Piping Loads on Rotating Equipment 3

[pstress]

When evaluating the piping loads acting on rotating equipment, I look at the G+T+P (Gravity + Temperature + Pressure) load combination and compare against the allowable loads permitted for that equipment. I also check the T+P (i.e. no Gravity) case. It is possible that T+P can be greater than G+T+P, depending on the piping arrangement. If the piping were connected to equipment such that there is no gravity forces or moments on the equipment nozzle in the 'cold' condition (i.e. no piping weight effects), then the equipment nozzle would actually see T+P only.


A couple of further comments...

-I suspect that, in reality, equipment nozzles would see somewhere between G+T+P and T+P since the exact approach to installation cannot be gauranteed.

-The difference between G+T+P and T+P seems to be more important for larger diameter piping.


I am interested in feedback on the above. Does anyone have comments?

 

[joshikv]

Well, I have not fully understood what you want to know. But let me give my opinion on the two points you have raised:

1) I dont know why you are saying like this. The piping GA drawing gives the arrangement and the layout is fabricated at site as per that. In operating condition the loads acting on the nozzle are sustained (dead weights + Pressure) and thermal loads. So how can it be in between G+T+P and T+P ? It will be G+T+P in operating.

2)Yes, it is correct. If you see the difference between G+T+P and T+P is nothing but the Gravity loads, which is more for larger diameter pipes / nozzles. But for any nozzles effect of bending moments is more, so they are the more critical factor to look for.

I also want to advice you to read about the primary stresses and secondary stresses. This is given in ASME Sect VIII Div 1 under stress classification. Because when we do analysis for stresses at nozzle junction, both categories of stresses are calculated and then summarised. First criteria is to design it for primary loads, then secondary and then for the combination.

Regards,
Kaustubh

 

[johnp]

See thread378-27395.
It is usual to calculate all three, G T & P, separately, as well as any sidesway effects from wind, or ground tremors if applicable. Then select the Design Code, and combine the forces and moments in accordance with the Code requirements.

I think what you are saying is that you can install the piping in a cold condition so that it will apply no forces to the rotary equipment nozzles. Maybe you can. But I have found that the self-weight reaction of the piping is a minor component of the load. The stresses really start building up when your equipment heats up, and the pipes start expanding.

The Codes are also very specific about "cold springing", and how much stress reduction you can allow by connecting the equipment cold.

 

[StressGuy]

I must confess, I am quite confused by the answers to the question. Stress is not the issue. Nozzle loads are for rotating equipment are set out in various standards (API-610, NEMA SM-23, etc) or specified by the vendor.

The truth is, the loads need to be within the vendor allowables whenever the shaft alignment is checked. That can be cold and hot.

The cases to check then, are W (pipe empty), W+F (springs at cold load), and W+F+T+P (springs at hot load).

You're never going to eliminate all deadweight. If nothing else, adding fluid, and pulling the travel stops on the springs are going to add load that is not related to temperature.
Edward L. Klein
Pipe Stress Engineer
Houston, Texas

All opinions expressed here are my own and not my company's.

 

[KernOily]

I have seen situations where W (deadweight) was nearly zero in both the cold and hot conditions due to the way the piping was supported near the equipment flange. So I just took account for that and did my analysis accordingly. A competent analyst would not just charge ahead and use W+T+P as his compliance hurdle, although I have seen that done.

An interesting case is if your applied load exceeds the manufacturer's allowable in a non-operating load case, for example, during hydrotest, assuming you did your hydro through the pump and had no blinds installed. Do you suppose the manufacturer will issue you new allowables for that particular case? I've never had it happen. Can't prove it, but I think that most allowables are developed for the shaft-deflection and alignment case rather than exceeding the allowable stress of the flange casting.

The bottom line is - look at your physical situation and make sure that your stress modeling (computer or manual) is as representative of physical reality as you can make it. This, plus your application of experience, judgment, and horse sense, and your firm's internal checking/QC process, will tell you which load combinations need to be considered when evaluating the loading conditions at the equipment flanges. Thanks!
Pete

 

[johnp]

Warning - we are talking about piping loads on rotating equipment in a very abstract way.

In the real world usually rotating equipment comes in pairs, or more. So do your vessels. Once you connect these up and calculate the reactions the whole picture can become a lot more challenging. It's not as easy as it looks to achieve the limiting nozzle loads stated. Then one starts asking questions such as "Is this for real?", "Have I dropped a couple of noughts somewhere?", and "This pump arrangement has worked on the Plant for years. Why does it not fit in with theoretical calculations?"

 

[KernOily]

"This pump arrangement has worked on the Plant for years. Why does it not fit in with theoretical calculations?"

This is an excellent point and probably extends beyond the scope of this discussion. A prime example is an air-cooled heat exchanger ("fin-fan&quot. If you do the usual analysis on the inlet (hot) side of a fin-fan that has several nozzles in the header box, which most of them do, and the nozzles are piped up such that they all come from a common header pipe, you will find that the nozzle loads exceed the manufacturer's allowable, every time. However I can't tell you the number of fin-fans currently operating that are hooked up this way, and I've not seen a single failure yet. I think that speaks to several things: (1) conservative allowables (2) inability of the analyst to account for the inherent flexibility of the header box and the nozzles (3) inability of the beam-column method used by CAESAR et al. to accurately account for the distribution of loads and deflections in the piping system.

As far as multiple/parallel equipment installations goes, the analysis algorithm is the same as for a single item of equipment, but it does take more effort, that's for sure. And it definitely is not as easy as it looks to acheive the alllowables. In my experience that's usually where the flexibility budget gets busted: trying to make the pump/vessel flanges work.

