×
INTELLIGENT WORK FORUMS
FOR ENGINEERING PROFESSIONALS

Contact US

Log In

Come Join Us!

Are you an
Engineering professional?
Join Eng-Tips Forums!
  • Talk With Other Members
  • Be Notified Of Responses
    To Your Posts
  • Keyword Search
  • One-Click Access To Your
    Favorite Forums
  • Automated Signatures
    On Your Posts
  • Best Of All, It's Free!

*Eng-Tips's functionality depends on members receiving e-mail. By joining you are opting in to receive e-mail.

Posting Guidelines

Promoting, selling, recruiting, coursework and thesis posting is forbidden.

Students Click Here

bolt pretension, AISC, and FEA
8

bolt pretension, AISC, and FEA

bolt pretension, AISC, and FEA

(OP)
We often analyze pretensioned bolted assemblies in FEA. In general these are bolted joints that can be loaded in the 3 force directions and 3 moment directions all simultaneously. I am trying to develp a method to extract the force from the bolt given by FEA and perform AISC bolt checks.

The issue is how to deal with the preload. AISC says not to consider the preload when checking the bolt capacities. But the FEA model has preload incorporated in the bolt force results. I tried a bolt joint stiffness approach but I wanted to see what others are doing in their work. Thanks.

RE: bolt pretension, AISC, and FEA

well, for one thing I wouldn't have fasteners reacting all 6 dof.

2, when you run FEA with preload, presumably you use the nominal preload. what about the extremes of scatter ?

3, your AISC is saying "assume the joint is gapped" which is typically the most critical state of the joint.

4, so change your FEA to have no preload (I mean you're doing something to preload the FEM, right?), and do the AISC checks. Run your FEA with nominal preload, check the fasteners again; did the MS go up ? (Sod's Law says go will be higher but some will be lower, so did the critical MS increase ?)

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

RE: bolt pretension, AISC, and FEA

(OP)
Our structures experience accelerations in all three directions; the joints are loaded multi-axially, there's no help for it.

If I run without preload the load change in the bolts is large and the load change drops as more preload is added.

If controlling pretension by torque we allow for a 30% scatter unless it's pressure containing in which case we use hydraulic tensioners.

RE: bolt pretension, AISC, and FEA

of course joints are loaded in "multi-axially" (or by a generalised force vector) ... but that is not the same as rigidly constraining all 6 dof.
Particularly if you constrain multiple points ... then your over-constraint causes fictitious redundant loads.

When you say "load changein the bolts" ... are you thinking of the fatigue of the bolts? Correct, but then you model without preload to satisfy your ASIC and with preload (minimum) for fatigue analysis.

Thining about it, preload effects in the FEM should be highly localised to the fastener ... so you should be able to extract the external load applied to the preloaded bolt which would be the load to use in your AISC calc.

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

RE: bolt pretension, AISC, and FEA

if the design is that sensitive to preload, you should be designing to two enveloping load cases - one with preload and one without preload. you really can't ensure preload over the life of the structure without some extraordinary measures.

RE: bolt pretension, AISC, and FEA

If you don't mind me asking: why the FEA check on a situation that can be resolved with hand calcs (at least as far as the bolts go) pretty fast? Is it something beyond the bolts you are worried about? (Maybe the frame being pinched?) Or is it a situation where the model is already set up and you just want to extract the forces?

RE: bolt pretension, AISC, and FEA

Outside of a PhD thesis or similar academic research project, I don't think I would ever encourage anyone to model individual fasteners like that. Maybe if they're trying for a "proof of concept" for a proprietary connection. But, that's basically the same category as a PhD thesis.

If you do go down this road, I have to wonder what kind of program you're using. Pre-tension would be related to increasing friction between connected parts and such. Does your FEA consider friction in connected parts? If so, that's a brutal model. So much effort.

RE: bolt pretension, AISC, and FEA

(OP)
It is very easy to export a 3D model to ANSYS and apply pretensioned bolts. That's what most of our guys, expecially the younger ones, are going to do. Sometimes we also need to do hand calcs.

Modeling mehodology or statements about PHd thesis was not in the scope of the question.

To repeat, the question is what force to extract from FEA to use in AISC bolt calc.

If I pretension a bolt in ANSYS, say to 10 kips, and after running the analysis under load my total bolt load increases to 12 kip, what load goes into the bolt hand calc? It's not the 12 kip because that includes preload effects. Nor is it 12-10=2 kip as this reduces the load to nothing.

RE: bolt pretension, AISC, and FEA

pylandj1,

Take a look at the information in the link: https://www.kitplanes.com/stressing-structure-13/

When you pre-load a bolt, applying a tension force (even a large one) marginally increases the tensile stress within the bolt until the clamping force is overcome, at which point the tensile force will continue to increase until you reach the ultimate tensile strength of the bolt.

So to answer your question in your provided example, the force used to design the bolt should be 12 kips. However because this involves FEA, there could be a multitude of things that could lead to erroneous results. As mentioned by others, what does a hand calculation show you should expect? You mention that this connection experiences forces in all 3 directions which might make it difficult to compare to hand calculation, but what if you modeled a simpler connection under a simpler stress state that you could easily check the modeling approach to something that can calculate by hand.

That will give you an indication of what the design bolt force should be.

