Minimum Web Thickness 3

[JpPhysics]

Ok, this should be simple, but I am having some difficulty.

I am trying to dictate to my designers the minimum amout of material below the head of a bolt based on the torque applied to the bolt.

T = (0.2)(Dia)(Force) <- Gives the Force due to the Torque

Then I model the plate as a simply supported beam, but when it is all said and done, I get a safety factor of 0.2.

Any suggestions?

 

[JpPhysics]

A bit of follow up, the plate is not supported directly under the bolt.

 

[desertfox]

JpPhysics

Can you provide a sketch of what you are describing because I cannot visualise what your saying. I am not sure modelling the plate as a simple supported without more information.

regards
desertfox

 

[Tmoose]

You want to prevent bending (yielding) the plate?

What is torquing the bolt accomplishing, and what does the bolt do?

 

[drawoh]

JpPhysics,

We are tying to understand your model.

It sounds like you have a plate with a hole in it. The plate is supported some distance from the hole. A bolt is inserted into the hole and tightened down hard.

Is this right?

JHG

 

[JpPhysics]

The drawing is a sample of what I am looking for. I am tring to solve in closed form what the minimum thickness is below the head of the bolt.

I don't wan't to use FEA for this if I can avoid it.

 

[JpPhysics]

Here is the file again as an eDrawing.

 

[rb1957]

if the flange under the bolt-head isn't clamped (butting up against something else), like in a conventional bolt, then i don't think the conventional bolt preload equation applies, 'cause there's so much more flexibility in the joint.

in a convention joint the bolt torques clamps the flanges together. in your case it sounds like the structure reacting the bolt preload is a flange in bending. your case, if i have it right, sounds a bit "funky" ... i can see that you might have an unsupported flange but i would either tighten the bolt to the flange (using an anchor nut) or let the bolt pass though the flange an tighten it up with a nut on a supported flange. that being said, you situation could be that the loads are known to be very low and you don't want the complexity of these other designs.

 

[JpPhysics]

rb1957,

We are using the bolt to apply a clamping load between componenets. We use grade 8 bolts to protect against fatigue failure because we cannot tighten the bolts up to proof load.

The design is "Funky" which is why I am having trouble.

Thanks for the reply,

John

 

[rb1957]

why do you need 6tons of preload ?

in any case, i don't think the conventional preload equation holds ...

 

[JpPhysics]

Because this is a poor design.

This is an antiquated system that I have to support and modify.

The goal is to get 2000 lb/inch at the left side of the drawing. Since my lever is backwards, I loose a lot of force going from the bolt to the target location.

 

[BobM3]

For sure you would need enough material to prevent shearing of the piece under the bolt head. You might also want to take a look at Roark's equations for flat plates with center loads.

 

[rb1957]

if you need the load at the LH reaction, and you're relying on the preload, i'd check your assumption ... maybe use a strain-gauged bolt to verify (and replace with a standard bolt for service). personally, i wouldn't bother with Roark ... this thing looks like a beam, use something like the minimum thickness to calc I ... should be near enough.

 

[drawoh]

BobM3,

Roarks's equations for flat plates are for flat plates.

The part does not look like a flat plate to me. I think the OP needs to tell his boss to choose between FEA and a new part.

Would this thing fail in shear around the head of the bolt?

JHG

JHG

 

[JpPhysics]

We have had some systems fail around the head in shear. Others would deform around the bolt head which causes a loss of clamping force.

The problem with modeling this as a simply supported beam is that I get very large bending moments on the order of 300 ksi which I know is not correct. I guess the problem is I need to use a simply supported beam, but I cannot model the bolt as a Point Load, it needs to be a distributed load. Any suggestions?

Like I said, FEA is an option, but what I am trying to come up with is some way to easily say: If you use a 5/8 bolt then you need this much material below the bolt head. If you step up to a 3/4 bolt then you need this much material...etc.

