Are you a mechanical engineer? If so...

would you possibly be willing to do a stress analysis for me? If so please PM for the details. Thanks

 

I think you are looking for a strucural engineer.

 

I ran a simple finite element analysis (FEA) on this home-built front axle for JPMACHADO for those interested. This analysis was strictly done for a static scenario to see what "hot spots" or high-stress areas might exist. I assumed the car weighs about 1,500 pounds and that 60% of the weight was up front (900 pounds up front, 450 pounds at each front corner). I applied twice the weight (2 G forces) upward on the axle king pin boss to simulate the wheel pushing upward during a pothole. I simulated lateral support from the 4 bar links pushing on the rearside of the axle (red plates in pictures) and a bolt simulating the vertical resistance from the coilover shock. I also put a 100 pound force pushing laterally backward on the spindle boss to simulate hitting a block in the road.
The approximate yield strength (where a material deforms and does not go back to original shape) of mild steel is about 53,000 psi and this analysis at 2G forces revealed a maximum stress of about 37,000 psi meaning that theoretically nothing will bend or break. The dark blue color is the lowest stress followed by lighter blue, greens, yellows and finally red with red being the highest stress areas. I noticed a little "hot spot" in a sharp corner on the topside of the axle (visible in pics) so I put in a little reinforcement plate (little yellow plate in pics) to see if I could get rid of that hot spot. Without the little reinforcement plate and at 4G forces (4 times weight), the highest stress in that little hot spot corner was 72,000 psi (failure of the steel at that point). With the reinforcement plate alleviating those sharp corners the stress was cut almost in half to 37,000 psi. Other than the high stress in those stress-riser corners, the axle appears to be well-built. This analysis is just an educated guess at what is happening in the axle and doesn't account for fatigue, weld quality or imperfections in the metal which could both cause failure. This type of analysis is good for finding high-stress areas and trying to find ways to make the parts stronger. It could also be used for finding low or no stress areas for cutting out lightening holes or removing metal that is not adding strength to the part. Anyway, here are some pictures and a link to a short video of the axle under load (the motion is exagerated because the axle moves very little and would not be very visible without scaling up the motion).

http://s156.photobucket.com/albums/t6/henruco/Axle FEA/?action=view&current=axle-fea-video.flv

 

very cool

 

I just wanted to take this chance to thank el chuco. The work he did for me is amazing. When I originally asked about someone analyzing the axle I had no idea I'd get something like this back. It's pretty awesome, and I'm very greatful.

 

which sae alloy did you use for your yield and ultimate stresses, and was it hot rolled or cold rolled?

 

I selected a predefined material (high strength, low alloy steel) from the material library in the software I used (Pro/Mechanica) with a Young's modulus of 1.11966e^10 to get me in the ballpark. Assuming that the material used for the axle is the ultra common steel supply store grade (1018 mild steel), I compared the maximum stress values from the analysis to the yield strength of that grade of steel which is about 53,000 psi to get a feel for the strength of this axle.

 

Hey, no problem, JP. Anything to try and help a fellow car builder understand this stuff. You should have a good feel for the workings of a car seeing as how you built a car from scratch!

 

53,000 just seems kind of high. The central steel book lists 1018 as yield at 35,000 and 50,000 ultimate for hot rolled, and 60,000 yield and 70,000 ultimate for cold drawn.
assuming ultra common supply store grade, I'd err towards it being generic a36, at 36,000 yield. Did you define the failure mode as plastic deformation, or complete mechanical failure?

the mechanica plot definitely points out the problem areas, but if any design decisions are going to be made, make sure you look at what you want your factor of safety to be (allowable vs. actual or estimated stress).

just for comparison, for fatigue life analysis on a particular brand of mining equipment, it isn't uncommon to raise the G level to 8 to account for impact loading.

 

just a general comment on the design.....I would expect the uprights to be fully boxed....

 

Thanks for the constructive feedback, fellas. JPMACHADO informed me that his upright plates are in fact A36. Other than the "hot spot" that appeared in the sharp corners, the rest of the axle appears to remain below the 36,000 psi yield for A36 at 4G static loading. If those sharp corners can be addressed with some reinforcement plates the axle will be stronger no doubt. I didn't tell the software to find the failure point and only asked it to solve for the stresses.

