Foundation of Genuine Need for Charpy Testing 2

[EngineerTex]

I have been conflicted for years on the subject of need for Charpy tests on raw material used in structural applications. On one hand:

1) It's required per codes and customers.
2) I can't deny the fact that the ductile-brittle transition exists.
3) In critical applications, I would rather be safe than sorry.
4) It can't be denied that Charpy testing does give an indirect test of mill/forge/foundry manufacturing quality.

On the other hand:

1) Bridges built in the 1920's & 1930's are still standing today. Charpy testing was probably never done on any of this structural steel.
2) Cars and trucks drive around Winnipeg all winter and I've not heard of frames, wheels, axles, doors and other steel members that have been manufactured by the millions shattering or breaking without bending. Do cars hit potholes in the winter and break axles or spindles? Do trucks run into cars in winter and have their bumpers break in half? Does the frame break in half? I've never heard of that, so I don't think it happens. Maybe I'm just in the dark about things and I can be corrected.
3) In the Siberian oil fields, I highly doubt that there is great attention paid to impact toughness of mill-rolled components comprising their derricks and service equipment.
4) I-beams, plates, pipes, tubular members, grating, expanded metal comprises all of the equipment my industry has used for decades. I *KNOW* it has not had any level of Charpy testing and I doubt the material would pass any testing if it did. This very old equipment is rigged up and down in -35F to -45F temperatures and used in daily service in those conditions. Some components are hardened alloys with carburized surfaces being used to ram against other steel elements. Others are mild steel which experience certain impact loads, while others are low alloys with various types of heat treatment. The materials run the spectrum of low to high quality and I have never seen or heard of the slightest problem in regards to this subject.

So all that said, I would be interested in:

1) Published papers on the ACTUAL need for fracture toughness on low-temperature structural applications.
2) Regarding components of structures such as radio towers, power transmission towers, above-water cold-weather marine equipment: How does one differentiate between parts which are exposed directly to an impact and those components that are indirectly loaded, i.e., if component A (a pipe, for example) is struck, but component B (an I-beam, perhaps) supports component A, are they both considered as having experienced an impact load? Or is only component A considered as such?
3) (Most importantly) Case studies of structural failure attributed to low-temperature brittle failures.

I've searched on the forums here and not found a great deal on exactly this subject. So any help is appreciated. If there are insightful books or papers anyone can refer me to, I'll gladly pay for them.

My big problem is that it is as if people regard low-temperature brittle fracture as some sort of mystic prophecy that is only known to the ancients, while the mortals who fill the ranks of regulators and customers see Charpy testing as a type of talisman that wards off the evil spirits. Call me a heretic, but I think that those who require extensive Charpy testing usually don't know what they're talking about.

Engineering is not the science behind building. It is the science behind not building.

 

[metengr]

EngineerTex;
Not sure what your long-winded post is regarding CVN impact testing or what axe you have to grind regarding impact tests. There is plenty of public domain information on Charpy impact testing if you conduct your own internet search on key words.

Good Luck

 

[EngineerTex]

The axe I have to grind is that I think CVN testing is requested and required far more often than necessary. I would like to know if is *ACTUALLY* needed or if it is just a collusory means to keep metallurgy labs open for unwarranted business.

In other words, if every metallurgist tells me it's needed, but I know that there are millions of applications where it was never performed, then I question the judgement of the metallurgists who say it's needed.

That's the more succinct and less polite way of asking what I asked in my "long-winded" post.

Engineering is not the science behind building. It is the science behind not building.

 

[metengr]

My short answer as an Owner/User Engineer of power generation equipment with no ties to research or production met labs - -YES, on occasion I need to have CVN testing. You can question my judgment until the cows come home.

 

[stanweld]

EngineerTex
Structures are most often designed well within the elastic range. As such there is little need for impact testing. However, most modern structures are welded, without stress relieving treatments. The high imposed welding shrinkage stresses are additive to the design stress and unfortunately not fully addressed in the Engineering design. If a crack like flaw, where stress intensity is maximized either process induced or service induced, increased sufficiently in size, failure will occur.

With regard to the riveted bridges made in the 1890's until welding became standard, the welding residual stress did not exist, and designs were often much more conservative relative to the yield strength of the material.

There are numerous case studies of failed structural steel components subjected to Seismic events, espcially so documented in the Northridge and Loma Prieta quakes in California and the Kobe quake in Japan. Much documentation was also produced from the transmission tower collapses in Germany and North Sea structure collapses.

 

[MikeHalloran]

See also "Liberty Ship"

and
IMHO
"Flxible Metro/Grumman 870"



Mike Halloran
Pembroke Pines, FL, USA

 

[toropt]

Impact tests are never on such a scale that can be considered too frequent to hamper developement. So do you also not like tensile tests? or chemical analysis of kinds? Charpy test is so easy and need such a basic equipment that it is not even worth thinking. How do you know that materials that you use did not go through charpy? Almost all materials go through rigorous test right at the mill that they are produced. You also do not seem to understant the nature of ductile to brittle transformation.

