For pump loads which fail API 610 is it possible to reduce the stiffness value of the nozzle connection from the standard anchor stiffness? I have never seen this done but at the company I am working for now they're reducing the stiffness to 1.E6 to pass API 610 allowables. I'm not sure that I agree with this method or whether it meets industry standards.

Hi mull211.
You should not do this. If you calculate a pump this way surely it would have alignment problems with the shaft.
You should add more flexibility to your system.
You could reduce the stiffness of a support, for example, if you do the proper calculation of it. But you can’t estimate this for a pump.

Regards,

OG

Are you nuts????? Engineer with Lurgi Indeed! Tsk Tsk...

Why not just change the modulus of elasticity to Tapioca Pudding!

:rolleyes: :rolleyes: :rolleyes: :rolleyes: :rolleyes: :rolleyes:

Ok.
Read up on why you are doing pump calcs.
(i cannot believe that you are doing stress and are not aware)
The whole purpose of a pump calc is to limit deformation of the pump, and hence avoid excessive wear on the rotating components.

jeysus.

If you think that is bad, I have seen a Ph.D. (definitely not sure in what field of study) reduce pump nozzle stiffnesses until the pumps pass! I told him not to do this and he never talked to me again.

Hi Jim,

You are certainly correct to question the practice of "softening" the boundary conditions at the pump. As you might have already gathered from the reactions of our esteemed colleagues (above) your client is paying for professional engineering and the practice that you describe is something less than that.

I will say however, that although my esteemed and very very much more senior colleague, John Luf, has suggested the modulus of elasticity of Tapioca Pudding in jest, that might be something to think about (albeit, Mr. Luf will have to supply the modulus of elasticity data - it is not to be found in the B31 Appendices). Remember that the B31 Codes require that you use the material modulus of elasticity at 70 degrees for stress calculations (Caesar II default). These calculations will give you higher calculated reactions at the pump than will really exist. You can legitimately use the material modulus of elasticity of the pipe at operating temperature to perform your equipment loading analyses and your hanger design analyses and this will provide more realistic calculated reactions. Just be sure to destroy the calculated stress output as it will not be "per Code".

Regards, John.

Agreed Mr. Breen.... as for Tapioca I'm not sure but google could probably come up with something.... I believe the proper term for a fluid would be the bulk modulus of elasticity.


Hot Mod E may not be of much help to you, steel takes a lot of heat before E drops appreciably and we are talking about a pump not a turbine!

Superpiper, your intent always seem to be abusing and ridiculing people rather than providing a logical answer. By scanning through your responses, I conclude that you are in your infancy as far as knowledge and maturity is concerned. However you tend to show off that you are a real expert.

I would appreciate if you refrain from answering.

The tendency in this forum is to try and show that you are smart, team up to shut a not so experienced user…..is that the purpose?


Going back to the original question, did it deserve all these criticisms?

Why one million lbs/inch is a bad number for a small size pump nozzle?
What is the basis of one trillion lbs/inch, used by Caesar as a default, that it has to be mandatory in case of a pump nozzle of any size?
Has anyone checked with a Pump vendor?
Has anyone checked with API ?

If API gives ridiculously low numbers, is it likely that the Pump casing will undergo huge deformation or misalignment the moment the forces are slightly exceeded?
Are not the values given by API overly conservative?
Has anyone checked with API or challenged them?

The methods used by most of the analysts to meet API loads, namely: Use springs, use loops and bends, use strange limit stops that are not practicable to be installed properly at site, use Teflon slides and so on, what it does to your piping when it actually operates?
Don’t you end up with piping that is too flexible and vibrate badly during operation, eventually affecting the pump itself?
How many times have you gone back to field after one year and checked that the operator has not put some strong U-bolts to stop vibration and the pumps are running fine?

Okay then lets say the default stiffness is too high, what do you propose the number should be???

Get the data from the pump vendor?... not in my or your lifetime. Is this a standard piece of data in API?... NO! Has such data ever been published and accepted by any or all pump manufacturers?... NO!