Boxing the upright plates is a great suggestion. I'll box them up in the axle model and re-run the analysis to see what happens. This info. is very valuable to JP and I think he will make these changes to his axle.

 

I've also modeled in the weld fillet, pro/mech doesn't like sharp corners very much and will always show a high stress there. With a weld part that you assemble you can also add the weld alloy material for the results. From one pro/mech user (and former traniner) it looks pretty good!

 

Messed with this axle analysis a little more for JPMACHADO. I added some round tubing cross-supports between the upright plates to the model that actually exist in JP's axle. I re-ran the analysis with those round tubes in there and the maximum stress was lower than without them (good design JP).

Then, I boxed the axle on the outer side (facing the wheel) and re-ran the analysis. Very interesting how by adding this outer plate, the stress is reduced in the upright plates and so there's more dark blue color in the uprights (dark blue=low stress).

Finally, I added an inner boxing plate on the other side of the upright plates which helped lower the stress a little more.
This analysis was done assuming an agressive 4G forces (4 times weight of car at front right corner--1,800 pounds). The highest stress value with the upright plates boxed was 26,720 psi which is below the yield strength of A36 steel (36,000 psi). Also, this highest stress is located in a small reinforcement plate added to the topside of the axle which is far better than the stress being in the upright plates or king pin boss. Here are some images showing all this and three links to some short videos. I need a cold cerbeza right about now .

http://s156.photobucket.com/albums/t6/henruco/Axle FEA/

 

How would that look on a bent tube axle?

I'm all for making stuff yourself, but this is looking like a ton of fab work when there are easier solutions out there.

A tube axle could be made in the home shop with a little fixture improvisation and would be a much more efficient use of materials (less unsprung mass, flex and better handling). I bet you could get your stresses to come out better, too.

Why do you think 4G is 'aggressive'?

 

I'm so confused, but it looks very cool.

 

Let me start off by saying I in no way take any offense to the comments in this thread. I really do appreciate the advise. However, the axle was built around Mustang II parts I already had. I needed a 7" drop with a king pin to king pin distance of about 58". I looked around and nobody had an axle like this. Also, I called some places with mandrel benders to have a true tube axle bent up, but none of the places could give me the tight radiuses I needed. Sometimes when you build a car you have to make compromises. I'm very happy with the way the axle turned out, but I can certainly understand it may not be everyone's cup of tea.

 

Thanks for doing the work for us all here I too would like to see stress tests on a tube style axle. Thanks again . Has there been some reported axle failures? That would be an intersestinf thread. John

 

I'll try to find a drawing of a tube axle and model one up to see what the stresses look like. I've seen an axle analysis thread quickly turn into a "well I want to see pictures of a failed axle" deal but that comes with the territory. Some of us don't accept someone else telling us how or why something is the way it is or that something won't work, blah blah blah...

I say 4Gs is "agressive" because I seriously doubt the car will ever experience that in normal on-road driving avoiding curbs and other heavy obstacles (school me if I'm wrong with some info.). In JP's case, he's using 465 lb/in spring rate coilovers (not sure about the preload he's running since he's using adjustable race-type coilovers). The coilover shock absorber is what's keeping the tire from bouncing upward freely. At 4Gs (1,800 lb) in a non-impact type situation the spring will compress 1,800 lb/465 lb/in = 3.9 inches which is a lot of supsension travel for this type of car. If JP hits a speedbump or curb while going pretty fast, the shock will become stiff as the wheel bounces upward because shocks are speed-sensitive but that's a whole nuther situation that I hope JP never gets into. The tires will take the brunt of the beating if he hits something in the road and then the suspension will get it after the tire is done deforming or blowing out.

 

That's a lot of work to set that all up--kudos to you for helping a HAMB'er out like this. I saw this original post and knew I didn't have time to set up the problem--looks like you've done a good job.

Keep in mind that yield strength only accounts for the first hundred or so times you apply the load. Beyond that, you've got to think about fatigue. The endurance limit (the stress at which the material will last forever) of most steels is about 50% of its ultimate strength, so if A36 has an ultimate of 58 kpsi, then its endurance limit is 29 ksi. Considering your analysis of the modified version, a 26.72 ksi peak load only gives you a 1.08 safety factor, and that's on a fabricated part with a few unknowns/assumptions.

