Not logged inAmerican Welding Society Forum
Forum AWS Website Help Search Login
Up Topic Welding Industry / Technical Discussions / Carbon Steel Piping Failure
- - By pipewelder_1999 (****) Date 02-11-2003 20:54 Edited 07-14-2014 13:01
I have question about failure of welds and surrounding pipe.


1) a weld failed in a brittle fracture mode and
2) the piping was above the minimum design metal temp and
3) the piping was verified to meet the original mechanical specification

Would the surrounding un welded pipe many feet from the original failure, show evidence of brittle failure or would it show some deformation prior to failure?

I dont know what failed first. I just thought it would be odd for piping to fail as shown in the pictures at just because of a weld failure.

I was just wondering if piping would fracture like this just due to another failure.

The pressure was around 200 PSI. The temp is unknown but the ambient was around 5 DegF. The pipe carried compressed inert gas.

For comparison, there are some very interesting pictures of compressed gas cylinders tested to failure at


Gerald Austin
Iuka, Mississippi
Parent - - By arcblue (*) Date 02-11-2003 22:16
What was the material composition? Can you calculate the Pcm or ceq values? What were the environmental factors like sour gas (hydrogen sulfide) exposure?? Or long time at elevated temps? Exposed to corrosives? The welds may have little to due with the pipe material failing in a brittle fashion if it was in an environment that was detrimental to the base material.
Parent - - By pipewelder_1999 (****) Date 02-11-2003 22:22
Material is A-53. The CE has not been checked yet. The service was inert gas pressureized to around 200 PSI. Mechanical/Chemical testing is in progress.

No information is available regarding the operating conditions at the time of failure.

Whats Pcm ?

G Austin
Parent - By arcblue (*) Date 02-11-2003 22:27
Short version : Pcm is another flavor of carbon eq.
Parent - By GRoberts (***) Date 02-11-2003 22:39
Carbon equivalent is supposed to be an indicator of hardenablity.

Pcm is supposed to be a measure of a steels tendency to crack. Pcm is calculated by: C + (Si/30) + [(Mn + Cu + Cr)/20] + (Ni/60) + (Mo/15) + (V/10) + (5B)
Parent - - By chall (***) Date 02-12-2003 09:37
If the piping was above the minimum design temperature, yet subject to excess stress above what it could sustain at that temperature (ie-impact), there is still the possibility of a brittle failure mode.

Not much, but the best I can muster at 0430.

Charles Hall

ps, Regarding leak testing, I tried to check the link you included but was only able to get to your home page. Nice picture of all those dogs. Which one is Will, Willie and Gerald?

Parent - By pipewelder_1999 (****) Date 02-12-2003 12:38
Wouldn't there be SOME area that would have deformed before failure?

G Austin

PS The link to the leak test pictures is fixed. I left a period at the end.
Parent - - By M-Squared (**) Date 02-12-2003 13:36
If I had to venture a guess, at this point I would say that the material dropped below its Ductile to Brittle Transition Temperature (DBTT). This is after looking at the photos on your web page. I have been trying to find the DBTT for A53 material, but no luck. You might suggest that during the metallurgical anaylsis that they test to see what the DBTT is for this material. For reference, “certain metals are tough at high temperatures but fragile at low temperatures. If a component is made from one of these metals is used at or below the temperature at which it becomes fragile, the component may fail unexpectedly in service.” (From ASM Handbook 8th edition). Save as many pieces as you can and try to look for the initiation site. I would start at the weld especial at the toe of the butt weld. See if there are any sharp notches that formed during welding (i.e. undercut). If you find nothing there work your way out. The other thing that makes me believe that the failure might be attributed to DBTT is the fact that it was carrying compressed inert gas. If I am not mistaken when a gas is compressed it undergoes a change in temperature. (in this case a drop in temperature). In other words the material being carried by the pipe was cold and then we add the ambient temperature.

I had a similar failure happen to a moment beam. The beam was hung with no external load on it. “Rat Tail” holes were cut into the beam web to allow access for welding. When the welders came back the next morning, the beam had failed and was on the concrete. Ambient temperature was 10 degrees F. During the failure anaylsis, it was determined that the DBTT was 70 degrees F!!!!

Also keep in mind that fatigue failures are very similar in appearance to a brittle failure. However, with fatigue failures there might be evidence of “Ratchet marks” or “Beach Marks” on the surface of the fractured parts. If you cannot see them with the naked eye, try a 10X magnifier, or have a lab look for this evidence.

Hope this helps.

Parent - - By jwright650 (*****) Date 02-12-2003 14:44
Those pictures are scarey. I hate to think what it was like for the mouse that was walking that pipe when it came apart.
I'm curious about this beam that failed also, mentioned by M-Squared. I haven't seen or been around anything that was destroyed like that. I guess I have been overly confident that nothing like that could ever happen to "our steel" once it's hung in the field.
Let me know what all you find out about this,
John Wright
Parent - - By M-Squared (**) Date 02-12-2003 15:58

The beam I had mentioned did not fail in quite the same manner (into hundreds of pieces). As I had mentioned, the beam was a horizontal moment beam connected to two vertical columns by both bolting and welding. The beam was ASTM A992 W30x132 wit h a web thickness of about 0.615 and a flange thickness of about 1 inch. The beam was welded to the columns using a single –V welded from the top of the I-beam. Backgouging of the root was performed to ensure full penetration. A “rat hole” was cut into the web to facilitate achieving full penetration at that location.

