Good afternoon. I have a question for the CII collective.

We are doing a project to install four 60' diameter API 650 atmospheric welded steel storage tanks. The design condition is 260° F and 0.5 psig. The problem consists of the design of the 24" vapor recovery header on the tank roofs. The header is 24" NPS sch10S 316L. The tanks are connected in series, in a straight line in plan view, to the vapor header, i.e. the header spans the roof of the four tanks before dropping down. The header is connected to each tank with a 4' riser and 24" butterfly valve. Pressure loss is critical in this system so no bends are allowed to increase flexibility of the piping.

We are trying to estimate the magnitude of the thermal displacement reactions imposed on the tank roof nozzles by the piping to ensure we have not overloaded the nozzles. As you know, API 650 provides a method to calculate nozzle flexibility BUT the published method only applies to tanks much larger than these, thus I am reluctant to extrapolate it to this particular situation. The situation can be fixed with flexible elements (bellows) but our client is adverse to the use of bellows. I can probably talk him into using a few bellows but not enough to relieve the large reactions.

However, if I can accurately model the roof flexibility, I can more accurately predict the magnitude of the reactions on the roof and hopefully leave out a couple bellows. Our initial thought is to use the RISA structural software to calculate all six stiffnesses of the roof nozzle and then place those into a CII flexible anchor.

Any thoughts/opinions/criticisms/cheap shots on this? Has anyone else used any other methods to calculate flexibilities of small thin-wall storage tanks?

Thanks ! ! !

Be very careful with this one. The API-650 Appendix P Nozzle Flexibility computation is for <em>low tank nozzles</em>, i.e. nozzles on the lower part of the shell.

The references for Appendix P discuss the instrumentation and full scale testing (with subsequent FE verification) of two large tanks, with diameters over 100 ft. The nozzles were located on the shell.

I do not believe you can apply Appendix P to your roof nozzles. I think you're going to have to build / analyze FE models to determine your flexibilities. Once you have these, yes you can input them into CAESAR II to model flexible anchors.

------------------
Regards,
Richard Ay (COADE, Inc.)

Just to emphasize what Rich has already said, it is not appropriate to use the Appendix P flexibility criteria for tank roof nozzles for the reasons he cited. It is not even appropriate to use these criteria for shell nozzles for smaller than 120 ft. diameter tanks or for other than "low-type" nozzles.

Using FEA to calculate the appropriate flexibilities and using a CII flexible anchor is certainly one way to account for the roof flexibility. That should help out a lot since both the tank roof and shell are pretty flexible structures. I assume that you would then use an ASME Division 2 type analysis when you get into evaluating the nozzle and shell stresses that result from the imposed loads.

I assume that the tank has a dome roof since you are designing for a small internal pressure. If this is the case, it might be worthwhile and simpler to try "fooling" the program (sorry Rich) and have CII calculate the flexiblility for you using one of the built-in flexible nozzle options.

The BS 5500 option lets you put in a spherical vessel directly. To use the WRC 297 option, you would probably have to first come up with a cylinder of equivalent stiffness under applied load (perhaps hand calcs. using Roark?), then use that in CII.

I also think that it would be worthwhile to test the pressure drop limit and the "no bends" requirement with the process engineers before resorting to using expansion joints. It's easy for them to "impose" that requirement, but they don't have to worry about bellows design, installation, and possible failure.
The roof flexibility still might not be enough to make things OK with the straight risers and the straight header run between the nozzles. How about offsetting the header from the nozzle centerline so that you can atleast get one elbow into each riser? The process guys might be able to live with that. With a valve in each riser, you also have to worry about any portion of that system being shut off and at ambient temperature while the rest is at 260F.

Hope this, in conjunction with Rich's comments, is helpful to you. Good luck!

Vince

------------------
Vincent A. Carucci
President
Carmagen Engineering
7 Waverly Place
Madison, NJ 07940

Visit our web site at http://www.carmagen.com

Thanks guys for your help and comments. As I said, I do not plan to extrapolate API 650 to this application. We will use RISA to calculate the nozzle and roof stiffnesses.

