4340 & Stress Concentrations

[drillrig]

Hello all,

This is my first post in this forum so please forgive my ignorance.

I am designing a large shaft (17.75" OD) made of 4340 alloy steel. See attached picture. The shaft undergoes both axial and torsional loads as well as internal pressure. My analysis indicates a stress concentration of 45,000psi at the fillet where the "flange" starts, as would be expected. The spec I am designing to stipulates that for stress concentrations I need a safety factor of 2.8 on the material's ultimate strength. My shaft would therefore have to be made from a material with 126,000psi ultimate, in this case 4340 in some sort of quenched and tempered condition.

With a shaft this size my experience indicates that even if I could find a vendor that produced a ~18" OD bar quenched and tempered, most of the heat treated steel would be machined away and the material at the 9" OD would be more like the annealed condition (~60,000psi yield, ~95,000psi ultimate).

My question is this: Just below the fillet in question we have to induction harden the shaft to 58/62 Rc. Is it possible to also heat treat the fillet to attain the required ultimate strength? Is this common practice? Can you directly correlate a yield strength with a hardness value? Most of the stresses in the shaft are very low, and I'd rather not rough machine/heat treat/final machine the entire thing.

Any comments would be most helpful.

Thanks in advance.

- Chris

 

[unclesyd]

What is the width of the flange?
Is the shaft hollow?
If hollow what is the thickness where the flange joins the shaft?
What is the thickness of the last few steps in the shaft?
Is there a radius where the flange joins the shaft?

 

[drillrig]

unclesyd,

To answer your questions:

2.880".
yes the shaft is hollow (3" ID).
wall thickness where flange joins shaft = 3" (9" OD).
the adjacent steps are 8.900" OD, 8.750" OD.
the radius where flange joins shaft = .375".

Thanks in advance.

- Chris

 

[CoryPad]

It is possible to induction harden the fillet region of your part.

It is common practice to induction harden flanged shafts, although I have no experience with parts as large as yours.

For quenched and tempered steels, there is a good correlation between hardness and ultimate tensile strength. ASTM E 140 Standard Hardness Conversion Tables for Metals Relationship Among Brinell Hardness, Vickers Hardness, Rockwell Hardness, Superficial Hardness, Knoop Hardness, and Scleroscope Hardness and ISO 18265 Metallic materials -- Conversion of hardness values are documents that tabulate the correlation. For Q + T steels with the composition and properties you mention, it is common to have a yield/tensile ratio of 0.9.

Regards,

Cory

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

 

[imcjoek]

Wouldn't hardening at a section change (the fillet) just increase notch sensitivity right at a stress concentration?

I ask because I am working on something similar and am worried about fatigue failure.

 

[CoryPad]

Notch sensitivity and fatigue failure are important concerns regarding hardened flanged shafts. However, it is a mainstream practice to use induction hardening of flanged shaft fillets, for example on automobile axle shafts.

Regards,

Cory

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

 

[drillrig]

The number of cycles at this loading is not expected to be greater that 1000, so I am not considering fatigue at this point.

So if the fillet is induction hardened to 58/62 Rc, is it safe to assume the tensile strength in that area is ~160,000psi? (I got that number from a heat treatment chart I found for 4340) If so, how deep does that hardness go?

- Chris

 

[TVP]

drillrig,

A hardness of 58 HRC corresponds to a tensile strength of ~ 300,000 psi; 160,000 psi is more like 35 HRC. The depth of induction hardening varies based on the inductor frequency & power, the scanning speed, and the quench process (media, temperature, flow rate, etc.). It is common for induction hardened shafts to have a case depth of several mm, so it could be as high as that or as low as 0.2-0.5 mm if desired.

 

[redpicker]

Make the part in two pieces. One, a hollow shaft. The other, a flange with a nipple about 6 to 8 inches long. Rough maching the flange with generous radii everywhere (0.5 inch or greater) to prevent cracking at heat treat. You might get lucky and find a oilfield wellhead valve body that might work, otherwise you'll probably have to go with a open die forging for the flange.

