JA,
I will give you a not-so-simple answer to a very good, simple question. The level of heat input that is correct for any welding application has to be determined by several items, including fabrication specification requirements (mechanical properties, requirement for impacts, etc), material (including thickness) to be used in production and welding position, amongst others. The final selection of welding parameters will result in a given level, or range, of heat input. Here is a quick example:
Material ASTM A572, Grade 50 1/2" plate to be welded in all positions.
Welding process/consumable - SMAW with E7018
As a start, the manufacturer's recommended parameters can be a source, but offer a very wide range for each electrode type/diameter (1/8 E7018 - Lincoln says 90-160 amps)
For a flat procedure (1G) starting parameters might be 130 amps, 24 volts, 8 ipm travel for a heat input of 23.4 kJ/in.
We will skip alternates for root and hot pass and illustrate fill and cap pass welding.
The weld is completed using stringer bead technique, the weld subjected to mechanical testing (per the fabrication code/Section IX) and all results meet, or exceed, requirements.
We now go to vertical (3G) welding with the same parameters. Goodness, we may have weld falling on the floor due to gravity, so we back off to 100 amps, 22 volts (tighter arc) and 4 ipm travel, using a slight weave. This combo yield a heat input of 33 kJ/in.
The weld looks good and all mechanical testing is acceptable.
As you can see, each position/material/welding process/consumable combination will require the same type of analysis. This should all be done prior to start of weld procedure qualification and is based on technical requirements and good old experience.
Throw in requirements for diffferent base materials, PWHT (some codes) and CVN impact testing and you may need significantly different welding parameters (and resulting heat input) to meet project/specification requirements, in addition to alternate welding processes and consumable.
Hopefully this adds a little insight to what others will add. As you might guess, I feel that a great deal of experience along with the proper technical welding engineering expertise should be used for any welding application. A bit of work up front may save lots of headaches later down the line. As applications become more and more stringent, the balance between experience and welding engineering input tips in the direction of the pre-engineering input. Experience is alway important.
Hope this helps and will probably stir responses from others.
For most non impact regime carbon steel applications heat input really isn't a primary issue. If it is, it is simply to ensure adequate fusion, or some practical welding end, is a by product of considering volts and amps, and not related to mechanical properties.
For the great bulk of Chrome Moly applications heat input is not much of a consideration either since grain growth for high temp applications, of which CrMo's are most frequently speced, is a good thing for high temp strength and creep resistance (though some CrMo's-AISI designate steels are speced for strengthin which heat input could be an issue).
An exception might be GTAW root passes with wide gaps (meanng possibly thin roots) where high heat inputs can lead to centerline cracking.
For nickel steels heat input is a critical consideration due to grain growth and the predominant application of nickel steels to low temp applications.
For austenitics in general it is hot cracking which forces consideration of heat input. This applies to nickel alloys, Monels, super austenitics, etc., but generally not for austenitics such as coppers and aluminums due to their high thermal conductivity.
There are 1000's of other alloys requiring heat input consideration but this is a general scheme.
In all steels it is desirable to try to avoid "excessive" grain growth because the strength and toughness of the steel will decrease. Grain growth in the HAZ of most stainless steels is typically not a problem, but is actually common. However, grain growth can be an issue with the ferritic stainless steels. The ferritic structure is much more prone to rapid grain growth than the austenitic and duplex structure. This, combined with the fact that ferritic grades start with a lower level of toughness make the ferritics vunerable to embrittlement in the HAZ. In the ferritics, as the grains increase in size, the toughness decreases and the DTBTT increases....Not good..
Chuck,
Excessive is a broad term. And to a certain extent I'm not sure I agree. Almost all industrially applied metallic materials have essentially a base line chemical strength range. Their strength can certainly be reduced to a manner in which they do not achieve minimum specified tensile strength, but to copmpletely destroy their strength I believe is quite rare. I personnally have never seen metallic material have its strength completely destroyed by high heat input weld procedures. Though I've never tried to accomplish this. Ductility yes. Toughness, certainly.
Grain growth has to really, really be excessive before it effects tensiles. I've run SAW welds on ferritics and bainitics at heat inputs well over 150 kj with no damage to the strength whatsoever. And still qualed for bends as well (20% outer fiber elongation). I would assume the grain size was huge.