RE: bolt pretension, AISC, and FEA

Quote (pylandj1)

If I pretension a bolt in ANSYS, say to 10 kips, and after running the analysis under load my total bolt load increases to 12 kip, what load goes into the bolt hand calc? It's not the 12 kip because that includes preload effects.

I'd presume it is the 12kip, unless you aren't modelling things suitably or your software is kinda screwy.

Quote (pylandj1)

AISC says not to consider the preload when checking the bolt capacities.

....Nor is it 12-10=2 kip as this reduces the load to nothing.
It seems like you are misunderstanding how bolt preload works. You seem to thing pre tensioning or preloading a bolt results in additive forces but we ignore the preload because AISC says to. THIS IS INCORRECT. The preload force in your bolt doesn't disappear it is still a load on your bolt. It is just that pretension forces are not additive.

Pretensioning a bolt by 10kips and then loading the bolted plies in tension by 4 kips will result in 10kips of tension in the bolt.

Pretensioning a bolt by 10kips and then loading the bolted plies in tension by 12 kips will result in 12kips of tension in the bolt.

There are various things online that explain this:
https://www.nord-lock.com/insights/bolting-tips/20...

RE: bolt pretension, AISC, and FEA

"It is very easy to export a 3D model to ANSYS and apply pretensioned bolts." ... oh dear ...
"That's what most of our guys, expecially the younger ones, are going to do." ...oh dear, oh dear, oh dear ...

Yeah, we're getting very lazy ... sure it is easy to import a solid into FEA and analyze it ... doesn't mean you should. It also matters how you do it.
Analyzing multiple fasteners as 6 dof rigid constraints is IMHO, wrong.
Analyzing the effects of preload inside FEA is fraught with issues.

But I hear you say "but preload will change the stiffness and so the FEA results". Sure, but then you have to commit to analyzing a myriad of possibilities.

I would recommend not analyzing preload inside FEA (no matter what FEA salesmen tell you) and to analyze preload effects (which are mostly beneficial for fatigue) by hand. I would contend that such modelling is conservative.

Typically I wouldn't model fasteners, but if I had to I'd use a finite stiffness (like a CBUSH).

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

RE: bolt pretension, AISC, and FEA

Quote (pylandj1)

It is very easy to export a 3D model to ANSYS and apply pretensioned bolts. That's what most of our guys, expecially the younger ones, are going to do. Sometimes we also need to do hand calcs.

Modeling mehodology or statements about PHd thesis was not in the scope of the question.

To repeat, the question is what force to extract from FEA to use in AISC bolt calc.

Tension in the bolt "element" at various stages of your analysis is the tension in the bolt that you will use for checking against AISC tension limits. It seems like a shockingly simple question.

Keep in mind that you're asking a forum of structural engineers what you should do. We're going to question your judgment or give you advice about what you're doing if what you're asking about is seems like an inherently a bad idea to us.

Honestly, if "most of your guys, especially the younger ones" are going to do an FEA analysis for something like this then I think there is something pretty wrong with the way your company is doing things. Period. I know this comes off as very harsh. But, it genuinely sounds like you guys don't know what you're doing, but are doing it frequently and with false confidence in your abilities to model and interpret the results of these models.

We're structural engineers. Bolted connections are almost always designed by hand using standard simplifications that are codified by AISC (or whichever governing body releases steel code design criteria in your area). To most of us (well, at least me) doing an FEA analysis like this is inherently a bad idea. Partly because it's a waste of time and energy compared to what a simple hand calc will do. Partly because there are way, way, way more things that can go wrong when you design something this way.

Quote (pylandj1)

If I pretension a bolt in ANSYS, say to 10 kips, and after running the analysis under load my total bolt load increases to 12 kip, what load goes into the bolt hand calc? It's not the 12 kip because that includes preload effects.

Yes, it's the 12 kips.... see what I mean when I say that so much can go wrong. It's not just the input of loading, or the material assumptions, or contact friction or 100 other modeling assumptions you have. It's also the INTERPRETATION of your results.

I suggest you run away from this type of analysis as fast as you can and return to basic principles and hand calculations before you create a huge problem / failure by having too much confidence in a flawed FEA analysis done by a new grad FEM jockey. If you think your liability insurance will cover gross negligence, then maybe it's safer for you to continue. But, I don't think it's truly safe...



RE: bolt pretension, AISC, and FEA

Quote:

To repeat, the question is what force to extract from FEA to use in AISC bolt calc.

If I pretension a bolt in ANSYS, say to 10 kips, and after running the analysis under load my total bolt load increases to 12 kip, what load goes into the bolt hand calc? It's not the 12 kip because that includes preload effects. Nor is it 12-10=2 kip as this reduces the load to nothing.

I'd assume it's the 12 kips. FEA models don't capture the relaxation accurately in many cases. So going with initial (envelope) forces may be best.

Another thing to keep in mind (when we are talking about comparing FEA results to AISC allowables): AISC allowables are built around nominal/average stresses......whereas FEA is typically spitting out all sorts of stress concentrations in the results. You can wind up with some questionable designs comparing the two. (You probably already know that....but since you are a mechanical guy, I thought I'd point it out just to be on the safe side.)

RE: bolt pretension, AISC, and FEA

(OP)
"Bolted connections are almost always designed by hand using standard simplifications that are codified by AISC"

We don't have simple connections that follow "standard simplifications." We don't have the luxury of picking the answer out of a book.

