tall vessel design -for seismic loading 1

[stephenebrady]

I am trying to use Pressure Vessel Design manual by Denis R Moss Procedure 3-8.

The period of vibration is s formula for imperial units, and i am struggling a little with my SI units.

Does anybody have a worked example (the one in Moss is no good, there's not enough info.), either in imperial, or preferably in SI units.

An alternative might be useful, I have heard of one in "process Equipment Design" by Brownell and Young.

 

[MJCronin]

I suggest that you consider a more detailed seismic analysis method for this important piece of equipment.

The simplified static seismic analysis methods described in Moss and Bednar are, in my opinion, dated.

A more modern and more appropriate method may be a detailed "time history" or response spectra analysis.

These will be more expensive and require the services of a structural engineering firm.

 

[stephenebrady]

Well it's a chimney, and the company has been building these for decades, as such theit company procedures are for now, set in stone. There are some other calcs I have come across where the chimney was put through a dynamic analysis in CAESAR II. But still, I would like to get to the bottom of this Periodic time eq'n.

cheers MJCronin for your help

 

[roca]

Does this mean then that the methods used by Compress and PV Elite (which I assume use methods similar to the Bednar and Moss) are now outdated and no good and we need to use the "time history" or response spectra analysis for tall vertical vessels / columns?

 

[dig1]

Roca,

You can do a response spectra in PVElite; I don't believe you can do that in Compress (at least not a year ago). I've used it before in PV and it works well.

If you do not have PVElite but have STAAD (or similar structural programs) you can model the vessel and obtain you shears and OTMs at various levels then use a spreadsheet to perform your longitudinal stresses. It isn't as bad as it sounds.

Patrick

 

[MJCronin]

roca...

You stated:

"Does this mean then that the methods used by Compress and PV Elite (which I assume use methods similar to the Bednar and Moss) are now outdated and no good.. ?"

I stated:

"I suggest that you consider..."

Perform the type of seismic analysis that the client desires and that is contained within the contract documents.

A suggestion is just a suggestion......a detailed reesponse spectra analysis may be able to actually save money in reduced member sizes and shell thicknesses over the more crude simplified methods.

I believe that my command of english and word usage is at least "on a par" with the average poster in this forum. Please pay attention to the words

Again, for the analysis and design an important piece of equipment that may cost hundreds of thousands of dollars, it may be wise to consider a detailed seismic analysis that will cost a few thousand. Detailed seismic analysis is more exact, and replaces methods that have been around for 50 years.

This is just a suggestion..... and my opinion only

- MJC

 

[vesselguy]

MJC,
I agree with what you're saying. If you're able to have access to a finite difference analysis for POV that can model a process column, then great and you should use it instead of the simple one liner formula or the summation of masses. I think most of us don't have such access to these expensive software. Even big companies are freaking cheap in buying software that don't justify an anuual cost benifit.

17 yrs ago, me and another guy was given the assignment by our Senior Eng at Fluor to dig up papers on vibration of tall columns and on vortex shedding loads on columns. We found a lot of papers but they were so theoretical that they were not practicle to apply. Long and short of it is that we concluded the closed form and simple methods given in the industry accepted bibles were good enough for "government work" and they worked. Of course one notable paper that everyone uses (including Compress) is the paper from Mr. Freese.

Periods of vibration in process columns is a very very complicated subject because the sloshing effect cannot be properly accounted for. Do you have any recommendation for a more recent paper that detail better POV analysis methods that we can use?

 

[vesselguy]

MJC,
Sorry, I didn't clue in that this discussion is about POV for seismic condition, and not about POV in wind condition, hence reference to Freese paper is not applicable to this thread.

So, is there a good paper on a speedier, yet detailed, calculations for determining POV of column in seismic condition?

Thanks.

 

[roca]

MJCronin
You “suggested” and I “questioned” your "suggestion” – note the ‘?’ at the end of my sentence.

Your command of English is pretty good and I did pay attention to your words – hence my post and question. (talking of English – please note that the word response does not have a double ee in it).

I have given your “suggestion” more thought and discussed it with other colleagues.

You “suggest”,
“Again, for the analysis and design an important piece of equipment that may cost hundreds of thousands of dollars, it may be wise to consider a detailed seismic analysis that will cost a few thousand. Detailed seismic analysis is more exact, and replaces methods that have been around for 50 years”.

