Simulating nut & bolt preload in FEA

[sprashanth]

I'm trying to do an FEA which involves an assembly of two steel plates which are bolted-onto each other. The preload due to tightening the nut & bolt is applied in the form of tensile force or pressure on the contact areas (acting in opposite directions to each other) between the bolt & plate and the nut & plate.

I'm wondering if the above inputs are enough for my FEA, or do I also have to include the forces that one plate applies onto the other due to the tightening preload?

I have attached a drawing which illustrates my question. Basically Option-1 is where I include only the preload force between bolt & plate and nut & plate (the arrow marks indicate the forces applied on the faces of the components). Option-2 is where I also include the forces that one plate applies onto the other due to the tightening.

So which option should I go with for my FEA - Option-1 or Option-2?

 

[amubashar]

Which FEA software are you using?

Aamir

 

[sprashanth]

I'm using Solidworks SImulation 2009 for my FEA

 

[johnhors]

Have you received any FEA training? Or are you a student?

 

[sprashanth]

I'm only a beginner at FEA & this is one of the exercises that I'm practising, but why do you ask so?

 

[rb1957]

did "they" just change the buttons (for reporting posts) ?

replaced with the weasle words "... student posting are not allowed ..."

 

[izax1]

Disregard the previous answers. The type of FEA SW you are using is irrelevant. But you should remember that the "bolted connection" is a highly non-linear problem, and finding "all" the answers will require non-linear analysis.

The best way to solve this problem is to apply negative thermal forces on the bolt. The bolt will then shrink, and compress the plates together. The full solution is available only if you define contact planes/elements between the plates. In that way you should get the surface pressures between the different planes.

An interesting test case for your FEA training.

 

[sprashanth]

Thanks for your valuable post, izax1. But unfortunately, the FEA software at our company does not have non-linear analysis capability & for the time being, we are stuck with being able to do only Linear Static analysis. Keeping these constraints in mind, which do you think would be the best linear static simulation possible for bolt & nut preload (if it is possible at all)? At our company, we

I would like to also clarify a few things for those who have mistaken my question as "student posting" - I am not a college student who's trying to cheat in his homework or exams, I'm an Engineer working at a Manufacturing firm for the past 4 years. The major problem is that our company is relatively new to the field of FEA & unfortunately, where we live there is very limited training resources available (more specifically, when it concerns FEA). The training we have recieved in FEA is very poor & there is virtually nobody to approach when we need help for complicated FEA problems. Other than the internet & specifically informative forums such as these, how else can I seek FEA knowledge for self-improvement?

 

[sprashanth]

Oops, apologies for leaving the first paragraph incomplete, so I'll fill in the missing details in this post (as I'm unable to edit my last post):

At our company, we have so far been following option-1, but we are not satisfied with the results - therefore, we are looking at alternate options of simulating the fastener preload that will give us satisfactory results. We're looking at option-2 as one alternative, since it accounts for compressive forces that each plate applies onto the other due to tightening of the fasteners.

 

[rb1957]

why do you want to simulate bolt preload in a FEM ?

the key point about preload is to control gapping of the joint under tension loads (and also to control the fatigue life of the bolt under tansion loads). i'd use the FEM to evaluate the tension loads in the bolt form the external loads, then i'd design the preload to account for these tension loads.

you Could apply thermal loads, but you Must understand the variability of preload; depending on how you control the preload (PLI washers, torque wrench, nut rotation, etc, size of the guy with the wrench in his hand, his eye-sight, and so on) you have a different variability in preload (up to 50%). you would be beguiled if you model one value of preload and think you've got a really accurate solution (because you're using FEA).

 

[Onda]

Hi sprashanth,

I think that Solid works could do a kind of iterative linear analysis. I'm not experienced in Solid Works but I think should be something similar to a linear gap analysis in Nastran with PARAM CDITER.

The solver iterate until all the gaps are in a know status. Giving a very high stiffness if they are closed and no stiffness if they are open.
Then the model is resolved as a normal linear model.
For some contact analysis you don't need a true non linear with large displacement or material non linearity.

In the link below you can find a very interesting pdf about bolt connection:

http://www.ansys.com/events/proceedings/2002/PAPERS/38.pdf

hope this help.
Onda

 

[johnhors]

sprashanth , I would strongly suggest that your employers obtain the services of a consultant FEA analyst and that you study his/her techniques. In many industries and countries there can be legal repercussions if a product failure results from analysis done by non proficient staff.

 

[izax1]

Of course you don't really need a FEM analysis for designing a safe bolt connection. You need to find the external loads on the connection and the required bolt preload to avoid separation of the planes. Then you need to calculate the required bolt torque to produce the required preload. The torque is very dependent on friction (underhead and in threads) and torquing method.
I you have the experience in performing bolt connection calculations, this should be straight forward. Don't bother with FEA if you can use bolts with sufficient preload (Bolts with large diam or hight strength capbilities)

 

[EdDanzer]

The type of problem you are trying to solve is common. Of the many ways to solve few will provide an accurate answer. The compressive forces of the fasteners may have little impact on the static analysis if the two parts are being pulled apart as the bolt acts as a spring. If the load is in compression applying a pressure to the washer contact area will simulate the fastener clamping load. If the bolts are in shear only a non linear analysis will work.

One thing to consider if the parts are being pulled apart is load spikes. We reverse engineered a part on a logging machine that kept breaking bolts and warped the plate. The first run we fixed the bolts and applied a load where the boom attached and the design was adequate. The second run we fixed the part at the boom attachment and applied a load on one washer equal to the breaking strength of the bolt. The results were the plate was too thin.

Ed Danzer

http://www.danzcoinc.com

http://www.dehyds.com

 

[rb1957]

ed,

i guess your joint was loosening, and vibration loads turned into impact loads.

 

[corus]

I'd do option 1 where you apply the pressure from the bolt to the surfaces. If you adopt option 2 then you're assuming something about the contact pressure distribution between the plates, which you don't know. Of course with option 1 you assume you know what the pressure distribution is from the contact from the bolts there, but that'll be a reasonable assumption to assume a uniform load.

corus