And when CrMo's as such are intended for high temp how much does a room temp tensile matter? Not a whole lot if the service regime is in the range of temps for that material where the large grain size actually increases strength. And yet, we still qual with room temps. My opinion is, heat input is something to be considered in general, but is not an issue for many alloys. Ferritics, bainitics, and martensitics intended for high temp would be among those where heat input maximums could be detrimental.
You always have to consider the metallurgical nature of an alloy, but the engineering application may render the metallurgical characterisitcs (such as minimizing grain growth) irrelevent or even damaging.
The beauty of this Forum is our right to agree or disagree. Personally, the statement regarding grain growth affecting ferritics more than austenitics and duplexes is a metallurgical fact. It's not an opinion, but a fact. It is also a metallurgical fact that the ferritics start with a lower level of toughness and are more prone, due to this lower level of toughness, to embrittlement in the HAZ. And lastly, as the toughness decreases in the ferritics, the DTBTT definitely increases. That is the reason I used the term "excessive" in that context. As in any steel, we try to keep grain growth to as much a minimum as possible, but in virtually every case we experience grain growth to some degree. It is just more damaging (for lack of a better term) in the ferritics. "I personally have never seen metallic material have it's strength completely destroyed by high heat input weld procedures". I never said that in my post. If you have any further doubts about the validity of the statements I made on the ferritics, please refer to the article entitled "Understanding Stainless Steel Heat-Affected Zones" written by myself and Dr. Jim Fritz, and published in the AWS Journal, dated July 2005 edition. If that's not good enough, maybe one of the trusted references we referred to, Ageen, G., Deverell, H. E., and Nichol, T. T., 1979, "Microstructure Versus Properties of 29-4 Ferritic Stainless Steel", STP 672, American Society for Testing and Materials, p. 367 will validate the points I mentioned. So, I would have to disagree that a heat input of 150kJ while welding on a ferritic had absolutely no affect on the ferritic strength whatsoever.
Chuck,
The next time your in Houston we'll probably disagree on where to have dinner. So why should welding issues be any different? Ha!!
The 150kj welds not losing any strength (and when I say strength I am talking room temp tensiles) was indeed a fact as well. I did the testing myself. With ASME SA-516-60/70 grade material (a fine grained pressure vessel steel).
I was running nuke quals and experimenting with HAZ responses(the old HAZ compared to unnaffected BM problem). And whereas my Charpys's, both impact strength and lateral expansion were all over the place at varying heat inputs, and not necessarily even consistent with heat input (meaning just becasue the heat input went up didn't mean the Charpy's went down), my tensiles were so consistent as to not even be able to detect any strength variance outside of the normal material scatter band.
I also think ASME recognizes this issue as such since heat inputs are only imposed upon impact regimes.
I'm glad we live in a country where we can disagree without consequences. I spent 4 long years helping the good ole USA maintain those rights. Now, the topic of my post was referring to ferritic stainless steels without the Ti or Nb stabilizers. Not arguing, but I cannot fathom welding a ferritic SS at 150 Kl/in and not see some, (even significant) loss of toughness due to coarse grain structure. From reading your post, it appears the tests you did were on carbon steel, grade SA-516-60/70. Is that correct? If so, I am referring to the SS ferritics (429, 430, etc. unstaabilized) I think that's what I referred to in my post.
Chuck,
Yes the high heat input welds were a carbon steel. I have never welded any stanless steel (austenitic, martensitic, or ferritic) with heat inputs of that magnitude. But it would be an interesting experiment.
And I agree, there would normally be a loss of toughness with increases in heat input. But not always(depending of course upon the alloy and the actual heat input ranges considered), since toughness isn't exclusively related to grain size. In some alloys high heat inputs can draw solution strengthening elements in the HAZ (such as V, Nb, C, etc.) into carbides that can actually counter the grain size increase phenomena. This would be especially the case with impact regimes involving lateral expansion as opposed to ft/lbs (which has a tendency to emphasize ductility as opposed to impact strength), where ductility increases due to this de-solutioning can actually improve results.
I suppose the jist of my argument is that grain growth though an expected phenomena, may or may not be detrimental. May or may not be relevent. With a loss of toughness engineering would ask, how much toughness is lost. Does it matter to this application.
I know your post was centered around SS's. But mine was for alloys in general. One other thing if I may, (not to tangent this thing too much) one can see a shift in DBTT without a variance in upper or lower shelf magnitude. Though certainly those temperatures close to the transition would see a variance. I think DBTT is intimately related to overall impact magnitudes but not directly. Similar to hardness and tensile strength. For example your DBTT could shift from -20F to -40F and yet your lower shelf value at -60F remains the same, as does your upper shelf energy of say 0F (part of the reason tey are actually a shelf in appearance-capable of moving horizontal on the Charpy curve with variances in DBTT). Conversely your shelf energies could vary and your DBTT remain the same.