How much do you think you will actually save by this detailed analysis?

I cannot see much of a saving on vessels / columns – maybe a small saving on shell / skirt thickness and on anchor bolt sizes.

On high pressure equipment you would only possibly save on skirt / anchor bolts, on vacuum and low pressure columns again you may not save anything on the shell due to the application of nozzle loads, deflection criteria, etc which would increase your base pressure thickness requirement.

I believe that for an average column there is no more than a couple of thousand dollars to save from such an analysis – which would be the cost of the analysis – so why bother?

Having a support skirt fabricated from 20mm carbon steel instead of the special analysis thickness of 16mm will cost maybe a couple of thousand dollars more but you will have a stronger construction – so why worry about sophisticated design methods?

How many of the columns that you have designed for seismic loading, using the 50 year old traditional method have actually collapsed?

Extra material will give you the extra comfort and there is no point designing right down to the bone. There is no worthwhile savings.

I believe the place for this special analysis method is on large structures – e.g. offshore platforms – where weight saving will always save money.

My opinion of course BTW

 

[gr2vessels]

Interesting subject and polemic. I am conservative and prefer the well known path.
However, even the oldies of ASME realised the the earth is shaking and new times are coming down, hard.
That's the real reason for up-dating the ASME VIII, because the new world belongs to the people who can save 4 mm of skirt thickness. The "special" analysis is not for the conservative people, it is for the new generation of engineers hell bent to make use of their engineering skills, rather than apply a questionable prescriptive design method. How smart one has to be to learn the Code backwards? You don't need a degree for that. Having said that, Roca and I will still be building safe columns for many more years to come.
MJC, show the young ones the real road..Don't give up and don't apologise.
PS. I'm always requesting the response spectra for the column analysis..
Cheers
gr2vessels

 

[dig1]

roca,

I agree with what you are saying for the most part; in general the more detailed analysis will not save any significant money. In areas where seismic loads can be significant, more detailed analysis can offer significant cost savings.

I recently worked on two projects in Chile that had vessels made from duplex and super austenitic stainless. The pressures were relatively low, 6 bar and with full vacuum. The savings significant because a response spectra was used in comparsion to a static analysis. The extra engineering was about USD1,300 per vessel but the savings involved reduced shell, reduced skirt, reduced reinforcing for skirt openings (two man doors per skirt plus piping openings), thinner base rings (they were still 80 mm inches thick), smaller anchorage (bolts and shear keys), resulting less reinforcing steel in the foundation, etc.

It is really not difficult to perform and I'm not right out of school. I will admit that in a vast majority of projects savings will be minimal, if any. However, one has to look past just the savings of the vessel (my company was the EPC contractor on these projects). It is really a case by case basis but one has to be open to see the potential savings on a project.

Patrick

 

[COSTASTSAPROUNIS]

Dear Sir
There are exist two seperate problems.
The first is the defintion of the first eigenperiod.
The second is the definition of the tower response to seismic excitation.
The approach of Moss and Bernard is very efficient for the definition of the basic period without sloshing effect.
The tower response can not be based on this approach except if you see tha the basic period is far away from the phasma of excitation.
In order to to have the repsonse of the tower in seismic response you can do two thinks
1. Pseudostatic analysis , applying the maximum g as a distributed load to all your construction . This leeds to overdesign of the tower.
2. Doing rersponse spectra fro the tower . This can not be in PVELITE because it will not use correlation between the eigenstructure of the tower and the excitation phasma.
In order to do this you need FEA analysis. A fast response spectra analysi can be done with SAP2000 or NASTRAN.

My opinion is to define the first 3 or 4 periods of the structure. A good estimation can be given by the book of Robert D. Blevins "Formulas for Natural Frequency and MOde Shape".
Having this values you can see where is your eigenstructure in comparison with the excitation phasma.
After this apply a hand made modal analysis (like pvelite does not)for the maximum g's tha corrspond to your 3-4 periods. Other wise use SAP2000 or NASTRAN.
This will give you a good estimation for the thikness of the various parts of the tower.
However you must remember that your savings in material will not be significant.
The main think is to clarify what we do for what reason.
1. Definition of eigenstructure to see if it conflicts with excitation phasma.
2. I we have conflict we do modal analysis
3. If the distance between eigenstructure and phasma is quite big pseydostatic analysis is enough.

Costas


Dr. Costas J. Tsaprounis