Can you help me in finding any written references or documentation in regards with limitatons and adverse effects, if any, of making multiple repairs by grounding or arc gouging deffects or unacceptable weldments and re-welding in the same area?
Construction codes (ASME, API, AWS, etc.) say nothing about this but some people have the opinion that there is a limit to multiple repairs in the same area, to three times or so, but no one can tell me a reason for that.
I would like to read any investigation or documented written material about the effects this practice could cause in the final welding, specially in carbon, low alloy steels and stainless steel. I will appreciate very much your help.
That is the million dollar question. I also would like to know the answer even though I doubt I will like the answer.
Sir,
I would like to address the stainless steel portion of your inquiry, if I may.
In stainless steel, and primarily depending on the family of the stainless, the number of repairs should be kept to a minimum. I, too, am not aware of any Code that specifically limits the number of repairs before the part is scrapped, but I could be wrong. I think that should be decided beforehand by the Engineering Dept. of the fabricator. For example, a 304 material would withstand more repairs than a 2205 Duplex. A lot of the answer would be determined by the grade, and the extent of the repair. The drawback of many repairs to stainless is the number of thermal cycles it is exposed to. Different families of stainless react to numerous thermal cycles differently. Depending on the extent of the repair, the stainless can experience excessive grain growth, sensitization, and in some cases, sigma. None of these are good. Numerous repairs in the same area can accelerate these problems if not careful. Generally, carbon arc gouging of the area of the S.S. is not recommended due to the possibility of carbon impregnation in the base metal. When speaking of repairs to S.S., one needs to consider if this a new fabrication or a piece that has been service for an extended time. Too many circumstances need to be addressed before determining just exactly how many times it can be repaired without degrading the part. Generally speaking, the number of repairs to a S.S. should be kept to a minimum and careful procedures should be observed.
Chuck
Good answer Chuck. I have also never seen the specific number of permitted repairs in Code but have in numerous specifications. I'll go back to what a good friend of mine said (who is infinitely more qualified than myself), that is, on many of the carbon steels, numerous heat applications tend to soften the area of the metal being repaired (how's that for not getting overly technical!!!!).
That's one of the many engineering questions that are answered with a "figure pulled off the air".
In fact, no Code or standard states how many repairs can be done on the same area. Mr. Meadows is right in saying that they should kept to a minimum, and we all agree. But what does "minimum" mean? One time, two, three? That's the question!
Jon pointed out a good clue: the maximum number is stated in many piping specifications. If the spec has been issued by a company that enjoys a good technical reputation, then go ahead with it.
Just for your information: back in my days of erector engineer, I took part in the construction of a vacuum oil refining plant for a first class oil company. That company (our client), allowed a maximum of three repairs to be done on carbon and low alloy steel weldings. Why three? Because they said that it was their experience and we (the contractor) never discussed that decision.
Giovanni S. Crisi
Sao Paulo - Brazil
In my other life when I was the W.E. for F.M.C., Inc. all of our specs. limited any repair to 3 times. That was for carbon steel, low alloys, and stainless. This was based upon what we considered good metallurgical and welding practice. "Minimum" can mean one number to someone, and another number to someone else. That was why I suggested that the Engineering Dept. set the number before welding, or fabrication, ever starts.
CM
Those are some good answers, and like everybody else I've never seen a hard number in any code. But in cross country pipeline work you will find that many times you only get one or two shots at a repair, not due to any code requirement, but based on economics. There comes a point where it's cheaper to make a cutout and weld in a pup, pure dollars and cents. Most companies know where that point is from years/decades of experience. So my take is that a lot of these requirements are based on how many repair attempts can be considered efficient, and at what point are you statistically throwing away time and money.
JTMcC.
What would the engineering department use to decide the number of times it could be repaired. Was there any evidence based on base metal, weld metal, process, defect type, service etc.. ?
