My first post:

I am looking at a 16" overhead line going from the top of a 80 foot tall, 1/4" thick 4' diamter Zirconium column to an Air Cooler.

A question I have always wondered. Is it accurate to model the column as deflecting in a wind case, with the Air Cooler remaining perfectly rigid? Does the structure itself tend to sway in the wind too?

Also, for API 661 qualifications of the Air Cooler, do I need to qualify the nozzle loads with the operating wind loads as well?

At just 1/2" column deflection, my Air Cooler nozzle API 661 allowables are exceeded. The nozzles pass in the thermal design load case. I modeled the Column in PVElite and at the design wind condition of 130 mph, I get 3 1/2" deflection. At 50 mph, I get the 1/2" deflection.

(I apply the deflection as equivalent forces acting on the Column in my model, so that I am not "forcing" the pipe to deflect at 3 1/2", but I am allowing it to resist the wind load.)

I suggest you model the equipment (the tower and the cooler). This costs you a few more elements, but then the model is always correct without you having to recompute forces or displacements (if the temperature or wind speed changes in the future).

For details on how to models vessels, see the examples in the Applications Guide.

I don't think you understand the question. I have the column and air cooler modeled perfectly fine. My question is, is it accurate to model the full "sway" of the column in the wind, the 3.5" deflection, while the air cooler remains perfectly still. The air cooler sits on a pipe rack structure that is independent of the column.

Also, do the API 661 loads include occasional loads? Are the load reactions I am getting in the wind case need to be below API 661 allowables?

I strongly suggest that you should talk to the Structural Engineer responsible for the AC support structure and find out how he(or she) expects the structure to act in wind of the same force as you have applied to the vessel.

Then you need to contact the AC vendor and ask them the same question.

Then you need to apply these two answers to your model.

I strongly suggest that you should talk to the Structural Engineer responsible for the AC support structure and find out how he(or she) expects the structure to act in wind of the same force as you have applied to the vessel.

Then you need to contact the AC vendor and ask them the same question.

Then you need to apply these two answers to your model.

That is a good suggestion Jop, I am limited on the structural engineer though, as it is an existing structure. It is a huge structure however.

I will call the AC vendor and see what they say.

Old structure!

Then I recommend that your company or the client hire a Structural Engineer Consultant for a short time.

Object - survey the structure and review any structure drawings that may still exist.

Goal - Prepare a written "opinion/report" on the stability of the structure and degree of flexing under the same wind conditions as projected for the vessel.

I doubt the client would want to pay for an exstensive analysis. Heck, mose stress jobs we get at my company are only budgeted for 25 hours each, and they never ask the analysis how much they would budget for...go figure.

And I was just told we are only doing analysis on the "pipe" not the equipment. HA! Got to love it. This is a replacement in kind line that they were having leaks at the welds (they wouldn't tell me exactly where) and they were wondering what was causing the leaks. The pipe itself proved to be fine from a stress standpoint, even with the 3.5" deflection of the column, but those Air Cooler loads sure were being exceeded! Also, I was never given Column nozzle details, I assumed 500,000 lbs/in moment stiffnesses.

It is amazing how so little most people in this industry know about pipe stress and how little respect they give it.

Gah, my spelling!
*extensive
*analyst

If you don't have sufficient data of existing structure or your client wouldn't agree to pay for an extensive analysis, you can do the quick estimate of steel structure's lateral deflection in wind or earthquake situations by using H/500 (or H/300 is the criteria use in SSE design) limiting criteria. This is the common practice of those guys in designing steel structures; "H" is the centerline elevation of your pipe from ground. But check also your plant site's steel construction code as some maybe using H/200 limiting deflection criteria-this is worst. There maybe some degree of differential displacement between the column and the structure; and we know that ASCE-7 (if this is your design code; or even UBC-1997) requires that differential deflection must be accounted in the analysis. I suggest to account the induced displacements on pipe due to independent support motion (of structure) in your analysis.

Regards,

The height of the air cooler on top of the structure is roughly 72 feet, so H = 72*12 = 864".

So I have between H/500 = 1.73" deflection or H/200 = 4.32" deflection.

My column deflects 3.5", so if I am understanding correctly, worse case would be if the structure was designed to H/500, which would give me a differential of 3.5" minus 1.73" = 1.77" of differential deflection.

In my model I would simply deflect my column 1.77" for my four wind cases (one case each for north, south, east, and west directions) and keep my air cooler and one support on my pipe "stationary".

Sound good to you?

My opinion is that maybe is not conservative. While 1.73" is the maximum the rack can deflect, nobody can say it will. If it will deflect less, the diferential deflection will be bigger than 1.77".
Also I expect that will deflect less longitudinally than laterally.

Regards,

So I guess we are back to square one and I am designing for 3.5" max. deflect! Which in this case, causes loads to go over API 661 allowables on the air cooler in the occasional wind cases.

not quite.

the aircooler structure has some flexibility. use this.

The structure may either deflect more in lateral than longitudinal direction, or even the opposite. Who knows? So I suggest you have to consider the most conservative and the worst-case scenario and use the maximum deflection values in lateral and longitudinal directions (unless you have the real numerical data of deflections from your structural group; or include the time consuming steel structure model in the analysis).

Assuming your design code is ASCE-7, it is not specific in the topic of relative displacements of structure due to WIND PRESSURE. However, let me relate ASCE-7 approach in seismic evaluation of section-13.
1.) ASCE-7 para. 13.3.2.1 - Displacements WITHIN Structures.
Accordingly if your piping is supported/restrained within the same structure, the relative displacements shall be the "algebraic difference" of the deflection between two connection points at different height within the structure.

2.) ASCE-7 para. 13.3.2.2 - Displacements BETWEEN Structures.
Accordingly if your piping is supported/restrained at two connection points on separate structural systems, the relative displacements shall be the "absolute sum" of the deflection between two connection points at different height in separate structures.

In your case, #2 is the applicable rule; and if you are using static analysis of wind load use the deflection of structures at every support point of piping (i.e. at column and A/C structure). Your load case should add absolutely the structures deflections.

I hope this helps.

Thanks Yhebostress. From what I am gathering for the equivalent seismic code, it is telling me that I need to add the max. deflect of the column with the max. deflect. of the structure, as if they were to move both max. distances in opposite directions! So, I would add 3.5" deflection of the column, to a possible H/200, or 4.32" of the structure for a total of 7.82" of differential deflection. Wow! That is a lot to design for.

Those Air Cooler nozzles will tear right off!

threeouts,

People have been designing systems like this with Columns and Air Coolers connected by piping for more than 100 years. Strange as it may seem the systems worked and they did not have sophisticated computer stress analysis programs to make the job easier.

I did not work on ALL the systems, on all the projects in the past but of the few that I did work on none of the nozzles ripped off that I know of.

Thanks for the all encompassing blanket statement. But, I am trying to tie-in reality with theoretical. I am not saying the nozzles will literally rip off. I am trying to build an accurate model, that reflects real world scenarios.

This particular system I am looking at does indeed have leak problems at the welds, and I am trying to figure out if wind loads could in fact be the culprit.