Bolted Circular Plates and Roarks

[Jimbo2010]

I have a bolted blank flange joint and I’m trying to discover an alternative way of finding the load in each bolt when an applied pressure is applied to the blank. Im currently using a circular plate model from Roak’s formula for stress and strain, is it possible to use the energy methods and the ratio of the deflection of the plate with and without the bolt load applied??

If Energy methods are not applicable, does anyone have any suggestions to find the loads I the bolt from Roark’s?


Jimbo2010

http://www.engcosolutions.com

 

[desertfox]

Hi Jimbo
Well you could follow a pressure vessel code such as PD5500 or ASME.
For a rough estimate I would multiply the pressure by the area of the flange, to get the resultant force and then divide that force by the number of bolts.
This link might help:-

http://books.google.co.uk/books?id=_8nVW5RMnjYC&pg=PA105&lpg=PA105&dq=bolt+loads+in+a+

bolted+blind+flange&source=bl&ots=JC6ou9m-oV&sig=lpJSMaf6FwEyklIjmvbzS4NppOM&hl=en&ei=CRhwTOKlNJOI
4gaTo53eCA&sa=X&oi=book_result&ct=result&resnum=5&ved
=0CCsQ6AEwBA#v=onepage&q=bolt%20loads%20in%20a%20bolted%20blind%20flange&f=false

desertfox

 

[westheimer1234]

this might help. design of monopole bases.

http://www.risatech.com/downloads/risatower/Technical_Manual_1.pdf

 

[Jimbo2010]

Thanks Guys

DesertFox thank you for your comment and I should have already mentioned that I have used PD5500 which I believe follows the Taylor-Forge Method and is much simpler then using Roarks.

Westheimer1234 Thank your for the document it is most interesting.

The reason for performing the assessment from first principles is that the Blank I am using is fundamentally a man hole cover and has two different cross sectional thicknesses, therefore is not of uniform thickness. The thicker section is n the centre and the thinner section is bolted to a coaming. A rough sketch has been attached.


Jimbo2010

www.engcosolutions.com

 

[dhengr]

Jimbo:
You still basically have (total pressure load)/(number of bolts) as an average bolt load, particularly if your sketch of the man hole cover is truly symmetrical and uniform. If, on the other hand, your stiffener ribs are at 12 to 6 o’clock and 3 to 9 o’clock and you have 12 bolts, then the bolts at 12, 3, 6 & 9, will be loaded slightly higher than those at 1 & 2 o’clock positions, etc. If the O.D. of the center thickened area is very close to the I.D. of the mating man hole, then you cold have a circumferential shear stress issue at the O.D. and you may have some prying load on the bolts due to bending of the thinner flange area of the cover. Take a 2" strip, O.D.+ long, supported at the 3 & 9 bolts and look at it as a partially fixed beam for a first shot at the stresses and deflections. Watch the reentrant corner if this is a cycled loading.

 

[desertfox]

Hi Jimbo2010

I've no idea how you would do such a flange with dual thickness, however if you calculate the bolt loads and bending based on a single thickness ie that of the outer rim, you could then use the code and assuming that the stresses etc are okay, I doubt you would need to worry if you increase the middle thickness of the flange,maybe the only thing to cause concern would be the stress raiser at the corner of the intersection of the two thicknesses.

 

[Jimbo2010]

dhengr,
Thank you for you insight. The manhole cover does not have any stiffeners within the plates. They are in the simplest form solid and perfectly symmetrical. I agree if it could be assumed the plate is thick enough to resist bending then it would be pressure load / no of bolts. However this is not the case and the effects of bending will have exert a force on the bolts in addition to the pressure load.

This is why I was trying to find and alternative method such as a work done analysis. If you find the deflection without the bolt load and the deflection with the bolt load (excluding pressure load) and work on this ratio times the circumference would the bolt load not be found?? (I have attached another pic for clarity). I have performed this analysis for perfectly flat plates before and compared it to PD5500 results and they are almost identical I’m just not fully sure why to use the method formally in practice?



Jimbo

www.engcosolutions.com

 

[Jimbo2010]

Desertfox,

In both Roark’s and Timoshenko “Plates and shells” methods are described for calculating deflections, stresses and shear values for plates with non uniform thicknesses both fixed and simply supported.. The problem is taking this and finding the loads that the bolts would experience. I agree that your method could be used to work out the worst case, but my system has some low safety factors so that’s the reason I’m striving to increase precision.


Jimbo

http://www.engcosolutions.com

 

[desertfox]

Hi Jimbo2010

Didn't realise Roark's covered varying plate thickness I'll have a look.
Another thought if you calculate the deflection of the blind flange without the bolts ie simply supported plate, then would it not be possible to use superposition and calculate the force required to reduce the deflection of the plate to zero, you would then have some idea how much total force you need to do this which could be divided equally amongst the bolts.
Failing that all I can see is an FE analysis coming on.

Regards

desertfox

 

[saplanti]

Please note that in the flange connections, if a gasket is used and depending on the type of the gasket, the gasket load for bolts might be a lot higher than the flange pressure load. Just follow PD5500 flange calculation methods, or check Taylor Method as suggested.

Does the connection have a gasket? If so, how are you going to calculate gasket loading on bolts? Are you using rubber o-ring type gasket that you can ignore the gasket loads?

If you can sketch it up, probably you will find more contribution.

Regards,

Ibrahim Demir

 

[rb1957]

i don't know your codes, but the bolt load should be higher than p*A/n, because of prying loads.

 

[cap4000]

Timoshenko and Gere's Mechanics of Materials 4th edition for your exact situation uses the Strain Energy Formulas for the bolts.

 

[Jimbo2010]

saplanti,

Thanks for your comments the gasket load will be concidered as per PD5500 and simply added to the loads due to shear and prying.

I have two systmens, one that has a double O'seal and one with a flat gasket.

Jimbo

http://www.engcosolutions.com

 

[Jimbo2010]

cap4000,
I have never seen this book before but will look into getting a copy. Is the example for bolts securing a plate such as this problem??

Jimbo

http://www.engcosolutions.com