Continuing the example I gave, we have been using the 12 kips in the AISC calc. I am the one trying to understanding this because you can fail the allowed bolt stress just by pretensioning it, and this does not seem right to me. We have a bolt pattern calculator based on excel which uses elastic assumptions. That is my preferred method, but some jobs just don't lend themselves very well to hand calcs.

I will definitely think twice next time before I get on here with a question. I think maybe my only poor judgement was to think I could ask a simple question without the accusations and denouncement.

RE: bolt pretension, AISC, and FEA

(OP)
It seems like you are misunderstanding how bolt preload works. You seem to thing pre tensioning or preloading a bolt results in additive forces but we ignore the preload because AISC says to. THIS IS INCORRECT. The preload force in your bolt doesn't disappear it is still a load on your bolt. It is just that pretension forces are not additive.

Pretensioning a bolt by 10kips and then loading the bolted plies in tension by 4 kips will result in 10kips of tension in the bolt.

Pretensioning a bolt by 10kips and then loading the bolted plies in tension by 12 kips will result in 12kips of tension in the bolt.

There are various things online that explain this:
https://www.nord-lock.com/insights/bolting-tips/20...

Yes, I understand this, except that the tension in the does go up a little bit in the first case you gave. The AISC design guide says about 10%.

RE: bolt pretension, AISC, and FEA

Quote (pylandj1)

I am the one trying to understanding this because you can fail the allowed bolt stress just by pretensioning it, and this does not seem right to me.
It is right! Have you not ever over tightened a bolt and broken it? You absolutely can fail a bolt just by pretensioning it.

Quote (pylandj1)

I will definitely think twice next time before I get on here with a question. I think maybe my only poor judgement was to think I could ask a simple question without the accusations and denouncement.
The thing is that it IS NOT a simple question.

If you think modelling bolt behaviour in FEA is simple then that is a problem. Wiser engineers are politely, and in a few cases more bluntly pointing this out to you. But you don't seem to be listening.

RE: bolt pretension, AISC, and FEA

Quote:

I will definitely think twice next time before I get on here with a question. I think maybe my only poor judgement was to think I could ask a simple question without the accusations and denouncement.

Are you kidding? You are getting free advice from some of the best I've talked to. Yes, sometimes we get off scope....but it's easy enough to get them back.

Don't blame us if you are doing something that seems strange.....and we note that fact.

RE: bolt pretension, AISC, and FEA

Quote (rb1957)

"It is very easy to export a 3D model to ANSYS and apply pretensioned bolts." ... oh dear ...
"That's what most of our guys, expecially the younger ones, are going to do." ...oh dear, oh dear, oh dear ...

yeah, this dump everything into an excessively detailed FEM (with every bolt hole modeled, because it is easy then to automesh the CAD model, and rigid element wagon wheels at every hole connecting to fastener elements, and all sorts of other insanity) is rampantly endemic in aerospace now, and it scares the living **** out of me. just look at some of the questions posted on the various FEA forums here.

RE: bolt pretension, AISC, and FEA

(OP)
https://www.eng-tips.com/viewthread.cfm?qid=511671....

OK, if I look at AISC steel code it gives a table of minimum preloads of 0.70 of bolt tensile strength. That's in table J3.1

But in table J3.2 the maximum allowed stress is 0.75 Tensile strength. Using ASD with omega = 2, this is an allowed stress of 0.375 of the tensile strength. So how can I be required to tension to 0.70 tensile in one table but be limited to 0.375 tensile in another table? That's what I mean by breaking the bolt just from preload.

RE: bolt pretension, AISC, and FEA

Quote (pylandj1)

It seems like you are misunderstanding how bolt preload works. You seem to thing pre tensioning or preloading a bolt results in additive forces but we ignore the preload because AISC says to. THIS IS INCORRECT. The preload force in your bolt doesn't disappear it is still a load on your bolt. It is just that pretension forces are not additive.

AISC says that you ignore it.... Because additional load / tension will tend to UNLOAD / relax the pretension stress in the bolt. You are saying that the end result was a larger tension. This suggests a) You're doing your analysis wrong (which is one of the things I was warning you about) or b) you don't understand the code that you're trying to apply.

Note: A number of comments have suggested (to me at least) that you don't generally understand how steel connections usually work and how AISC works. Maybe I'm wrong. But, I can only go off of what you've said in this thread.

These are not meant as an "attack" on you personally, but as a sincere expression of concern that you are operating without knowledge in a field where a mistake could cause serious problems. Maybe injury or death.

Quote:

We don't have simple connections that follow "standard simplifications." We don't have the luxury of picking the answer out of a book.

We've all dealt with some weird and wacky connections over the course of our careers. If you want us to give you advice on your actual connection (rather than just on your suggested method of dealing with this connection), then we should probably see what the connection actually looks like. That might be the best course of action.

In fact, I apologize for being harsh in my previous critiques without making this rather obvious suggestion. I have assumed a lot, probably too much. Perhaps seeing what your actual connection looks like will allow me to better understand why you want to got the FEM route or why it is a good choice for your project.