The beam failed catastrophically into two separate segments. The failure was the result of cracking that extended through the full cross sectional thickness of the flange and along the full length of the flange. The distance of the crack from the weld ranged from 1/8 to ¾. Anaylsis showed that the crack originated from the “rat hole”. This was evident during visual inspection of the beam. Chevron marks on the fracture surface pointed to a notch that was formed during cutting of the “rat hole”. The rat hole was not ground smooth after welding, as indicated in AWS D1.1 Section 5.17 nor was it inspected. However there was some argument from the contractor erecting the job that this was not a Group 4 or Group 5 shape and was not subject to the requirements of section 5.17.

During the failure anaylsis, it was determined that the beam DBTT was 70 F. Temperature during welding and erection was 10 degrees F. Since that failure we have tried to eliminate the used of “rat holes” where we can, and if they are used the must be ground smooth and inspected for any cracking. In addition, we now order ASTM A 992 using the supplementary requirements. (i.e fine grain and we have Charpy V-notch done).
Parent - By jwright650 (*****) Date 02-12-2003 16:37
Will cutting the rat hole with a plasma torch rather than a OA flame help with cracks initiating from the rat hole?
I had some W24x104's crack from the copes after galvanizing. I was certain when I left to go inspect these beams I would find the copes not finished properly or something wrong with the way they were cut. When I saw them, I had to rethink what went wrong. The copes were smooth and a nice radius had been cut, nothing notchy or jagged for cracks to start. The copes first had the web cut and then the flanges were cut putting the flame tip into material that was being wasted rather than the beam. All nicely done but cracks!!!!!!!!!!! I contacted the AISC about this and they recommended welding a reinforcing bead around the edge of the copes on both sides of the web before galvanizing. We did that also and it cracked at the copes right through the weld and web. We switched galvanizers and haven't seen anymore cracking(coincidense?). Note all the mill test reports were pulled and the beams were from four different mills, all cracked. Two truck loads of galv'd beams cracked. The galvanizer that we first used had a different pickling process(at least that is what I was told). We have been fabricating steel from this plant for 35 years and haven't seen the first beam crack like this. We fabricated these beams at another plant and they used a galvanizer that was close to them. The AISC acts like it happens all the time, but this was the first we had seen of it...
Thanks for your reply
John Wright
Parent - - By pipewelder_1999 (****) Date 02-12-2003 17:47
Isn't one of the the purposes for the weld access holes to not only allow access but to reduce residual stress and reduce the likelyhood of brittle fracture due to excessive residual stress? (D1.1 Commentary C5.17.1)

Wouldn't eliminating weld access holes where required/useful vs doing them correctly be like nose bleeds and tourniquets?

G Austin
Iuka,, Mississippi
Parent - - By jwright650 (*****) Date 02-12-2003 18:27
I have encouraged my welders to give themselves plenty of room and make the rat hole genorous. I feel like you run the risk of trapping something if you aren't careful. Just grind em smooth and free of any notches. I was always concerned when we had to fill in rat holes after welding the flanges up, all that filler cooling in different directions. Maybe I was concerned for nothing, I don't know.
John Wright
Parent - By pipewelder_1999 (****) Date 02-12-2003 18:41
There seems to be some logic on that concern.

"Welded Closure of weld access holes is not recommended" AWS D1.1 Commentary

Good day

G Austin
Parent - - By M-Squared (**) Date 02-12-2003 18:55
Well that’s not the way I see it. First, as I explained before this beam was not a group 4 or 5 shape, (it was a group 2 shape) and it did not have a web thickness greater than 1.5 inches. Therefore, it was not subject to the stringent requirements of Section 5.17.2. (i.e grind to bright metal, MT or PT inspect for cracks). Section 5.17 has some dimensional requirements and states that copes and access holes shall be SHAPED free of notches or sharp reentrant corners. It does not mention grinding smooth and inspecting. The worker did a decent job on the cut with no grinding. This is considered by D1.1 as doing it correctly for this type of beam. However, the kerf marks from the flame cut acted as initiation sites for the crack (excessive surface roughness).

As far as reducing residual stresses, this may be true. However thermal cutting itself can induce stresses and research has shown that oxygen cutting of steel leaves a thin layer of hard untempered martensite at the cut surface. So now, you have residual stresses and a martensitic layer on the surface of the access hole induced from flame cutting. There are numerous accounts of beam failures associated with weld access holes.

Elimination of weld access holes is not uncommon. After the Northridge earthquakes in CA (Jan 1994) and various earthquakes in Japan it was found that most of the cracking and/or failures were associated with backing bars and weld access holes. In Japan, they currently have eliminated the use of weld access holes and it is growing in popularity in the U.S.

