CVN and Fracture Toughness 3

[Koz]

Hi Guys,

I'm really bugged by this CVN thing. Acknowledging ASTM Standard E23, A6, and A992 for wide-flange "I" beams and we all run these tests correctly with all of the properly calibrated instrumentation. Oh yes, and the chemistry variations ie product versus ladle have been understood for decades. And in the mild steel business we like 20 ft-lbs @ -40 deg F as being "tough".

The location of the coupon plays such an important part in the data gathered and the engineering decisions that are made! I have records of tests on wide-flanges that show CVN variation from from 222 ft-lbs @ 70 deg F at the flanges (ASTM A6 standard location), 60 ft-lbs @ 70 deg F at the web, and 3 ft-lbs @ 70 deg F at the "K" region intersecting the web to flange. These are not of the same heat of material, but all of these wide-flanges were mill produced, taken from fabricated products, and meet all ASTM requirements. I believe similar values would be encountered in any single wide-flange section produced today. A992 is not that different from A572 GR 50 and indeed A36 wide-flange shapes are triple certified to these standards.

And how do these values work in the Barsom-Rolfe fracture toughness equations in AASHTO and BS7910? For the flanges, I get a Kmat of 280 ksi rtin and the "k" region 19!!

Also, how can the E23 standard 10mm x 10mm coupon accomodate the increased restraint at the "k" region. Aren't we getting into a plane-strain condition? Once we machine away the restraint, we should have nice ductile material again, right? This is similar to thick-to-thin regions in castings or forgings.

Please fill me in, because I am having a hard time with CVN and its application in an engineering sense. 100 deg in Miami is a lot different than 100 deg here in Utah. I will appreciate any input.

Koz

 

[Carburize]

The Charpy test has no real meaning in terms of fracture resistance, the specimen size,notch geometry and strain rate bear little to no relationship with actual failures which happen in large structures, from sharp flaws under psuedo-static conditions - it is simply a screening test for situations in which known failures exist.
The original use to sort "good" and "bad" material from the Liberty Ships and T2 tankers was based on a reasonable sample of failed and un-failed ship plates.I don't believe there has ever really been sufficient study in any other application to justify the use of Charpy testing as applied by many codes and standards today apart perhaps for the work done on Wide Plate tests and Charpy relationships and used in BS 1515 and BS 5500.
Correlations between Charpy properties and Fracture Toughness values are approximations at best.

 

[stanweld]

What you have experienced regarding wide spread CVN variations reflects steel manufacturing methods; e.g., chemistry, fine grain vs coarse grain melting practice, deoxidation practice, inclusion control, degree of hot work (especially in light of continuous cast billets), thermo-mechanical processing, thickness/vs Charpy specimen, etc. CVN was/is a simple way to determine toughness (strength and ductility) at the minimum exposure temperature under load and is used to control the steel manufacturing process to assure a greater consistancy of the materials used for a particular design. Noting that typical brittle fractures in bridges occured in materials whose CVN was < 5 ft-lbs at the exposure temperature, higher values were required to &quot;assure&quot; operation in the elastic-plastic or fully plastic regime.

Although not representative of fracture toughness, correlations have been provided in a number of studies for pipe systems - most notably by Battelle Institute in the '70s and '80s.

CVN may also be misused by design engineers for static designs - resulting in added costs of compliance with little real benefit.

 

[Koz]

Hi Guys,
I'm still pondering Wells and Wide Plates. But with the Battelle studies, don't we still have a configuration problem?

These are pipes not wide-flanges. The grain flows are much more uniform for pipes. Even the discontinuity at a welded long seam is less discontinuous than the joint of a typical structural moment frame connection. Are we permitted to extrapolate engineering data for fracture toughness from pipe studies into the warm, fuzzy structural studies performed after Northridge? I guess I'm asking is the actual fracture toughness of low carbon steel pipes determined by ASTM E399 or E647 the same for low carbon wide-flange shapes? Common thought in the fracture toughness literature I have read seems to think so for steels with YS less than 100 ksi.

The almost 1200% variation I have seen in the CVN toughness for wide-flanges are for W14 x 190. That variation, with some granted allowance for variation between heats, happens in 7-inches through the width of the section! Obviously ASTM A6 drives us to obtain coupons in the area most favorable to the producing mill. A992 offers some other location options for group 4 & 5, but these are much thicker sections than the 190# columns often used in buildings.

And I am still stumped by the Barsom-Rolfe equation and its variations over the years. Where is the upper shelf? Barsom has used his equations to calculate Kmat from CVN and CVN from Kmat. The structural industry is hanging a lot of importance on these equations.

And won't the fact that we are joining low carbon HSLA wide-flanges with low carbon weld filler metal having significantly higher yield strength than the wide-flanges affect the actual fracture toughness of the joint? It shouldn't according to the common thought for low carbon steel. Kmat is a physical material property. I am not sure which part of the material I am supposed to focus on, however, ie wide-flange, weld metal, joint, and which part of all of this found in the 56 sq in area that is called the connection.

If CVN is a screening tool, like for the Liberty ships, and KIC is a material property. Do either of these describe the synergy of the welded joint? Wells worked on large welds, but, again, these were plates not shapes. And are the CTOD values determined by the Brits applicable to shapes? I wish I could get into the TWI industrial sites!

Thank you so much for diving into my morass of data and scavanged literature search. I've been screwing around with this for several years now and this is the only web site that even scratches the surface of my questions. Hallelua for Eng-Tips!!