 

[Wrenchbender]

Tex,
Ductility is favored by low strain rate, low carbon, high temperature, and high chemical purity. Charpy V-notch (ASTM E23) evaluates ductility under a high strain rate. Steel and BCC crystalline material exhibit a ductile-to-brittle transition temperature in CVN tests that depend on the previous variables. A hundred years ago steel mills could not control their chemistry as well as today's mills. More junk elements = lower CVN; however the design margins for old structures was adequate in most cases. (Exceptions might be the rivets and hull plates of the HMS Titanic, a molasses tank in Boston, and other well, and not so well documented case such as broken railroad rails mentioned by Boris Pasternak in Dr Zhiavago.) A useful reference on the subject is a textbook by Hertzberg, which includes some case studies.

Is 'extensive' CVN testing needed today? How extensive depends on your product. If you do it once at first and occasionally thereafter to certify that your chemistry and processing are in control, well…. On the other hand its great CYA against Murphy's Law, and easily performed (can break ~100 bars/hr). Its also a QA and legal issue.

 

[blacksmith37]

There is the basic approach ; If it fails, add toughness requirements for the next one !
Quite a few old oil storage tanks have failed (much like the famous molasses tank failure.) I would get Weather Bureau summaries and show the managers how the failures occurred on the coldest days ; even a manager could understand, and approve Cv req'ts to specifications.
As Bestwrench indicates; there were many brittle failures but Cv reqts of the last 50 yr have greatly reduced the frequency.

 

[EngineerTex]

Thanks for all of the quick responses.

For clarification of the conversation and where my problem lies, we are currently required to perform CVN tests for every heat for 100% of the material we use. I believe that this is overkill, but I'm willing to be swayed away from that belief.

And btw, this is not for the aerospace or nuclear industry. Think of a crane or a mobile service derrick, which is not exactly what I'm dealing with, but it's close enough for the discussion at hand.

My biggest problem with the notion of CVN requirements is that customers demand CVN testing at service temperatures, which, for new equipment usually never causes a problem, since material usually comes from the mill with this test performed already. But for equipment that was built 40+ years ago, is still in use and has provided successful service in the low-temperature environments, I find it unreasonable to take a sample out of every I-beam, pipe, tube and plate to prove notch toughness.

Under worst-case loading, I know that certain parts will experience high strain rates and others will not. This is yet another reason that I am not impressed with a one-size-fits-all, blanket requirement for 100% CVN testing for every member, especially for existing equipment. This is why I said that I believe that those who require it (regulators and customers) probably don't know what they're talking about. (Again -- someone might change my mind on that.)

Bestwrench, thanks for the info on the textbook. I'll check it out post-haste.

Blacksmith, "If it fails, add toughness requirements for the next one!" I'm not opposed to this notion, it's just that my group has hundreds of enormous welded steel structures, zero incidents of this type of failure and ever-tightening CVN requirements. If the CVN testing is genuinely required, that means we have to scrap hundreds of millions of dollars worth of historically suitable equipment and start over.

Engineering is not the science behind building. It is the science behind not building.

 

[redpicker]

Engineer Tex, you have some valid observations, but you ignore some other obvious ones.

Concerning the structures built nearly 100 years ago, when processing was not likely to produce materials with high toughness, those structures that had a need for high toughness have long since collapsed. What you see still standing is a sub-set of the original production that was built with materials and methods that were successful. It is, in fact, the higher toughness of the materials used in the structures that resisted collapse that have resulted in the modern use of CVN testing.

Concerning automobile components that are routinely subjected to low temperature shock loading, these components are specifically designed for these applications (and by design, I mean not only the mechanical design, but material design as well). If a design is developed that is susceptable to low-temperature failure, the design is quickly changed to address the issue. It is not like it was 100 years ago when the causes of low temperature failures were not well understood.

However, I do tend to agree with your thesis that CVN tests tend to be over-specified. I find the minimum specified values tend to be highly inflated, particularly for the test temperatures involved. I believe this results from the idea "if a little is good, then more is better". I understand how this happens.

A partiaular component that had no impact requriements fails (Part A).
Analysis of the failed material reveals, among other issues, the toughness is very low.
A replacement is produced with the highest toughness from the available material (Part B).
The replacment lasts much longer, but still exhibits a short life because part of the whole problem is the environment is particularly severe to begin with.
The next replacement (Part C) gets designed for a toughness requriement that 1.5 times what was used for Part B. Even if Part C does not perform any better than Part B, the higher toughness requriement becomes the requriement since it obviously can be obtained and nobody is going to risk going back to the original material.

On top of this, the 1.5 times requirement gets applied to low temperature applications when the component is needed for a low temperature applicaiton. This is then used for all applications. Case in point, nearly all equipment used for offshore drilling in the Gulf of Mexico is qualified at -20C.

Another issue you bring up (which I disagree with) is
"It can't be denied that Charpy testing does give an indirect test of mill/forge/foundry manufacturing quality."
This is not really true. While materials with the highest toughness will generally come from the higher quality producers and higher quality producers are more likely to control their processes, the testing does not indicate any particular quality level. Likewise, a higher toughness for a given material does not indicate a better quality, provided the requirements for that material are met. There is a lot of perception that a higher toughness translates to better fatigue resistance and/or corrosion resistance. I commonly will see specificaitons written that try to improve the quality simply by requiring higher toughness levels.

rp