ASM Online Member Community

 View Only
  • 1.  Improving first pass yield for HT of low alloy steel

    Posted 01-19-2021 09:27
    Hello,

    I am working with some new specifications that call out very tight hardness and tensile ranges, for low alloy steel (8630 and 4340 large-ish forgings, among others). Our HT crew is trying to hit a specific Brinell number on every HT load, but they often miss and have to either temper longer, or perform a full re-austenize and quench prior to re tempering. 

    I would like to improve our first pass yield, and I think reviewing the chemistry from each bar of material and adjusting our tempering parameters accordingly will be key. The issue I'm having, is that my background is titanium and nickel, and I only have a basic understanding of steel chemistry vs HT time and temp. 

    I am starting a log to capture chemistry vs HT time and temp vs hardness outcome, but while I gather all that, can anyone recommend resources (books, class, articles, etc.) that may help my efforts?

    I appreciate your time. Thank you,

    ------------------------------
    Vanessa Molina
    Chief Metallurgist and Lab Manager
    AFGlobal Corp
    ------------------------------
    Education courses


  • 2.  RE: Improving first pass yield for HT of low alloy steel

    Posted 01-20-2021 10:14
    Heat treatment of steel forgings is a dance between many variables - component size and shape, the size / BTU throughput of your furnace, uniformity of your furnace, your racking layout, the size / agitation of your quench tank, your quenchant, how fast a transfer time your charger can achieve, your temperature, your time at temperature, your means of measuring time at temperature... it can feel like herding cats sometimes trying to optimize every parameter but a chain does fail at its weakest link. I would offer a few thoughts on what I consider to be the most consequential variables:

    • Uniformity of your furnace is critical. If you survey it to a recognized standard like AMS 2750, that will be a good start towards seeing whether you have hot/cold spots and also how quickly your furnace gets to temperature. If your furnace is gas-fired, the orientation of the burners will affect the convection currents within the furnace so you want to make sure they're oriented in such a way to circulate the air without impinging directly on the parts. You also want to make sure your furnace is well-soaked when the parts go in - the thermal mass of that refractory contributes to the temperature stability of your furnace, and above a certain temperature (this is more applicable to austenitizing/normalizing than tempering), radiant heating from the refractory contributes a greater fraction of the heat going into the parts (the rest being convection).
    • How the parts are laid out on the tray is important. You want enough spacing between them (2" minimum but ideally more) to allow air/quenchant flow around them. You also want to make sure none of the parts are "shadowing" each other from radiative heating (don't stack them). If your parts are bored out, you want to orient the bores such that the quench tank agitators are blowing into them. The trays should be resting on ceramic pylons in the furnace so they are raised a good foot or two off the floor to allow air flow underneath.
    • Pre-HT rough machining to a near finished shape will improve your hardenability response, especially in thick sections. That said, if you do machine, I recommend shotblasting the parts before heat treatment to dull the surface. If they go into the furnace shiny, they will reflect radiant heat back until they scale up and this can cause your furnace control TC to think the furnace is hotter than it is and overshoot.
    • Quench tank agitation is super important. Your want your quenchant to look like it's at a rolling boil. Depending on the position of your agitators in the tank, you may want to install baffles to re-direct the flow more evenly into the parts. Transfer time into the tank is also important - some specs say 60 seconds max, some say 90, but the faster the better. We almost always make 45 seconds in our shop.

    These are, of course, all heat treat shop equipment considerations and I appreciate that you may not be able to control some of these (especially if they require costly investment) so here are my two cents on what you can control:

    • How you measure temperature (furnace TC, contact TC or heat sink) will have a bearing on what your soak time should be. I actually posted a thread about this on ASM Connect last year where people gave good input on which they prefer and why but the boiled down version is that rules of thumb like "1 hour per inch" aren't always optimal. If you have any scrap parts lying around, perform an experiment by taking one and drilling a hole through the thickest section to embed a TC and running it through a standard heat treat cycle so you can compare what you're seeing on the furnace control TC with how the part is responding in the thickest section. If you can run multiple channels at once, put a contact TC or two on the part and compare those as well. Personally, I think the best way to measure temperature is with a heat sink (maybe even 2 or 3 throughout the furnace) of the same size as the thickest section of the parts. Once they are at temp, you only need to hold for 30-60 minutes for phase change operations like austenitizing or solution annealing (tempering can be prolonged strategically, more on that below).
    • When it comes to tempering, temperature wields the big stick but time wields a small stick too. You definitely want to have the parts in the furnace long enough to totally soak so that you don't have a significant difference in time at temperature between the interior/exterior of the part but if you're really trying to hit a very tight hardness range, you can play around with tempering a bit colder but holding for a lot longer. Refer to the Larson-Miller parameter to dial in your hold time.
    • If you do have issues with furnace uniformity or quench tank capacity, you might have to reduce your HT load size. Operationally this is of course not desirable but it's easier to get a good soak and quench on one forging than ten.
    • If you are the designer of the parts being heat treated, you should evaluate whether they are designed with heat treatment in mind. Large variations in section thickness and improper alloy selection for the thickness of the part are two cardinal sins in the steel heat treatment world. Obviously we don't always have the luxury of redesigning parts but if the design really doesn't make proper heat treatment possible, it's not a good design.
    • Regarding what you were saying about the chemistry, it matters but I don't think that's going to be as impactful as the other variables I've described and it won't be a definitive factor unless you're talking about a truly lean heat vs. a truly rich heat. That said, I should qualify that statement by pointing out that many customers (my company included) specify narrower ranges, or even wholly different ranges, than the standard AISI grade to obtain more consistent properties. This can be done by proxy by specifying minimum values for Di, CE, Jominy band, etc. or specifying particular elements. For example, standard 4130 has 0.40 - 0.60% Mn but in the oil and gas world, some companies have been specifying Mn in the .70 - .80 range to improve hardenability as component section sizes get heavier. Whether this material is "really 4130 anymore" is a subject of perennial debate. You might improve your consistency to HT response by imposing tighter chemistry requirements on your meltshop rather than just playing the hand you're dealt.

