FEA fastener advice for an upright

[andrewkeisler]

Hello folks,

I'm running into an issue using fasteners in Solidworks FEA- specifically bolting a hub assembly to the upright. When I noticed excessive stress around and under the bolt head surface, I decided to run a simple study with tried and true components- A C5 corvette hub and upright vs. my more complicated design using bolts and spherical bearings at all fixtured locations. Both designs displace excessive stress under and around the hub mounting bolts. Below I have information regarding the basic corvette suspension study-


I'm using a remote load from the tire contact patch on the road surface to the hub face. Simultaneous 1G lateral, 1G longitudinal, and 2g bump force.
The lower ball joint taper is fixed in all 3 axis
The upper ball joint mating surface is fixed on X and Y axis- allowing for bump movement
The steering tie rod mating surface is fixed on only the Y axis
The mating surfaces between the two components are set to no penetration.
I'm using the counter bore screw fastener option for mounting the hub to the upright. I've selected the threaded portion of the hub assembly for bolt contact and the bolt hole circle on the upright for the bolt head mating surface. The 3 M12 fasteners are set to 96 ft lbs (GM spec)
I've selected the hub to be rigid so that I may focus directly on the issue I'm finding on the upright.

With a von mises static loading scenario, you may see the area around the head of the bolts have reached yield while the rest of the upright has minimal stress throughout.

I understand many of you may not use solidworks FEA, but I was hoping to receive some insight as to how you would potentially run this study differently, or if the study seems accurate and I just woke up on the wrong end of the bed today.

https://files.engineering.com/getfile.aspx?folder=97e7c125-d415-4f79-aa03-a1caeef7e53d&file=FEA1.png

https://files.engineering.com/getfile.aspx?folder=a2a3cfa5-0160-40ac-8029-ecd31b6dd5c8&file=FEA2.png

Thanks guys

 

[andrewkeisler]

Here is a photo showing the fastener selection method-

https://files.engineering.com/getfile.aspx?folder=ea32e7f9-120d-4e79-8691-2aa6d33bf81a&file=FEA3.png

 

[GregLocock]

" Simultaneous 1G lateral, 1G longitudinal, and 2g bump force."

That might work as a first pass load case but it is important to look at each direction separately, and combine them in realistic ways, in both directions.

Also your loads are way off what a road car sees, and are low for even a circuit car , which can often be successfully designed using 1 2 3

3 off 12mm bolts to hold the hub unit to the spindle seems a bit light on, but if that is the actual Corvette design I won't argue.

Sorry can't help you with FEA, fastener modelling is not my thing. It is worth back checking your expected axial stresses in the bolts with your loads, which will obviously be easier if you concentrate on uniaxial loads.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[GregLocock]

The fit of the hub unit in to the spindle may be a press fit, in which case the loads are taken directly, not via the bolts.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[gruntguru]

Local contact stresses on a small scale (as you are seeing) can usually be ignored. If that area does yield, the load is immediately distributed to other regions of the bolt-upright contact surface. If the stress in the shank of the bolt is well within limits, the underside of the bolt head is not likely to fail the upright.

It might be worth posting in the FEA forum. Link

je suis charlie

 

[andrewkeisler]

Greg,

I appreciate the insight.

I understand the simultaneous forces aren't exactly ideal. At minimum, a 2g bump with 1.5g lat or long force should be far below yielding the upright. It's fairly common to see TT cars, corvettes included, experiencing 1.3 -1.5 g lateral acceleration with hoosiers and big aero. Mix with aggressive curbing on the rumble strips can give 2g bump. Said OEM uprights withstand these forces lap after lap digging deep into the S-N curve without failure. This leads me to believe the bolt connector modeling is simply not accurate especially given so much stress concentration around the bolts while the rest of the upright shows to be well over engineered (better yet under engineered hub mounting bolt size).

The corvettes are a bit of an odd ball to use 3 12mm mounting bolts. I've found 4 10mm fasteners mounting the hub assemblies to uprights are more common on Asian cars (Mazda, Toyota, Honda, etc)

Great point on the press fit. This style of hub assembly typically has a .3"- .5" boss which resides in the upright with an approximate .005"-.01" clearance on the diameter. Many moons ago I worked as a suspension technician. When replacing these hub assemblies on cars with steel uprights, they were almost always seized together on removal. Once I'd get them separated and the mating surface cleaned on the upright, the new hub would drop right in the bore. That's not to say under abnormally high loading scenarios that the hub and upright wouldn't deform enough to allow the hub boss to come in contact with the upright which may reduce stress on the mounting hardware and surfaces.

Hopefully someone will chime in regarding the fastener settings. I played around with selecting the thru hole in the upright as a threaded "joined" feature to the fastener. When running the FEA this way, the results seem much more accurate regarding the stress displaced throughout the upright instead of concentrated around the bolt head. Mehhhh

 

[andrewkeisler]

je suis charlie,

I appreciate the reply. I went down a rabbit hole last night reading about the misguided affects of local stress around bolted surfaces. I'll post in the FEA forum for some added advice.

https://www.eng-tips.com/viewthread.cfm?qid=237265

https://www.eng-tips.com/viewthread.cfm?qid=127139

Best regards,
Andrew

 

[SwinnyGG]

As far as I can tell the only places you have stresses which approach or exceed yield stress are right on the corners of your fasteners. Two things should ease your mind here:

1) gruntguru's point about contact stresses is 100% correct. Few FEA systems are really good at dealing with stresses approaching hertzian conditions. Solidworks' solver is not one of them.

2) these stresses are impacted by the perfect, sharp edge on the head of each fastener. In the real work, these perfect 90 degree edges do not exist.

 

[BUGGAR]

Holes in FEA have always been a problem as you found. Now I check my models for overstress at the holes as a means of verifying that my model is correct. The Corvette hubs are a tight fit but not a press fit.

 

[CWB1]

Solidworks does a pretty poor job of modeling surfaces period, that's why its typically a limited-use shop floor tool and not used by engineering nor for analysis work. As mentioned, it also appears that you're using really simplified fastener models that don't have the underhead chamfers or bearing surfaces modeled correctly.