Koz

 

[Carburize]

Just to complicate your life a little further! Have you researched the CEGB R6 approach which might offer more relevance to ductile materials such as those you are dealing with? I think one of the responders on another forum raised the issue of plastic collapse as the real failure criterion rather than crack propagation and the R6 method offers ways to deal with this situation.

 

[Koz]

Hi Carburize,
I have no idea what you are talking about!!! Fun!!!!
If you could elaborate, or just point me in the right direction, I will dig into it. I'm a voracious seeker.

I hope it isn't proprietary, as I don't have the right connections. I'm cobbling together this fracture toughness thing on my own and I find it fascinating.

Thanks,
Koz

 

[Carburize]

Back in the late 70's the Research Division of the Central Electricity Generating Board (CEGB) in the UK came out with a document called &quot;Assessment of the Integrity of Structures Containing Defects&quot; the report was numbered R/H/R6 rev.. and thereafter the method of defect assessment in the report became know as the &quot;R6&quot; method.
Part of the assessment includes a method to evaluate whether the structure is likely to fail by fracture propagation or by plastic collapse.
I believe copies of document should still be available.
On the specific topic of wide flange beams - fracture tests on full size sections containing flaws were undertaken back in the 70&quot;s by the British Steel Corporation at their Scottish Laboratories to try to correlate Charpy, CTOD and full size behavior I don't know whether that data was ever published.

 

[Koz]

Hi Carburize,
Thanks tons for the references. The game is afoot!
Koz

 

[chrisbullough]

The R6 defect assessment method is alive and well, and after the split of the CEGB was maintained by staff at Nuclear Electric, now known as British Energy Generation Ltd. I think the main centre for continued work is at British Energy, Barnwood, Gloucester, UK.

Some elements of R6 were adopted in British Standards in BS 7910 : 1999, and possibly other defect codes, although I believe that R6 also remains in current use, particularly in the UK nuclear industry.

If you search on Google for &quot;R6&quot; and &quot;defect assessment&quot; or &quot;nuclear electric&quot; or &quot;british energy&quot; you should find plenty of background material.

 

[Koz]

Hi Folks,

You guys are so much fun! Thanks for the update on R6. I'm currently looking over Barsom's Fracture and Fatique Control, but I give much more credence to the British work.

I asked about configuration and fracture toughness and have gotten no response. Is the question too lame or is it not being considered in these days? I'm stuck comparing the work on ductile butt joints in piping and what I am sure is plane strain in the &quot;k&quot; region of HSLA wide-flange shapes. KIC is a physical property so configuration should matter. But I get such a huge variation in CVN within one section of material. So how do these screening correlation equations have any relevance (ref AASHTO and BS7910)?

It may be a question better left unanswered. Regardless, I appreciate the references and I will continue to dig into the dirt. It's fun stuff.
Koz

 

[CoryPad]

Koz,

I believe the reason you got no comments regarding &quot;configuration&quot; is that we don't understand what you mean. You terms are somewhat ambiguous, and some of us may not be familiar with your product form(s). I think you are referring to orientation, i.e. crack plane relative to product (longitudinal, through-thickness, etc.) and geometry effects (thickness, notches, etc.). If you can restate your question, I am sure we can supply more theory and practice for you.


Regards,

Cory

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[Carburize]

There are significant differences in fracture toughness values - on data I have in terms of CTOD - between the locations in the type of product you describe. Generally speaking the order is the same as those observed in Charpy tests, web highest, flange next and finally the flange/web interface area. There is however no consistent relationship between the two types of test that could be reliably applied.

 

[Koz]

Hi Folks,

I had a vision this weekend and I think I have answered my configuration problem.

With respects to Corypad, wide-flanges are the technical name for a group of I-beam structural shapes. CVN coupons are taken from the &quot;flanges&quot; which make the top and bottom of the &quot;I&quot;. It is assumed that the material is homogeneous in design and that the &quot;web&quot;, the upright portion of the &quot;I&quot;, is the same as the flange. Very little work is done at the intersection of the web to flange, what is called the &quot;k region&quot;. These shapes are often joined in 90-deg. angles to form frames that must resist seismic and balistic events. Welded joints failed during the Northridge earthquakes in CA originating at the bottom flange and in the vicinity of the k region of this typical connection. It blew everyone's mind.

So my vision came that the severest restraint of this connection is at the above crack initiation site. The plane strain condition should give us a brittle crack initiation site which then zippers into the more ductile regions of the connection at the outer edges of the flanges or into the web. That crack, as C.E. Inglis cautioned us over 80 years ago &quot;has got such a hold on the plate that no amount of ductility will prevent the crack from advancing.&quot;

There have been other geometric and fabrication problems with this connection but they have largely been solved and are well understood.

To Carburize: My test data shows the flanges have the highest CVN values which, coincidently, is where ASTM takes their standard test values. Hmmmm, I wonder why?

So I can see that configuration is not important for determining the stress intensity factor for my mild steels. My joint configuration has an area of plane strain and the rest is predicable. And I believe the anisotropy of the CVN values within the wide-flange really makes the Barsom-Rolfe correlation equations meaningless for this shape. I have gotten their book and am digging into this further. I may change my mind.

I have begun to dig into the R/6 search you folks recommended and this is interesting. I would really like to find data specific to wide-flanges but will continue to search. That may come back as a specific thread request on this site. Thanks again for all of your insight and recommendations. We stand on the shoulders of giants.
Koz