Stress exerted on pipe column face by base plate stiffener??? 4

[ST04]

Hi

I was looking for a reference to determine the Stress exerted on column face by base plate stiffener, the base plate stiffener that I'm talking about is similar to this:

http://www.fhwa.dot.gov/bridge/images/sign36.jpg

I have only found an equation in "Design of Monopole bases" page 77. But the result is relatively small around 30 N/mm^2.

The problem is that when comparing the hand calculation with the Ansys FE model I'm getting larger stresses and ends up with the upper edge of the stiffener connected to column face is overstressed as well as the column surface at that point.

For the constrains in the Ansys I have tried both cases of making stiffener connected to the column as an assembly and then put "Bonded" contact constraint and tried to make the column and stiffeners as one part then connect them to the base plate, but in both cases I'm getting what is stated above.

So either the hand calculation is not conservative, or the Ansys analysis is not giving accurate values.

Any suggestions?
Thanks

 

[corus]

You'd expect the end of the rib stiffener to have a very high stress as it acts like a point restraint (or point load) on the column. You'd probably see the same thing around the holding down bolts (depending on how you modelled the restraints there). These are peak stresses rather than average membrane/bending stresses that your hand calculation may give. These peak stresses would be checked against fatigue design curves if the loading was cyclic. If you just have a static load then you'd need to look at the stress distribution in your model around these peaks, and remove that peak stress component to give you the nominal stresses in that region. The nominal stresses would be compared against your design stress limits, related to the yield stress.

Tara

 

[ST04]

Thanks corus for your quick and informative reply

I have found small but important error in my hand calculation, and now I'm getting higher stress values than before.

Regarding to what you mentioned "and remove that peak stress component to give you the nominal stresses in that region."

I'm dealing with static load.
How should I remove these peaks?

There was a suggestion regarding the over stressing points - by a colleague in this forum - to use a stiffening ring at the place under consideration, it worked some times but didn't solve the problem in others.

The only solution was to use stiffening pipe extending from the base plate to few centimeters above the top of stiffeners. but that will end up with really thick section (45 mm) while the column it self should be only 27 mm to resist the applied loads.

Regards

 

[corus]

To remove the peak stress component you can plot the stresses up to the maximum stress. You should see a non-linear/parabolic curve up to the maximum, starting from a more linear curve. Extend the linear portion of the curve up to the position of the maximum stress to get the nominal stress there. This will cut off the peak/non-linear part of the stress disatribution. The method is a little subjective as to how you see the stress distribution, but I've not seen any clearly defined way of doing it.

If the stress at the end of these stiffeners is still of concern then you could put an annular plate around the column, welded to the top of the stiffeners. like so
! !
_______! !
!_______! !
/ ! !
/ ! !
/ ! !
/__________!_!

This will take some of the peak stress out but you'll still have high stresses due to the sudden change in section. I'm not sure what another pipe over the stiffeners would help as you'd still have the peak stress from the edge of the stiffener plate acting on to the tube. It might reduce the nominal stress as you've increased the overall bending stiffness, but how do you access the holding down bolts with a pipe blocking them?



Tara

 

[Ron]

Increase the height and thickness of the stiffener....spread the load a bit.

 

[hokie66]

Another way is to prefabricate the stiffeners and base plate together, with the stiffeners crossed, and slot the assembly into the tube. Takes away a lot of the discontinuity.

Just curious...in the photo, why the large bolt adjacent to the stiffener, then another smaller bolt which looks to be centred between the stiffeners?

 

[ST04]

Thanks dear colleagues

For an axial loaded base plate (Compression) I'm getting almost the same results in hand calculation and Ansys, but for the same base plate if the load is (Tension) the maximum principle stress will go up to 441 MPa (N/mm^2), however, the Von-Mises stress is almost exactly the same for the case of compression =340 MPa.

