Sleeve Nuts 1

[RonJeremy]

I have been wracking my brains over the theory behind sleeve bolts (some people call these compensators).

It is a sleeve to extend the stud for high temperature service.

Every time I try to do the theory I seem to come up with a different answer. Example: one time I try to maintain constant gasket stress by holding gasket strain between ambient and operating (bolts colder than flange); another I will try to ensure that the gasket does not unload in shut-down. Every time I come up with different answers, some of which do not even seem realistic.

However, sleeve nuts have been used countless times, so where am I going wrong???

In the 1990s I had a book called 'Theory and Design of Pressure Vessels' by Harvey, which had the theory behind this (about page 435 - I used to catalog pages when I was young and keen!). Lent it to my Client's engineer-weasel and never saw it again - Thank You Shanghai Chemical! - Hope you enjoy it.

In a similar vein, I am trying to work out whether Belleville washers make sleeve bolts redundant (another subject I think addressed by Harvey).

Tough for me to ask for a freebie on this, but with the downturn in O&G I cannot really face buying another copy of Harvey just for this one question.

Thanks for your time.

 

[TGS4]

Bolt sleeves increase the grip length of the bolt, which in turn decreases the stiffness of the bolt - total elongation is Δl=l*F/(A*E), where:
Δl is the total elongation due to the axial force F
l is the grip length of the bolt (i.e. from inner nut face to inner nut face)
F is the axial force in the bolt
A is the effective cross-sectional area of the bolt
E is the Young's Modulus

During events which result in a change in temperature from installation, the thermal expansion imposes its own Δl. Depending on the situation, this may increase or decrease the axial force on the bolt. But, the longer the bolt, the less that the value of F will change.

Belleville washers work on a similar principle, by adding some springy-ness between the flange surface and the nut face. The goal in both is to maintain as constant a bolt force (and hence gasket contact force) as possible, throughout a temperature change.

 

[RonJeremy]

Excellent! Many Thanks!
I wanted a way to get a sleeve washer to equal Bellevilles, but came up with three different answers (depending on parameters). Your answer appears to work. Thanks again!

 

[r6155]

a) Pressure Vessel Design Manual 4th Ed., by Dennis R. Moss and Michael Basic

Studs. Studs used for high pressure applications are integral to the overall design, the performance of the joint,
and functioning of the vessel. For these reasons, it is imperative that every design detail receive the appropriate
attention.
Typically the threaded shank of the stud creates a stress concentration. Numerous tests have shown that this is the most likely area of failure. To preclude this, and minimize stress concentrations, a generous radius is machined into the shank of the stud at the location of the first thread. This is critical for cyclic service or high temperature operation. Any time high temperature or fatigue are involved, the studs should have a generous radius machined at the root of the first thread.

Sleeve Nuts. Sleeve type nuts are utilized whenever normal spacing of nuts based on wrench clearances cannot
be met or is not desirable. Sleeve nuts are a smooth sleeve of metal that is internally threaded. The nuts may have a nut machined to the top of the cylinder to accommodate hand or wrench tightening. However, they are not designed for manual tightening.
Most high pressure applications do not utilize wrench tightening anyway. So why base the bolt spacing based on
wrench clearances?
In this case, the nut is not there to create the stud tension, but only used to secure the elongation of the stud. Sleeve nuts achieve this goal with a minimum of metal. Sleeved nuts may have holes drilled in them to facilitate the turning of the nut during stud elongation.
Sleeve nuts have longer threads to develop full strength as opposed to nuts. Sleeve nuts should have a minimum of
10 threads to ensure full strength.


b) AD 2000 MERKBLATT B7 BOLTS

2.2 In order to make a bolted connection as elastic as possible, it is recommended to design it with neckeddown bolt to DIN 2510. Neckeddown bolts should be applied if the design temperature exceeds 300ºC or if the design pressure is higher than 40 bar. Care has to be taken that adequate effective bolt length is available which can be increased e.g. by sleeves to DIN 2510. The length of the bolt shank must be at least twice the thread diameter
2.3 Neckeddown bolts are regarded to be bolts having a shank diameter ds≤0,9 dk or having dimensions according to DIN 2510. Bolts with no reduced shank are regarded to be rigid as far as design is concerned.

Regards
r6155

 

[RonJeremy]

Thank you r6155,

I wrongly referred to 'sleeve nuts': what I meant was 'sleeve washers' which are not internally threaded - they are simply plain pipe with ends machined parallel to each other, as described by Harvey (mentioned above) and EAD Saunders in 'Heat Exchangers: Selection Design and Construction'.