    As far as books go, "Republic Alloy Steels" is a tremendous primer on steel metallurgy and very user-friendly. "Steels: Heat Treatment and Processing Principles" by Krauss is also great but more academic. The ASM "heat treaters guide" and "practical heat treatment" series are very useful and user-friendly.

    Hope this helps.

    ------------------------------
    Sean Piper
    Product / Process Metallurgist
    Ellwood Texas Forge Houston
    Houston TX
    773-524-8985
    ------------------------------

    Education courses


  • 3.  RE: Improving first pass yield for HT of low alloy steel

    Posted 01-21-2021 17:51

    Vanessa,

     

    Correlating chemistry to mechanical properties is quite complex particularity if you are an end user.  Most master-melters have slightly different chemistry aims for a given alloy based on their equipment and available feedstocks.  Adding to this variation is the heat treating, machining and testing of samples.  These all add to the scatter in the data analyzes.

     

    Having said this I do not want you to get the impression that it is impossible nor worthwhile to attempt.  It is necessary to determine what data is worthwhile to spend your time and efforts to analyze.  Since every situation is different I can only suggest the following guidelines.

     

    • Review that data and its source to identify blocks of data with the greatest integrity and potential to yield maximum economic benefit.

     

    • Perform initial statistical analysis to identify data errors and outliers.

     

    • Perform multiple regression analysis to identify the effect of elements, suppliers, heat treatment, etc. on the mechanical properties of interest.

     

    • Sit back and think about how your conclusions match up with metallurgical theory and practice.

     

    • Perform some Statistically Designed Experiments (DOE) in your manufacturing processes to verify what you found is actually significant.

     

    • Integrate your results carefully and slowly into production and quantifiably track any benefits/problems.

     

    I have had successes and failures correlating mechanical properties with composition.  I recommend starting off small, learn as you go and above all enjoy the journey of discovery!  If you would like some additional insight into statistical analyzes and some real-world examples then check out my website: Planish-Inc.com 

     

    Best Wishes,

    Robert Werber, PE

    Planish-Inc.com 



    ------------------------------
    Robert Werber
    Oregon City OR
    (503) 557-2906
    ------------------------------

    Education courses


  • 4.  RE: Improving first pass yield for HT of low alloy steel

    Posted 01-22-2021 09:10
    Vanessa,

    If you do decide to pursue chemistry more, I recommend you focus mainly on carbon content. Steels get their hardness mainly from carbon​. For a high-hardenability alloy like 4340, if you specify that your melter keep the carbon within a range of .02 (like .40 - .42), I really think most variability should be eliminated. The other alloying elements (Ni, Cr, Mn, Mo, etc.) contribute solid solution strengthening but their primary function is achieving other attributes like hardenability, toughness, high-temperature strength, etc. Strong carbide formers like V, Ti, and Nb can contribute a decent bump in strength but it also depends on your heat treatment because there are several mechanisms through which they strengthen, and specs will often limit their addition anyways. Boron is a potent strengthener (even in the ppm range) but most specs don't allow its intentional addition unless it's specifically a boron grade. Deoxidation practice during the melt (whether it's silicon or aluminum killed) will affect your predisposition to coarsening at normalizing/austenitizing temperatures, which will in turn impact strength.

    If you're interested in learning more about the effects of the various alloying elements, "Alloying Elements in Steel" by Bain and Paxton is pretty good, although somewhat academic.

    Something else I recommend you look into is your actual hardness testing procedure - in my experience, getting "perfect" Brinell readings is difficult because it's very sensitive to lots of variables - your surface prep, how deep you grind, how perpendicular your indenter is relative to your part surface, whether the part and indenter are secure or whether they "rock" some when force is applied, etc.. You'd be surprised how much grinding 1/16" vs. 1/8" can affect your results because it can make the difference between getting through your surface (de)carburization layer or not. Unless your Brinell procedure is totally dialed in, it will create so much variability that you'll never be able to draw useful conclusions from data on other variables.