Here is a screen shot for the Maximum principle stress in MPa or N/mm^2

http://files.engineering.com/getfile.aspx?folder=24d46f00-7ec6-4100-9485-c94c43279936&file=p1.JPG

And I can't find any valid reason for this other than that the effect of deformation in Tension causes the top edges of the stiffeners to be overstressed!!!

If that is right, so I can't get these results for tension by hand?


@ hokie66
Maybe these are combination of holding down (anchor) bolts and ordinary bolts.


Thanks again for your valuable tips
regards

 

[corus]

The other thing to consider in FE results is that you have a 90 degree juncture between the stiffeners/gussets and the tube, with no fillet radius between. This gives an infinite stress concentration factor there, so your results at that point are solely dependent upon mesh density. Use a finer mesh and you'll get even higher stresses. Again, you'd have to consider the stress distribution up to that point to make a judgement on how much of the stress is due to the modelling of the stress concentration there.

The results from your loading may depend on the boundary conditions you've imposed. I would guess that in compression the whole of the base plate is restrained axially, whereas in tension only the holding down bolts provide restraint. This would give different results.

Tara

 

[ST04]

Thanks corus
I thought of adding fillet radius between them but I felt that it may not be correct to include the fillet weld material to the stiffeners.

I will try to add fillet radius to the stiffeners edges equivalent to the fillet weld dimension!!!

I have tried to add fillet weld but it only makes the situation more complex and more uncertainties comes up.

 

[corus]

Unless you can guarantee that the welders will make exactly the radius you want, with no defects in it, then I wouldn't bother. The normal practice is to just model the sharp corner and ignore the very high stress you'll get at the corner from your model. In fatigue assessment the calculation is based upon the nominal stress at that feature, which is away from the corner. The method of extrapolating stresses to the corner, I outlined earlier, is generally taken to be the way to assess the stress there. In your case the loading is static anyway, so you don't need to worry about peak stresses at sharp corners.

Tara

 

[csd72]

Is this stress resulting from bending or just plain compression in the tube column?

If it is just for axial force then blodgett design of tubular structures gives some furmula for the induced ring bending. They also give good formula on the annular ring mentioned above.

 

[paddingtongreen]

Do you have the elasticity of the concrete base in the model? I'm perplexed at the different stresses in the pipe and the stiffeners, and the effect of the different strains.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[ST04]

Thanks again corus I will put your comment on mind.

Dear csd72
These are results for the case of tension applied to the column, I did that for the sake of verifying the manual solution with the FE because it is easier than the one with moment.

Thanks for the reference, if you have a link where I can get it as a pdf I will appreciate it because I couldn't find any link.

Dear paddingtongreen
I don't have it but (I'm using fcu=40 N/mm^2), I have assigned the bottom surface of the plate to be compression only support (To allow for the upward movement which will only be resisted by the fixed washers) also I used the general support and made necessary settings to make it similar to compression only but with planar constraints.

Also regarding the Elastic support, I have noticed that it restraint the upward movement!! Should it?

regards

 

[ST04]

Using round or chamfer for the edges didn't solve the stress concentration, so there is no any way to overcome that situation?
Even though it is related to geometry, but it gives the feeling that the structure is overstressed :]

 

[dhengr]

Ansys FEA software, and its brethren, are wonderful analysis tools; but they seem NOT to give us better or more efficient designs, they just give us confuseder designers. The general arrangement your photo shows has been working for years, with hand calcs., it’s only when some turkey sold you an expensive bit of analysis software, that calls the stresses von Mises stresses or principal stresses which no one seems to understand, and then shows these stresses in pretty colors and the red color scares the hell out of us. You should go looking for some truly bad connection details, or maybe just average fabrication, model them, and see how much red you see, and they’re still standing. Look at the photo you show, it appears there is a primary nut/washer at ever other stiffener and unsymmetrical to that stiffener. What does the intermediate stiffener even do, how many loaded stiffeners does that pipe column have?

