Acceptable External Nozzle Loads calculation for Pressure Vessels and Tanks

[Vikoll]

I am a piping stress engineer currently developing a company procedure Acceptable Nozzle Loads for Process Equipment. We need this procedure because in the past we have seen equipment such as tanks and vessels supplied without any allowable external nozzle loads, with unreasonably low allowable nozzle loads and with “zero” allowable external nozzle loads. This procedure is meant to oblige the potential equipment suppliers to follow standards and good industry practice and supply robust equipment capable of withstanding reasonable piping loads with the minimum values indicated in this procedure.
Another use of this document is to allow piping stress analyst to approve nozzle loads that are below those specified in the document without further vendor approval or FEA analysis.
In my past experience I have come across many different company standards where vessel allowable nozzle loads were calculated based on some empirical formulae (such as in attachment) and tank allowable nozzle loads were given in a tabular form without any explanation on how they were calculated. In many cases provided values and formulae differed quite substantially and it’s not clear for me whether different safety factors, nozzle flexibilities or approaches were use.
I have done research reading again Section VIII Div.2; API-650, some ASME papers, forums etc, but could not find any universal rule for calculating minimum acceptable allowable nozzle loads.
My question is how to calculate minimum allowable loads imposed by external piping on tank and vessel nozzles so they will be reasonably high to not create a problem for pipe stress analysts and designers and at the same time guarantee acceptable and safe level of a local stress?
Please provide your opinion or direct me to a reliable source. Thanks.

Cheers,
Vikoll

 

[MJCronin]

Vikoll...... an excellent question !

In 1977, while at an ancient company called Stone & Webster in Boston, Mass, I was given virtually the same task for Nuclear Power equipment. The task was reassigned over and over to new engineering personnel and it outlived at least 30 years in the engineering mechanics division.

There are no hard and fast rules...

Many companies develop tables of reference loads...... to be given to the vessel fabricator at the time of purchase. These loads are used at the time of piping completion and stress analysis to "pre-qualify" the nozzle.

There have been many threads on this forum regarding this subject.

Good luck.....please keep us informed about your final decision !

http://www.ast-forum.com/ast_forum_tree.asp?master=11209

http://cr4.globalspec.com/thread/18641/allowable-nozzle-loads

http://www.eng-tips.com/viewthread.cfm?qid=75687

http://www.lixanindustries.com/engineering/info/lore_view.asp?ID=16643

http://www.pipingsoft.com/bbs/viewthread.php?action=printable&tid=979

 

[jtseng123]

I am not the fan of nozzle loads because it is misconception in the first place, even our company and many other companies have table and formula for it.

Question is: what are those loads ? sustained ? thermal ? operating ? I haven't seen any company well define what it is.

In CAESAR, sustained and thermal loads will be compared with allowables, not operating. You may have more thermal loads then sustained, vise verso. Sustained stress may fail but thermal stress will pass, vise verso. And when you run the local stress analysis, either WRC107 or FEA, you need to identify which is sustained, which is thermal, because they have different allowable stress per Div 2, Part 5. Any local stress analysis by using "operating loads" has not much meaning per B31.3 or Div. 2 since both codes require separation of sustained (including pressure) and thermal loads, and then compare with the specific allowables.

Many pipe stress engineers I have seen, do not understand the significant difference between sustained and thermal loads. They will give me nozzle "operating loads" for me to run local stress analysis. I kick it back and ask them to identify " sustained " and "thermal" from CARSAR II output. I don't need any operating loads.

Pipe stress engineers will claim their loads will be within the table or formula. I am just laughing and would not want to argue with them. Are they treat these loads as 'operating" loads ? Most likely, but they miss what code says.

I give our load table and formula to vessel vendor, and tell them these are "sustained loads" so they can only compare with 1.5 S, not 3S, for a conservative nozzle design.
You will see some vendors compare with 2S, 3S, or whatever. The truth is, as long as vendor provides calculation, regardless what allowable stress they use, many vessel engineers just let it go. No one dare to bother.