Relating to grain growth, all of my posts were centered on SS ferritics. When you mentioned that you used 150Kj when welding "ferritics and bainitics", I assumed you were talking about the SS ferritics, as the post had been mentioning. Under no circumstance will I debate the CS's with anyone. My knowledge is relating to SS. As far as experimenting with 150Kj when welding a ferritic SS, save your time. Too many publications of documented experiments have been published to certify the excessive grain growth of ferritic STAINLESS STEELS definitely causes significant loss of strength in the HAZ, to the point of unusability. With all respect, I think this is my last post on this subject...It's too late in the week to debate this issue further...Almost time for that 1st glass of "the fruit of the vine".
One last referring to ferritics and the detriemntal affects of grain growth to this family of stainless steel. Please allow me to quote from TWI (The World Centre of Materials Joining Technology.
".......Ferittics....a course grained HAZ will have poor toughness."
"In thicker material, it is necessary to employ a low heat input to minimize the width of the grain coarsened zone and an austenitic filler to produce a tougher weld metal. Although preheating will not reduce the grain size, it will reduce the HAZ cooling rate, maintain the weld metal above the DTBTT and may reduce residual stress."
The way to minimize grain growth, since on is not going to completely eliminate grain growth in ferritics, is to use a stabilized grade such as 439 that contains Ti and/or Nb.The resulting carbides and/or nitrides in the icrostructure will pin grain boundaries and slow the grain growth during welding.
I do not mean to sound arguementative, but ferritics are totally different when it comes to detrimental affects of grain growth than compared to any other family of SS. The less tough HAZ, due to grain growth leads to embrittlement quicker than any other SS. Check it out...
What an awful job of spelling I did in this last post...I just got in from Arizona and my brain and fingers are not in conjunction yet. Sorry.. Jon, Deb says Hi.
The big problem with heat input is knowing when you are too high or too low. Here is a link to a technical paper on the JFLincoln Foundation website. I think you will find it interesting in relation to your question.
http://www.jflf.org/pdfs/papers/keyconcepts2.pdfI often use the formula shown for approximate fillet leg size [w = sqrt (heat input/500)] to check WPS's; to make sure the parameter ranges will result in a "doable" weld. For groove welds, I use the same formula and then calculate what the equivalent fillet size would be for an equivalent groove cross sectional area.
A quick check for a WPS is [heat input = (fillet leg size) squared X 500]. If the WPS parameters don't calculate out to the needed heat input, the welder will probably struggle with making a proper fillet at the required size, when using those parameters. I can't say the formula works for every situatuon, but it has worked out to be fairly close for everything I have tried it on so far.
Chet,
You realize than when you post 'keyconcepts2.pdf' that it just leads to keyconcpts 3, 4, 5, and etc. pdf? Next thing you know the whole morning is shot. :>)
By tom cooper 
Date 05-04-2007 16:04
Edited 05-07-2007 09:08
Chet-
That is an interesting use of Mr. Funderburk's equation for heat input. Since a groove weld is built up one bead at a time, I suppose you would use an average bead width to calculate your heat input instead of groove cross section?....... a limitation in Funderburk's equation is that it doesn't account for different alloys. I don't think the method will give you a reasonable number for SS or alum heat input, but then again his article seems to be written to the context of D1.1 & D1.5 only.
I think Chet's reply comes closest to answering JA's original question which was basically "What reference value can I compare my calculated heat input value to? " This is similar to the question which has been torturing me recently, which is " What recommended heat input value can I use to establish travel speed for new WPS's?"
You can search to the end of the internet (like me several times), call steel companies (like me), call metallurgists (like me), go ahead and call the Engineer (that should be fun!), go through the entire bookshelf on welding at the University and I can tell you no one has the answer EXCEPT for Chuck Meadows who has given me wonderful info on SS. A couple of Lincoln's texts will show some amp/volts and ipm for mild steels which is really what you're after, but that is it!!! Please post your references if you know otherwise!
There is a lot of suggestion that trial and error as well as previous experience is required to dial things in and I know this to be true but it is difficult to tell your management (or your customer) to pay for all the effort that goes into just one PQRT and a full battery of destructive Lab tests more than once because you just don't know the right heat input to start with.
Thanks