Have there ever been any technical studies done for fitness for service of welds subjected to multipple repair cycles ? That would be some interesting reading.
Thanks
Gerald
Hi Gerald,
In my opinion, each repair should be decided prior to any fabrication. To just say that 3 weld repairs is acceptable is not always the best approach. At FMC, Inc. it was the Engineering Dept. that set the number and it was incorporated in the QC Manual. We also incorporated the applicable WRP (Weld Repair Procedure) in each weld package along with the PQR and WPS for each grade of steel. Naturally, each WRP differed for each grade of steel, (stainless steel, low alloy steel, carbon steel, etc.) We would use the same repair procedure and allowable number of repairs for new construction the same as for old weldments, which I never agreed with. Like you mentioned, grade of steel, weld process, extent of repair, the service the weldment was exposed to, and the type of repair should all be considered independently.
In our metallurgical lab there was a number of analysis done on multiple repaired weldments. We found that Duplex 2205 would not stand the multiple repairs that a 304/308 could stand, due to the possibility, or probability, of sigma formation. There were many studies done, but the findings were not published, but kept "in-house" for the company welding dept. In spite of the findings, the number of allowable repairs remained at 3. Go figure....
Chuck
3 sounds about right although I have worked several stainless steel jobs(both 304L & 316L) and never had I seen a weld repaired more then once thru radiograhic examination. I have noticed though that the repair areas certainly did not weld the "same" as a new weld so I would guess that there is some carbide percipitation going and multiple repairs of the same area would not be a good thing.
Even when a weld was cut-out and the same fitting re-used....the weld puddle seemed resistant and it was required to let the metal cool down (alot more than a new joint) to produce a sound weld.
But you know, when an examiner determines a weld (to be good), he/she simply signs off: "Acceptable" and you never hear about it again.
One sometimes wonders though............
Some very good replies to JRESTREPOB's post.
When I worked for ARAMCO, their standard specifications limited repairs to 2. A 3rd repair required approval from an ARAMCO rep. on the project in the position to do so. A brief case study is performed before the 3rd attempt to repair. If this failed, then as noted by JTMC a pup piece is typically used to replace the repair area.
For structual applications, it depends where the repair occurs. Many times I have seen the RDP permit 1 repair only. This is determined by where the repair occurs in the structure. If it is part of the SFRS, then typically repairs are kept to a minimum (1 or 2). I have also seen where the RDP did not permit a repair and required the member or section of the member to be replaced, again as noted by JTMC.
For aerospace applications, it is not uncommon to have repairs prohibited. This typically applies to Class A and sometimes class B welds. When repairs are made, D17.1 denotes: "Repaired parts shall not be considered to fully meet engineering drawing requirements."
It really depends on the application.
Gentlemen,
yesterday I've gone through the "Requirements for pre-fabrication and erection of piping" and the "Piping general specifications" issued by M.W.Kellog Company and The Lummus Company.
I don't know if they are still in business (things have changed so much in the last times), but years ago they were the more prestigious project engineering firms in the world in the petroleum and petrochemical field.
Well, I couldn't find anything speaking of how many repairs are allowed in the same weld.
Taking into account all that has been said here by a number of people who have a long and knowledgable experience, I've made up my mind as follows:
1. No Code states clearly how many times a weld can be repaired. However, as Jon 20013 says, it's written in Specifications issued by several companies. As G.S. Crisi (me) suggests, if those companies enjoy a good technical reputation, their specs can be followed with no problem.
Question to Jon: would you let me know the names of a couple of those companies? I'm just curious.
2. Chuck Meadows, who worked for FMC, says that FMC allowed three repairs either in carbon, low alloy and stainless steel. G.S. Crisi (me) says that when he worked as a contractor for Repsol YPF, the oil company he mentioned, the client allowed three repairs on carbon and alloy steel. Stainless steel was not mentioned because there was no stainless steel in that job.
