Fatigue Design of Bolts per AISC

[structuralecstasy]

I am trying to understand how to design A325/A490 bolts in tension in fatigue applications using AISC 360-05 (13th ed.) and I'm looking for some expert opinion. The ASD 9th edition seems simple enough, but this edition and the subsequent one are not quite as straightforward.

Background: I have an agitator gearbox (vertical agitator shaft) that is supported on horizontal steel beams. High tension loads & high cycle rate (3 sec period), so fatigue is a major concern and the bolts have to be in tension. Even though this is for mechanical equipment, I prefer to use structural bolts since they can be properly pre-tensioned, as required by AISC and RCSC for bolts in fatigue applications.

My issue is determining the correct allowable stress range or allowable stress (understanding there is a major difference between the two) for the bolts. The problem is there are several places in the specs where this is addressed, and they all give different answers:

1. Per AISC 360-05, Appendix 3, Section 3.4 (b): For all bolts (does not differentiate between pretensioned and non-pretensioned), use Cf = 3.9E8 using Eqn. A-3-1, with threshold value Fth = 7 ksi. For 1,000,000+ cycles in this application, the threshold value controls, so the allowable stress range is 7 ksi. This doesn't seem right though, since Item 8.5 in Table A-3.1, which is specifically for non-pretensioned bolts, also uses the same design values. There should be a benefit to pretensioning the bolts.

2. Per AISC 360-05, Appendix 3, Section 3.4, last paragraph: Talking about pretensioned bolts with tension loads, there are two options given. A) An analysis of the stiffness of the assembly can be used to determine the allowable stress range due to the service loads plus prying, or B) The allowable stress range in the bolts can be taken as 20% of the service loads, not including prying loads. The difference between the two seemingly has to do with an analysis in A to determine the effects of prying due to pretensioning, where in B the effects of prying are accounted for in the 20% estimate. But they both say the allowable stress range is based on the applied service loads. This is where I get lost, as I don't understand how the allowable stress can be based on the applied load. I think I'm wrong here, but it seems like AISC is saying the allowable stress range is equal to 20% of the total applied load. The cyclic portion of the bolt tension load (almost 100% in my situation), would then be compared to this stress range, so I'm always overstressed by a factor of 5. Something's not right here either.

3. Per RCSC 2004, Section 5.5: For pretensioned bolts, the allowable bolt stress is dependent on the bolt grade and the number of cycles, per Table 5.2. I'd either use 31 ksi for A325 bolts or 38 ksi for A490 bolts. Note that this allowable stress is not defined as a stress range, so I'd compare the total applied load, not just the cyclic portion of the load, to this allowable stress. This value seems to make the most sense, but unfortunately it doesn't seem to line up with the other 2 AISC methods above. Also, AISC states in Section J3.1 that the RCSC should be used except where AISC states otherwise. This is one of those situations it would seem, so it's not clear if I can or should use the RCSC values.

Thanks in advance for any help!

 

[NS4U]

I'm not sure I follow questions 2 and 3. I didn't have time to dig into them. But for question 1: For the same applied load, the stress that a bolt "feels" is different whether you pretension the bolt or not. For example say you apply a 5 kip cyclic load, the stress range in bolts of a non-pretensioned connection may be 10 ksi, but if those bolts of pretensioned to make pretensioned connection the stress range may only be 1 ksi (I've made up the numbers). So, you can apply more load to pretensioned connections which will actually stress the bolts less than if you applied less load to the same a non-pretensioned connection. See Salmon and Johnson they have a good example of this for hanger connections. It has to do with the fact that you are reducing the compression between the connected parts and because of strain compatibility the bolt elongates very minimally and therefore doesn't see a big stress change.

I'll bet the guide to the design of bolted and riveted connections will have the answer to question 3, may be question 2 also.

[URL unfurl="true"]http://www.boltcouncil.org/files/2ndEditionGuide.pdf[/url]

 

[structuralecstasy]

Well, since my original thread, I think I figured out the intent of the AISC allowable values, parts 1 and 2 of my original post. For a large number of cycles (more than 1 million cycles), the allowable stress range is indeed 7 ksi for both pretensioned and non-pretensioned bolts, which at first doesn't seem to make sense. But the difference is that for non-pretensioned bolts, all of the design tension load must be accounted for in the bolt design. For pretensioned bolts, the bolts do not see the full tension load, only a portion of this value, since the applied tension loads acting on the bolt are offset by the stored energy in the pretensioned assembly (the compression of the steel plies from the pretension load creates stored energy in the plies). The literature says that pretensioned bolts see about 5% to 10% of the applied load as additional load, but AISC suggests using a value of 20% in Appendix 3.4. This means the actual stress range of the bolts can be taken as 20% of the applied load, and compared to the allowable stress range. So for pretensioned bolts, you essentially get to design for 20% of the loads that you'd have to design for if the bolts were not pretensioned, so there is a large benefit. In my OP, I mistook this 20% value as being the allowable stress range, rather than the actual stress range.

This makes AISC line up more with RCSC despite the difference in the approaches, but there is still a significant difference in the design values, since AISC requires using net tensile areas (bolt area at the threads) and RSCS allows using nominal areas (bolt area at the shaft). Designing A325 bolts produce about the same result, within 15% to 20%, but designing A490 bolts produces a design about 35% to 40% heavier using AISC instead of RCSC. These two specifications are intended to line up, but they do not for the design of A490 bolts in fatigue applications. I'd suggest sticking with the AISC approach over RCSC since it is more conservative, but more importantly because it is the code referenced by various building codes, not RCSC.

Side note: since AISC does not differentiate between A325 and A490 bolts when designing for tension fatigue, you might as well go with A325's and forget about A490's if fatigue is controlling the design.