These tanks are cone-roof with external rafters. As you suspected, I had a tough time trying to strongarm the CII nozzle flexibility engines (e.g. WRC 297) to fit this model. I have been around the horn with the client and the process folks and they are unyielding (how's that for a pipe stress play-on-words ;^) ). They do not plan to operate these valves except on plant-wide shutdown. We all know how that goes, however... All I can do is make my case and then try to make it work. I'll let you know what we come up with on the stiffnesses.

Thanks again for your help.
Pete Chandler

------------------
Thanks,
Pete
-----------------------
Thought for the day:
Good judgment comes from experience;
Experience comes from bad judgment.

Dear Pete,

One final piece of advice, having been in your shoes before....

You have not done as much as possibly could be done. FEA will show your high D/t tees to be a bit more flexible than the simple code formula, also detailed FEA of your exact tank geometry is more reliable than your approximation, therefore qualify your presentation.

After doing so your probably on safe ground stating that despite these shortcomings the magnitude of your loads are such that further refinement won't "make it work".

Therefore your "stress analysis" which is really a stress justification, for a too ridgid layout if it does not meet required loading criteria is not your fault, if the parties involved do not allow for added flexibilty they should be put on notice that this action places them and only them as being responsible for the long term structural integrity of the system as a whole.

The bottom line .... everybody likes to be a back seat driver but if there is an accident its always the drivers fault! If you are the engineer of record then your driving and if your driving you need to use sound engineering judgement and compliance to the pertinent codes in your design. if it breaks and you have not done this an outside expert will be hired to provide testimony that you acted incorrectly, meanwhile all the back seat drivers will have left the scene of the accident!

I hope this help, I love this work but I hate the politics, and the fact that people involved in the desicion process are usually ingnorant.

As a final tool, if another engineer is providing design advice (or restrictions) and if he or she is practicing outside their field of competence (EE, CheE talking about structural integrity for instance) THEY ARE IN CLEAR VIOLATION OF THE NSPE CODE OF ETHICS! Remind them of this as a last resort, but if all else fails this usually does the trick!

Good Luck,

John C. Luf

"Many can cut few are surgeons" T. Paulin

If you cannot achieve the needed flexibility by modeling nozzle flexibilities or more accurately modeling piping flexibility, you may want to consider using "thick wall expansion joints". Though much stiffer than bellows expansion joints, they are almost as safe as piping in my opinion, and may provide enough flexibility in your case.

Dear Pete,

I've been out of pocket a bit and am just catching up.

I noticed something in your later comments that is unrelated to pipe stress. In case you haven't thought of it (apologies if you have), since the tank cone roof has external rafters, the roof/shell joint cannot be considered "frangible." Therefore, there must be adequate emergency venting capacity installed on the tank.

I had a client that ran into this. As I recall, the required venting capacity for their particular applications was so high that it was unrealistic to provide the needed capacity. In some cases, they rethought their "need" to have the rafters outside. In other cases, they did a "risk assessment" to rationalize being OK with the venting capacity they were able to reasonably get in, supplemented by equipment spacing and operational procedure reviews that made them feel comfortable.

Regards,

Vince Carucci

------------------
Vincent A. Carucci
President
Carmagen Engineering
7 Waverly Place
Madison, NJ 07940

Visit our web site at http://www.carmagen.com

It will take time to do FEM analysis. For most of such cases,we do not resort to FEM analysis.

If I am in your shoes, I will calculate nozzle flexibilty of WRC 297 in a conservative way.

Firstly, calculate the equivalent diameter of the cone roof. then get the equivalent thickness of roof consdering rafter effect.
The equivalent parameters shall be obtainned reasonably conservative.

Normally, I try do do a accurate analysis if case is critical or profits can be made by doing this way.(e.g. save piping materials or big support or no-reinforced nozzle). But if the activity of accurate analysis take so many man hours, I just make conservative proximation. However, you may carry out a FEM analysis if cost and time are not problems.

Regards

Alvin Zhu