Ellipsoidal Head Thickness Calculation Inquiry

Not sure, but could it just be because Mandatory appendix-1 (SUPPLEMENTARY DESIGN FORMULAS) paragraph 1-4(c) supersedes UG-32(d)?

I make (6.059*2406.0000*0.997)/(2*137892.00*1.00-0.2*6.059) = 0.0527 so there's a factor 100 out. It's because P is in bar (answer 10x too big vs N/mm²) and S in kN/m² (answer 1000x too small vs N/mm²).

I don't think the formula is correct as it stands anyway. In (2*S*E-0.2*P) S and P should be in the same units (N/mm² preferred). In your case it doesn't matter, as the pressure is low. Doing it your way 2*S*E = 275000 vs 0.2*P = 1.2, << 2*S*E, by factor > 200000. Using N/mm², 2*S*E = 275 vs 0.2*P = 0.12, still << 2*S*E, but by factor > 2000. So 0.2*P is negligible vs 2*S*E in both cases.

But for P say 1000bar = 100N/mm², 0.2*P = 20, and your formula gives thickness ~ 10% too small. 1000 bar design pressure may not come up too often, but the code has it there for a reason! Be interesting to know know PV-ELITE handles it.

Incidentally, I assume your data relates to ASME VIII. My (fairly old) copy of BS5500 says t = P*D/(2*S - P), so the effect of P in the bottom line is much greater.

Please ignore the last paragraph of previous post. Just realised the query was about a dished head, the formula I quoted from BS5500 is for cylindrical shell thickness. The BS doesn't give a formula for head thickness, but uses a table. If it's of interest, for 2:1 head, and your P/S = 0.6059/137.982 = 0.0044, it gives t/D = 0.0028 so t = 0.0028*2400 = 6.8mm. That's before the 3mm corrosion allowance.

Dear codemaster ,

firstly , thank you for the quick reply to my inquiry , I was about to send you the same inquiry on a message yesterday before I decide to put it as a thread, that`s when I saw that you already commented , appreciate...

I just wanna focus on the main core of my question before getting to the units , you mentioned in the first paragraph that appendix 1 supersedes UG-32(d)? ,frankly me also don't know and I didn't see that clearly in the code of ASME SEC.-VIII-DIV.1, besides through surfing the code and some other external references I knew that we are using UG-32(d) in the calculation of ellipsoidal head proportions of ratio=2:1 , for any other proportions other than 2:1 ( 1.9:1 for instance ) appendix 1-4 (c) is replacing UG-32(d) and you can see the factor " k " in this case which can be obtained from table 1-4.1.

The main reason of my question was that the pv-elite used appendix -1 directly instead of UG-32(d) , thought the vessel`s head input data introduced to the software was simply " elliptic " , my main focus not about the values of the equation thought you raised a valid point regarding the units , it`s about the concept behind the usage of appendix 1-4 and not using ug-32(d) , as you can see the value of factor "k" is 0.997 which is almost 1 , but the concept of using the equation was my main core…

The only conclusion I am up to now , is that the software is applying the safer case of not considering the elliptical head as 2:1 as the dimensional proportions may be something else which has not been identified in the input data and directly applies appendix 1-4(c) for more precise results… I don't know it`s only a guess , correct me if I am wrong please..

I went through the report again to check the units used for both P , S and I found the user used bar and kPa , I can send you the report if you wanna have a look on any details if you like no worries , I am not using the BS standard for this am only using ASME code so please don't feel sorry for any wrong paragraph :D

All in all I appreciate your help.

Cheers,

Glad if it was some help!

As you probably realise, my comment about UG-32(d) being superseded was just a suggestion, I'm not very familiar with ASME VIII and I haven't got it to hand, (even an out of date copy!). I haven't opened the links in SolarEagle's #4 yet but hopefully he has cleared it up. Is it possible to "force" PV-ELITE to use UG-32(d) instead of appendix 1-4 (c) and see any differences in results?

I was going off on a tangent about units, but I'd just add there's no problem using bar and kPa as long as all the factors are multiplied or divided, provided the right multiple of 10 is put in at the end to get the right answer. But for 2*S*E-0.2*P, the units of S and P must be the same. I wouldn't have thought the guys who wrote PV-ELITE would have made that mistake, but you never know. If I had it to hand I'd be inclined to also put the formula on XL and try some figures, including high pressures, to see how they compare.

