Energy Release Due to Tensile Failure

Thanks, this type of discussion is really what I am hoping for. If you think of this a a tensile test to failure, that's close enough for now.

The equation I'm using is an approximation of the area under the under the stress strain curve, but may not subtract for the inelastic portion after yield. In as much as that increases the energy to be considered I'm willing to allow that cushion.

I understand that my entire system will be acting as a spring. I'm thinking there is a lot of math ahead in estimating the total stored energy and then how it will be re-absorbed into the system without breaking everything. What I am struggling with is how do I get from this release of energy back to forces and deflections in the structure. We have an FEA simulation program but I'd like to understand what it is doing and be able to do a manual estimate. At these magnitudes I feel that just trusting the software and in an adequate factor of safety is not enough.

It would be nice to perform a battery of actual destructive tests, but it doesn't sound like that is in the cards.

.

It is hard without knowing the set up what to expect, but with those constraints, you could consider the potential outcomes of scenarios where all the energy released accelerated just one of the components, or some portion of the broken part.

.

If you want to be very conservative (and very simple), consider what happens for each piece and each of various hypothetical broken parts, if a force equal to the tension at failure, were applied over the distance equal to the total elongation/strain for the system.

.

If things are around (or reasonably could be around) that won't take kindly to being impacted by any piece that happens to break or swing free, considering some restraints would be a good idea.

.

I'm not sure that is going to be that helpful. I feel a little like I'm blindfolded, whacking at a pinata.

Destructive testing is not in the cards, or at least not in the budget.

It seems like you and Welderman and I may thinking along similar lines. All that potential energy will convert to kinetic energy when released and the force of impact is the thing I'm looking for. Although I'm also looking to size shock absorbing devices.

I'm sure this is scalable and it would be interesting to measure the reaction forces at one end of a tensile test rig. Again, not in the budget.

We are considering restraints, or something, to protect the things around the failure as a separate issue. In this I am trying to make sure we don't break any million $ parts.

As for the pinata analogy, it really is as simple as I've described. We put a part in tension as part of our process and we have to assume it could break. The structure is (will be) designed to handle the static loads with a reasonable factor of safety but I'm afraid the forces resulting from breaking at these extreme tension loads will exceed that FS.

Thanks for the input.

JRW

Perhaps consider a slightly loose jacket for your weak part.

Maybe a jacket of two material like nomex and kevlar. The jacket should be loose enough so that the fibers aren't under tension until the part has elongated the maximum it normally would w/o breaking.

If the fibers were just getting tight when the part fails, the kevlar could elongate another 2 or 3% before failing itself, absorbing some of what would be release when the part fails. Smoothing and spreading out the transfer some.

Nomex will typically elongate >20% before failure, so unlike the kevlar, it won't dissipate much energy in the main mechanism, but it would be useful in catching any pieces from the failed part.

Perhaps two birds with one stone... or bag actually.

Are you by any means trying to establish/design a rated break point?

You will have to determine what the failure mode is for that break point and what loading there is for any of the other parts. You will have to model this with a finite element analyses, me thinks. But me also thinks that the energy will be used in breaking the part and is "wasted" so it will not be a problem for the other parts. At least that what I get a predetermined breaking point is for.

No, I'm not designing a rated break point. I have a theoretical worst case and am trying to estimate the damage, so to speak. Ideally we should never experience the break, but we also know from experience that it will happen sometime. Hopefully just not at the designed level.

See my response to "truth..." above. We have an FEA system I will use, but I should have a good idea what the result should look like before I run it.

Thanks for the response.

JRW

Let me try to understand:

You design a system with components that are suspectible to loading and if ever the load is too high you want a certain component to break and not the other ones and also you want to prevent the other components to take damage during the incident.

The first part sounds still like a rated breaking point, the second question will have to be answered by the FEA by looking at the whole system.

But I dont know your system. For example a chain will break at its weakest part. But it also will make that system chain none-functional. Do you need to retain functionality in the sytem?

What is the currently considerd safety factor and why? Making it higher will most likly cost more but will lessen the probabilty of system failure.

Or are you saying you take damage into account and your system is not designed with a safety above the theoretical worst case?

Hope its not a bridge!

Can the main load bearing component; if that's what this is, be contained in a fluidic environment?

Similarly,if a pump-shaft breaks will the driving motor attain higher speed and cause damage to its cooling fan,bearings etc?.

Not knowing what the part is I have no idea how to quantify what 3.2 mil ft-lbs actually means to it, but it is a substantial stress load in any industrial context, so if this is a turning shaft that's getting all this as torsion then it comes down to time. The faster it goes from static to full stress, the more likely a stress wave can be formed. Failure of a solid component in that scenario returns 30%.

Maybe this will add clarity to my OP.

The system I am considering is very much like a tensile tester. I am assuming anyone with a mechanical engineering education has seen one. My school used an (old even then) Instron vertical tensile tester and sometimes the whole building shook when a sample broke.

My system will be horizontal and is not intended to stress the tension member to breaking, but has to be designed to survive and be usable if it does break. The tension member is not flexible, so very little of the recoil will be absorbed by deformation.

So, without giving up proprietary information or being too specific about what I am building, I was hoping someone on CR4 with experience doing a similar analysis would recognize it and lend some advice.

There are two components to mechanical energy, potential and kinetic. The potential energy in (for example) drawing a bow is simply the integral of force, and distance of draw. It is measured in ft,lbs. The kinetic energy of the arrow depends on its mass and velocity(mvsquare). It follows that the reaction of your mechanism upon failure will have to be calculated as kinetic energy.

I like your analogy, as I imagine my structure will be flexing like a bow, but with the relatively small deflections you'd expect from a massive steel structure. I can estimate the structural deflections and add the stored energy of the tension member, so I could have the components to calculate the potential energy.

Looking at the kinetic energy of the structure slamming back into the foundation sounds reasonable.