If you are only looking for 130 KSI ultimate, you might get 4130 to work, which can be water quenched.

Stub the two pieces together with 45,000 PSI sub-arc wire. Preheat both sections to 400-450 F prior to welding. With the section size and thickness, that should be fine to handle any torsional or bending loads in the constant diameter section. Post heat the entire assembly 50 degrees below the lowest tempering temperature and finish machine.

If you don't know any processors to do the work, look at oilfield wellhead service companies.

rp

 

[metengr]

I would agree with redpicker using two piece fabrication for this component geometry. However, I would add bit more caution as to the filler metal selection and weld joint geometry.

Either submerged arc (SAW) or flux core welding (FCAW) could be used for a 3" wall thickness at the flange, based on your information below

2.880".
yes the shaft is hollow (3" ID).
wall thickness where flange joins shaft = 3" (9" OD).
the adjacent steps are 8.900" OD, 8.750" OD.
the radius where flange joins shaft = .375".

The recommended filler metal would be one of the high strength filler metals that match close to the base metal strength for Q&T. Also, you need a full penetration weld versus a fillet weld to assure adequate torsional and axial load carrying capability.

If you are careful in your selection of filler metal, you won't need to induction harden the fillet itself.

 

[EdDanzer]

Are you sure you have applied all the loads as seen in use? If you are sure your analysis shows 45,000 psi there are less expensive materials than 4340 or 4130 that will provide a better part. The problem with machining the flange from the solid is grain structure. Most flanged axles are upset forged for grain structure. If you cannot justify the cost of forging this part make it from two pieces of 8620. This is a less expensive material with better machining and welding characteristics than 4130. Weld with a 70,000 yield wire using proper pre and post heat. For additional strength heat treat the 8620 to 80,000 yield and weld with .035 E80SD2 solid wire with no pass larger than ¼” and keeping the heat under 600°F.

Ed Danzer

http://www.danzcoinc.com

http://www.dehyds.com

 

[mfgenggear]

Ed

Are you sure 8620 is better than 4340?

This stuff (4340) is used a lot for Aircraft.
Good heat treat through hardness, Hardness 50 HRc or higher.
very stable at heat treat.

8620 is a like 9310 but not as good.
to my opinion, it more a carburizing steel.
It will distort more through Heat Treat.

You guys are the structural experts, but I worked a lot with
steel manufacturing.

Gentlemen
I would recommend a one-piece manufacturing.
Welding can be a pain. only weld when it’s your only choice.
Welding can also cause issues.

 

[redpicker]

I would agree with redpicker using two piece fabrication for this component geometry. However, I would add bit more caution as to the filler metal selection and weld joint geometry.

Weld wire selection is not my specialty, so I defer to those with better knowledge. In my experience, however, as you go with the higher strength wires, toughness drops. I have seen using a higher strength wire than is needed often cause problems. With over 53 in^2 of cross-section, if you stay away from stress concentrations, the service stresses should be low enough so I wouldn't think tensile strength to be an issue. You do want a high enough toughness so that weld defects do not result in a failure under loading conditions.

Yes, a full penetration weld is called for, in the cylindrical section, away from section size changes. I agree either FCAW or SAW could be used; I prefer SAW for these type of stub welds as I think you get faster deposition rates and end up with a cleaner weld. In either case, proper preheat and post heat would be required (even with 4130). You'll have to post heat, both to lower the hardness of the heat affected zones, and to relieve residual stresses from welding.

I would not recommend trying to induction harden the fillet. I don't think it would accomplish what you are trying to do (provide strength at the section change) and it would be difficult to accomplish. While a one-piece design would be desireable, it isn't as pratical.

rp

 

[drillrig]

All,

A belated Merry Christmas to you.

