Tube failure on new bundle

[MaterialsDude]

So I am trying to figure out why 2 tubes failed (via leaking) on a new tube bundle that was recently installed in the exchanger. Right now I suspect 3 possible root causes: 1) Pressure surge: Shortly before the leaks were discovered in a hydrotest, a PSV valve popped. 2)Velocity: These are admirality brass tubes and they have a design pressure of 7 fps. The operational data shows that shortly after being placed in service, they were exposed to ~12 fps velocity. 3) Thermal shock: Both tubes that failed were in the first pass (bottom pass) and on the outside section of the bundle (i.e. along the circumference of the bundle).

I have listed other root causes on my root cause tree as scaling,SCC, pitting, erosion corrosion, and mechanical vibration but ruled them out because the new bundle was only on service since February 14, 2010. The psv failed around March 3, 2010, and the leak was suspected on March 13, 2010 (confirmed the next day during a hydrotest).

The tube diameter is 1 in and has a 16 BWG (0.065 in) thickness.

 

[metengr]

MaterialsDude:
You can speculate until the cows come home - the root cause is determined by proper failure analysis. Get your hands on the failed tubes. Do not rule anything out, this is the first step of a proper root cause analysis.

 

[MaterialsDude]

Metengr, that's what I was trying to tell my boss because I believe that we have to take it out and send it for analysis but she told me that we won't do that. Obviously it makes my job a lot tougher especially considering that I am not that experienced with failure analysis.

 

[metengr]

MaterialsDude;
There comes a time in each person's technical career where you have to carefully manage the boss. What I would do is lay out a comprehensive root cause investigation strategy where you have several internal or, if necessary, external subject matter experts that review design, manufacture, operation and material analysis of the failed tube sections.
You can also ask your boss after you present your strategy, and she is still hesitant on not removing a sample, why? Most likely, it could be the following;
Unfamiliarity with a detailed root cause investigation
Cover-up of a mistake in design or testing
Technical incompetence

Good Luck.

 

[rmw]

First off, velocity isn't probably the culprit. While 7 fps is a good reasonable design velocity for brass, 12 fps won't hurt it in the short run. In the long run, it will cause accelerated tube wear.

However, that said, I would be concerned with the pressure that caused enough pressure differential across the bundle to produced the 12 fps velocity you mention.

On the other hand, you might just have a case of what is referred to in the business as "infant mortality." It is not terribly uncommon for a tube or two or three to give up shortly after going into service, even though they did pass the initial certifying hydrotest.

Since your boss babe won't allow you to send the tubes out for analysis, you might try running a bore scope down the failed tubes to see if you can make any determinations about the nature of the failure.

Do you know if your tubes were rolled full depth of the TS or not. Is there any possibility that they (or some of them) were rolled pass the end of the TS?

rmw

 

[metengr]

The post by rmw got me thinking and after reviewing your second post regarding your boss not wanting "to take it out" could be the bundle instead of the failed tubes.

You realize that the tube bundle does not need to be removed from the shell. The failed tubes can be locally extracted using a tube puller and the tubesheet plugged. I just wanted to be sure because removing an installed bundle is not necessary.

 

[Guest102023]

I concur with the above posts. Barring a factory defect (the factory hydrotest can be superficial), I would not expect the cause to be corrosion or erosion corrosion in such a short time.

The tube can indeed be removed without removing the bundle.

And Ms. Boss needs to appreciate the value of F/A in averting potential recurrence.


A tip for any future service failures: be careful to minimize damage when extracting tubes. I have done about a dozen F/A on such tubes - most were by corrosion or SCC, and most of those happened within a few inches of the tubesheet. Erosion corrosion is more likely at the tube inlet and U-bend.

 

[rustbuster]

MaterialsDude,

What is the consequence of failure? Than should drive the amount of effort for RCFA and failure analysis. If the consequence is low just plugging tubes should be acceptable.
If the consequence is high then not understanding cause is risky.

If your boss doesn't support the FA, she is accepting the risk. At best your RCFA can only arrive at un-validated hypothesis.

Eddy current and IRIS are also capable of assisting in determining cause.

Best of luck!

 

[MaterialsDude]

Hey guys,

Thank you all for your replies. I talked with the boss this morning and she said that she did ask if we could pull out the tubes but that others in the company don't want to for whatever reason. The latest info I got was that the PSV valve popped after the condensate was contaminated with the water inside the pipes. That and the pressure data we have does not really tell me that pressure had much to do with the failure.

We are now looking into the possibility of water hammer being the possible cause either from the high velocity or from the condensate.

As some people here mentioned, I feel that it could be just a factory defect that could've led to the failure. Rustbuster, there aren't severe consequences and we just plugged the tubes. I have attached a drawing showing the bundle layout and the locations where the 2 tubes that were plugged are located. The 2 plugged tubes are the highlighted ones in the bottom pass.

 

[EdStainless]

To start with we need to remember that at 7 fps this is not a long design life bundle. You will get significant inlet end erosion.
Plugging a couple of tubes has no impact on performance.

The bigger issues is what impact leaks have on the system and how ofter you are willing to put up with them.
The reason to do a FA is that the solutions will depend on the cause.
If these look like mfg defects then you do a ET and plug a few suspicious tubes and life is good.
If they look to be from pressure, flow, impact or some such thing then you have to address operations.

I am guessing that you guys had no QA oversight on the tubes or bundle.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[MikeMet]

I recommend you try to characterize the leak as much as possible without removing the tube (although I do believe extractiong for PFA is the best choice) At the very least you inspect the ID using flexible boriscope and if possible use the same flexible scope to examine the OD. You might find something, you might not, but at times it can be a valuable too for determine failure cause without removing a tube. You could also using eddy current or IRIS UT to find the leak and determine if the leaking area is localized or thinning over a given length of the tube. I don't think you can rule out vibration and leaks at baffles. I have seen this happen in a very short time. What on the shell is close to the failed tubes? Nozzle or other than may be a source of impingement on the tubes? Keep thinking.

 

[MaterialsDude]

Mike,

Good points, U know what's interesting is that I ws looking at the tube plugging pattern of all 6 heat exchangers we have, and noticed that the outside tubes are more often than not the ones that get plugged first or even the most. That was the trend for at least 4 of the 6 exchangers. I wonder if it's because the outside tubes are subjected to more vibration and steam than the inside ones!?

 

[Robertmet]

Short term failure ? I wish I could remember the exact time but we had a short term failure of copper or copper alloy tubing.After all the tubing was replaced we found by chance that the crew that installed the heat exchanger originally used it as a urinal -too lazy to walk a short distance to use a portable toilet !!! That was 40 years ago so I don't remember the details but we were surprized that the whole exchanger could fail so quickly.We never thought of the source !!

 

[Guest102023]

Robertmet, I'll bet these guys were making the usual dumb@ss cracks about engineers while doing it ...

 

[stanweld]

Robertmet,
Your story coincides with the cause of one of our short term failures of admiralty tubes, where some of the local fauna used the open tube storage box as their urinal.

 

[EdStainless]

You not only need to look at the tube sheet plug maps but you need to check and see where along the tube the leak is. Then look at the results in three dimensions.
Could just be poor inlet flow distribution.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube