Bolt assembly stress vs. allowable stress

[FFS-EU]

When assembling flanged joints, a typical target bolt stress is 40-70% of its SMYS, see e.g. ASME PCC-1 Appendix O. If you look at the allowable stresses for bolt materials (e.g. SA-193-B7) in ASME II Part D, you will find that these are much lower. That is, when assembling a regular flange connection, we stress the bolt (way!) beyond its allowable stress.

This makes me wonder:
1. Do you agree with the above: i.e. we do not consider the allowable bolt stress values when calculating/applying the target bolt assembly stress?
2. What is the significance of the allowable bolt stress values specified in ASME II Part D, for which calculation should we use them?
3. When designing a flange, the bolt stress is one of the loads that matter: It leads to circumferential bending of the flange. Which bolt load would you enter in the flange design calculation? Equivalent to:
(a) bolt allowable stress (which is less than actual),
(b) target bolt stress from torqueing,
(c) same as (b) but increased with a conservative safety factor?

Any other thoughts on the subject - please feel invited to share.

 

[desertfox]

Hi FFS-EU

I don’t have access to the ASME standards but I will try to answer as best as I can.

1/ you should always consider the allowable bolt stress because if you ignore it there is a good chance the joint will fail.

2/ cannot comment on the ASME however the significance of the allowable bolt stress will play a huge role depending on the joint function and whether it is subject to fatigue or not. Usually using bolt stress at around 40-75% of SMYS is for joints that may be taken apart and reassembled and using up to 90% SMYS are for joints that are not meant to be taken apart but if the are then new bolts and nuts should be used.

3,/ if you can post a flange design calculation I might be able to comment further however from memory I would derive the bolt preload force based on the joint pressure or external loads in order the ensure the joint did not fail, once I have the bolt preload force I would look at the bolt stress that this preload caused in various bolt sizes until I hit my required allowable stress and then look at the stress due to bending the flange, bolt head embedding etc. and adjust the joint design accordingly.

“Do not worry about your problems with mathematics, I assure you mine are far greater.” Albert Einstein

 

[drawoh]

FFS-EU,

I don't have access to ASME specs either.

In the past, I have had problems understanding bolt stresses. Think instead of bolt strain. You strained the bolt when you installed it. If you apply force to a flange insufficient to pull it apart, then you have applied no additional strain on the bolt. That means that there is no additional stress either. This is why for a lot of applications, we torque bolts down to 90% of yield or proof stress. The fun, and the metal fatigue, starts when the flanges separate.

--
JHG

 

[rb1957]

I also ... don't have ASME, and likewise will comment.

Allowable stress for the bolt material is IMO for non-bolt applications, including safety factors applicable to general purpose designs. Bolts are I think "special". The guidance to "wind them up", to preload to 70% fty is typical, to prevent bolts gapping under load. The "problem" with bolts is that preload and applied load interact on the bolt (when it is ungapped, or clamped up, as we'd like it to be). So you can limit your applied loads to the material allowables if you want and then add preload (with an unfactored allowable or using the bolt allowable strength). Eg the applied load can be limited to 40ksi (or whatever you have for common bolt steel), but the preload + applied load should be less than the bolt allowable load ... and that'd be preload factored up for probable scatter).

another day in paradise, or is paradise one day closer ?

 

[Scuka]

@ drawoh

Good explanation. It's useful to think of a bolt as a spring (i.e. think in terms of strain as you say)

https://upload.wikimedia.org/wikipe...svg/440px-Bolted_joint_spring_analogy.svg.png

 

[desertfox]

Actually the bolt is strained further when the external load is applied to the joint however the additional strain is quite small because the joint stiffness is usually much greater than the bolt stiffness so most of the external load reduces the compressive stress on the clamped faces but equilibrium must be maintained and so as the joint faces relax due to the decrease in compressive stress the bolt as to elongate to accommodate that movement.

“Do not worry about your problems with mathematics, I assure you mine are far greater.” Albert Einstein

 

[FFS-EU]

Many thanks for the interesting comments. Allow me to react:
- Indeed we think of the bolts like springs. The figure on wikiuploads is very instructive to visualize this first principle. @Scuka, I gratefully downloaded it. Yet, I also agree with desertfox that you have to consider the flange as a spring, too. Often the flange's spring is less flexible than that of the bolts, but hardly ever negligible.
- @desertfox In my opinion answers 1) and 2) in your first post contradict each other. It is this very contradiction that lead me to posting this question. If you limit your bolt assembly stress to the ASME II Part D allowable stress, you end up below 40%. This would be bad in practice as many other posts explain that the pre-stress is to compensate for the in-service loads to avoid flange opening.
- @desrtfox Our flange calculations have to respect various limits also mentioned by you. The gasket seating pressure knows a minimum required and a maximum allowable, similar for the bolts (however, these limits are roughly 40-70% YS and not allowable stress as per ASME II), and a maximum stress in the flange due to bending. The calculated values will depend on materials, geometries, and operating conditions.
- @rb1957 I think that you indeed hit the answer to my basic question. I also think that the ASME II Part D allowable stresses are meant for general constructions - but flange assembly. However, I do not know any reference to this. Your opinion shared helps a bit in this respect.

 

[desertfox]

As previously stated I haven't got access to the ASME standard but lets say I needed the bolt stress below 40% of the yield stress, on that basis I would initially start as follows having established the external force on the joint.
Divide the external force by the number of bolts I wanted to use to obtain the amount of external force each bolt would see/, Then use the cross sectional area of the bolt I wanted by dividing it into the portion of the external force I obtained from the previous answer, if at that point the stress in the bolt was to high I would either increase the number of bolts or the size of bolts to accommodate the required allowable stress, finally I would check the amount of increased strain using the approximate stiffness of both joint and bolt and again adjust bolt size or quantity if the stress went over the allowable. If you are using flanges to a standard where the number and size of bolts are fixed in value then usually in the same standard there is a maximum preload or torque figure given for that particular flange arrangement.
So that's my interpretation of using the allowable bolt stress so I don't a contradiction.









“Do not worry about your problems with mathematics, I assure you mine are far greater.” Albert Einstein

 

[iangineer]

I have access to the standards.

1. You want to use Table Y-1 for the yield strength values of bolts. That 40% to 70% is percent of yield strength. yield is different than the max allowable.
2. Not sure which table you are referring to, but Table Y-1 is the one for applying to PCC-1 Appendix O
3. When bolting a flange, your main concern is seating the gasket and maintaining that seat during operation while avoiding damage to the flange from the moment created by the bolts from the gasket.
(a) Use Table Y-1 Yield strengths
(b) You need information on the stresses needed to seat the gasket. You want to stress the bolts from torqueing enough to stress the gasket. The flange thickness needed for a particular bolt load can be found in ASME VIII-Div1 UG34(c)(2)eq(2). I check that equation first to determine how much stress my bolts can have and not bend the flange. Then I can determine the torque based on bolt stress from there.
(c) the 70% of yield is already a safety factor.

Determining bolt torques is currently a major part of my job, so feel free to ask additional questions and I'll try to guide you in the right direction when I have time.