Thanks for the excellent responses. I have to admit, making this a two-piece shaft and welding together scares me. Perhaps this fear is unfounded, but at this point I'd rather keep things "simple" and stay with a single machined piece. In the past we've never made any of our drive shafts as a weldment, and this project is just in the prototype stage. Also, the shaft has to withstand a 5000psi internal working pressure, so if it was welded wouldn't the weld root cause issues (with a stress riser and a wear point for the internal fluid)? Lastly, the spec I am designing to imposes additional requirements ($$$) on the part if welding is involved i.e. radiography, etc.

An additional note: the material for this part must be Charpy tested @ -4deg F to withstand 31 ft-lb.

- Chris

 

[metengr]

Chris;
Your OP seems to be going full circle. It is reasonable to have some apprehension regarding a two piece welded fabrication. Just to clarify your main concern regarding the weld root; having a weld root that is properly fused with the base metal is not a concern for the internal pressure in your design. I work with many high pressure piping systems and we use full penetration circumferential butt welds. The weld roots are deposited using the GTAW process only to assure best quality for radiographic inspection requirements.

Fortunately for circumferential butt welds the service stress from internal pressure is only 50% of what a long seam butt weld would be exposed to from internal pressure. So, I see absolutely no concerns with a properly deposited and fused weld root for this application.

Welding does require expertise and from your post I would assume your organization has little experience with it, so I fully understand a one piece shaft.

 

[EdDanzer]

I would agree an upset forged 4340 part would be the best but for a prototype part that has 45ksi stress I would chose 8620 for weldability. The chemistry of 8620 is similar to A514 plate without the heat treat. Many small passes on a welding positioned using proper heat control will provide a part with better strength than cutting from a solid round at a reduced cost. Our experience building high stress (70ksi+) products with 8620 and A514 welded with E80SD2 solid wire with 90/10 shield gas has been very good.

Ed Danzer

http://www.danzcoinc.com

http://www.dehyds.com

 

[CoryPad]

I can understand your reticence regarding welding. Even if you decide to forgo forging or welding, there are techniques available to maximize strength at the fillet. Fillet rolling or roller burnishing can be used instead of just cutting the fillet. How close are you to forging and heat treating sources? Working closely with them will produce the best results.

Regards,

Cory

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

 

[unclesyd]

I agree with all the above in recommending a forging in lieu of machining from a bar.
As it seems that you are leaning towards using a bar I would recommend that you start with a preheat treated bar with a 135,000 tensile strength. You should be able to get this bar from the AM Castle Co.

http://www.amcastle.com/pdfs/AlloySteel_QG.pdf#

There are a lot of forging companies that could forge a blank. Click on the flanges section.

http://www.libertyforgeinc.com/

What is the overall length of the part?

 

[drillrig]

All,

Thanks again for the excellent responses.

I am seriously considering the welding option, but perhaps the biggest challenge will be convincing the other engineers here with 30+ years experience versus myself (3 years). I'm sure this would be well outside their "comfort" zone.

unclesyd,

Thank you for the AM Castle link. I didn't know they sell 4340 heat treated bars in that size. If I get an 18" OD 4340 heat treated bar, what do you think the properties would be at the 9" OD?

The shaft length is about 37". Since this is a prototype, is it worth the cost of getting a forging?

- Chris

 

[EdDanzer]

Drillrig,

You have not provided any information about the application, types of loads, safety factors or type of operating environment, nor a budget.

We had a flanged axle forged from 4140 for a tracked logging machine final drive (similar to an Army tank track drive) and found 10 pieces to be the smallest cost justifiable quantity. The one machined from the solid did not perform any better than the welded part but did cost all most as much as the forged part and neither machined from the solid or welded lasted as long as the forged part. All of these manufacturing methods broke at the beginning of the radius do to load induced fatigue.

Ed Danzer

http://www.danzcoinc.com

http://www.dehyds.com