RE: bolt pretension, AISC, and FEA

(OP)
Well that's interesting. I have two sources with the same formula for the load in a pretensioned fastener. Both sources say the bolt load is the pretension plus the joint factor multiplied by the external joint load. And they say it is the compression in the layers of the joint that decreases with the application of joint load, and that the bolt tension increases. And the joint factor is often in the range of 0.1 to 0.2, so that increase in bolt tension is less than the decrease in layer compression.

Are these two sources wrong then?

RE: bolt pretension, AISC, and FEA

Quote (pylandj1)

Well that's interesting. I have two sources with the same formula for the load in a pretensioned fastener. Both sources say the bolt load is the pretension plus the joint factor multiplied by the external joint load. And they say it is the compression in the layers of the joint that decreases with the application of joint load, and that the bolt tension increases. And the joint factor is often in the range of 0.1 to 0.2, so that increase in bolt tension is less than the decrease in layer compression.

Are these two sources wrong then?

No, they are not wrong. They are however not applicable when applying the AISC code (which is what you said you were trying to do). As you noted yourself, AISC says that you ignore the increase in bolt force due to the joint factor. In it's simplest form, AISC is saying to make your joint factor equal to 0 (basically assume that the compression layers are infinitely rigid and that the strain is zero). You are getting so much push back on this because the idea of a joint factor is foreign to structural engineers, all of our codes ignore it so most are not familiar with it. In the mechanical realm (especially in pipe flange joints with gaskets), this is an everyday calculation, it is not in the structural world.

There is no good way to directly compare your FEM results (which do capture the joint factor) to the allowable forces given in AISC. The conservative thing to do is what you have been doing in the past, using the highest force calculated.

RE: bolt pretension, AISC, and FEA

Quote (pylandj1)

but I wanted to see what others are doing in their work. Thanks.
I think the answer to this is relatively straight forward based on the above. Others are not using FEA and modeling preload when doing their work.

RE: bolt pretension, AISC, and FEA

Hmmm.... This does get confusing. Care to cite your sources?. I still think you're wrong. I believe when the clamping force goes down that's because the pretension is getting taken out of the bolt. But, I'm perfectly willing to review your source and see what it says in the fear that I might have just said something really stupid.

That being said, I'll say this, which is based on years of experience in the structural engineering field. Much of it for petrochemical plants where all the connections must be field bolted... because the owners will almost never allow a field weld in a facility with so much highly flammable or explosive materials.

The usual assumptions for bolted joints that I tend to design are the following, which is conservative:
a) The bolts that resist shear are assumed to be incapable of taking any tension.
b) The bolts that resist tension are assumed to be incapable of taking any shear.

This is the recommendation for a number of different variations of bolted end plate moment connections. This is not a hard and fast rule for ALL connections, but you use it whenever you can simplify it this way.

I will re-iterate my suggestion of showing us your actual connection if you want to get better advice and less criticism.

RE: bolt pretension, AISC, and FEA

Quote:

OK, if I look at AISC steel code it gives a table of minimum preloads of 0.70 of bolt tensile strength. That's in table J3.1

But in table J3.2 the maximum allowed stress is 0.75 Tensile strength. Using ASD with omega = 2, this is an allowed stress of 0.375 of the tensile strength. So how can I be required to tension to 0.70 tensile in one table but be limited to 0.375 tensile in another table? That's what I mean by breaking the bolt just from preload.

Well first off, you aren't going to break the bolt with either of those values. And secondly, only one of those values is used in determining the allowable shear. The available slip resistance per bolt (see Sect. J3.9) is figured by a ks factor. A variable in this factor is that value in Table J3.1. And as I have alluded to before: I don't know that I would count on the full preload value always being there. These things relax.

By the way, IIRC, you still (by code; at least by the 13th ed., which is what I am quoting from) have to check it also as per bearing strength. (Whether it is slip critical or not.)

RE: bolt pretension, AISC, and FEA

Quote (dauwerda)

I think the answer to this is relatively straight forward based on the above. Others are not using FEA and modeling preload when doing their work.

Yes. Excellent point.

For what it's worth, if you (the OP) don't like what Structural engineers have to say about your question, maybe it's not entirely on us. Maybe you provided very little information about your connection. And, you are doing something that Structural Engineers do NOT generally do.

I think maybe you wanted to get a simple confirmation that what you're doing is "correct". Well, to us Structural Engineers it's not really correct. That's just not how we do it.

I get that you're getting defensive. I don't blame you for that. I know I was being harsh in what I was saying. I have seen other structural engineers working on projects where they went down an FEM rabbit hole that was totally unnecessary and a waste of their time energy and budget.

Maybe there are mechanical forums (and codes) meant for the types of connections you're designing?

RE: bolt pretension, AISC, and FEA

For what it is worth I regularly use FEA for structural steel. But as most people know it can be a dangerous tool in the wrong hands, garbage equals garbage out. But I rarely model bolts for the many reasons and problems that others have describe above.

That said I have occasionally modelled bolts. Many times unsuccessfully and giving up and other times successfully. I'm not sure if my attempts at modelling in preloading worked successfully or not, I remember having trouble with it. Recently I've found Ideastatic very useful FEA which models bolts mostly suitably. Though like everything FEA you need to be switched on to make sure you aren't doing something stupid.