So what are we left with?.. Meeting the PUBLISHED ACCEPTED BY ALL LOAD values, using a default and conservative boundary condition, or venturing off into terra incognito? I vote conservative and accepted versus terra incognito.

As for how to make the piping adequately flexible and stll not vibrate excessively I have been doing this quite successfully for many years, these are after all centrifugal pumps not pd type pumps.

Or one could guess at a Krate and make the loads go away with the mathematical trickery....

I saw an example of such nonsense on a hot oil system once. The output report said everything was within allowed values yet the pump seized tight after 20 hours of run time... I guess it couldn't read the output report.

So trick up things in the model based on guesses and unfounded opinions or do the job correctly, I personally don't care because like the 20 hr pump I will probably be well paid to fix up the result of the magic stress analysis....

As for "If API gives ridiculously low numbers" the numbers are what the manufacturers based their design on. Someone may second guess that they are too low but its only their guess.


While I type this response up the story of the tragic Adrea Doria sinking is on PBS.... more poor human judgement at play.

Hello shorustress and all,

I will say without pointing fingers that I agree in a general sense with your assessment of abuse and ridicule. Strong disagreement can be communicated without resorting to rudeness. As I read the original question, Jim is of the opinion that the practice (of some of his compeers) of "adjusting" the pump stiffness is a form of "wishful thinking" and he was looking for reinforcement from the community - abuse is never deserved and in this case his skepticism should have been applauded. We should not want to discourage the free exchange of ideas here even in those instances (not this one) where they might seem to be naïve to us.

I think that the API method of calculating the allowable forces and moments gives piping engineers something to try to "shoot for". The API Standard actually tells the pump manufacturers what loadings their pumps (for specific suction and discharge sizes) must be able to accommodate if they want to claim they make API-610 compliant pumps. That leaves the FEA evaluation of the pump casings as the responsibility of the manufacturer. The API limits on pump loadings were not set arbitrarily, they are the result of a study that included significant FEA analyses. I see the API-610 Standard as a "place" where the pump industry and the piping engineer can come together in the interest of good design. I think as of the seventh edition of API-610 we have reasonable allowable loadings from the Standard.

Structural analysis of piping systems is an approximate science. There are many unknowns (or "variables") that enter the modeling process and I believe we are lucky to get within 10 percent of the true loadings on pipe supports and terminal equipment (just the amount of inadvertent "cold spring" that is built into every piping system assures inaccuracies). There must be some design margin in the piping systems to assure that they will have adequate useful service lives. There is enough margin for the piping stresses thanks to the piping Codes. The strain sensitive rotating equipment (e.g., pumps) to which the piping is connected deserves a little tender loving care too.

I think that if we consider the pump to have six degrees (3 translational and 3 rotational) of absolute fixity we are somewhat conservative but if with this boundary condition we can work with our piping design to achieve compliance with the loading limits of API-610 we will have better pump bearing life. If the pump casing and nozzles are more flexible, it is necessarily at the expense of a flexing casing (and/or nozzles). The more the casing flexes the more misalignment we have of the bearings and the coupling to the prime mover - not a good thing.

The thing that makes us piping engineers is being competent to use our skill to supply enough flexibility to assure the limits on pump loadings are met but at the same time addressing all the many other design challenges that come with adding flexibility to the system. I say all of this in support of my opinion that we as piping engineers should make every effort in our designs to satisfy the limits on pump loadings that are prescribed by API-610, while assuming an absolutely rigid pump..

We do have the option of communicating with the pump manufacturers and submitting to them the lowest set of loadings that we can achieve and sometimes we can get approval for those loadings. Also, some pump manufacturers will suggest spring loaded pump bases as an alternative to reduce the loadings. This I think should be a last resort (even after expansion joints).

Regards, John.

Shorustress,

If the results change dramaticaly if you change the support (nozzle) stiffness, then the evaluation of that model is very sensitive to that stiffness.