Question to Chuck: Rather that a single company, FMC is (ar at least was) a large holding that owned several other companies. One of these was Link Belt, who were the best specialists in the world in solids handling. Do you know if FMC and Link Belt are still in business?
3. DGXL says that when we worked for Aramco only two repairs were allowed. A third one needed Aramco's approval, which was not granted easily.
4. Taking into account that both FMC and Repsol YPF are companies that enjoy a good technical reputation, I would say that three repairs is a safe number that can be adopted.
As per Aramco, I'd say that they are more stringent than FMC and Repsol, perhaps unnecessarily.
5. Now, this is a general guidance for process and power plants that doens't apply to all cases. As JTMcC says, they do different in cross country pipelines because of economic reasons. Also, as DGXL points out, in aerospace applications repairs are not allowed and we can easily understand why: aerospace is possibly the most stringent industry that exists.
Also in process and power plants special cases may arise that need to be treated in a different manner.
6. Question to QCCWI: did I answer the question? If so, I'll e-mail you my bank account number so you can make a deposit of one million dollars.
Giovanni S. Crisi
Sao Paulo - Brazil
Professor Crisi; Most notable of the companies I have seen with specifications limiting the number of repairs was with Bechtel Construction company. Morrison-Knudsen (now Washington Group International) as well as Stone & Webster (now part of the Shaw Group)had similar. Additionally, I have worked with numerous electric utilities which have adopted similar limitations. P/S: I also worked for FMC in a past life. FMC remains in business although I believe many of their holdings have been dispersed.
Thank you Jon. I happen to know well all the companies you mentioned and all of them enjoy a high reputation worlwide.
We can safely follow what they say in their specifications.
Giovanni S. Crisi
"Have there ever been any technical studies done for fitness for service of welds subjected to multipple repair cycles ? That would be some interesting reading."
Yes, there have been, but I don't have the reference. Contact Krishna Verma at the FHWA Office of Bridge Technology (krishna.verma@fhwa.dot.gov) and ask him about a study on fatigue performance of welds in carbon steel after multiple repair cycles. I remember nothing about who did the study, but it was Krishna who showed me a copy of it.
My recollection is that they went up to about 7 with no deterioration of fatigue performance.
That said, everyone senior to me where I work remembers there being a 3-repair limit in the past, although it's not in our specs now.
Hg
I don't have any technical data anymore but when I was a pup, the general thought was that three welds total was the limit before serious concerns were raised about potential Hydrogen Embrittlement. I don't know if this is accepted thought any more but it seems that the same rules of thumb seem to be held but without the metallurgical reasoning behind it. IMHO.
Bruce
Why would hydrogen embrittlement be an increasing problem with multiple repair cycles if the same practice is followed each time?
Hg
Hydrogen embrittlement would not be a concern in the stainless steels. I, too, was not aware of any hydrogen embrittlement when doing repairs. Usualy the areas of repair are smaller and less heat affected than the original weld. Also, usually the areas of repair are where the original weld was, and any hydrogen related problems were addressed during the initial welding process. I'm not disagreeing, I just never heard of this only on repairs.
Chuck
It is my understanding that hydrogen embrittlement is not a permanent condition. If cracking does not occur and the environmental conditions are changed so that no hydrogen is generated on the surface of the metal, the hydrogen can rediffuse from the steel, so that ductility is restored.
I would expect the project engineer would have some say in the number of repair cycles permitted, if there is a need for any such limitations.
The number one consideration would have to be "what base metal are you welding?" Plain carbon steel should not raise a major issue. Stainless steel could be a problem with repeat thermal cycles and sensitization. Heat treatable aluminum alloys could see major reductions in tensile strength, etc.
Most issues could be minimized if the repair includes removal of the HAZ if there is a concern with grain coarsening, etc. That in turn could cause additional concerns with distortion due to the increased volume of weld deposit and the increased thermal history that would be experienced.
I don't believe the magic number should be set at 3 unless there is sound engineering reasons for the limitations. In some cases, three attempts may be too many.
