API 650 - App P.3 Calculations Puzzle

[JohnGP]

When I take the time, I have been developing a Mathcad worksheet to perform the evaluation of tank nozzle loads as described in API 650 Appendix P.3. So far so good, but when I plug in the same numbers that are used in P.3.10 Sample Problem No.1, my Stress Factor due to radial load (fr) is nearly 10% higher than indicated in Table P-5 (0.968 vs 0.884).

This seems odd as the other factors I obtain are pretty close to those shown in the table, and in fact the answer I get is consistent to an interpolation between Figure P-8B and P-8C. Has anyone else come across this little anomaly?

Cheers,
John

 

[IFRs]

John - if you care to share the mcd, I would be happy to try to find where the divergence is. In my profile should be an email address.

 

[JohnGP]

IFRs,
Thanks for that, but I don't believe that the problem is with Mathcad - it does still need a bit more work, but I have also performed the calculation by hand (well by using a calculator) using the calculations in Table P-2 in this instance, and end up with the same result.

i.e., for the given example, d/tn=36, t/tn=2, u=0.95. For d/tn=30, I calculate fr=0.943, for d/tn=50 fr=1.0232. No way can you interpolate those to obtain the 0.884 that API 650 comes up with. As I noted originally, interpolating the curves in API 650 also is at odds with the tabulated result in Table P-5.

This suggests to me that the equations and plotted curves are consistent in API 650, but the example calculation seems to have gone astray for that one factor. Provided the equations given are Ok, then I guess there is no problem, but it just troubles me when something is inconsistent like that.

 

[DSB123]

JohnGP,
Have you considered that the Example calc in the Code is wrong? It's not unusual. You should see the errors in the BS EN 13480!!! Don't always beleive that the example calcs given are correct especially if it's a new addition to the Code.

 

[jte]

Folks-

There is an error in sample problem number 3. I understand that the API 650 committee will be meeting soon (or already have...) and it'll be fixed. However, I've not heard of problems with the other sample problems...

The paper on which Appendix P.3 is based is available from ASME. A related paper published in 2002 is available from ASME electronically at

http://store.asme.org/product.asp?c...Papers&category_name=&product_id=PVP2002-1279

Unfortunately they don't have earlier papers available electronically.

jt

 

[JohnGP]

Yes, I think I had some differences with the other problems as well, but will need to sit down some time and work out where the variances occur.

 

[JohnGP]

I had better wait until the API 650 committee sorts out Appendix P. In checking my variances with the Code examples, it seemed that I was also at variance with the graphs, but on closer examination, the labelling on graphs P-8E, P-8F, P-8G, P-8H, is reversed, i.e. lines for t/tn=0 are actually the plots for the equations for t/tn=10, etc.

I recall discussion some time ago about the skewed labelling on the "u" axis of all the graphs, but don't think mislabelling of individual lines was touched on.

Does anyone know when the next Addendum is due for this Code?

 

[JohnGP]

Well, Addendum 4 has been published (thanks IFRs for the link), but Appendix P still has that "early days" appearance. While I can sleep easy now that the Stress Factor due to radial load (fr) in Table P-5 agrees with my 0.968, other issues have been created.

For example, half the equations in Table P-2 have been omitted (don't throw away Addendum 3 just yet). The labels on graphs P-8E, P-8F, P-8G, and P-8H have been corrected (except I believe the top line on P-8E should be for t/tn=5, not 10), but unfortunately the designations on graphs P-8A, P-8B, P-8C, and P-8D have now also been switched around so do not correctly represent their associated equation. Could be a bit confusing for someone trying to use.

I can see that Appendix P will be really useful once the bugs are ironed out - 11th Edition maybe?

Cheers,
John

 

[JStephen]

John, one of the appendices in API-650 tells how to submit a request for interpretation. Itemize these things and send it in to API.

Several years back, I was looking at the flush drain nozzle table, and discovered that some of the dimensions were just physically impossible- you just could not build it with the dimensions, flange hub sizes, and weld sizes that were in the table. But that table had been in there for several years- and either no one ever noticed, or no one bothered to let API know. They can't fix it if they don't know it's wrong.

 

[JoeTank]

You might want to download latest addendum to 650. There were some changes to Appx P.

Steve Braune
Tank Industry Consultants

http://www.tankindustry.com

 

[JohnGP]

Steve,
Yes, my last post was about Addendum 4 and my observation that it seems to have gone part way in correcting some of the errors, but unfortunately introduced further ones. In one of my infrequent spare moments, I will take JStephen's suggestion and send off a collation of my issues with Appendix P. I don't think this is really a request for interpretation but rather a list of apparent problems for consideration by the committee, who are hopefully aware of them anyway (but I take the point they may not).

Cheers,
John

 

[JoeTank]

JohnGP,
Please send your specifics to API. It will be useful for them to know where the problems are located.

Steve Braune
Tank Industry Consultants

http://www.tankindustry.com