Slender Pipe Column Compression 1

[StrucEng864]

I am working on the preliminary design of a structure supporting a piece of equipment.

The support structure will be a 60" diameter pipe or circular plate with a thickness of 3/4". The height of this pipe (if it's fair to call it that) is 15' tall. The load magnitude is 35 kips and 322 ft*kips, applied simultaneously. The vertical pipe will have a cap plate to distribute the equipment loads to the walls of the pipe.

I have modeled it using RAM elements as a flagpole, and created the custom properties for the 60" shape.

Using A36 steel, I can verify the calculations. Using A992 (material may or may not be A992) the Fy changes the shape to a stiffened, slender element, and my design capacity plummets. (I have already contacted Bentley to discuss the calculations)

Does anyone have experience with a similar style structure? Is there a better approach than the one I am taking? Would a different software to check calculations be appropriate? Any other recommendations? Considerations on why a frame would be more appropriate?

Thanks for any feedback provided.

 

[JoshPlumSE]

I've seen this similar things happen for monopoles. Not quite the same but cases where the capacity cannot be calculated because it is beyond maximum code slenderness ratio. In those case, I believe, it is common practice to de-rate the material to a lower Fy.... Low enough so that the code capacity equations are valid. Then you can calculate capacity using this reduced Fy. The Canadian code actually codifies this procedure in the S16 Hot Rolled steel code.

My belief is that this procedure is generally more conservative than calculating reduced effective section properties due to local buckling similar to what we do for most cold-formed codes. I've hardly done a comprehensive study, however. That is just my impression based on a few tech support questions that have come in over the years. For that reason, I would agree that it seems odd that the A992 material would come out with a much lower capacity.

 

[CANPRO]

I've experienced this with the Canadian Code. Using the reduced Fy does tend to produce more conservative results (in my particular analysis anyway). In my case it was because the lower Fy is applied to the entire section where as if I had reduced the section properties to meet the slenderness limits I would have only penalized the portion of the cross-section that did not meet the slenderness limits (compression zone). If your section is entirely in compression it may not make a difference.

The end result is that I used the lower Fy because it was easier to write into my spreadsheet...but I am aware of the conservative number and I can keep it in my back pocket if needed.

 

[dhengr]

StrucEng864:
Josh and I are saying essentially the same thing; his from a very good FEA and software background, and mine from the older guy, background theory standpoint. But, rather than a monopole, try looking at this like a steel tank or chimney, or some such. You have a fairly thin wall, large OD, cantilevered structure, and you will likely have local plate buckling, or plate stiffening issues. Thus, it is likely that you will not be able to take full advantage of the Fy of A992 over the lower A36 Fy before buckling at a lower stress becomes the controlling issue. Look at pipe column and local plate buckling theory to sort this out.

Take a look at Timoshenko’s several different texts on Theory of Elasticity; Timoshenko and Gere, “Theory of Elastic Stability;” Timoshenko and Woinowsky-Krieger, “Theory of Plates and Shells;” F. Bleich, “Buckling Strength of Metal Structures;” all from McGraw-Hill. Any, number of Advanced Strength of Materials and Theory of Elasticity texts. And, “Tubular Steel Structures, Theory and Design,” by M.S. Troitksky, from Lincoln Arc Welding Foundation.

 

[SAIL3]

A992 may not be available in pl. Anyway, R/t=40 is not a thin-walled vessel. There are many well-established formulas for buckling and the allowable equivalent compressive stress, as reference by dhengr.
Also, ASME has a design guide on steel stacks , ASME STS-1-2000(probably a later version is available). For the most part, FEA programs do not handle buckling very well and one needs to be careful in using them and interpreting the results.
You really do not need an FEA/computer program to analyze this, but if you want to go that route, there is a program, if I remember correctly, called "compress" that you may be able to use...this program is used in ASME applications.

 

[JStephen]

I thought A992 was purely a structural-shape specification. There are plate steels available in 50ksi. A tube of that size would normally be rolled from plate, and A36 would be the most common. It may be cheaper to bump the thickness to 7/8" than it is to increase the grade, as far as that goes.

There are a number of standards that include allowable compressive stress for tubes of that configuration. The problem is they have varying factors of safety built in, so you can use 4 different standards and get 4 different answers.

I don't remember how AISC handles this, but I didn't think it fit under the "stiffened thin element"- double check that. There should be a separate t/R criteria for tube/pipe.

 

[SAIL3]

meant to say computer programs that are geared for prismatic members do not handle plate buckling very well , unless, the program is dedicated to do this. Anyway, it is the responsibility of the engineer to verify the results obtained from a computer program and understand the underlying theory of the formulas they are using in that program. AS JS mentioned , A992 is a strucctural shape specification.

 

[fegenbush]

Again, JStephen is correct that ASTM A992 covers only shapes. However, ASTM A572 has several grades of plate that exceed the Fy of ASTM A36. There is little cost increase with using these grades, as they have become more common in modern construction. Often, they are available at local service centers in common thicknesses.

 

[racookpe1978]

So, solve the problem by using the inside of the 60" dia (5 ft) "short" piper by welding an array of ribs inside the pipe at 20 to 30 degree intervals.

The size of each rib is unknown right now of course, but probably 1/4 to 1/2 thick (3/8 maybe) by 3 to 4 inches high, stitch-welded to the pipe wall.

 

[StrucEng864]

Thank you all for the responses.

I originally used A992 as an attempt to run the calculations considering a higher strength than A36. I appreciate the responses. I will be certain to specify A36 or A572 to clarify any confusion.

Because of my unfamiliarity with this type of application, I wanted to back check my calculations with a a software program. I have recently been notified that the calculation that caused the compression strength to 'plummet' is a software error. Based on the other resources I have used, along with AISC, I feel more confident about the magnitude of my hand-calculated results.

DHENGR- Thanks for the references. I will likely use them along with the ones that I have used to this point.