Tapered Hollow Section 1

[ars13]

I have a situation where I have to design steel tapered hollow columns which have both axial and bending loads. Does anyone know where I can find some good information on tapered members and their design? Also, can tapered members be entered into Risa? Thanks

 

[JoshPlumSE]

AISC and MBMA are coming out with a design guide that is geared towards tapered members. I haven't looked at it in awhile. It was really based on the type of tapered wide flange members that you use in metal buildings. However, my impression was that the theory and design procedures presented could be extended into tapered pipes or tubes as well.

At the current time, RISA does not have tapered tubes or pipes. Only tapered I-beams. You can always create a model for your tapered column by breaking it into a number of smaller untapered sections.

 

[paddingtongreen]

In a distant galaxy, distant in time, I divided the column into six segments and entered each with the properties at the middle of the segment into STAAD. I thought that would be close enough.

Before that, I used a similar process to find stiffness's at each end of the column for use in moment distribution or in slope deflection analysis.

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

 

[rscassar]

What is driving the use of the tapered HSS columns and what are you intending for the taper detail. Are you providing the industrial welded end plate at every taper location or is it intended on being more "aesthetic" with the detailing.

For the analysis, I would subdivide the member and provide elements with the appropriate section properties for each taper.

 

[ars13]

Maybe I should clarify myself and say they are tapered octahedron members that are about 60' tall. They will be used for an industrial application to support electrical equipment. Maybe I should be asking about mono pole design? Does Risa have mono pole design with tapered members?

 

[JoshPlumSE]

RISA-3D does not have tapered monopoles.... But, the program formerly known as RISATower (now called TNX Tower) does have tapered monopoles.

That program breaks the tapered monopole into about 20 segments of prismatic members to approximate the taper.

Twenty sections is probably overkill. I think something like 10 segments would probably be enough. However, structural stability effects (i.e. 2nd order / P-Delta effects) usually control these types of slender monopoles. So, overkill on the analysis is one of those "better safe than sorry" type of things.

 

[dhengr]

Ars13:
I suspect BAretired would agree with me, that this sounds like a good problem for Newmark’s Method of Numerical Integration. If you are using TNX Tower as JoshPlum suggests, I would go with the 20 segments, for the improved accuracy of representing the cross section, since you are not doing the actual number crunching. BA and I probably would have gone with 10 or 15 segments when we were doing that problem by hand, although it’s just a bookkeeping problem, requiring a wider piece of graph paper, to hold more columns of calcs. for each segment, once you decide to tackle it.

 

[paddingtongreen]

I only used six. The paper I used showed that six was very accurate for a constant section member, so I figured it would be close enough.

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

 

[BAretired]

As dhengr has stated, the problem could readily be solved with the use of Newmark's Numerical Methods. The number of segments chosen is a little arbitrary, but using 10 segments would seem imminently reasonable to me. Selecting more segments would not alter the results in any significant way.

A closed form solution using a continuous function for moment of inertia, bending and axial load may be possible but prone to mathematical error and not in the least attractive for the majority of practicing engineers, including me.

BA

 

[IDS]

This might be of interest:

http://newtonexcelbach.wordpress.com/2010/08/09/buckling-of-columns-with-varying-cross-section/

but if you want to analyse a framework with tapered members, rather than a single column, the subdivided memebers method should be fine.

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[BAretired]

Doug,

If you want to analyze a framework with tapered members, I suppose you could determine stiffness and carry-over factors for each member in the frame and carry out a Hardy Cross Moment Distribution. Alternatively, I believe you could carry out a slope deflection analysis with this much information in which your results would be theoretically, correct.

I have never done it because I have never had occasion to analyze a framework with tapered members.

BA

 

[dhengr]

I suggested 10 or 15 segments because they would give nice segment lengths of 6' or 4', and you use these lengths many times in the calcs., so you don’t want some weird a$$ed length like 6.67' because you selected 9 segments. This clean length selection would be true with the computer solution also. But, BA is right, in our hand calced. solution, there would not be much improvement in results by going from 10 segments to 15 segments.

 

[oldrunner]

Ars13:

The Guide Specifications for Design of Metal Flagpoles is a good place to start. (ANSI/NAAMM FP 1001-07). This guide has sample calculations for designing a 55 foot tapered pole. In this sample calculations there is also has a procedure of how to analysis the pole in tabular form such as a spreadsheet. Note that this method also takes into account the P-delta forces by estimating the second order effects. Starting with this guide you can follow the Calculation Procedure in Section 7.

