Depressurizing time calculation

[Chem2020]

Heat transfer fluid in preheaters @ 600 F boil water to 1500# steam. The Psv at the outlet of HTF has been sized based on tube rupture in steam site. During tube rupture all control valves will be closed. Steam Gen will be isolated in 30 seconds. The question is how to calculate the time which take to depressurize 1500# of steam in the Preheater?

 

[sshep]

Since the leak is liquid, the pressure in the steam side of the exchanger is basically constant until all the liquid inventory is gone. The time is the mass of water in the drum divided by the leak rate.

The API basis for a tube rupture is basically that the tube shears-off to give 2x the cross-section of the tube to depressure into- all of which will be seen as a steam load at the hot oil side PSV. This is obviously the highest design rate, but most tube leaks are not that extreme. Whatever you calculate for a minimum depressure time based on relief rate, it could easily take longer. Flashing of the water into steam, helped by the sensible heat of the hot oil may also reduce the rate, leading to a longer depressure time.

This is the way I see it from what you described. Others may have a different view.

best wishes,
sshep

 

[MortenA]

sshep,

FYI a fairly resent example from offshore UK experienced a total delamination of the tube sheet that resulted in the shell being blow of the head. So a worse than 2x1 tube cross sectional area also exists. The case involved some tricky metalurgy, sea water cooling and some other factors - but it can happen.

Best regards

Morten

 

[sshep]

Thanks Morten,

Your story showns that often whatever the scenario that we use for relief design, we can come up with something even worse.

Today I got an email from one of my old sites where tower internals were wrecked by putting the heat to the tower when it was 2/3 full with liquid (bad level)- i.e. a BP Texas City style boo-boo, but fortunately not a boom-boom.

best wishes,
sshep