RE: bolt pretension, AISC, and FEA

(OP)
"Hmmm.... This does get confusing. Care to cite your sources?. I still think you're wrong. I believe when the clamping force goes down that's because the pretension is getting taken out of the bolt."

I have a total of four sources:


1. Introduction to the Design and Behavior of Bolted Joints, John H. Bickford

2. AISC Steel Design Guide 17, High Strength Bolts, A Primer for Structural Engineers, Geoffrey Kulak

3. High Strength Bolting for Canadian Engineers, GL Kulak.

4. Mechanical engineering design, Shigley and Mishke, fifth ed.

This is from Design Guide 17:


This graph shows the bolt tension slowly increasing until seperation of the joint, from the same source:


The Canadian deign guide is pretty similar to the AISC one. The first two are hard copy books which I can try to scan if desired.

RE: bolt pretension, AISC, and FEA

(OP)
oops, I meant 1 and 4 are hard copies.

RE: bolt pretension, AISC, and FEA

(OP)


The two little red arrows are the loads. The in-plane load is 48.5 kip, the out-of-plane load is 2.4 kip.

There are a total of 7 different structures with different arrangements of members. Some have members framing in underneath, others like this one don't. The padeye can be bolted in three different positions. The 48.5 kip load can be at 0, 45, or 90 degrees to the top plate.

Why I did this in FEA as opposed to hand calcs:
1. The prying forces through the three different angles and the three different plies of material.
2. The beams are short and stubby compared to the length, so simple beam theory may not be applicable.
3. The top plate was only attached by welds at the edges so quantifying its effect on the joint was not clear to me.
4. How much the top plate contributes to the strength of the underlying beam-what effective width does it have.

RE: bolt pretension, AISC, and FEA

in that structure, I'd expect the lowest MS to be in the little brkt/lug, or the two fasteners (holding the clip onto the table).

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

RE: bolt pretension, AISC, and FEA

JoshPlumSE,

If the plate is considered to be rigid, an external tension force on a pretensioned joint will result in a reduction in the compression in the plate without any increase in the tension in the bolt, since no movement will occur due to the rigidity.

If you consider the actual stiffness of the plate, an external tension force will result in a reduction in the compression in the plate. That reduction in compression in the plate will cause it to be thicker, which means the bolt needs to elongate. This results in some increase in the tension force in the bolt (about 5 to 10% per Design Guide 17).

For typical connections, the plate ends up being much stiffer than the bolt because of its larger area, so the rigid plate assumption is usually acceptable.

RE: bolt pretension, AISC, and FEA

pylandj1 -

Thank you! I am perfectly willing to admit that my statement in that one post wasn't correct. Additional tension does NOT relieve the pre-tension as I stated. But, merely relaxes the clamping force as you suggested.

However, that image and excerpt from DG 17 confirms what my belief was. But, which I stated incorrectly.... That the bolt tension does NOT significantly increase until you get to separation of the components. Meaning that the clamping force needs to be overcome before the force / stress in the pre-tensioned bolt increases in any meaningful way.

Thanks also for showing your structure. I think this is actually a pretty simple connection. The calculation of force / stress from the Pad eye into the welds that connect it to the plate should be pretty simple. So, should the force transfer into the bolts. Even when considering prying action. Not much different than a typical bolted end plate connection.

Now, I think you're letting your concern over how to analyze the ENTIRE structure lead you into thinking that a FEM of the CONNECTION will be useful / productive. I still think the connection should still be designed by hand.

2) I don't think these beams are all that short and stubby. They're short enough that shear deformation may be a consideration though. So, I certainly understand the desire to model them rather than use hand calcs for them.

3) How thick is that top plate compared to the beam flange of the pad eye connection plate? My tendency would be to basically ignore it in the FEM model completely. I might even ignore it in the connection design if it's thickness were small compared to the other elements. We usually ignore shim plates. And, I genuinely don't think it will contribute to prying action compared to the main items (beam flange and pad eye plate) that are sandwhiching it on either side.

RE: bolt pretension, AISC, and FEA

Looking at that connection and the force vectors - is there even any tension applied to it (I know Ansys force vectors can be misleading)? It looks like shear in the bolts in orthogonal directions.

RE: bolt pretension, AISC, and FEA

(OP)
JoshPlumSE,

The top plate is 1.25" thick. The beam flange is 0.495" thick. The top plate is that thick because of large moments on the top plate center hole during operation. The padeye load is during installation so the two loads do not occur simultaneously.

RE: bolt pretension, AISC, and FEA

I think if the two diagonals are working very hard they should be rotated 90° so the webs are vertical and aligned with the webs in the beams under the top plate and the webs of the two right hand vertical posts/columns. I predict a nice increase in stiffness, along with improved fatigue resistance if the loading is cyclic.
https://weldtalk.hobartwelders.com/forum/equipment...

Simlarly the beams under the top plate need stiffeners to create a load path to/from the flanges of the two left hand vertical posts/columns. That is if you are counting on much lateral stiffness from the two two left hand vertical posts/columns.

RE: bolt pretension, AISC, and FEA

Bit late to the discussion, but for the work we do modelling bolts with preload is essential.

For bolted flanged joints (with gaskets) the change in bolt load due to mechanical interaction of the joint components causes the bolt load to either increase or decrease with the application of load, albeit a small percentage of the preload. For structural applications, the bolt load will not decrease but may increase to a small extent due to mechanical interaction and prying action. For the most part this is generally ignored, and it is simpler to say that the bolt load doesn't increase until the preload is exceeded.