So, what are the nozzle loads you want to define ? Sustained or operating ?

 

[JStephen]

Ideally, the vessel/tank designer calculates the stiffness and unloaded deflection of the nozzles, reports that to the piping designers, who calculate the loads, and those loads are then used to check nozzle stresses by the vessel/tank designer. In practice, this is not often done.

If you specify a set of loads based on nozzle size without taking stiffness into account, it is not so clear that you are accomplishing anything. A small-diameter thick-wall vessel will easily "pass" the nozzle load calculations, but it is the type of very stiff nozzle that will in reality generate the highest loads. A large-diameter thin-wall tank will be most likely to fail the calculation, yet it is also the most flexible nozzle that is least likely to experience those loads. Then the solution to "fix" that nozzle will involve making it stiffer, which will actually increases the loads on it.

Also keep in mind that for a vertical tank, there can be radial deflection of nozzles, vertical deflection, and rotation, even if there is no applied nozzle load. On a large tank, the radial deflection can be over an inch. Also keep in mind that the information available to calculate deflections and stresses is rather limited n some cases.

 

[DSB123]

jseng123,
I read with interest your reply. In particular "Many pipe stress engineers I have seen, do not understand the significant difference between sustained and thermal loads.". If this is the case then those Pipe Stress Engineers "you have seen" are not really pipe stress engineers except possibly by name at that particular Company. If they do not know the fundamental difference between sustained and thermal loads then have no right calling themselves "pipe stress engineers" and are earning money fraudulently. I think your quote is a bit "tonque in cheek" and rather offensive to most reputable Pipe Stress Engineers but I must add there are an increasing number whom could fall under your criticism but again they are not "real" pipe stress engineers except by name only.

 

[TGS4]

jtseng123 - your description of the different piping categorization of the stress: sustain, thermal, etc having different allowable levels when evaluated to the stress classification pertaining to ASME Section VIII, Division 2, Part 5 is completely and demonstrably incorrect.

ASME Section VIII, Division , Part 5, Table 5.6 says that the local membrane stresses from external loads (without reference to whether they come from piping weight or restrained free thermal displacement) are to be considered primary local membrane and the local membrane-plus-bending stresses from external loads are to be considered P+Q.

Anyone who is performing a WRC107/537 or WRC297 evaluation and making different stress classification depending on the origin of the piping loads is doing it incorrectly. In actuality, when the piping stress engineer provides you with operating loads, they are slightly more knowledgeable about that Code than you are.

For Vessel Component = Any shell, cylinders, cones, sphere, and formed heads, For Location = Near nozzle or other opening, For Origin of Stress = Net-section axial force and/or bending moment applied to the nozzle, Type of Stress: Local Membrane = PL, Bending = Q, and Peak = F

Accordingly, from Table 5.3 (assuming that you are using WRC107/537 to calculate the stresses in the shell), you need to evaluate Design Load Combinations (1) P+Ps+D, (2) P+Ps+D+L, (3) P+Ps+D+L+T, etc (there are 8 Design Load Combinations; you ARE evaluating ALL of them, right??). Based on the description from Table 5.6 (above), all of the piping reaction can be considered to exist as either D or D+T in Design Load Combination (2) or (3) - and I would argue based on Table 5.6, they are all D. Therefore, D (or D+T) are the coincident loads on the structure. They may be design or operating - the Code is silent on that (for the time being - hopefully I can get that changed for the 2015 Edition). You won't go wrong if you use the normal operating loads.

However, note that the check for PL is very different from the check for P+Q. PL uses the stress value of a single application of the loads. P+Q uses the stress range in the determination. For the P+Q ratcheting check, you need to do a whole lot more work to figure out what the operating cycle is, so that you can determine the stress range. If you aren't doing this, then you are doing it WRONG.

 

[jtseng123]

TGS4, I can't argue with you, as I have said very early Part 5 is too complicate, until I retire, I will still say I don't know what it is, and many vessel engineers are no better. Why ? because it is not needed for daily task.

