Full Vacuum vs Half Vacuum 5

[sshep]

My Friends,

I know that this subject has come up in forums several times, but my research of the site does not give a satisfactory explanation so I am posting the question again.

Is there any API, ASME, or other documentation which rationalizes why a half vacuum rating would be sufficient for vessels subject to steam out?

We are designing several new plants and there is a push to design vessels for steam out using a half vacuum condition rather than full vacuum. The saturation pressure of steam at half vacuum is in excess of 180F, so it is a hard thing for the process engineer to swallow to rationalize how half vacuum rating is sufficient protection for all the possible ways to collapse steam during a steam out activity. Early in my career I saw a vessel collapsed during cooldown after steam out, and myself and most of my peers have always specified full vacuum ratings. Now we are being asked to accept a half vacuum rating on the basis of "thats the way alot of companies do it", which inspires no confidence.

best wishes,
sshep

 

[Latexman]

The three companies I have worked for use(d) full vacuum.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[MJCronin]

I believe that it is reasonable and prudent to design pressure vessels and piping to "full vacuum" if there is any vacuum.

All vessels I have specified in 42 years of engineering follow this philosophy and it has been incorporated into company guidelines.

There is no significant capital cost savings for true(Coded)pressure vessels and piping for a partial over a full vacuum design

I do not believe that you can justify that during steamout conditions approximating a "half vacuum" will occur.

Only an MBA would try to develop design rules using a justification such as ""thats the way alot of companies do it"

____________________________________________________________________________________________________________________________

Exception: API-650, 620 flat-bottomed tanks are sometimes designed to part vacuum.

I would add to this exception, a partial vacuum may be considered for dished-end horizontal tanks with a diameter greater than 60 inches and D/T > 90


MJCronin
Sr. Process Engineer

 

[sshep]

My Friends,

I really appreciate these answers, and they support what I have been doing my whole career. Still, I know that this half vacuum thing has a following because other threads on this site reference it.

Please keep weighing in. I need all the ammunition I can get.

best wishes always,
sshep

 

[SeanB]

I am not sure of the exact reason but it is pretty typical to only specify half vacuum for vessels that are only steamed out for maintenance. That is what is specified in our design manuals. Maintenanace procedures would dictate that the vessel is not to be blocked in during steam out. My guess would be it is a matter or dollars and cents. Half vacuum likely means a thinner walled vessel and a cost savings. If we used full vacuum for all vessels there are likely clients out there that would say we are over-designing and wasting their money. Half vacuum is probably a happy medium from a risk perspective.

 

[moltenmetal]

While a vessel designed for 150 psig will likely inherently resist full vacuum, a large vessel designed for less than that could require vacuum stiffener rings etc. which can add substantial cost. Pipe rather than ducting is generally inherently vacuum resistant up to at least 12".

If a vessel is not designed for full vacuum, vacuum relief is required. If the right vacuum relief system is selected, installed and maintained properly, 1/2 vacuum is plenty of driving force to actuate it reliably.

Where people screw up with steam-out cases is forgetting about condensate accumulation. They try to install vacuum relief valves down low somewhere so they can be easily maintained, and the pipe connecting them to the headspace of the vessel fills with condensate and the head pressure and/or corrosion prevent the valve from opening. I had a paper which showed a picture of a huge coking drum which collapsed during post steam-out cooldown for this reason.

 

[MJCronin]

I have a few more comments:

SeanB: "Half vacuum likely means a thinner walled vessel and a cost savings" -

I disagree...Cost savings are not significant unless they are for materials more expensive than carbon steel AND for diameters and the dimension ratio as referenced above.

SeanB: "it is pretty typical" and "everybody's doing it" are not good enough for me. Got any ASME papers, studies, experiments.... or anything ? How about a Chevron/EXXON/NORSOK standard or guideline specifying your half vacuum criteria ?....anything ? anybody ?

Why I inherently "feel" and agree that a steam-out condition will not generate a "full vacuum" (a condition that is difficult to achieve in a lab), IMHO the best course of action is to recommend a full vacuum design to a client AND LET HIM MAKE THE CHOICE FOR THE VESSEL THAT HE IS PURCHASING...

MM: I respect and concur with almost all of the things you have contributed over the years to these fora.

But, any CS/SS pipe (not duct) up to 36 inch diameter and 0.375 wall has full vacuum protection ( under the rules of ASME VIII). This means smaller Schedule 40/80 piping has full vacuum protection

Any other contributions out there



MJCronin
Sr. Process Engineer

 

[MJCronin]

Oh.....and another thing about vacuum and piping systems....

