line to guarantee the tank was inert.
“Short circuiting” should be minimized when designing the tank and laying out the connections (fill on one end/side and exhaust on the other). In my experience “short circuiting” has not been much of an issue. The minimum oxygen concentration for most flammables ranges from 8 to 15% O2, which is not difficult to achieve starting from 21% O2. It’s about the same order of magnitude; not that much reduction. Just give it plenty of time, and let Brownian motion nullify the affects of any “short circuiting”. You can test for O2 two or three times to see if the trend is consistent. That should give you some confidence there’s no pockets of air remaining. And then, the equipment should be grounded, and the area electrical classification should preclude any ignition source, so chances are it’s not going to see any ignition.
I do recall we had the capability to inject N2 into more than one connection on some tanks, so that helped too.
Thanks for the input Latexman! I wasn’t sure about the diffusion time scale, I thought it wouldn’t be practical in tanks.