Two commonly used programs, CAESAR II for pipe stress and PV Elite for pressure vessel, both are designed by the same COADE company. Assuming they are programmed per Div 2, Part 5 correctly, they need you to separate sustained load, thermal load, occasional load (wind or seismic) in order to run WRC107, and it will output stress into P, PL+PB, and PL+PB+Q, and compared with S, 1.5S and 3S, very simple. There is no room to input "operating loads".

According to the program, you can fail in PL+PB if you have large sustained load but PL+PB+Q will pass if you have small thermal load. All kinds of things will happen if you play around the loads. If this program is wrong, then so many companies around the world using it are just wrong for decades.

Your explanation is very profound, but I haven't seen any pressure vessel engineers or pipe stress engineers willing to go into that details to fully understand what it is. The code to too deep, let expert doing their job, and we engineers just do a quick and dirty calculation by running WRC107 with separate loads from CAESAR II output, but never operating loads. Most companies around the globe doing the same thing. I never provide operating loads to vendors. I always separate into sustained, thermal, and occasional loads directly from CAESAR II because playing WRC107 for 2 decades, I know operating loads can pass, but when loads are separated into Sustained and thermal , PL+PB can fail and requires additional reinforcement.

The topic for this post is how to specify some nozzle loads so vendor will provide additional reinforcement in addition to pressure area replacement per code. If the loads is define as operating loads, PV Elite can't run WRC107. However, it will be OK for Compress because it has only one load input, does not care what load category is. I do not think Compress is programmed correctly at least for one thing: allowable stress for wind and seismic can go 20% higher, and Compress won't recognize it.

 

[Vikoll]

Let me thank every responder for provided expertise. In an engineering field one can learn even from different opinions if a good engineering judgement is applied.

At the same time I would ask again a question about acceptable “benchmark” allowable nozzle loads for PV, heat exchangers and API-650 tanks that can be found in design standards of many well established EPC companies. I cannot adopt any of these without a clear understanding where these allowable loads are coming from and what formulae/approaches were used as a basis for its calculations.
It is my understanding that these loads have to be conservative enough to guarantee acceptable stress level regardless of the load type and reasonably high to not cause a trouble for piping designers and stress engineers.

I personally check nozzle stress using FEA software such as NozzlePro by PRG or ANSYS. Former one requires input of SUS, OPE and OCC loads from ASME benchmarked pipe stress analysis software. I have to use FEA to prove that piping design is safe even though the manufacturer provided unrealistically low allowable loads or did not provide any
I want to establish design criteria with loads in a tabular form for variety of sizes, ratings and materials (CS/SS) that will be submitted as a bases for design to equipment suppliers. I believe that this can be accomplished with some help from the industry experts and ask you again to provide your opinions on this.

Thanks and Cheers
Vikoll

 

[jtseng123]

Vikoll,
NozzlePro is in line with the input WRC107 in PV Elite. THERMAL =OPE -SUS that NozzlePro will be able to extract. But according to TGS4, they are incorrect. However, many companies and vendors have no better cheap and quick tool, and CAESAR II does produces Sus and Thermal loads, we will still use them as input to WRC107.

The problem of load table or formula is: it misses to address the temperature. Applying the same loads to 100F vessel and to 1000F vessel does not make sense, knowing that it shall be pipe stress engineer's responsibility to provide additional flexibility in high temp service.

MjCronin has provide links with some formula
Ours is no better (for vessel) :
150#-300#
all 3 forces (lb) = 300xD, nozzle dia. in inch
all 3 moment (ft-lb) = 57.8xD^2+2.6D+9.8
600#-2500# , increased by 84%.

These are treated as operating loads. That comes to tricky part in NozzlePro or WRC107 in PV Elite. If SUS has no input, the entire load become thermal.
Some of my vendors use PV Elite, so I ask them to input as SUS, that will generate a very conservative design. Perhaps a 50/50 split is a better choice.