 

[rb1957]

how would the guy tightening the bolt know when he has applied 1 ton/in at the reaction edge ? maybe they're over-torquing the bolt ... maybe you specify a torque, but it gets loose so they tighten it up more ??

analytically its a pretty easy problem ... a single load, a simple reaction, ok a little fussy about the section in bending. distributing the load under the head of the bolt probably won't change things by more than a few %age points.

 

[drawoh]

JpPhysics,

You cannot model your structure as a simply supported beam unless it is a beam. The beam equations in the handbooks are based on double integration method. Read up on it and make sure you understand all the assumptions.

JHG

 

[desertfox]

JpPhysics

Thanks for posting the drawing.
I do not believe you can use simple beam theory because the length to depth ratio of the component looks to small. Secondly I am not sure how your calculating your bending stress but you would need to know the position of the neutral axis for that component shape shown on the drawing assuming beam theory was applicable.
Thirdly your supporting reactions would depend on the stiffness of the supports and those of the component shown within the supporting region.
Finally the central bolt, you don't say by what method your pre-loading the bolt but if I assume torque wrench then this method as a error of +/- 25%, so on one clamp it could be 25% higher than you need and on another 25% less, not including thread tolerances, lubrication, and surface finish on bolt and mating thread, embedding of bolt head etc.
I think if you want an accurate answer then do some FEA with the parts modelled correctly so that you can obtain deflections of the clamp and its supporting components.
If the clamping force needs to be accurate then you need to do some practical tests and come up with a method of assembly that reduces the error in your bolt pre-load rather than just relying on the torque wrench.
If you have had failure's of the clamp component around the head of the bolt, I fail to understand why you use a grade 8 bolt to protect against fatigue failure when you make no mention of bolts failing or have I missed something?
What is the cyclic load on the clamp? maybe you should be looking at a fatigue analysis on the clamp instead of a static analysis.

regards

desertfox

 

[JpPhysics]

desertfox,

Thanks for the reply.

This is what I get for posting half a question.

The drawing is simply a sample of one of the many products we offer that use this type of clamping configuration. All of these products have been used for more than 20 years. In the early days, these products were not "Engineered" they were simply "beefed" up until they held together. The old trial and error method.

Enter 2009... how can we save some money while maintaining system stability.

These bolts are torqued to 120 ft-lbf. We have done enough tests to determine that the clamping force we get at the tip of the clamp is sufficient.

These components spin at anywhere from 1100 to 2500 rpm and take an impact load equivalent to 1000 lb per inch all day long. (2 to 3 times per day the system is shut down for exactly 1 hour.) Under these conditions, the system shown will last for a couple of years. In fact, the clamp will wear out before it fails. (Now you can see why we use Grade 8 Bolts. Additionally, these components are sharp so having one go flying around and hitting someone is not generally a good idea. Better safe than sorry.)

The bolt is removed one to two times per day then re-torqued depending on the severity of the work being done. The bolt is not replaced on a regular basis, only if it breaks or is visibley worn on the head.

Moving forward, we are designing new systems that do not allow for a lot of material under the head of the bolt. I am the only person capable of running the FEA software at my work, so I become a major bottleneck when we have 10 to 20 new projects running at a time.

My thought was, what if I could get my designers (not equal to engineers) a chart or equation that they could use to get close to the correct amount of material below the head of the bolt thereby reducing the number of simulations that I would have to run.

Does anyone have any suggestions or rules of thumb that we could use or am I simpy doomed to run simulations all day long?

 

[hydtools]

Is this a machine tool hold down clamp?
What is the mode of failure you are using to calculate your FS or design margin?
Were the bolt failures really shear failures due to overtorque or in fact tension failures?
The resistance to bending will depend on the overall clamp thickness as well as the thickness under the head of the bolt. How have you figured the overall clamp thickness into your calculations?
You can use the bolt torque-tension equation to estimate the pull down force.

Ted