    ------------------------------
    Sean Piper
    Product / Process Metallurgist
    Ellwood Texas Forge Houston
    Houston TX
    773-524-8985
    ------------------------------

    Education courses


  • 5.  RE: Improving first pass yield for HT of low alloy steel

    Posted 01-22-2021 19:25
    The comments by Robert are very thorough.  Some additional comments that have come to mind:

    A classmate of mine worked on producing a computer program to do a hardenability calculation, based on alloy % to hardenability relationships in our textbooks, back in the late 1970's!  So such equations based on compositions were available even then.  (And were no doubt based on regression analyses done in the previous couple of decades!)  There are multi-elemental synergistic interactions as well, but I am sure that the original poster's concerns have been shared by others, and that there must be commercial packages that estimate heat treat response, and equally importantly tempering response.  THESE ARE NOT NEW PROBLEMS.  In fact they are old enough that the solutions to them may have been incorporated into work processes and the reasons for these solutions may have been lost to time and retirements and corporate re-organizations.  I have seen this sort of knowledge loss within my career.  "Oh, it's because we have always done it this way."

    The grade designations you describe don't have the H suffix.  I'm pretty out of date here, but the idea was that the compositions of the H grades had been controlled somewhat more tightly to avoid co-variations (for instance, low Ni in the same heat as low Cr) that would both swing the hardenability in one direction.  This has the effect of narrowing the possible variation in the heat treat response that one could get. A raw steel supplier's interest may start at providing a product that meets the letter of the spec while minimizing alloy costs, which can result in "low-side" content of all of the elements added.  The heat treater's interest is to provide a part that meets property spec.  The H grades, and any other requirements that you may impose on your supplier, are your agreement with them to supply what YOU want.  You can always negotiate to buy a product with a specific narrower hardenability range within the boundaries of, say, 8630/8630H.  This increases your chance of hitting a heat treat target, and still meets your own client's chemistry requirement.

    Likewise to hardenability, tempering response varies with alloy content, and slightly differently than the hardenability response.  Again I am out of date on this but as I recall an element that is not a carbide former but affects quenching response, such as Ni, will not have the same degree of effect on tempering response as a carbide former, which affects the decomposition of the alloy carbides present.  This type of relationship should be included in any calculational package that you might find to help predict holistic response to the combination of quench and temper.

    Finally, you can ask your client why a narrow range of hardness is specified.  Some but not all designers will be able to articulate the reasons for what the target is.  XXX strength minimum but with a fracture toughness of YYY minimum?  If you show that you can meet fracture toughness at a slightly higher hardness, would the part still be able to be accepted?  Maybe fracture toughness isn't the limiting factor, it could be avoiding an embrittlement range or environmental effect.  There might be wiggle room somewhere.

    I hope these are illuminating.

    ------------------------------
    Paul Tibbals, PE
    ------------------------------

    Education courses


  • 6.  RE: Improving first pass yield for HT of low alloy steel

    Posted 01-23-2021 17:41
    Quite right Paul, what was good and worthwhile in the 70's is still worth doing and Microsoft Excel is a great platform for doing hardenability calculations. In 2016, I had my heat treat class build workbooks for hardenability calculations.  ASTM A255 was the starting point.  In the past 5 years my class workbook has evolved to include temper hardness calculations and grain size effects. If anyone is interested in getting a copy of this workbook, just drop an e-mail to me at dcva@mst.edu.  A great deal of "old school" metallurgy went into this workbook and there is no cost.  It's great for exploring the effects of alloy, heat treatment temperatures and quench severity on final tempered hardness in a quenched round bar. Best regards to all.

    ------------------------------
    David Van Aken
    Missouri University of Science
    Rolla MO
    ------------------------------

    Education courses


  • 7.  RE: Improving first pass yield for HT of low alloy steel

    Posted 01-22-2021 21:44
    Hello,

    The hardenability of each heat can be described in terms of DI (Ideal Diameter), surface hardness (J0), and the hardenability curve calculated from the chemical composition using the Grossman method. If you are capturing a log it would be useful to register some of these values that are directly related with the chemical composition and describe the heat treatment response of the alloy. Also, austenitic grain size has an impact in hardenability: the bigger the grain size, the higher hardenability.

    The effect of chemical composition in temper response is similar to that for hardenability, but in some books I've read the term "temper resistance" to describe the softening behaviour of an alloy as a function of any element. For this issue I do not know a mathematical model as for hardenability, but there are curves comparing the effect of some elements in the hardness decrease after tempering at different temperatures.

    In my experience, many of the issues were mainly related to the quenching system (quench media conditions such as temperature, cleanness, agitation, etc.), nonetheless, we were heat treating only one alloy.

    However, my recommendation is to record DI and the most critical elements for hardenability such as C, Cr, Ni, Mo as well as the as-quenched hardness obtained and compare it to the calculated theoretical hardness, and then adjusting the parameters accordingly to DI variation.

    Finally, I would recommend George E. Totten's "Steel Heat Treatment", the Timken "Practical Data for Metallurgists", and the ASM Handbooks on Heat Treating.

    ------------------------------
    Jose Mariano Flores Herrera
    Research Intern
    Ternium Mexico
    San Nicolas de los Garza
    5218717957482
    mariano@flores-h.com
    ------------------------------

    Education courses