My other recommendation : even the nozzle just meet the minimum code reinforcement requirements, it can take certain amount of piping loads due to 1.5S and 3S or 2Sy high local allowable stress per Div. 2, not "zero" as claimed by vendor. Why not asking vendor to provide additional reinforcement , says 20%, 30 % beyond minimum code requirements instead of table and formula.

 

[C2it]

I must agree with DSB123.
Quote jtseng123:
“Many pipe stress engineers I have seen, do not understand the significant difference between sustained and thermal loads.”
Unquote.
jtseng123 … Your firm must employ some rubbish ‘stress engineers’. Perhaps they are just software monkeys ?

The fundamental need for clear allowable nozzle loads is to facilitate parallel engineering to a tight schedule. For example, confident design of a long lead vessel, prior to stress analysis of the connected pipe.

In this respect, what is required is a set of realistic loads to be used as maxima by the stress engineer and minima by the equipment designer. Reliable loads for most common equipment configurations are provided in the Norwegian document, Norsok R-001. This is available free of charge on the web, and has been used in North Sea area design for about 20 years.

 

[jamesl]

I agree with TGS4. I think jtseng123's opinion on nozzle load analysis is little out of date. Several papers have been published by vessel engineers since 8-9 years ago to not use the sustained / thermal classifications in B31 codes on vessel nozzle load analysis. They suggested treating membrane stresses developed from all nozzle piping loads, regardless of whether their origin is sustained or thermal, be treated as primary for the purposes of evaluating the corresponding shell stresses.

 

[JackSam]

Try NORSOK Standard R-001 Para 5.1.5. NORSOK standards are available to download for free from:

http://www.standard.no/en/

 

[Vikoll]

Let me thank all participants for provided guidance and opinions. I downloaded a copy Norsok R-001 but still have a spread +/- 100% for loads recommended by various company standards. My personal opinion Norsok R-001 allowable loads are very conservative and may be difficult to comply with by conventional piping design means. If you have any other thoughts/suggestions or can share your valuable experience your input will be truly appreciated.
Cheers
Vikoll

 

[C2it]

I disagree about Norsok being conservative.

The pressure vessel allowables are a math model of the tabulated values used for years by major contractors.

Allowables on pumps are very generous, up to 10x API 610, and equally with compressors at 4x NEMA SM-23.

Like I said previously, Norsok has been used in North Sea development for about 20 years without complaint so far as I am aware.

 

[saplanti]

Please do not forget that the arbitary nozzle loads issued by contractor companies or standards are for communication between pipe stress analysts and vessel manufacturers. Therefore, the values are targetting both sides to comply, and are the opinion of that company company or standards to make the project equipment workable and reliable. In some industries, as indicated by C2it the higher loads are selected, and these come from the experience of the contractor company in that field. Therefore there is no right or wrong nozzle load, but the implementation sometimes is very difficult on the piping side if the selected nozzle loads are small.

I have to agree with jtseng123 on the evaluation of the loads. The contractor company should provide the acceptable load for each load case (I have seen one company is doing it). I saw many engineers use the allowable loads in their favour since there is no proper definition of type of the allowable load. Some engineers, this includes me if I do not have the vessel to design in addition to the piping, use the allowable loads on the conservative side. But this makes the vessel more expensive than required.

Believ me if you do not see the type and the allowable load in the specification, the contractor company or its engineers do not knon what they are doing or they are not experienced. This makes the job very difficult if the piping and vessel side engineers are not experienced and do not know what to ask and whom to ask. So the jobs most of the times left to the risk analysis of the contractor companies.

Kind regards,

Ibrahim Demir

 

[saplanti]

Some cases, the owner companies want vessels survive without repair a lot longer than the piping. In this scenario the owner or contractor company requires larger allowable loads (perhaps with higher shell wall thickness) on the nozzles to design the vessel to meet this requirement. So we cannot see this directly from our office unless we involve in this type applications or get information about the reason officialy or unofficialy.
Please forgive me I did not check the spelling in my previous comment.

Regards,

Ibrahim Demir

 

[jtseng123]

Finally, saplanti understands what I am talking about.