If you want to know the actual 'bolt load', excluding preload, you would have to measure the change in clamping force / bearing pressure at the bolt location. If we exclude interaction effects, this is because the reaction decreases as the load increases, while the bolt load remains 'unchanged'. As the clamping force drops to zero the bolt load starts to increase. This reaction would be your bolt load for connection design, not the actual assembly load, which is often close to proof load for the bolts.

If you do have multiaxial cyclic loading, I would look to use shear pins or similar, so that the cyclic shear force don't loosen you bolts.

RE: bolt pretension, AISC, and FEA

(OP)
There is no fatigue loading, the buoyancy load only acts for a couple of hours and is pretty constant. It rotates once through the 90 degrees during installation.

Quote (BJI)

If you want to know the actual 'bolt load', excluding preload, you would have to measure the change in clamping force / bearing pressure at the bolt location.

The current preload is 41.3 kip, and the load in the bolt after external loading is applied is 44.7 kip, so the change is only 3.4 kip.

RE: bolt pretension, AISC, and FEA

the change in bolt load is not the external load applied to the bolt, see fig 4.1 in your post 13:38 7th

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

RE: bolt pretension, AISC, and FEA

If I'm understanding correctly, the preload in the bolt is 41.3 kips and the external tension force applied to the bolt is 2.4 kips. Based on the excerpt from Design Guide 17 posted above, the increase in tension force on the bolt would be 5-10% of the 2.4 kips, or .12 to .24 kips, so the total tension on the bolt increases to, at most, 41.54 kips, which is why the Design Guide says in the portion you highlighted, that it is "neglected in practice", and goes on to say in the sentence after what you highlighted that the increase in tension is assumed to be insignificant.

RE: bolt pretension, AISC, and FEA

Quote (pylandj1)

The current preload is 41.3 kip, and the load in the bolt after external loading is applied is 44.7 kip, so the change is only 3.4 kip.
I suggested measuring the change in clamping force / bearing pressure, not the change in bolt load. However, it sounds like a hand calculation would be sufficient to check your bolt capacity.

RE: bolt pretension, AISC, and FEA

Hi,

I found this topic is interesting and I also sometimes having difficulties to find the "Design tensile force" for verifying against tension resistance, where I am used to use Eurocode EN 1993-1-8. Normally I use hands calculation but with more complex structures, have to use FEA.

It's not always correct but to extract reaction forces I sometimes put in zero preload and then use axial+shear from there to check with EC3.

Then another load case with put in preload and I use total force from this to check against equivalent stress, or thread stripping at base plate.

Below charts are samples of bolt M30 - grade 10.9, preloaded at 70% yield capacity. Tapped thread joint with thickness of clamp plate is 30mm. Bolted joint stiffness using Shigley's approach. Two charts with same preload having stiffness constant 0.25. The last chart with higher preload at 75% yield has stiffness constant 0.14 (the same bolt & plate dimension, grade but using VDI 2230 to check for stiffness).

It can be seen that if using total load on bolt to check against tension resistance, then the external load can be applied is much lower than using only external load without including of pretension.

I'm writing from phone so has some limitation but if you find any concern about charts or incorrect assumption, please correct me.

Engineering is a journey of thousand miles begins with a single step

RE: bolt pretension, AISC, and FEA

(OP)
Hi, thanks for your post.

I have found that the joint stiffness method has complications. The shape of the frustrum assumed by Shigley's method is not constant. Plus, the contact area between the plates changes.

Also based on comments above the joint load is the sum of the change in bolt load plus the change in the joint load. These values can be extracted from the FEA analyis. It works for plate to plate connection but am not sure if you have something like a gapped flange with a gasket.

RE: bolt pretension, AISC, and FEA

Hi Pylandj1,

Sorry I'm not familiar with gapped flange with a gasket

Engineering is a journey of thousand miles begins with a single step

RE: bolt pretension, AISC, and FEA

(OP)

I meant a raised face flange with a gap where the bolts pass through.

RE: bolt pretension, AISC, and FEA

Quote (JoshPlumSE)

Honestly, if "most of your guys, especially the younger ones" are going to do an FEA analysis for something like this then I think there is something pretty wrong with the way your company is doing things. Period. I know this comes off as very harsh. But, it genuinely sounds like you guys don't know what you're doing, but are doing it frequently and with false confidence in your abilities to model and interpret the results of these models.

We're structural engineers. Bolted connections are almost always designed by hand using standard simplifications that are codified by AISC (or whichever governing body releases steel code design criteria in your area). To most of us (well, at least me) doing an FEA analysis like this is inherently a bad idea. Partly because it's a waste of time and energy compared to what a simple hand calc will do. Partly because there are way, way, way more things that can go wrong when you design something this way.

I'm very late to the party, but I agree completely with the bold part.

Depending on the fine details of the FEA, it could give almost anything for the bolt forces and stresses. Younger guys -- who didn't even understand the general concepts a few years ago -- going straight to FEA seems like a disaster waiting to happen.

To use one of these approaches, I would need to have experimental results. If the FEA prediction is accurate, then I could use FEA on very similar parts with some confidence. Otherwise, I'd consider the FEA results suspicious at best.

