This problem souns interesting. (Probably frustrating to you though.) Obviously, it is difficult to make any conclusive remarks about something that one has not seen yourself, but I have a couple of thoughts that may be of help.

The most common cause of in-service failures of Austenitic Stainless Steels (AS/S) is Stress Corrosion Cracking. (SCC) You therefore have to check for this as a first priority. SCC in AS/S can be identified by it's very characteristic branch-like transgranular cracking. Have you had these failures checked out by a metallurgist?

Obviously, for SCC to occur you need to have a chemical species present that will cause this. Typically this is Chlorides. Do you have any chlorides present in the medium? If so, your temperature is just great for getting SCC.

From where is the crack starting? The inside of the line, or the outside? Again, a metallurgist should be able to tell you this. If it is from outside, then you must find out if the line can be contaminated externally by chlorides or other SCC causing species. Is your plant located near the sea? If it is operating at 500°F, then it is probably lagged. Is your lagging of a low chloride content? Some lagging is not. Does the line's temperature rise and fall through the dew point, so that moisture forms under the lagging?

You seem to indicate that all the failures were at areas of previous repairs. Why were the repairs necessary in the first place? Was it due to hot cracking of the weld? If so, there could still have been some residual contamination in these areas. This is unlikely however. Much more likely is that these areas have the highest residual stresses. This means that these will be the first ares to crack in the case of SCC. If this is the case, then it is only a matter of time before the rest of the line also experiences this cracking.

Another issue is the fact that you perform R/T and U/T without finding anything, and then it cracks. If it is SCC, then it is very possible that radiography will not show up anything because the cracks are so small. U/T testing on lines are in general a problem, because the lines are generally too thin to get reliable results. How thick is your line? Another problem is that some AS/S materials have a very coarse grain, leading to such a large amount of "noise" when performing U/T that SCC cracking can easily be missed. Have you spoken to the U/T technician? Is he obtaining a noisy signal?

Another possibility is that you have gotten some sensitization in these repair areas. As you are using Ti stabilised AS/S, this is unlikely. Again, a metallurgist will be able to tell you is you have some sensitization in these welds. Cracking due to sensitization will also be intergranular, and should be picked up by radiography as it tends to be much coarser.

If you have some answers to the questions I raised, let us know. Maybe I can then try to give a less "fuzzy" answer.

Hope this helps

Niekie Jooste

Answering to this question makes me feel the Sherlock Holmes of welding.
The puzzling thing is this: why did the cracks appeared exactly where the primitive welds were found defective and repaired? Attention though! This is also a clue that may lead us to the reason of the problem.
You say that "....... local repair was made by complete removal of the weld at defected areas ........". How were the welds removed? You don't say, but I imagine it was by means of a grinding wheel. What kind of grinding wheel you used? You don't say, but chances are you used a commom grinding wheel made of silicon carbide (Carborundum is a usual commercial name). These are the most common and cheap wheels you can buy at any tool shop. If this is so, hurrah!!, we've found the answer to your problem. Silicon carbide (Carborundum) grinding wheels MUST NOT be used for stainless steel because they contain carbon, which will contaminate the stainless steel. And what happens to stainless steel contaminated with carbon? You're right, intergranular corrosion develops, leading to cracks. For stainless steel, you must use grinding wheels made of alumina (aluminum oxyde) which contain no carbon in their chemical composition.
If my supposal is correct, please tell me. What a marvelous Marmaduke Surfaceblow story it'll give on Power magazine!! Don't you gentlemen know who Marmaduke Surfaceblow is? Shame on you!! He's the best engineer humanity has seen in the XXth Century. We're in the XXIst, right, but he's still active. Get Power magazine (a McGraw Hill publication) and read Marmaduke's stories, they're delightful.
Giovanni S. Crisi
Sao Paulo - Brazil

THANK YOU ALL FOR YOUR REPLY. LET ME ANSWER YOU QUESTIONS

A.TO NIEKIE 2

THE FIRST FAILURE WAS CHECKED , THE CRACK WAS STEPWISE CRACK STARTED FROM THE TOE OF THE WELD.THE CRACK STARTED FROM INSIDE.
I DIDN'T HAVE A CHANCE TO DO ANY INVESTIGATION ON THE SUBSEQUENT FAILURES. BECAUSE OF DOWN TIME LIMITATIONS , EVERY TIME I HAVE TO REPAIR THE LEAKY WELD , SO SAMPLING WAS NOT POSSIBLE. HOWEVER THIS TIME I HAVE CUT THE SPOOL CONTAINING THE LEAKY WELD FOR DETAILED INVESTIGATION. YES WE HAVE CLORIDE IN THE SYSTEM.
REPAIRS WERE MADE DUE TO DIFFERENT REASONS SUCH AS LACK OF PENETRATION, SLAGS,POROSITY.....ETC.
I AGREE WITH YOU THAT FINE CRACKS CANNOT BE DETECTED BY R/T. ALSO U/T ON SS IS NOT THAT EASY.

B. TO G.S.CRISIS

WELDS WERE REMOVED BY GRINDING. THEY SHOULD HAVE USED WHEELS SUITABLE FOR STAINLESS STEEL . BUT FRANKLY SPEAKING I CANNOT CONFIRM THIS , AS YOU NEVER KNOW THAT THE WELDER OR THE FABRICATOR HAS USED THE CORROCT WHEELS FOR THE JOB , UNLESS YOU GO AND CHECK IT BY YOURSELF EVERY TIME WHICH IS NOT THE PRACTICE.
BY THE WAY WHAT IS "MARMADUKE SURFACEBLOW" NEVER HEARD ABOUT IT.SEND ME SOMETHING.

C. TO DPWEBER

WHAT ABOUT HIS DAUGHTER?