Recent brittle behavior of carbon steel (A105/A106/A234/A53) 2

[marty007]

I've been reading a few reports and bulletins on recent brittle failures of carbon steel components that have occurred during hydrotests. Materials discussed include A105/A106/A234/A53. Here are a couple of links to bulletins:

NCPWB Technical Bulletin May 2016
ABSA Information Bulletin No. IB16-018

Reference is made to failed impact tests with absorbed energy values as low as 3 ft-lbs at 70°F, where the material otherwise appears to meet all requirements of the specifications! It sounds as though one of the major contributing factors is the direction microalloying has taken at the mills in recent years, with the Mo:C ratio being highlighted.

These low energy values are concerning, how do you foresee the code committees reacting?

 

[metengr]

Well, regarding BPV I (SGM) we are taking a wait and see attitude because BPV VIII has taken the lead. Frankly, I have not heard this being a really serious problem throughout industry. The number of failures has been very low.

The MN to C ratio that is being proposed is not that simple to explain away the problem. I believe this is a local sourcing issue rather than a wide spread materials problem. Once BPV VIII decides on a plan going forward we can pursue under BPV I. Otherwise, nothing will be done. Impact testing will not be used by BPV I because boilers operate well above any DBTT for carbon steel. If anything, ASTM should do something regarding chemical composition limits or on tramp elements.

 

[marty007]

Do you know if any testing program has been initiated anywhere to try and dig down to the root cause?

 

[metengr]

I know of no material testing being initiated by ASME B&PV Codes and Standards, other than a bunch of CVN impact testing and chemical analysis correlations. Being a metallurgist I do believe there is a significant impact of reducing Manganese content and using other alloying elements to achieve similar strength properties. Again, this appears to be a local sourcing problem of material rather than some universal metallurgical phenomenon.

I am certainly not trying to downplay this thing but until I hear from others on my code committees and subgroup, I intend to gather more information before I jump on the revision band wagon.

 

[weldstan]

I have seen A-106 with 0.40% Mn (as low as 0.37%) and have commented previously in this regard. I would stay clear of these manufacturers (Chinese and Eastern Euroean). It was customary in the US and Japan to use a 0.60% to 0.90% Mn aim range in manufacturing A-106 and A-105.

 

[weldstan]

I would also note that additions of B, V and/or Nb were not recorded on the MTRs.

 

[jte]

This is an issue which has been reasonably well known in industry for about a year now. Which is a very short time in terms of "Code Time". Some outside info on the topic available here.

I don't pretend to be a materials engineer and so won't comment on the science. I will say that I'm now familiar with several occasions in recent months in which flanges for new construction of pressure vessels were destructively tested. The test results did not meet energy values which Section VIII would have led me to expect. Taking a "usual suspects" country of origin approach does not explain the failures. The jury is definitely out on this. I'm looking forward very much to either B16 or B31 or BPV to getting some more definitive results out on this. This issue is being taken rather seriously by some.

 

[Shmulik]

API 570 - 2016:

5.11.4 Test Temperature and Brittle Fracture Considerations

At ambient temperatures, carbon, low-alloy, and other steels, including high alloy steels embrittled by service
exposure, may be susceptible to brittle failure. A number of failures have been attributed to brittle fracture of steels
that were exposed to temperatures below their transition temperature and to pressures greater than 25 % of the
required hydrostatic test pressure or 8 ksi of stress, whichever is less. Most brittle fractures, however, have occurred
on the first application of a high stress level (the first hydro test or overload). The potential for a brittle failure shall be
evaluated by an engineer prior to hydrostatic testing or especially prior to pneumatic testing because of the higher
potential energy involved. Special attention should be given when testing low-alloy steels, especially 21/4Cr-1Mo,
because they may be prone to temper embrittlement.

To minimize the risk of brittle fracture during a pressure test, the metal temperature should be maintained at least 3 °F
(17 °C) above the MDMT for piping that is more than 2 in. (5 cm) thick, and 10 °F (6 °C) above the MDMT for piping
that have a thickness of 2 in. (5 cm) or less. The test temperature need not exceed 120 °F (50 °C) unless there is
information on the brittle characteristics of the piping construction material indicating a higher test temperature is
needed.

 

[jte]

Shmulik-

It is not clear what your intent was with the quote you posted. Some commentary from you would be helpful when posting stuff. If your intent was to draw people's attention to the known and well documented issue of MDMT or MPT and that codes and standards exist which address this then great. But the issue in this thread is not solved by the application of known MDMT design. That's precisely the problem: New (so no in service degradation) steel has been failing in a brittle fracture mode at temperatures well warmer than Code determined MDMT.