Watch out for that worthless, useless, bad schedule 5S piping under any type of vacuum conditions... !!!!

This is piping that MBAs love so much....and will require you to incorporate in your designs.

After the project is built, and after several important client meetings,(that you were not permitted to attend) a crazed MBA will come to your desk and explain to you how you must perform a "stress analysis" to prove to him how his bad Schedule 5S piping system will sustain full vacuum.

Oh yeah, he needs that calculation on his desk by tomorrow.... :0

(yeah, its a big chip on my shoulder...)

MJCronin
Sr. Process Engineer

 

[moltenmetal]

MJCronin: I don't think we disagree- I haven't done the vacuum calcs for larger pipe, so I couched what I said, limiting it to 12" NPS. As far as I recall, up to 12", even sch5S stainless is good for full vacuum to a pretty high temperature. The only time we use sch5S is for thin shells for tanks, because we don't have heavy slip rolls to work with sheetmetal- we NEVER use it for piping. We haven't found sch5S to actually be cheaper than sch10S, though it probably is if you're buying mill-run quantities, so why would we select it? Definitely a bear to deal with at the larger sizes- so thin that butt-welding requires all sorts of jigging etc. We use sch10S stainless extensively, are quite used to it, and as long as you don't weld out the full bevel of 3000# o-lets on it, it really isn't a problem for us from a fabrication/welding perspective.

 

[sshep]

My Friends,

Thank you for the further discussion. It still comes down to two camps. Those who operate will want the most inherently foolproof system, and those who try and save cost. Please keep posting as I really wish to know.

In the meantime I will try and give you something for fun (no one hurt, but very expensive for the owner) which shows even a lightweight pipe can collapse. This was a compressor suction, "protected" by an instrumented system.

In nearly 30 years I have seen many types of procedures and systems fail with serious consequences or lucky near misses. I will always like that which is inherently foolproof.

best wishes,
sshep

 

[moltenmetal]

sshep: the plant that doesn't get built because it is too expensive is also inherently safe.

Nothing is foolproof- the best you can hope for is fool-resistance. While protection by inherent design is sometimes possible without a whole lot of extra cost, that needs to be validated on a case by case basis rather than merely being assumed. I've seen client specifications which are a long list of little things that couldn't possibly increase the cost that much, but actually do- sometimes enormously, and for very little if any meaningful safety benefit. Unfortunately, the people who put those little suggestions into the spec have all long ago left the company, and the kids left behind know only that they are not to question the wisdom contained in the holy books.

We don't design tanks for full vacuum, even though that would be inherently safer. Some people design all stamped vessels for at least 150 psig and full vacuum as a matter of course, irrespective of the process's requirements: it makes them easier to operate, reduces relief-related issues, and makes them more useful in the boneyard after the project is over. We love customers like that- they make us lots of money for which we are very grateful- when they're actually able and willing to pay the resulting cost of course. But we don't assume that everyone wants their plants designed that way, and if we did, we'd be soon out of business.

 

[sshep]

Thanks Moltenmetal,

I know you are a solid engineer and active in this forum. I really appreciate all the advice you have given over the last ten or more years. You know I will gladely give a star to everyone who can tell me by what reasoning did half vacuum gain a following.

If we do this as a role play, it typically goes down like this:

> Designer: We recommend designing the vessels for steam out using half vacuum rating.
> Client: Why?
> Designer: It will save alot of money.
> Client: How much?
> Designer: We don't know.
> Client: Knowing that steam condensing can pull nearly a full vacuum, what is the rational for half vacuum?
> Designer: We don't know (or in my case "Thats what some of our other clients do").

It is at this point the client begins to wonder about the quality of the service he is getting.

By the way, the reason we accept tanks to be designed for atmospheric is because API 2000 provides a very quantitative inbreathing protection standard rather than leaving it up to each individual client. If half vacuum were backed up with similar reasoning, I probably wouldn't have many questions.

Please give me a few more posts.

best wishes always,
sshep

 

[SeanB]

To MJCronin:
I think you should go back and take a read of what I wrote.

"I am not sure of the exact reason but it is pretty typical to only specify half vacuum for vessels that are only steamed out for maintenance. That is what is specified in our design manuals."
"Our" being the EPC I work for and have worked for many years. Exxon design practices also state to consider vacuum condition after steam out - but it does not state either half or vacuum should be used.

I presented a statment of fact based on what my company does. I don't know the reasoning behind it. But this is supposed to be forum where we can bounce ideas and opinions off one another. Not to personally attack one another.

 

[MJCronin]

I am sorry...but

No attack was meant,....however, I am sure that you agree that, for true progress and learning to occur in an engineering forum, more than just an opinion must be offered.

sshep and I are on the same page...