If you wish to "play with the numbers" to get a certain value, that's your business. But, in my opinion, that does not replace goood engineering.

By posting your issues/questions on this forum, you are asking others opinions. Perhaps you should seek whatever value lies in even those critical opinions.

shorustress.
Sorry you feel that way.
My european sence of humour seems to be missed by the rest of the world.
No offence meant,
Yes i am in my infancy.
No i am not smart.
No i don't (like alot here) profess to be a stress engineer.
I don't have a degree (although i went to UNI)
But i do have 18 years of experiance in piping
Hence my tag of superpiper.
I dO read and i do try to investigate a problem before posting on this board.
More so, i realise my limits and would never try to engineer something i am not capable of.
I am constaqntly coming across so called stress engineers who have not a notion of what piping is, of the consequences of something going wrong.

During this last year, i am noticing an increasing number of posts from people who plainly have no idea what they are doing and have no supervision. I find it worrying and allarming.

My last analysis of a glass lined system(as discussed on this board) was sent to a satellite office for review and check by an independant engineer. This was at my request because i wanted to be sure that i had done the work correctly, so no i am not an expert.

My humour is often missunderstood on this board
Jeysus is tonque in cheek here in ireland
so dont be insulted.

asking how to ask how to fudge a code is wrong
wafling around stack calculations is wrong
wondering why caesar does not overstess Op cases is wrong
The list goes on.

In my view, and i am pushing this on the company i am working for, it should be manditiory that anyone wishing to perform an analysis on Caesar should have attended the training course first.

People doing stress analysis of the back of a degree only,without experiance,without supervision,without piping knowledge,without reading codes, without using the search button, without morals, are in my opinion immoral and should be flogged.

pax to all....

Simple question to answer the original comment. On Small Pumps (12" diameter nozzle and less) who does not think the pump will move one (1) inch if a force with a magnitude of one (1) million pounds is applied to the nozzle? Assume in only one direction, but in reality it could be all three.

Because if you feel 1 million pounds per inch is too small for a stiffness, then you must also believe that the pump will with stand a million pounds of force. If you find that pump, let me know. Thanks.

of intrest:
My latest analysis is for Thermal fluid system
Using EN ISO 5199 pumps.
The forces on these pump were all ok except fort the vertical force on the discharge nozzle, which was a factor of 2.5 times overloaded.
The Design team insisted that bellows/springs/routing changes were not possible and that the pumps should be ok as the standard pump layout had been used for years without problems.
After much arguing, i re-read the code and noticed a clause that the forces could be whatever you wanted them to be, as long as the supplier agreed. So for the first time in this companies history, i created a force matrix and passed it onto the pump people.
Neadless to say,the forces were rejected and I am still stinging from their reply.

SIHI Reply in full.............


"the question is not easy to understand. Someone need to contact the customer and ask what the calculation shows. Our pumps are designed in accordance to ISO 5199 which content is the same then DIN ISO 5199. Overload is not allowed. They compare the ISO load limits, reduced by factor 0.7 with a "Caesar" calculation. What is that? The load are extremely high. Even if the pumps would withstand the load, the hole machine will become deflected and sealing failure should be expected. We were never faced with this problem. Please explain what Caesar describes"

The layout has had to change!

So for Jim, i would suggest he suggests to his company that the API code is an absolute. and that way, nobody is at risk.

PAX?
A non-profit, non-governmental Catholic peace movement working on a global scale on a wide variety...????

Best of luck

I don't know that saying "can we add a flexibility factor to a nozzle in a pump" is necessarily "fudging the code" - as long as there is a good basis for the factor.

I have been known to mutter "these loads must have come from hand calcs" every time a pump fails when I have a gap on my guides and passes if I don't.

At the end of the day, I don't think that there's a huge difference between asking "can we model the pump nozzle flexibility" and "do we have to put that X$#%#$# gap in the guides".
(The answers may differ, but the basic question is the same).

I have also seen printed advice like "ignore friction at the first support from the pump"... is this "fudging" or "sound engineering"... depends on who you ask I guess.