Your para The only conclusion I am up to now , is that the software is applying the safer case of not considering the elliptical head as 2:1 as the dimensional proportions may be something else which has not been identified in the input data and directly applies appendix 1-4(c) for more precise results… I don't know it`s only a guess , correct me if I am wrong please.. Sorry I don't know enough about ASME or PV-ELITE to comment, apart from saying the ratio clearly must be an input, and it seems odd to handle 2:1 one way and any other ratio differently. There won't be a sharp change in design thickness as you go through 2:1.

..."In Pv Elite, Appendix 1-4 is used when you are designing ellipsoidal d/end with proportions other than those given in UG-32. this has to be designed as per App. 1-4 only."

http://65.57.255.42/ubbthreads/ubbthreads.php?ubb=showflat&Number=56375

make sure your software is up to date....

http://support.ecedesign.com/support/solutions/articles/6000016564-pv-elite-release-notes

Dear SolarEagle,

Totally agreed with what you mentioned above , as it is the same I read in appendix 1-4 in the code ,but the change in the ellipsoidal proportions from 2:1 to anything else is the decision of whom basically? the designer or the client? in other words why we change the proportions if the 2:1 ellipsoidal head is the most economical among the other heads ( correct me if I am wrong please if what I read is right or wrong ) , if we put the software aside a bit and focused on the procedure of the solution itself , can it be the following? :

1- Defining the design conditions ( P , T , ..... etc.)

2- Directly applying appendix 1-4(c) equations and considering K factor = 1 if the vessel`s head is 2:1 or any other value from table 1-4.1 if the proportions was something else (again this is dragging me to the same question raised above , who will decide these proportions?

Another clarification worth to be mentioned also here , in some design references for the ellipsoidal head 2:1 , I saw the calculations before going to UG-32(d) compared the ellipsoidal head thickness to that of a seamless spherical shell , ( attached below for reference ):

1-"Value of 0.665 SE=

2-Since the value of 0.66SE > P, Use Thin Wall Equation for

3-Calculating the Min. required Thickness of Head, t:

t1 = PD / (2SE-0.2P) [UG-32(d)] (1)

4-Compare to Thickness of Seamless Spherical Shell :

Ps = 0.665 SE=

5-Since P < Ps, Calculate Thickness for thin Wall Spherical Shell:

t2 = PRo/(2SE+0.8P) [APPENDIX 1-1] (2)

6-For thin Wall ellipsoidal 2:1 head :

t = PD/(2SE-0.2P) [UG-32(d)]

7-Ellipsoidal Head Thickness due to Internal Pressure, t = MAX ( t1 , t2)"

Any idea about the reason of this comparison between ellipsoidal and spherical heads ?

Sorry if the questions were many …

Well when seeking answers it usually pays to follow the money...

process design

thanks for the reference ...

2- Directly applying appendix 1-4(c) equations and considering K factor = 1 if the vessel`s head is 2:1 or any other value from table 1-4.1 if the proportions was something else (again this is dragging me to the same question raised above , who will decide these proportions?

The R/h ratio of the head is usually determined for structural or process reasons, depending on the vessel application. Eg if you want to maximise the vessel straight side as % of overall height, you would go for a shallow head. But that would likely be thicker and may cost more than a deeper head, so as in most aspects of engineering you would weigh up the importance of various considerations and reach a compromise.

Any idea about the reason of this comparison between ellipsoidal and spherical heads?

Your formulas for t1 and t lower down for 2:1 ellipsoidal head, PD/(2SE-0.2P) are the same and basically correct for cylindrical shell thickness (the -0.2P could be debated, but it's negligible for low design pressure). BTW that's inside diameter.

t2 = PRo/(2SE+0.8P) is correct for a spherical vessel. As you can see (Ro in place of D) t2 is half t or t1, so t = MAX ( t1 , t2) always = t1. So no, I don't see any point in this comparison.

In practice, plate thickness for the dished head is often greater than shell thickness, due to thinning when the head is formed.

codemaster , thanks for the intrest in relplying my post again....

your analysis seems logic , specially that when i go to UG-32(d) , i`ve never found this comparison in it , so i think i`ll go with what you mentioned ...

thanks again gents....