This is a great discussion, no? It's one of my favorite topics (have you figured that out yet? ;-) ) and always engenders a lot of lively discussion amongst flexibility analysts. Thanks!
Pete

 

[StressGuy]

"This pump arrangement has worked on the Plant for years. Why does it not fit in with theoretical calculations?"

Yes, well you, as the stress engineer have to take these kinds of statements with a grain of salt. One operating company may think that getting a three month run out of a set of bearings is "working."

In most cases where I've reviewed existing pumps and found the arrangment to be far in excess of allowables, it turns out that the maintenance people have to fess up that they do have a lot of maintenance on the pump.

While I do my best to educate plant engineers when I can, I find that the piping system is not something that maintenance engineers think about checking when their pumps are breaking down all the time.

74Elsinore: we're getting off topic, but I'm curious to know what the usual method of modeling a finfan is that results in overloading the nozzles? If it's one header supplying one header box, I don't usually have a problem. Now, if you meant a common header supplying a bank of air coolers I can see where those are a problem since the outer coolers have to slide with the growth of the header and the force to overcome friction on the cooler does get the loads up pretty quickly. Edward L. Klein
Pipe Stress Engineer
Houston, Texas

All opinions expressed here are my own and not my company's.

 

[KernOily]

Edward - The situation to which I was referring consists of a header pipe serving a cooler inlet such that the header pipe outlets are an o-let or reducing tee and then the nozzle's companion flange, with no straight pipe in between. This provides no flexibility due to the very short length, the SIF, and relatively high stiffness of an o-let or red tee plus flange. There are gobs of those in operation out here just like that (Calif. oilfields) and I've not seen one fail. Yet. My point was that the S=Ee methods we use are inherently conservative. An FEM or rigorous elasticity analysis would likely show that the hookup is OK.

I was just involved on a forensic project where I was asked to look at a piping system for two 150hp vertical multistage pumps operating at 160° F. They were experiencing very short MTBF on the shaft seal and the shaft bearings. I did the analysis and it turned out the flanges were loaded substantially over the mfg. allowables. The plant engineer was astute enough to realize that pipe strain can cause this type of failure, so they called me in to take a look. I think as plants mature and as staffs become more sophisticated and embody the principles of TPM and life-cycle cost, the old methods of use-a-comealong-to-spring-the-pipe-in-place are finally going away. Thank goodness. Thanks!
Pete

 

[pstress]

Thanks for all your responses. I have found them interesting. My question about nozzle loading actually revolves around 'real world' concerns initially raised by 'johnp' above...

If I have a situation where the nearest support to the nozzle is, say, 5'- 10' away, the stress analysis program will report forces and moments acting on the nozzle for the 'gravity only' case. Depending on the piping arrangement, it is possible that these forces and moments may act in the opposite direction to those generated by the thermal expansion of the piping. This effect would make the loads on the nozzle for the 'G+T+P'case (Grav.+Temp.+Press.)lower than 'T+P', by itself. Now...suppose the installer 'inadvertently' connected the piping to the equipment such that were no loads on the nozzle in the 'cold' condition (i.e. 'stress-free connection'). (This could be done by adjusting or shimming supports or minor alterations to the piping). Under this sceario, the nozzle would actually see T+P and NOT G+T+P during operation .... This is why I am suggesting that it may be appropriate check if T+P is higher than than G+T+P.

Actually, my suspicion is that, in the real world, the nozzle would actually end somehere between G+T+P and T+P, depending on approach the installer used.

Also,... yes I am aware that the 'Gravity only' case must be checked for the equipment, as well.

I hope my explanation above is clear. Any response?

 

[KernOily]

"If I have a situation where the nearest support to the nozzle is, say, 5'- 10' away, the stress analysis program will report forces and moments acting on the nozzle for the 'gravity only' case. Depending on the piping arrangement, it is possible that these forces and moments may act in the opposite direction to those generated by the thermal expansion of the piping. This effect would make the loads on the nozzle for the 'G+T+P'case (Grav.+Temp.+Press.)lower than 'T+P', by itself."

Exactly right. You, as the competent analyst, will recognize that this is happening and account for it it your analysis.

"Now...suppose the installer 'inadvertently' connected the piping to the equipment such that were no loads on the nozzle in the 'cold' condition (i.e. 'stress-free connection'). (This could be done by adjusting or shimming supports or minor alterations to the piping). Under this sceario, the nozzle would actually see T+P and NOT G+T+P during operation .... This is why I am suggesting that it may be appropriate check if T+P is higher than than G+T+P."

As the analyst you can only account for "the normal foreseeable operating condition" unless you have a very good reason to do otherwise. This condition you report would not be a normal condition because a competent millwright/mechanic will not install the pump in this manner, thus you would not try to account for it during your analysis. THis assumes a regular design process, i.e. pre-construction. A forensic analysis is something entirely different.

This reminds me of when I took Failure Analysis in college. The prof told us the story of when he had been called in on a lawsuit where a small child had climbed up and stood on an oven door that had been lowered to the horizontal position. Well the door collapsed and injured the child so the parents sued the oven manufacturer and everybody else. The jury found that a kid standing on the door was in fact a foreseeable design condition and that the manufacturer should have anticipated that happening and should have designed the door thusly. So they lost their collective rear ends on that one.

The point is that you can only design for what you think is the usual loading conditions that the pump will see: normal operating, startup, shutdown, hydrotest, maintenance events, seismic events, support settlement, steam-out, etc. It's not possible or reasonable to anticipate eveything else. The manufacturer's published allowables are what they want you to hold to when the pump is operating to prevent things like excessive shaft deflection. Not sure I answered your question...

Thanks!
Pete