RE: bolt pretension, AISC, and FEA

271828 -

Remember when the Dallas Cowboys practice facilities collapsed during a thunder storm (which is a frequent and expected event in that area). That was in 2009 and the structure had only just been built (2003). It was designed by a company called Summit Structures.

I remember reading that the person who designed it was NOT an engineer, but a physicist or mathematician or something. He was quoted as saying something like "Engineers aren't very good at understanding how to do complex FEA analysis". I wish I could find the news article which has the actual quote. But, it was shockingly arrogant.

It was exactly this type of false confidence in their analysis ability that caused the company to build structures that collapsed so spectacularly under such a normal event.

These are the types of structural failures that we MUST teach to every generation of young engineers. In this day and age, it's probably an even more important case study than the Hyatt skywalk collapse:
https://en.wikipedia.org/wiki/Hyatt_Regency_walkwa...

RE: bolt pretension, AISC, and FEA

Well, I found the Eng-Tips thread on that collapse!

https://www.eng-tips.com/viewthread.cfm?qid=244250

The quote I was thinking of was from a Dallas News article linked to by ChadV towards the bottom of the thread. The link is now dead, but JAE gave this excellent quote / summary:

From what I can tell, Jeffrey Lawrence Galland was hired by the Cowboys to assess their facility because they knew there were problems. The only problem was he was physics major who never graduated from college. But, he was 100% confident in his ability to assess this structure with virtually no expertise on the subject. I have highlighted the section that I remember and which is relevant to this thread.

Quote (JAE)

Timbie believed that "the original [engineering] analysis of the structure was inaccurate," Galland recalled. He said he had no copy of Timbie's report to the Cowboys and did not recall its conclusions in detail.

Generally speaking, Timbie's conclusions "were accurate but not important," he said. "He's not an expert in this style of buildings."

In his e-mail Saturday, Galland said he was "interested in the science behind engineering. Engineering in itself is of little interest to me as it tends to simplify that which is not."

He stressed that the tentlike structures he analyzes are nonlinear, making his physics skills more useful than a "static engineering approach. In general, engineers do not excel in physics or mathematics."

It's this type of false confidence that gets people killed. He was crapping all over engineering because the "science" isn't as simplified as engineers assume. That's 100% true. However, there is a reason why it works. Those simplifications are PROVEN effective by the regulatory agencies over decades. That's how they get into the building codes in the first place.

Back to this thread, there is a good reason why bolted connections are usually done by simple hand calculations. I have no problem using FEA to test out some of our hand calculations to see how "accurate they are" and to help us extend those calculations into configurations that don't perfectly conform to our hand calc assumptions. That being said, you should still perform a simplified hand hand calculation that you can use as the starting point and a conservative minimum for your design.

I will now climb down off my high horse and cease to harp on this subject.



RE: bolt pretension, AISC, and FEA

Quote:

From what I can tell, Jeffrey Lawrence Galland was hired by the Cowboys to assess their facility because they knew there were problems. The only problem was he was physics major who never graduated from college. But, he was 100% confident in his ability to assess this structure with virtually no expertise on the subject. I have highlighted the section that I remember and which is relevant to this thread.

Ol' Jerry can't even make the right decision when it comes to hiring a consultant (never mind a coach). smile

RE: bolt pretension, AISC, and FEA

Quote (pylandj1)

Also based on comments above the joint load is the sum of the change in bolt load plus the change in the joint load. These values can be extracted from the FEA analyis. It works for plate to plate connection but am not sure if you have something like a gapped flange with a gasket.
Refer to my post of 8 Sep 23.
The below paper was one of the first publications of the mechanical interaction equations for bolted joints with ring gaskets. But you should be able to search for more recent publications of similar formulas. These flange methods are currently under development for inclusion in the ASME BPVC.
https://asmedigitalcollection.asme.org/fluidsengin...

There are also numerous publications on metal-to-metal contact outside the bolt circle, have a look into the work of Schneider and Waters, as well as publications by Rodabaugh and Moore.

For flanges with full face gaskets, there are again many publications available but I am not sure if there is any consensus on an accepted method.

RE: bolt pretension, AISC, and FEA

(OP)
In building construction there are well defined predetermined joints which have been thoroughly tested. If this is your situation you are lucky. It is not our situation. Our structures have nothing to do with buildings. It is very rare that we have anything like a “standard connection”. There is no testing data and there won’t ever be because of the variability of our interfaces with other components. There are an infinite number of possible connections.

I normally do some kind of hand calculation as verification. The first thing I did was a hand calc of the bolt pattern and subsequent loads and that’s where the size of the bolts was selected.

Hand calculations can be just as wrong as any FEA can be, even for the same reasons: a lack of basic awareness into the mechanics of the problem. An example: there was an open C-shaped semicircular bellmouth shell which was loaded in torsion and bending. The hand calcs that were done completely neglected the extra stresses induced in this type of section due to torsional loading. When it was checked in FEA the extra stresses showed the quite glaring inadequacy of the hand calculations. The author of the hand check did not consider the proper load actions.

As I mentioned, the load directions we encounter are usually highly variable. To take this example we have been talking about, the load can vary from 0 to 90 degrees. You can find information on checking a beam end connection, and you can find information on a T-stub design, however, what do you do when the load is at 45 degrees? Then it’s part T stub and part end connection. If anyone can show me where this load interaction is found in the steel code or other literature, I’m all ears.