It would be interesting to see your hand calcs. and exactly how you treated this connection in that analysis. Why do you suppose the stiffeners are wide at their base and taper to almost nothing up at their top? Might you reduce your FEA stress concentration by shaping the stiffeners differently that a straight slope to 1" wide at the top? Should you weld up the sides of the stiffener, and then across the top of the stiffener (its thickness) or just up the sides of the stiffener? What sort of stress raisers and stress concentrations have you introduced by you welding details and weld starts and stops? These are probably more important, as failure mechanisms than the stress you think you are seeing, however you model it. If you could model them, every weld start and stop or imperfection will cause the same kind of stress concentration you think you are seeing, and many of these will be at a point where the steel has been hardened by the welding process. Of course, you will see very high stresses where ever you have stresses flowing around a reentrant corner, like at the top of the stiffener, as Corus suggests @ 14FEB, 8:38.

RE: “How should I remove these peaks?” 14FEB11, 4:15: Maybe what you should do is, once the structure is all welded up, take a 1" core drill and centered on that high stress point, drill right through the stiffener and pipe wall, leaving a 1" hole with zero stress. But, be careful not to leave that piece of high stress inside the pipe. Or you could use a little engineering judgement and common sense and consider the fact that a little, very localized, yielding might take place and relieve this high point stress. I would be inclined to weld over the top of the stiffener to keep water out, this won’t change the apparent FEA stress concentration though. I’d have the welder do this by using a smaller size weld for the last inch of height of the stiffener and over its top, and I’d have him grind the top of the stiffener to a circular shape with radius equal .5(thickness of stiff.). This way there will be much less chance of leaving notches in top edges of the stiffener, another stress raiser, at a critical location.

 

[BAretired]

I like hokie's suggestion of making the stiffeners continuous through the tube.

BA

 

[hokie66]

I think dhengr has hit the nail on the head. If you don't know how to interpret FEA results, and I don't, why use it? Having said that, I wouldn't worry about those red spots, but wonder how there could be a green band between two yellow bands when the load is pure tension.

 

[dhengr]

BA & hokie:

That’s just a fit-up night mare without much real improvement or advantage, as I see it. With more than four stiffeners, their fit-up on the interior is a mess, and what is their function on the interior other than dead weight, the base plate loads are on the exterior, from the A.B’s. Then their good tight fit-up in the slots in the pipe wall becomes a real tolerancing problem for a good weld at the root. You’ve lost the continuous weld btwn. the pipe and the base plate and traded that for a discontinuous weld with starts and stops at the highest stress location. Finally, you don’t really improve the weld termination up at the top of the stiffener. There’s still a 50/50 chance that the welder will weld around the top of the stiffener and leave notches on the edges of the stiffener or else leave a water catching opening.

It seems to me you have to make a decision about where you want the continuity, and I would prefer it btwn. the pipe and the base plate. The stiffeners then distribute the concentrated A.B. loads sufficiently up into the pipe. One would hope that the stiffeners and the A.B’s. are symmetrical w.r.t. each other. And, a bending moment and the potential of fatigue certainly complicate our design worries, but the stresses don’t change much one way or the other.

 

[dhengr]

Hokie:
Don’t be so humble, you do too know how to interpret FEA results, and you just proved it. You know enough to ignore the pretty little red spots (you’ve probably already had chicken pox too, but have you had your ‘shingles shot’?) and wonder how the hell a stress field can just jump from one magnitude to another without any apparent inducement. Having said that... Usually you can fix that dumb green stress band (a lower stress field) within a higher stress field (the yellow) by trading one indeterminate boundary condition or restraint for another indeterminate boundary condition or restraint. We’re never quite sure which ones are the right ones anyway. The scary thing is that when I increase the size of his link so I can read it, the size of the high stressed area gets larger too, and I’m wondering if the whole darn thing will become over stressed at some point. Alternatively, you can just photoshop the picture to change the colors to your liking.

 

[hokie66]

dhengr,
My suggestion of stiffeners through the column was based on the original photo, which only showed four stiffeners, I think. That's not a difficult fabrication.