Tower Design Vacuum

[sshep]

I am currently working on at a site where large low pressure and atmospheric distillation towers are built with the following design pressure: 400kPa and not rated for vacuum. While I even consider 400kPa on the weak side due to utility pressure concerns (N2, steam, etc), I have never specified a tower, or any significant process pressure vessel to be rated for anything less than full vacuum. I have seen several old pressure vessels sucked in from the usual causes- injecting cold water while steaming out, draining or pumping out, etc. This site relies exclusively on high volume N2 vacuum breaking, but I always have concerns during commisioning, shutdown/start-up, and other unusual times.

The reason I am given for not having a full vacuum rating is cost (even in carbon steel towers). I am not a mechanical engineer and have always assumed that some stiffening rings are all that is usually needed. How much are my full vacuum specifications costing my company? What do others consider normal practice for tower minimum pressure design? Thanks, sshep

 

[StoneCold]

sshep
At one time I worked for a large engineering company in the oil industry. All columns were designed for at the minimum 50psig and full vacuum. This allowed the service of the column to be changed if the day ever came and helped reduce the size of relief valves as well. It is a long term vision approach.

Regards
StoneCold

 

[JStephen]

You might run that question by your vendors the next time you specify one.

If the diameters are small and design pressures are high, you may not have any extra cost. If diameters are large and design pressures are low, it may be a different issue.

What we find in the tank industry is that in many cases, it is more economical to increase shell thickness than to add stiffener rings. So you could be paying for costs that you aren't aware of.

I think the pitfall in your approach is that you could just as easily apply it to overpressure, too. Somewhere, you're probably aware of tanks or vessels or boilers that have failed through overpressure. So do you arbitrarily design every vessel for double or triple the anticipated pressure, just in case? At some point, you have to make some reasonable assumptions about the anticipated service conditions, and those may or may not include full vacuum design.

 

[jay165]

sshep:

To answer your question, in addition to stiffening rings, you can add some extra wall thickness to cover the full vac case. I've always specified a full vacumm rating for any ASME pressure vessel I've bought. What I've found is that by the time you design for wind/seismic/vessel load out, etc, you've effectively "bought" your rating for nothing extra because of the extra wall thickness those conditions require over that of the pressure rating and corrosion allowance. In the case where none of those factors govern (which for a tower would be rare), I can't see it costing more than 2-5% extra for a full vac rating. You can settle the argument by asking your bidders to quote the tower both ways.

 

[sshep]

Thanks, I will gladly take any more replies. If I make any costing studies I will post the results.

With reference to JStephans comment about fattening up the overpressure side, specifying 2 to 3 times the required MAWP would be an unneccesary rationalization for any experienced engineer because vessels have overpressure protection in the form of relief devices. In any event super-stout construction usually just moves the problem downstream to some other weak link, or makes any loss of containment result that much worse. However, I have seen several towers use up their entire supply of spare rupture disks during commissioning and start-up because the tower pressures are lower than the N2 supply pressure so some intelligence is required.

 

[steveen]

sshep,

Relief devices are rated by upstream pressure, and DERATED by built up back pressure.
A lower MAWP for a vessel can require large diameter piping downstream of the relief device. You should consider the cost of the system, not each piece.
Stonecold's observation of 50 psig and full vacuum does not apply for all large oil companys. Our vacuum tower was 42' in diameter (design error) and only rated for 20 psig..... The flare header system was found to be woefully inadequate and had to change out components of the system to get it up to API standards.

Overpressure design was and may still be based on a design of 1/4 failure. That is to say - - a vessel with a 100 psig MAWP should fail at 400 psig. Code does not allow overpressure greater than 121 %for the design of relief systems. This is why you see so many more vacuum failures than vessel ruptures.

A HAZOP review and flare rating should be done prior to specifiying design MAWP of the vessel.

Also, when the designer calls for a certain wall thickness, the fabricator will buy the next thicker plate. Have the designer RECODE the vessel based on the fabricated wall thickness rather than the design -- less corrosion allowance of course. Fabricated MAWP... This is almost free MAWP...
Also consider using a Div 1 shop for welding. This gives a higher welding efficiency. Somewhat more expensive MAWP, but still on the inexpensive side.....

For vacuum failures, expense is the reason why all vessels are not designed for full Vacuum. Each system needs to be carefully reviewed with an eye towards unusual activities, startup, shutdown, power failures and even draining water after a pressure test... etc...

 

[Omama]

What is the operating pressure?
Nowadays oil refinery Vacuum towers operate at less than 1 PSIA so you must design for full vacuum

 

[sshep]

Hello Omama,
Thanks for your reply. Please keep coming back.

This old thread was questioning the practice of not making some large atmospheric towers capable of handling a full vacuum at a site where I was working start-up duty. I generally specify all such equipment to be capable of full vacuum because of the hundreds of ways it is possible to pull a vacuum. That site did not specify a full vacuum rating because of the added capital cost. Instead reliance is given to a large volume N2 breaking system (valve, pipe, multiple transmitters, ect) and administrative control.

My question was basically: how much cost is being saved by not rating an atmospheric column for full vacuum? Alternatively: how much have I cost my company over the years with my conservative approach? I think the total project savings of not making an FV rating is being over-represented and believe one day that site will suck a column in (and then how much was saved?). In any event this forum wasn't alot of help on this question except to show that opinion varies. The next time I am in this situation I will get a quote both ways and just see for myself.

Thanks for the interest. See you around- sshep

 

[roker]

hello,

in our refinery most of the equipment is designed for full vacuum: pressure vessels, heat exchangers, air coolers,
reactors, etc.
the demages caused by the collapsing equipment to piping, steel strucures, instrumentation, electrical and because of it loss of production justify to design also for full vacuum.

regards,
roker

 

[SeanB]

Normally one would look at all the operating conditions and evaluate whether a vacuum could be created during normal operating condtions. For example, rapid cooling and in turn causing condensing. You would also look at maintenance reasons. If your vessel is steamed out for maintenanace reasons, the vessel is normally designed for a minimum of 1/2 vacuum. Or at least that has been my experience.