Dear forum,
I am looking at 32” HP steam line in 15Mo3 operating for the last 20years. The line is operating at 435oC. Beginning of the last year, the line went through a hammer (water / steam no one knows) and dropped of the supports. Temporary rigid supports have been provided since the incident. Full revamp of supports is planned for shutdown in September.

We have also bought some piping and elbows if we find damage on the existing pipe.

In the short time of the shutdown, we are about to do limited no of replicas at the elbows that we think are most likely to be highly stressed and confirm if they need replacement.

The tricky part is to predict the locations of the highest stress points at each elbow. We have done FEA with arbitrary loadings in typical directions and the internal pressure load, to discover that the high stresses were most likely to occur at the inside surface of the elbow.

On the other hand, I learned that most of the creep failure happen on the elbow extrados, but our material is not operating in the creep range.

I would appreciate opinion of the experts, with experience in this field, please throw some light.
Regards
Ranka

An old Tube Turns Process Pipe manual talked about fatigue failures in welded elbows. That report showed consistant failure points at the inside of the bend, right where the pipe and elbow were welded.

Erosion metal reduction is greatest (remaining metal thickness the thinnest) at the outside arc of the elbow where the steam flow impinges the metal. Shouldn't be too much erosin on the inside of the bend.

A single water hammer event isn't really like a long-term fatigue failure, but I'd recommend UT & PT (for thickness and weld cracks) at those two points: if erosion over the years has occurred you will find it - even if there isn't a water-hammer crack yet. If the weld was tending to fatigue fail, then you'll find that at the inner weld area when you look for the larger crack from the water hammer.

Try those for a starting point to the discussion.

Hello Ranka,

It is an interesting topic. Some random thoughts occur. As has already been implied, you will have to examine the long term deterioration (creep, erosion, corrosion, et. al.) separately from the short term damage (hammer event). Presumably, a thorough "walk-down" visual examination and damage inventory has been done.

The hammer event was most likely a steam hammer (did you have a load rejection turbine trip?) or perhaps a slug flow event (in which case you would want to look into water damage to the turbine). It is very improbable that the pipe was filled with water so there was NOT a water hammer. Check you pipe pitch (slope) and check your drip legs and traps. I would disagree with your statement that you are not in the creep range of this material.

For background information ("as an exercise"), have a look at the B31.3 (yes I know it is a B31.1 application) minimum thickness for bends equation in B31.3, 304.2.1. For pressure loadings the intrados is the most highly stressed portion of the elbow. It is interesting to note that if you check (by calculation) many B16.9 elbows you will find that they are too thin according to Code minimum wall thicknesses but they are excluded from these equations because they are B16.9 standard elbows (104.2.2). An interesting bit of insight when you look at the calculated results of an FEA analysis.

Curved pipe (bends) is generally thinner on the extrados due to the bending process. Most pipe mills here work to the buyer's minimum thickness specification for power plants, so to assure they make or better the "min wall" at the extrados they start with thicker pipe. You will find other reasons for vagaries in the overall pipe thickness (mill tolerance, et. al.). Unless you already have base-line thickness (as built at the time the pipe was new) for your piping, you will not be able to accurately determine how much material has been lost to thinning processes as a function of time. However, Mr. Cook's suggestions are certainly very sound advice.

Most of the highest (triaxial) bending stresses will be found at the extrados of the bends and those are the stresses that will be of interest in your creep assessment. Of course the bends will also have seen some permanent plastic deformation due to the initial start-up loadings and "relaxation" (shake-down). Because pipe is not perfect either in geometry (round and hollow) or in metallurgy (there are "hard spots" and "soft spots" as the ductility will vary around the bend - not withstanding the "work" (strain) hardening at the "relaxed" areas), it will indeed be "tricky" to predict the places of actual highest stress. While it might provide some insight, your FEA model will not be of much value here since it is a model of a perfectly round (cross section), even thickness wall, perfectly curved and homogeneous structure, not a typical curved steel pipe in a power plant.

When your piping system experienced the hammer event, it will have had relatively large displacements (as confirmed by the disengagement of the support(s). You may get some clues regarding the event pipe movement by looking for damage to the insulation and damage to the pipe hangers. If you have not already done so, it may be enlightening to model the event with Caesar II (either applying the assumed displacements or applying the calculated hammer forces at changes in direction) and determine the locations of the largest hammer event bending stresses. Then you would want to apply the scheme of NDE procedures suggested by Mr. Cook at these locations. I might suggest that much of the plastic deformation stress due to the hammer event will have had a chance to partially relax by September. You can help the process complete the beneficial relaxation by bringing the steam system down slowly for the September outage.

Good luck with your project.

Regards, John.

Dear John and Robert, thank you very much for your valuable advice and comments.

I just want to add a few more notes:

Details about the event and possible causes were never published by the plant...

The hammer occurred year and a half ago and the line has been having tough times since. The anchor was badly damaged and line was experiencing low frequency vibrations for duration of +/-8 months.

On the temporary RIGID supports, the line was already once shut down and started up, apparently “slowly”.

Now, in operation, only few of the supports actually carry the weight. I believe that the primary stress levels are very high. But no one knows how high…

Apart from the replicas, we are planning to remove all insulation, to do visuals, do the surface crack testing and wall thickness checks. Level surveys have already been done.

Original plant was built to DIN standards. They do fail ASME checks, as John points out, on intrados. The client insists on ASME codes for all mods.

John, you disagree that the material is operating in the creep range. Please elaborate. I have assumed that equivalent material to 15Mo3 is A335 P1 and they are in the creep range only at temperatures over 468oC…

We do know which elbows might have experienced the highest stress, but the “tricky” part is the exact location on each elbows. At the moment we are going for three sections 0o, 45o and 90o. Each section will have replicas taken at 12o’clock (extrados), 5, 6 and 7o’cock (intrados).

I am a bit nervous about taking only one point on extrados, especially if John is right in saying that the material is in the creep range.

Thanks again for you comments.

Regards
Ranka

Hi Ranka,

To my knowledge there is not a microstructural equivalent ASME material to 15Mo3. In the past A204 Gr A was used as a rough equivalent after a chat with the metallurgical engineer.

For interest sake, you mention that the client wants mods done to ASME. Did he by any chance specify B31.3?

Guys,

Some additional info that may be useful.

DIN 17175-15Mo3 is identical to EN 10216-16Mo3 (material no. for both is 1.5415). Both the DIN and EN quote creep rupture properties starting at 450°C.

Yes, Berndt, it is to ASME B31.3.
Stan, thanks for the information.

Regards
Ranka