Of course there’s more to this than just initial sizing. We know that you have some additional loads from equipment. At this time I assume that these are cantilever poles. The second step then would be to use AASHTO’s Highway Signs, Luminaires and Traffic Signals standard specifications. Here you will find the rest of the parameters that have to be considered such as vortex shedding, buckling, fatigue etc. This standard also provides width-thickness ratios for octagonal sections when checking whether the section is compact, non-compact and a maximum limit. It also provides the allowable bending stress for octagonal members. Note that this code uses ASD and not LRFD. One nice thing about both the Guide and the Specification is that they contain excellent commentary.

One nice thing about using a spreadsheet and dividing the tube into sections is that for a 60 foot member, you can use 10 – 6 foots section or 6 – 10 foot sections. I like to vary the thickness of the sections, thickest at the bottom and thinnest at the top. With a well organized spreadsheet you can play with the input and “tune” the total member.

One or two sections? If you are going to have two sections, then you’ll have to design mating flanges somewhere. For this area you will have to be paying close attention to local buckling near the flanges.

If you have never done a pole like this before, make sure it doesn’t end up like this:

http://www.youtube.com/watch?v=tB82mqEyBBg.

This pole was only up about 24 hours before they cut it down.

If you want to see a pole that’s been up for a few years:

http://en.wikipedia.org/wiki/Aqaba_Flagpole.

This pole has mating flanges using large A354 bolts that had to be tightened with tension meters.

Good luck.

 

[SAIL3]

The link provided by "oldrunner" to the pole oscillating
is really interesting.It brings up a few questions.
This occurred with wind at 15mph which results in a steady
state loading of about 0.5 #/sq. ft....
If the pole was designed for say a design wind of 90mph, it
would result in a loading of about 16#/sq.ft and say an
allow. tip deflection of 6" per 100ft=3.30x6= approx. 20"
Mag. factor would be 16/0.5=32 to get a def. of 20" due
to oscillation at 15mph.From the video it looked like
a lot more than that.
I have never encountered that high a mag. factor and leaves
me reassessing what I thought I knew about this problem.
I would not have expected such a large deflection from such a low driving force.I am sure there are alot engineers out there who are more knowledgeable in vibration than I am and
look forward to their input.

 

[paddingtongreen]

There was no flag to show wind direction, but I think it was perpendicular to the sway. I think we saw vortex shedding, albeit at a low wind speed, but the pole obviously had a low natural frequency.

If it had been in a near constant wind, it would not have been oscillating, especially not in a harmonic mode.

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

 

[dhengr]

I agree with Paddington’s first paragraph, and disagree with his second paragraph. I believe constant or near constant wind velocity and direction is often when vortex shedding occurs, particularly around cylindrical shapes. I think he meant to say if the wind was not constant in velocity and/or direction, the oscillation would have been much less likely. Although, I do not consider myself the last word in structural dynamics.

 

[oldrunner]

Well, the problem was that the pole was moving in the same direction as the wind.

Go figure.

 

[paddingtongreen]

I agree with you, dhengr, I phrased it badly, I meant if it was simple bending, under a steady wind force, it would not be oscillating.

Vortex shedding does require a steady wind.

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

 

[transmissiontowers]

We use tapered tube members all the time in the Transmission line business and also in the design of substation structures. ASCE 113 is a design guide for substation structures which includes tapered tubular sections for the columns.

There are several specialty programs for analysis but we use PLS-POLE from Power Line Systems which will handle free standing single poles and frames that include tapered poles for columns. I also use GTStrudl which has a command for generating a tapered pole but you cannot use it for a nonlinear analysis (the are working on making it work). You can also generate the properties in a spreadsheet and make the step tapered prismatic members.

Also take a look at ASCE 48 which is the design guide for the tapered tubular poles.

All the big manufacturers of tapered poles like Thomas & Betts, Valmont, FWT and a few others I forget right now have their own programs and will design the poles for free.

I would get ASCE 113 and read up on the process. ( I was on the committee that wrote the guide )

_____________________________________
I have been called "A storehouse of worthless information" many times.

 

[paddingtongreen]

I would have tried a rope, helically wound around it, that modifies opposite sides and spoils the shedding, but I would have done that from the beginning.

Oldrunner, that wind velocity must have been oscillating at or near the lowest natural frequency of the pole for it to bend with the wind in that manner.

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