Threaded Rod Design

[sierraechozulu]

Greetings!

I have been doing structural analyses of a pushrod assembly of a mechanism. The pushrod assembly has two rod ends and an adapter in between. Rod ends with male threads are attached to the adapter with female threads via cut threads with UNJF profile.

After fine tuning the length of the pushrod, there is some threaded zone out of the adapter at both ends. My rod will probably have some eccentricity of load application or eccentricity due to maching, tolerance mismatch etc. Therefore, a compressive force will also aresult in bending forces, which in the end results in column buckling phenomenon. So the section is under combination of compression and bending. Since the threads are not included in FEM and the threaded region is a simple cylindrical region, diameter being the representative major diameter of the profile, the stress values acquired in the software will only represent a simple cylindrical section.

Shigley states that it might have a stress concentration of 2-4 depending on material grade, forming type etc.

Should I limit the stresses around cylindrical region so that they are below 1/3rd of the ultimate strength? (given a Kt=3.0) Or is it too much conservatism from a static analysis point of view?

Thanks in advance!

 

[Stress_Eng]

Just pointing out, for fatigue loading I would think the kt factor will apply to the threaded region only, where the bending should be much less than the maximum (in the middle). For your pushrod, are you modelling the buckling as a simply supported beam with applied friction moments at the ends? Eccentricity can be included by assuming an initial sine displacement curve (some may say end moments). In your modelling, is your pushrod stepped? Obviously the diameter at the ends can be conservatively based on the shaft diameter at the thread root radius. Is the longer middle section a larger tubular section? You should check on policy / methods when it comes to analyzing your pushrod under ultimate loading. Is kt usually just a fatigue issue?

 

[SWComposites]

your model should use the thread minor diameter, and yes you need to apply a stress concentration factor both to static and fatigue.

have you performed a beam-column buckling analysis? and if buckling margins are not high, should be running a validation test.

 

[sierraechozulu]

Just pointing out, for fatigue loading I would think the kt factor will apply to the threaded region only, where the bending should be much less than the maximum (in the middle). For your pushrod, are you modelling the buckling as a simply supported beam with applied friction moments at the ends? Eccentricity can be included by assuming an initial sine displacement curve (some may say end moments). In your modelling, is your pushrod stepped? Obviously the diameter at the ends can be conservatively based on the thread root radius. Is the longer middle section a larger tubular section? You should check on policy / methods when it comes to analyzing your puchrod under ultimate loading. kt is usually just a fatigue issue.

Actually both rod ends and the adapter part in the model are meshed with very fine tet10 elements. contact region is modeled with Tie contact in, which is similar to Glue Contact in Nastran. The imperfection is given to the model via use of first linear buckling mode shape displacements. as per some ESDU document, a 0.1%, 0.5% and 1% of the total rod assembly is given to the rod as imperfection. Then a nonlinear buckling analysis is carried out using arc-length method along with elasto-plastic material model created using Ramberg-Osgood material model. In order to be a bit more conservative, an elastic-perfectly plastic material model is induced later on.

The model however does not include the threads physically. Therefore, the plasticity around the cylindrical region which is normally threaded is not represented correctly. However, the level of loading is ultimate as you just mentioned.

 

[sierraechozulu]

your model should use the thread minor diameter, and yes you need to apply a stress concentration factor both to static and fatigue.

have you performed a beam-column buckling analysis? and if buckling margins are not high, should be running a validation test.

Yes, I have carried out a nonlinear buckling analysis with arc-length method (a.k.a. Modified Riks) along with an induced imperfection. A validation test is to be carried out. However, I expect some plasticity and maybe premature crack or even a total collapse of the structure since we did not include the threads.

What if I apply different material around threaded region so that the yield strength is say 1/Kt of the original material so that the threaded region is much more realistically represented?

 

[rb1957]

1) the OP asks "from the static point of view"./ Ok, we can say "you should consider fatigue", but we've (and he has?) no to little idea of fatigue loadings so it's only something to consider in a complete design. For all we know they have a separate group doing fatigue analysis.

2) this is about eccentrically loaded column ... why talk fatigue ? Personally I'd develop a moment allowable and combine compression and moment loads. Super conservative says apply the additive loads as a compression (apply the peak stress as a full perimeter stress)

3) how on earth do eccentrically loaded threaded columns work ? there is clearance around the thread ... there is a stack of stiffness in the compressed (under preload) flanges ... would the flanges react the eccentric moment ?? they would be much stiffer than the bolt.

 

[Stress_Eng]

For a buckling analysis, I’ve seen the MS based on compression and bending allowables (P/Pcrt + M/Mu = 1) where Mu is the moment allowable (e.g. moment of rupture). I’ve also seen the buckling MS based on an allowable stress level, such as yield. Check with methods what MS is to be used. If you want to do a thread analysis, you could extract the moment and axial load seen at the threads from your FE model and do a standard ultimate condition thread analysis using hand calc’s, or even a hand calc socket analysis. If shoulder contact is involved, then you've got multiple load paths to consider (through the threads and abutment at the shoulder).

 

[sierraechozulu]

To be more specific, the rod is actually a two-force member. So, the perfect geometry has no imperfection. However, there is manufacturing tolerances, hole tolerance between the centerline of the threaded hole and the centerline of the adapter. As the structure is imperfection sensitive, the initial imperfection induced greatly affects the secondary moments generated due to eccentricity. Therefore, the amount of imperfection has an affect on the section normal and bending stresses. This is another point I seek for help.

How much imperfection is required for such an assembly?

 

[rb1957]

get it, by design a simple compression member. But what about manufacturing tolerances ?

Well, what about manufacturing tolerances ? yes, of course, they mess with the perfect geometry. So you could consider adverse tolerances, to give you the "maximum" that the design would permit. A less rational (and more conservative) way is to say "assume the load vector is skewed 10 degrees (1 degree ? 20 degrees ??) or eccentric by 1/10th diameter (1/4 diameter?) ... pick a number, the higher the more tolerant your design is to imperfection, and the "better" (and heavier). Some amount imperfection will fail any structure. You could say that these extreme imperfections are coupled with limit load (rather than ultimate load).

BTW, if this is a threaded rod, then it is less likely to be a perfect two force member ... the threaded ends have much more stiffness than a true pinned joint, and much more likely to react moment and allow the structure to carry transverse loads.

You're going to be testing this so the proof is always in the pudding (what an odd expression ...).

 

[WKTaylor]

Related... my company specifies push-pull rods or struts... with one end fixed, other end adjustable... have very special design requirements...

Special procured extruded tube... very straight, round with consistent/constant wall thickness...
and
Concentric rod-ends mounted at each-end of the of tube section...
make-for
very consistent axial loading with minimal centroidal shift/off perfect axis for maximum load bearing.

[RuRo moved this line to the Mishap forum]

 

[WKTaylor]

Push-pull rods... and could-be adjustable strut-rods... to take-a-look-at....
NAS91-thru-NAS119
NAS354-thru-NAS373

NAS37 Rod-End, Spherical Bearing, 3/16 x 5/16 - 24, Interchangeable Type - Rev 1
NAS660 BEARING, ROD END, PLAIN SHANK, SELF ALIGNING, ANTI-FRICTION

NASM21-37 BOLT, CLEVIS

 

[rb1957]

quality comes with a price ... good quality is when you pay it, bad quality is when you pay the lawyers ...

 

[Ng2020]

The typical practice in aerospace structures analysis is to assume an eccentricity of L/400 to account for initial imperfections.