Now, is the axle going to see that extreme 4G max load a million times in its life? Doubtful, but it's still a pretty low safety factor. I don't know enough about suspension design to know what a realistic "normal driving" load is, but even if it's 3G, that safety factor probably isn't above the 1.5 I'd use as a minimum if it was my car. But it's not my car.

 

Just a few comments:

You should do a fatigue life analysis on that axle, as it's surely going to be experiencing cyclical loading. Having the maximum stress below yield doesn't necessarily make this a good design.

I don't think a purely static analysis is sufficient either. I think the unsprung mass of the chassis will come into play as the springs compress. The suspension will have to control that mass as it accelerates up and down under load.

Are the coilovers progressively wound? I suspect they are, which means that the suspension may not travel the 3.9 inches that you've assumed. The spring rate will increase as they compress. You said yourself that the dampers will stiffen up under an impact load. Keep that in mind too.

Have you confirmed that the car weighs 1500 lb, and that your assumed weight distribution is reasonable? I know from previous post that it's just a homemade roadster body on a frame, but it's a big one. Much larger than a Model A.

Is a 100 lb side load reasonable? That seems low to me.

Have you looked at the deformation of the axle to determine how it will behave under load? Will it experience large camber changes as it flexes? This could cause handling issues.

Finally, I would suggest backing up your FEA work with some hand calculations, if you haven't already. FEA is a marvellous tool, but it's very sensitive to input variables and therefore can easily turn into a "garbage in, garbage out" situation, where the initial assumptions are out to lunch and the result is a very pretty but meaningless solution.

You'll need to pick a design factor to cover the variables that can't be accounted for. A good start though. Keep at it.

 

Watch your stresses on the welds. Weld material does not have the same qualities as the parent material. It is true that the weld is typically stronger.... from an applied load point of view.... however from a fatigue standpoint your in trouble. I would dissagree on the 4g load being agressive. In my experience a railroad crossing or big pot hole can generate 15g + loads on the axle. Higher if you measure it at the wheel end. That would be 15g's of the weight of the axle. The vehicle itself would be in 4 g range. It all depends on your approach for the analysis.

Just my 2 cents.

 

This is all very good information. Originally I had written Ron Covell, the famous metal working guru, and he had suggested this course of analysis. However, he did say it would be helpful to model a normal axle, with known capabilities, to for comparison. Does anyone have an axle model? If we had a model of a traditional I-beam or tube axle (even one of the simple tube axles Speedway sells) we would have a workable base line

 

I have modeled a CE axle for referance for my tudor build. It does not have the mounting holes in it but i could throw them in and run a quick analysis if you would like.

 

I tend to agree. Even if an analysis shows that it won't suffer a catastrophic failure, it certainly isn't an elegant solution. This is more of a "throw metal at it" approach.

If I'm not mistaken, the axle is already built.

 

First off, I would look to say good job. I had originally signed up to do this analysis as well, but my current workload has been keeping me way too busy.

Here is one question and one comment.

Q. What modeling software are you using? I had planned to use either UG or Ansys to run this simulation. If it is in either one of these formats is it possible to get a copy of your input file?

C. I agree that fatigue life must be addressed, but obtaining a random vibration input (amplitude vs. frequency) curve would be a challenge. I know some guys who do vibration testing on bikes and I might be able to squeeze a generic curve out of them to use as an input for this model. (If I could get a curve could you use it?)

Thanks for keeping us geeks busy...

 

any stock front axle from the 40s-50s is a well engineered part, that's what I would be copying

 

This would be great. Will the analysis be simiar to the others? I just want to compare apples to apples. I'm not smart enough to understand it otherwise.

 

It's a pretty thick thread, but I posted my reasons earlier. Thanks

 

I agree -- Impact loads are going to be far more than 4x. What are the forces required to raise a wheel assembly and the unsprung weight when the wheel drops into a pothole 1" deep and hits the far side of the hole at 60 mph? The whole assembly has to be raised 1" very quickly (like in .006 seconds at 60 mph). Naturally the tire absorbs a lot of it. But that's not the worst pothole you'll encouonter, either.