Parent - By pipewelder_1999 (****) Date 02-12-2003 19:07
That sure would save some fabricators, industry, and end user some money if some of these AWS requirements/recommendations could be relaxed or at least alternative methods mentioned in the code.

Has any destructive testing been performed comparing the two methods or any refernces you know of. I find this interesting and I want to learn more.

Thanks for the information.

G Austin
Parent - - By jwright650 (*****) Date 02-12-2003 19:27
Does FEMA 350 or FEMA 353 address these issues? I'm not as familiar with those documents as I should be.
John Wright
Parent - - By M-Squared (**) Date 02-12-2003 20:06
FEMA 350 does address weld access holes. They have some very specific requirements when using access holes. This section does mention that connections made using access holes frequently fail as a result of low cycle fatigue of the beam flange material at the toe of the weld access hole resulting from strain concentrations introduced by this feature (access hole). This was taken from I should mention that it mentions when using tough filler metal in conjunction with the access holes.

Interesting to note that FEMA says that moment-resisting connections should utilize weld access holes. It does not say SHALL use weld access holes.
Parent - - By rpoche (*) Date 02-12-2003 21:09
FEMA 353 - June 2000 has added some very interesting "suggestions" we can probably execpt to see in future code revisions.

1. Group 3 shapes with flanges thicker than 1 1/2" should be considered as Heavy Sections

2. Preheat of 150 Degrees for weld access holes with MT & PT testing (heavy sections)

3. Welder certification for CJP Groove on 1 1/2" plate (bring a lunch)

4. Non-Heavy Sections (FEMA 4.5 Part I) Prior to acceptance, the weld access hole shall be inspected using MT or PT and shall be free of cracks.

Their website is fications and II Recommended Quality Assurance Guidelines. It is worth reading.

As I understand this document it is all supposed to a Recommendation to the governing bodies to decide on how to implement.
Parent - - By jwright650 (*****) Date 02-13-2003 12:40
I've noticed more and more reference to FEMA in the specs on the jobs we are bidding. It seems the spec writers are putting these refrences in the specs without realizing what they are actually asking the fabricator to do. The last job that came through here, the eng on the job wanted us to follow FEMA 350 and 353, but he asked this after we had bid the job and was awarded it. He didn't like the change order when he saw it and changed his mind. He also didn't realize what he was asking us to do(changing all our shop drawings, extra hours in the shop for all these full pen stiffners, etc..). When we started asking questions about FEMA he didn't know what we were asking. He hadn't designed the moment connections to reflect any of the FEMA requirements. He had typical moment conn details that were std AISC conn's. All the field welding was costing more to remove backing bars at the bottom flanges and adding OH(4G) fillet welds at the backing bars at the top flanges. The erector didn't have his people qualified for OH(4G) because they seldom weld OH. The eng. backed off on alot of the FEMA requirements. I'll have to learn more about this as it seems the way eng's here on the east coast are trying to go now days.
John Wright
Parent - - By rpoche (*) Date 02-13-2003 13:24

This subject probably deserves a post of it's own at some future date. The recommendations in FEMA 353 will greatly impact on Design, Fabrication and Inspection costs - as you found out. The report shows several photographs of Northridge Moment Frame connectors - typical failures. On some frames where the CJP weld held, the Column web tore - Heavy Sections. Most of the weld failures occurred at the root joint at the lower flange to column flange. This was contributed to backing being left on, Low Toughness of FCAW filler, excessive deposition rates, position of welder etc.

Check out their website - you can download/print the report (+125 pages) and it makes interesting reading.

Parent - - By jwright650 (*****) Date 02-13-2003 14:46
Sorry Gerald,
I should have probably asked these questions on a post of my own like Robert said. This post was originaly about brittle fractured pipe. I get carried away typing sometimes and wasn't thinking.
John Wright
Parent - By pipewelder_1999 (****) Date 02-13-2003 16:24
I'm glad you did. I may have looked over some very useful information. Thanks to you all for your input and time.

Have a nice day
G Austin
Parent - By Niekie3 (***) Date 02-12-2003 18:16
Hi Gerald

From the pics, I must agree with some others that have suggested that the material was below its DBTT. Usually a brittle failure starting in an area such as a weld will continue to propogate in a brittle manner even into the "ductile" base material, but at some distance away would change to a ductile failure mode and eventually arrest with a very distinct buckling of the material. If you did not see this, then we have to believe that the base material itself was also failing in a brittle manner.

It is obviously possible that there was some other embrittling "corrosion" mechanism at work. If this is the case, it usually becomes obvious quite early in a failure investigation. Excluding this possibility, I would have to think that it was due to the temperature of the base metal.

5°F is cold, and many a C/Steel will fail in a brittle manner at this temperature. Typically, in South Africa, we would not want to expose a C/Steel to this temperature without impact tests. In addition, we need to remember that autorefrigeration is always a posibility with compressed gasses. Was the failed area close to a flange or some other point that could possibly have been leaking? If so, you need to keep this mechanism in mind.

Niekie Jooste
Up Topic Welding Industry / Technical Discussions / Carbon Steel Piping Failure

Powered by mwForum 2.29.2 © 1999-2013 Markus Wichitill