BTW, MM, I have also found that 5S SS piping is more expensive than schedule 10S..... and far more difficult to weld ! I agree

Also, with regard to decisions about vacuum rating for piping and equipment, see the referenced NORSOK document:

(Section 5.5)

http://www.standard.no/pagefiles/1212/l-002.pdf

MJCronin
Sr. Process Engineer

 

[moltenmetal]

sshep: here's your argument, restated:

Designer: We recommend partial vacuum design for vessels which must be steamed out.

Client: Why?

Designer: Because we need the vessel to be able to sustain enough vacuum during condensing of the steam without collapse to reliably actuate a vacuum relief valve.

Client: Why not design for full vacuum?

Designer: If you'd like us to design for full vacuum, we certainly can do that. It will eliminate a vacuum relief valve which you'll need to maintain and recertify. It will make the vessel inherently safe against vacuum, but will not guarantee that the system will be vacuum-tight- what leaks back in will be air from the atmosphere rather than the gas we'd feed to the inlet of the vacuum relief valve. It will also require us to either significantly increase the thickness of the shell or add stiffener rings to the vessel to resist full vacuum.

Client: How much extra will that cost?

Designer: I can get the vessel quoted for you both ways so you can make an informed decision. But my gut feel is that the extra cost could be substantial because the vessel is large.

Client: OK, show me both costs and I'll decide.

Vessels made from small pipe are going to be vacuum resistant inherently, so full vac design costs you nothing. The issue is for large vessels with low MAWP. Full vacuum rating on an a large diameter low MAWP vessel is not a trivial additional cost. That's the trouble with design by rules of thumb- thumbs come in different shapes and sizes.

 

[sshep]

Hey Moltenmetal,

I will give you a star for the effort, but when did you guys add the vacuum relief valve?

What I was really talking about is half vacuum for steam out, and no vacuum relief valve. The question was how to rationalize half vacuum (or any other partial) vs full vacuum. Even after this discussion, half vacuum still seems arbitrary.

I sort of believe that there was a historical reason from ASME PV Codes which drove half vacuum to gain a following, but perhaps these reasons have been superceded and are now lost in time to the audience in this forum.

best wishes,
sshep

 

[130684]

Hi,

According to Shell DEP, "If the pressure may drop below atmospheric pressure due to utility failure, instrument
failure, low-load conditions, mal-operation, blocked-in situations or any other cause, the minimum possible absolute pressure shall be stated in the documents for mechanical design.
Steaming-out and draining operations need not be considered since these shall be covered
by adequate operating procedures."

Cheers

 

[jte]

Well, I suppose I'll jump in to the fray.

How many Section VIII vessel failures related to steamout can we point to (specifically, not "I heard once upon a time")?

Of those which failed, what fraction were designed for at least half vacuum?

We could, of course, design every vessel which is subject to steamout for full vacuum. But why require FV when experience doesn't demonstrate that it adds value?

Part of the problem is that there is very little hard data that I am aware of which either camp can point to. Given the lack of data, the argument comes down to "I'm conservative and most careful" vs. "I'm including sufficient design margin and sufficiently careful." I'd certainly like to see something like an API survey related to this issue... For now I'll stay in the "half vac for steamout" camp. Naturally, I also expect a max allowable external pressure to be determined and this value to be stamped on each vessel.

Next... We'll be discussing what temperature said FV / Half Vac condition should be designed at. That should provide for some vigorous discussion as well.

 

[sshep]

My Frineds,

If you work in production for a long time, and over many sites, you have probably seen vacuum failures. This is usually just going by to see the result of some unfortunate story which sounds completely idiotic in retrospect.

My first vacuum experience was a steam out incident early in my career. In the thread above I have some pictures of something equally dumb, where humans failed. Humans and their procedures fail more often than our office bound peers think. If you wish to see more, I have a nearly 30 year career of this stuff.

Anyway, thanks for all the input.

best wishes,
sshep

 

[jte]

sshep-

I have worked in the same business as yours for a long time (you have a few, but not many, years on me), and I support sites globally.

I don't doubt that external pressure collapses occur, and more often than we'd like. However, this is not what you were asking about. Simply put, a vessel which is designed for half vac would be expected to survive a full vac condition - the design margins are sufficient. [The premise here is that we're talking about ASME VIII vessels.]

Thus, I will reiterate my question:

Of those which failed, what fraction were designed for at least half vacuum?

I'm sorry that my perspective does not support yours... But that does not make either of us wrong. I'd simply like to see more data before pushing for a change.