Need help analysing a failure 2

[MJMinear]

Hi, I'm new here. I just found this site and it looks like a great place to kick some ideas around.

I have a project I've been working on for some time now and I thought I had the solution worked out but ther was a failure recently and I'm having a hard time figuring out why. Hopefully, if I can figure out why it failed, I'll also be able to figure out how to modify it to prevent future failures.

Ok, heres the situation. I have a part that I designed for my race car. It's a wheel hub. The OE parts were not up to the task of road racing with big sticky tires. The part I designed is forged out of 4340 steel, machined, through hardened to ~RC52 and then the critical dimensions are finish ground. The part has the wheel flange and shaft forged from a single piece of material. By my calculations, the maximum stress this part will see is 190,000 psi. The stress peaks at the radius between the shaft and the wheel flange as you might expect. According to the material data sheet I reviewed at my fabricators shop, the material used at the heat treated condition is supposed to be 250,000 psi. The failure appears to be a sudden break. Although there is some discoloration and what appears to be bits of red rust and patches of black oxidation in the break. This might mean that the part cracked earlier and eventually completely failed. But the hammered "ring" around the perimeter of the break that I would expect to see if that were the case is not there.

Questions:

1. Any ideas why this part broke?

2. Would induction hardening instead of through hardening be tougher?

3. Is the material just not strong enough, either due to an inferior piece of 4340, or is 4340 the wrong material?

Thanks in advance for any insight you may be able to provide. I'm sure there is more info you might need to help, so feel free to ask for any info I left out.

 

[swall]

I must disagree with tr1ntx. Higher fatigue strength comes with higher tensile, higher yield. This drives you to a fully hardened part. A non hardened part is the wrong way to go.

 

[Wrenchbender]

Refering your questions:
1. Part broke, even tho the applied stress in the bulk section was less than yield, because either there was an undetected crack, or the fracture toughness was inadequate (and allowed a crack to nucleate underthose conditions).

2. Don't know. 4340 steel is pretty 'deep-hardening' so you may still get thru hardness even with induction hardening and an air quench. Maybe something like 4140, 4130, would give a softer and tougher core. Check the literature, see what spindles are typically made of.

3. Is 4340 not strong, inferior or wrong? No, no, and maybe.

I think you need to do any combination of:
[a] inspect for pre existing surface flaws, esp. in the high stress areas (e.g., mag particle or dye penetrant). From the photos, looks like there may have been a crack that grew by high cycle fatigue.
eliminate high stress conditions thru redesign with large fillet radii, if possible.
[c] use material with a higher, plane strain fracture toughness. 4340 at 52 HRC will have very low K_Ic because of the inverse relationship. You really want toughness in this application. Do then drop the UTS and hardness. Otherwise you may have to go with another alloy like maraging steel ($).

 

[MJMinear]

Thanks for all the great replies. I've been without a PC for a few days so I just got a chance to read all of them.

The radius I'm using is pretty well maxed out within my space constraints. Besides, modeling shows that even doubling the radius results in only about 5% reduction is stress.

From the input above, I think the best thing to try next would be induction hardening instead of through hardening. Does anyone have any reference material on how to calculate the resulting strength and how to calculate the desired depth of hardening?

Thanks,

 

[swall]

You have to consider the section size of the material that will be hardened (the case) as well as the section size of the unhardened core. If you were to choose 1045, induction hardened, you would have a case that would run approximately Rc 53, which converts to 269,000 UTS. For 1045 hot rolled, which would be the core, an old handbook from Republic shows 45,000 psi tensile (typ) for this material.

 

[redpicker]

52 HRC is just too high for 4340 in a rotational bending application. At that hardness, the toughness is so low that even small imperfections will propagate and lead to failure.

Try dropping to hardness to 38/42 HRC and shot peening the radius afterwards. Look up Metal Improvement, Inc., for a shot peening vendor near you. This is much less expensive than induction hardening and avoids problems induction hardening could cause.


rp

 

[MJMinear]

The only problem with the HRC ~40 is that the listed tensile strength is right at my predicted peak stress (about 190 ksi), ie no safety factor at all. This level of stress is only reached momentarily, but may be reached many times per lap.

I studying the section, I think a 0.05" increase in the radius diameter can be achieved if I can find the appropriate seal. That would help a little. That coupled with the increased toughness at a lower HRC might be enough.

HRC 45 has a tensile strength of 228 ksi, Do you think that might be a better compromise between toughness and strength that will still maintaining enough UTS?

 

[metengr]

MJMinear;
If you absolutely need strength greater than 190 Ksi based on current design constraints, and you can't reduce stress concentration effectively, you may need to investigate using a maraging alloy, like Maraging 300. The UTS for this alloy is 290 Ksi min, with a yield of 285 Ksi, and it is tough material.

 

[tr1ntx]

I looked at this briefly and am currently at the opinion of saying you're not going to be able to make a part that won't break. You're going to be able to increase the time (load cycles) until it breaks using suggestions above, but at 190 ksi, it's going to break. I don't know about the Maraging Alloys though, maybe that will work.

This phenomenon is why well funded race teams throw way "perfectly good" used parts.

 

[MJMinear]

Are the specs on 300M you list for the base material or are they heat treated numbers? If 300M is required, that's what I'll do. Per my fabricator, it will add about $100 per part. They are already pretty expensive, so if it takes an extra $100 to make it right, that's right thing to do.

 

[metengr]

MJMinear;
What I mentioned is not 300M (a modified 4340), it is a different alloy containing much higher nickel and cobalt. It is more expensive than you think, but it will work under your conditions. It is strengthed by aging and not by conventional heat treatment for alloy steels.

You should look at 300M, but I am not sure that you will have the necessary through thickness mechanical properties and toughness that is needed for this service duty.

http://www.steelforge.com/alloys/?alloy=Maraging 300

http://www.dynamicmetals.net/

 

[unclesyd]

These papers discuss the effects of aforementioned shot peening on high strength 4340 steel.

http://www.metalimprovement.com/PDF_DOCS/The Effects of a Machining-

Like%20Scratch%20on%20the%20Fatigue%20Life%20of%204340%20Steel.pdf

http://www.shotpeener.com/library/pdf/2002006.pdf

http://www.gurteq.com/RAYSTRESS DOWNLOAD/RAYSTRESS 005.pdf

Paper on variable amplitude loading

http://www.eurojournals.com/ejsr_29_2_02.pdf

Here is an off the shelf Astralloy V bar material that I've used to overcome numerous problems with fatigue, especially in very high loads and load reversals. I have in-service highly loaded shafts that are over 25 years old.

The people at Astralloy should produce the numbers for the comparison data for Astralloy V versus other alloys.

http://www.astralloy.com/pdf/astralloy_v_plate.pdf

 

[rob768]

To me, all these measures as discussed will have influence on fatigue life. Considering the function of the part, it WILL fail through fatigue at some point.

 

[MikeHalloran]

Have you done an analysis with the mass of the spacer integrated with the mass of the stub axle?

I.e., I don't think the spacer buys you anything. At the thrust face of the bearing, put a radius just a tick larger than the bearing edge radius, and a shoulder big enough to pick up the thrust. You can still pick up whatever remains of your big radius, so on a macro scale, the shoulder becomes a bump on the radius.



Mike Halloran
Pembroke Pines, FL, USA