RE: bolt pretension, AISC, and FEA

Quote (pylandj1)

If this is your situation you are lucky. It is not our situation.

Not sure how you arrived at that conclusion. Did you review the AISC design guide on bolted end plate moment connections. Let me just repeat what I said back in September:

Quote (JoshPlumSE)

Thanks also for showing your structure. I think this is actually a pretty simple connection. The calculation of force / stress from the Pad eye into the welds that connect it to the plate should be pretty simple. So, should the force transfer into the bolts. Even when considering prying action. Not much different than a typical bolted end plate connection.

RE: bolt pretension, AISC, and FEA

What's so unusual about this? I agree with JoshPlumSE this looks like a bolted end plate.

Quote (pylandj1)

In building construction there are well defined predetermined joints which have been thoroughly tested. If this is your situation you are lucky. It is not our situation. Our structures have nothing to do with buildings. It is very rare that we have anything like a “standard connection”. There is no testing data and there won’t ever be because of the variability of our interfaces with other components. There are an infinite number of possible connections.

I think you are over-estimating the complexity of your problem....talk about death by FEA.

RE: bolt pretension, AISC, and FEA

Quote:

Hand calculations can be just as wrong as any FEA can be, even for the same reasons: a lack of basic awareness into the mechanics of the problem.

I think it boils down to familiarity. By the same token: FEA can predict things will be no good that AISC will allow. (And we aren't talking running with stress concentrations either.)

RE: bolt pretension, AISC, and FEA

WARose -

I tend to use FEA for connections that way. I start with some basic hand calculations for the basis of my calculation. Then, if I'm concerned that the connection doesn't match the assumptions that were the basis of my hand calculations, then I'll use FEA to justify a beefier design. But, I won't usually use it to justify a lesser connection. Not unless there are other publications or testing that justifies it.

RE: bolt pretension, AISC, and FEA

(OP)
Well, maybe someone found a better example, but when I searched bolted end plates I never found any that accounted for all three: bending, shear and axial force.
This is my load case that will have all three components.


All the examples I could find only accounted for bending and shear.
For example:

RE: bolt pretension, AISC, and FEA

(OP)
There's been no "death by FEA." I'm only still here because of the possibly naive assumption that I might learn something.

RE: bolt pretension, AISC, and FEA

Quote (pylandj1)

bending, shear and axial force. This is my load case that will have all three components.

Bending, shear and axial force. Very common for bolted end plate moment connections. Here's how it's treated (which is all elaborated on in the design guides).

1) The design guides focus mainly on bending moment. You calculate the axial force in the tension bolts using their procedure. The assumption is that the tension side bolts resist the entire moment and the compression side only resists shear.

2) When you have axial tension on the connection, then half of this tension just increases the tension in the tension side. The other half reduces the compression on the compression side and is mostly ignored.

2a) However, the tension and compression sides are not ignored for the local force transfer effects into the beam (or column) the connection is attached to. So, you still have to check local flange bending, local web yielding, crippling and buckling.... Basically all your requirements from section J10, though the equations are often modified slightly to account for the end plate helping to spread out the forces.

3) Since the compression side bolts don't carry any axial force, they are assumed to resist the full shear on their own. Though if you get enough axial force to overcome the compression force from bending, then you'd have to check those bolts for tension / shear interaction.

Does that make more sense to you now?

I don't see anything in your connection that is substantially different from a simple bolted end plate moment connection. I feel like anyone who is an expert in steel connection design should feel essentially the same way.

RE: bolt pretension, AISC, and FEA

(OP)
I have AISC design guide 16 and 39.

What I remember is that in both of these guides, all of the examples and cases had an I-beam with a tension and compression flange. In my connection the loaded plate has no flanges. Am I looking in the wrong place?

RE: bolt pretension, AISC, and FEA

The Design Guide 16 and 39 formulations could be adapted to the bracket shown above.

The AISC Manual Part 9 "Prying Action" formulation might be easier to adapt to your situation.



In our world, we would use some sort of rational statics calculation to determine forces on the bolts. There might be prying action, and that's included in the Design Guide 16, Design Guide 39, and Part 9 formulations.

That has been done about a million times, so we know about what we'll get from this approach. That seems like the "standard of care" to me. If I went and did my own analysis using FEA or whatever other atypical method, I'd better be right! Even more so than if I'm doing this how other structural engineers would typically do it. Just giving a little insight into how we look at this sort of thing.

Red Flag This Post

Please let us know here why this post is inappropriate. Reasons such as off-topic, duplicates, flames, illegal, vulgar, or students posting their homework.

Red Flag Submitted

Thank you for helping keep Eng-Tips Forums free from inappropriate posts.
The Eng-Tips staff will check this out and take appropriate action.

Reply To This Thread

Posting in the Eng-Tips forums is a member-only feature.

Click Here to join Eng-Tips and talk with other members! Already a Member? Login



News


Close Box

Join Eng-Tips® Today!

Join your peers on the Internet's largest technical engineering professional community.
It's easy to join and it's free.

Here's Why Members Love Eng-Tips Forums:

Register now while it's still free!

Already a member? Close this window and log in.

Join Us             Close