The engineer may want to be consulted after the first repair was found to be unacceptable. What is causing the problem? Changes in the procedure may be justified if the problems are not related to welder technique.
Best regards - Al
Al says-"I don't believe the magic number should be set at three unless there is a sound engineering reason".
Al there are more things to consider than "engineering reasons", primarily economics. The engineers may say you can set up shop and repair that weld until the year 2020. But long before then the contractor will be out of business. There comes a time when you have to do something different, IF you plan to make money on the job. Most companies have a real good handle on when that time is.
Looking at the number of repairs allowed on a defective weld goes beyond grain coursening, HAZ and reductions in tensile strength. Someone, somewhere, needs to make a profit on the job and somewhere the owner needs the job finished.
JTMcC.
There are some pretty technical answers here. The original question of "some people have the opinion that there is a limit to multiple repairs in the same area" may not be. API 1104 19th sec. 10.1.2 states "A qualified repair procedure is required to be used whenever a repair is made to a weld using a process different from the original weld or when repairs are made in a previously repaired area."
Essential variables are to include:
Method of exploration of defect.
Method of defect removal.
Examination to confirm removal of defect.
preheat and interpass treatments.
welding process.
requirements for interpass NDT.
It's possible that management may conclude that removal is cheaper than developing a procedure at that point. And that message may be mistaken to mean only one repair is allowed for quality reasons.
Gentlemen,
This discussion took place in this Forum as far back as 2005, i.e., six years ago.
In it, pipewelder (Gerald) asked whether a research had been carried out to establish the maximum number of times that a weld can be repeated in the same location. His actual words weren't exactly those but the meaning of the phrase was.
The only answer came from HgTX, who said that once, a gentleman by the name of Krishna Verna, who worked for the Federal Highway Administration of the USA, showed him (HgTX) a study on the fatigue behavior of carbon steel welds after multiple repairs cycles. HgTX didn't remember neither the title nor the author of the study, so he suggested to whoever was interested in having more information to contact directly Mr. Verna, whose electronic address followed.
Shortly after HgTX's posting I sent an e-mail to that address, but it came back with a legend saying that the address didn't exist and so my e-mail could not be delivered. I tryed to contact Mr. Verna through the FHWA, but the answer was that he no longer worked in it.
So my question is the following:
Do anyone of you, Gentlemen, know whether an actual research has been carried out on how many times a weld can be redone in the same location, since this original thread first appeared in this Forum?
I'm curious to know. Thanks to all who will reply.
Giovanni S. Crisi
Sao Paulo - Brazil
Thank you, Henry, your information has been very useful, as usual.
Incidentally, the third citation (sciencedirect) is an article authored by two professors at the Sao Paulo State University, Campus of Guaratingueta (difficult to pronounce, isn't it?). Guaratingueta is a city of about 100,000 people at 100 miles from Sao Paulo. It's an Indian word that means "the wolf's refuge".
I've been in contact by e-mail with the first author and have exchanged ideas with him.
There are plenty of Indian words in Brazilian geography, as there are in American one. Massachusetts and Tallahassee are just a couple of examples.
Giovanni S. Crisi
Ladies and Gentlemen,
I have five patents in welding, graduate engineer and CWI since 1993. A good friend messed with and ended up bolting. You might over look code reference, at the discretion of the engineer. The engineer is supposed to have a view that the metal etc might not allow for a welded repair! I hope this is helpful.
Sincerely
Len Andersen weld@spemail.org
212-839-6599 8-4 New York Time , 4042 FAX , Co-worker 6381 / 914-237-7689 (H) / 914-536-7101 (Mobil)
POB 1529 / NYC 10116-1529 ( $1090 per year Caller Box GPO NYC )
www.lenandersen.com
http://www.steel.org/~/media/Files/Autosteel/Research/Safety/ahss_repairability_phase1_study2.ashx