Fortunately on at least some of my stupid questions I have asked them of people privately so managed to avoid public ridicule, though I'm sure some of the people I've asked the questions of still snicker once in a while.. I'd rather see a question asked and the asker learn that "it's not that easy" than have them scared to ask. Remember to the person asking, it's usually the first time they've asked, to the person answering it's the 100th time they've seen the same question.


Shannon
PS, it's not just your Irish sense of humour, mine gets misunderstood all the time too... online is always worse than in person.

pax = latin for peace....

Latin courtesy of the Romans is the most common of all western languages and portions (words) can be found in most of the Western languages.... exception being German / Deutsche....

So Kerwin please bless us all with the official Kerwin krates for pump nozzles... I intend to print the data out and add it to my next calc file.

Thanks for ethical responses.

I have had cases where the Pump vendor has accepted loading higher than API Appendix-F (2 times).

I have also had cases where we were to make a tie-in to an existing system so we had analyzed the existing pump lines for completeness of the model and found that the nozzle loads are way out( of the order of 3 to 4 times of API). When we told the Client about it, they said that they were running those pumps for last 15-20 years and had no problems with these pumps.

My feeling is that whatever method has been used to establish the tables in API, there has been a very generous factor of safety used on the allowables.
This does not only make the life of the Piping Engineer difficult, but make the piping system very much prone to vibrations.

I would like have opinion on this from the forum members.


I personally do not change the stiffness values but when a question in those lines are asked, I do not immediately conclude that the user is ignorant.

Numbers and rules are black and white. The B31 codes have built in factors of safety... a competent engineer does not promote a design outside of given guidelines.

Ok lets say we use a factor of 3 on the vendor or API allowed values.... and maybe things just don't work in your favor... maybe they build the pipe with a little cold spring, maybe the pipe wall is a little heavy, and maybe the pump case are is a bit thin on that production run...

So the pumps get whacked out guess what I would say if asked to review your dandy design!

Rules and standards should be followed and not trivialized and minimized... but hey go use mod e = 10 e3 and a krate of 500 #/ in your work, its your reputation, client base and maybe license....

As Forest Gump once said Stupid is as Stupid does.

Mr. Luf you are not saying anything new here. Everybody knows that one has to design within the published code. No questions asked.

You have misinterpreted the question.

It is just a dicussion as to wheather the code in this case is unreasonably stringent that results in unnecessarily flexible piping.

If a qestion is asked just for discussions sake, does not mean one would implement it.

(Rest assured that all the Engineers in the world follow the code. They certainly know what they are doing.Do not violate the code of ethics by not showing respect to your fellow Engineers. )

hi there mull,

you have said that you analysis failed on API610, but I am not sure if you have tried to check it 2 times the API610 code allowables... then also, you must pass it according to appendix F....,my advice is to read more about the code so that you will not be lost on your profession.

Best regards...

ian

Hello all,

The API-610 limits on loadings provide the manufacturer with the MINIMUM loadings their pumps must be able to accommodate if they want to claim they make API-610 compliant pumps. For many years prior to the present API-610 limits being adopted some pump manufacturers simply chose their pump dimensions (strength) by trial and error (there was no FEA then). Sometimes they wanted to use the same tooling to manufacture pumps of various materials and for design of the tooling they used the dimensions that would result in an acceptable strength regardless of the material. I don't doubt for a moment that SOME pump manufacturers (not likely the majority) are building (designing-in) additional robustness in their pumps. The old saying "when in doubt, build it stout" comes to mind.

So yes, I imagine there are pumps out there that for various reasons have a significant design margin (i.e., they are "overbuilt") and we will wonder why they perform with excess calculated pipe loadings. But, the pump (and valve) industry is changing and more pump "designers" are compelled to "make the most efficient use of material" now. So while I will be delighted when a pump in an existing system continues to perform with calculated excess loadings, I will endeavor to comply with API-610 in the design of new pump suction and discharge piping systems.

Regards, John.