I do have a copy of Moss (3rd) and, before posting my query, had been spending a while manipulating fig 2-24 'Typical Joint Diagram' (ibid) for different bolt grip lengths. In other words: a longer grip length would result in a shallower 'elongation of bolt line' as shown in Moss. However this did not answer my question - I merely ended up by shifting the point at which the line of bolt load and the line of gasket load meet to the right. So, just based on increasing bolt length does not have any effect on the design of the flange according to Moss (3rd) [which is true in a way, since bolt length is not a consideration in BPVC VIII-1 Apx 2]. I do not know whether Moss (4th) gives any further clarification of fig 2-24 that does answer my OP, which you seem to suggest it does.

TGS4 has stated above that an increased grip length compensates for thermal expansion, which is true, but only (assuming identical coefficients of expansion for flange and bolt) if the bolt is considered to be at a lower operating temperature that the stud. This is because DELTA.L(f) is proportional to L.F and DELTA.L(t) is proportional to L.DELTA.T.
Where,
DELTA.L(f) is displacement due to bolt force
DELTA.L(t) is thermal expansion
L is grip length
F is bolt load
DELTA.T is difference in temperature between operating and installation.

So, since both are directly proportional to L, L cancels and therefore the use of sleeve washers has no net effect in this case.

HOWEVER, with uninsulated bolts (say, B16 studs and 1Cr0.5Mo flange and sleeves, both of which have coeff of exp as shown in Column C Group 1 of Table TE-1 of ASME II-D) it can be seen that, if (say) the operating temperature of the flange is 500 degC (linear thermal exp = 6.9 mm/m from ambient) and that of the stud and washer is 275 degC (linear thermal exp = 3.4 mm/m from ambient) then having a sleeve washer that extends the grip length to roughly twice (6.9/3.4) of the overall thickness of the flanged joint will ensure that the load on the gasket under operating conditions will be the same as that at installation.

 

[SwinnyGG]

So, since both are directly proportional to L, L cancels and therefore the use of sleeve washers has no net effect in this case.

This is only true in a case where the steady-state condition has such high thermal conductivity into the bolt (or low conductivity out of the bolt) that the flange and bolt are always the same temperature

In the real world this is unlikely.

The other point is that in the real world, you have to account for the transient condition. Even if the actual service of the system in question involves being brought up to temperature and then living at that temperature until the end of time, the connections still have to survive the temperature gradient between ambient and full service temp.

 

[RonJeremy]

Yeah, that is why I followed that paragraph with:

"HOWEVER, with uninsulated bolts ..."

I accept what you say about the transient conditions, which are sort of covered in U-2 and UG-22 of BPVC.

 

[r6155]

See Roark’s Formulas for Stress and Strain
Impact and Sudden Loading

When a force is suddenly applied to an elastic body (as by a blow), a wave of stress is propagated, which travels through the body with a velocity
V = square root(g E / δ)

where E is the modulus of elasticity of the material and δ is the weight of the material per unit volume.
When one end of an unsupported uniform elastic bar is subjected to longitudinal impact from a rigid body moving with
velocity V, a wave of compressive stress of intensity

σ= v E /V = v . square root(δ E / g)
is propagated.

The intensity of stress is seen to be independent of the mass of the moving body, but the length of the stressed zone, or volume of material simultaneously subjected to this stress, does depend on the mass of the moving body

For example: bolt d= 25mm x L= 100 resist the same static load than a bolt d= 25 mm x L= 150mm, but resist less impact load . Hence, a sleeve is required to increase the length of stud-bolt.

Regards
r6155

 

[RonJeremy]

Thank You r6155,

Well, as I said in my original query: 'Theory and Design of Pressure Vessels' by Harvey did discuss the theory behind sleeve washers. I have only hazy recollections, but can categorically state that it had nothing to do with impact loads.
I fail to see how "sudden impact loading (as by a blow)" is relevant to flange bolting (even for sudden changes of pressure).
Moreover, your quote from R&Y "The intensity of stress is seen to be independent of the mass of the moving body, but the length of the stressed zone, or volume of material simultaneously subjected to this stress, does depend on the mass of the moving body" comes under the heading in R&Y 15.3 "Bar with free ends" which is what a flange bolt is not.
In any case, R&Y state in 15.5 that: "it is improbable that in any actual case of impact the stresses can be calculated accurately by any of the methods or formulas given above" (which includes yours).
Therefore, I have to conclude that this particular avenue of discussion appears to be a dead-end.

Many Thanks
RJL

 

[r6155]

See more in

https://www.hindawi.com/journals/sv/2017/2829783/

"Failure Characteristics of Joint Bolts in Shield Tunnels Subjected to Impact Loads from a Derailed Train"

Regards
r6155