use of expansion joints 1

[steve1]

Are there any established guidelines or rules of thumb for when an expansion joint should be used in a piping system? It seeems to me that more often than not expansion joints are requested when they are not really needed. I have also heard that expansion joints are required to be maintained. What is involved with that and how costly is it to maintain them? My questions are intended to give me some ammuntion when dealing with clients who think that expansion joints solve all manner of problems. Thanks in advance for your assistance.

 

[KeithMB]

I work as mech eng on a oil refinery so quite often get involved with piping expansion joints.

As a general rule we consider expansion joint as a last resort as due to there construction they are a weak point in the piping system (they are thin and vulnerable) other disadvantages are cost and lead time.

Generally the requirement for expansion joints arises when the proposed piping system is modelled using stess analysis software and this indicates piping stresses are greater than code allowable (our relavent code is ANSI B31.3). We would normally attempt to reduce stresses (to within code allowable) by a number of means prior to considering expansion joints, typical solutions may include: change route of line to increase flexibility, install expansion loops, change pipe supports/anchors, consider springs supports - constant or variable.

If we cannot solve with above "tricks" and do install bellows, we create an inspection schedule in our maintenance management system to inspect them. Inspection methods generally consist of visual examination, MPI crack detection, radiography. For very critical services you can buy (I never have) multi ply joints which have an alarm system that detect increase in pressure between the ply's which indicates failure. We have an in house inspection department who generally undertake inspection so it does not cost us a lot but I believe some of the expansion joint manufacturers can provide this service. Typically we inspect them every 6 years but this depends on service and consequence of failure.

 

[MJCronin]

Steve1,

Expansion joints are frequently specified and used where there is no good justification for them. You have to be specific about the types.....vulcanized rubber or metal bellows.

Keith (above) gives good information about EJ usage from a stess engineers point of view. Metal bellows expansion joints are an analysts last resort and are frequently used when there is no room to place expansion loops in the piping system.

Equipment subject to vibration is another good application. Examples include reciprocating or positive displacement devices. Expansion joints should NEVER be used to make up for piping misfabrication......this is just a failure waiting to happen.

Most new balanced centrifugal pumps and compressors do not require expansion joints. I am also aware that the much lauded "AIA MASTER Specifications" require expansion joints at virtually evry piece of equipment.

The "Piping Handbook" by Nayyar contains a chapter on expansion joints ( written by Charles Becht) Charles Becht has written an Appendix to B31.1 about EJ application and usage. You might want to look at it....

Becht also has published an ASME book about B31.1 Piping with a lot of good EJ information....

Good Luck

MJC

 

[petern1]

I have to agree with MJCronin that expansion joints are often used where they are not required. They require a sigificant amount of maintenance and are expensive. I many cases, not used properly they are worse than not having them.

In my refinery days we used them in fluid solids (i.e. coker) service where there was high temperatures, lots of expansion and you needed straight lines.

In slurry pipelines we use them in long (i.e. 1000's of feet) runs to take up expansion and allow maintenance (i.e. shut down a line at 90 deg C, flush and unbutton it. let it cool to -40 deg c then try to re-connect). Slurry lines tend to be afraid of expansion loops.

For most rotating equipment (pumps) properly designed piping is beter than expansion joints in most cases.

 

[davefitz]

If the piping analysis included a fatigue analysis that is based on the expected vibration response of the pipig system and connected machinery, and that analysis shows adequate fatigue life of the piping and the nozzle at the machinery, then you don't need the flex connectors.

Problem is , very few people conduct such an analysis. To not provide a flex connection on a nozzle at a piece of machinery that can take down the whole process if the nozzle or machinery fails sounds like a risky decision.

 

[Scipio]

Another place I've seen people wanting to put expansion joints (metallic bellows) a lot is on my rotating equipment, primarily high temperature pumps. When stress runs indicate that high thermal expansion rates exceed piping limits, or allowable nozzle loads on the equipment itself, I've seen expansion joints pop up to solve the problem of movement due to thermal expansion between the pump & the pipe. While not really a problem, it should only be a last resort, as everybody else has stated. A couple additional reasons, bellows style expansion joints near the suction of a pump can cause added pressure losses because of the flow over the convolutions. There's also an axial thrust load introduced because of the annular area between the top and bottom of the convolutions that can exceed nozzle load limits on the pump - a lot of people think these things eliminate all loads, but they can actually cause an increase in loads because of this in certain situations.

Another big problem I have with expansion joints is in installation. Expansion joints, at least bellows style joints, are designed for a specific number of cycles over a fixed range of elongation. For example, a joint might be good for 100000 cycles of +/- 0.25", but only 1000 cycles of +/- 0.5". And axial vs. radial displacement makes a big difference too. I once saw an expansion joint designed for four inches axial displacement, and a quarter inch radial, installed in a system (to accomodate thermal growth of a 80' tower made of stainless steel), and be pulled radially an almost two inches to match piping to a vessel nozzle. Piping wasn't where it was supposed to be, vessel nozzle wasn't where it was supposed to be, but there was an expansion joint, so nobody worried about it, despite the fact the fitters had to use a chainfall to pull the joint into place. It was just a fluke I happened to walk past and notice it before commissioning.

There are places where they're unavoidable, but I've found far too many people say 'just drop in an expansion joint' and forget about it. If you've got one shown on a piping isometric, I'd recommend strongly indicating on the iso the actual limitations of the joint, and if possible even wiring a stainless tag with allowable deflection on the joint itself.

 

[davefitz]

scipio makes a good point, and these crude measures of fitting pipes is not limited to cases where there is an expansion/ flex fitting. These measures ( using a chainfall, or even more amusing, heating the pipe with a torch to force it to grow to fit the mating flange on a pump- quick! bolt it before it cools) are one of the causes of rapid failures of pumps.

 

[tothepoint]

The primary purpose of expansion joints is to isolate and prevent the transmission of vibration, NOT to reduce/eliminate piping stresses. It's prudent to place an expansion joint on the discharge side of a centrifugal pump- thus isolating vibration from the piping.
Like it or not, there are pipe sections that do not mate together well. Quality expansion joint manufacturers (Garlock is one), will fabricate the joint to "fit" the existing angular, torsional and lateral off-sets that exist between two piping sections. Don't take an "off the shelf" joint and try to make it fit between off-set piping sections. The stresses on this joint will cause it to fail prematurely.
As for maintenance of expansion joints, periodically inspect the joint for tears or undue stresses in the jacket. Visible wear/damage indicates that the joint requires replacement- and probably that the joint was not sized/constructed properly for the joint.

 

[Scipio]

tothepoint,

I don't know, if I had a centrifugal pump generating that much mechanical vibration that it was shaking the discharge piping, I think I'd look into properly balancing the pump & ballasting the baseplate before putting in an expansion joint. That wouldn't help correct an acoustic vibration, say pulsations induced by the vane passing frequency of the cutwater, but at the same time I can't see an expansion joint correcting that problem either.