Neglecting Pre-Tension in Bolt Failure Calculations 6

[Taz99]

Hi All,

When considering steel bolt shear and tension interaction failure I use the Rt² +Rs³ < 1 failure criterion. Where Rs and Rt are the ratio applied load / allowable load in shear and tension. I have never considered the pre-tension load in the bolt in this calculation.

Recently the question came up about the effect of Pre-tensioning the bolts to 80% of yield strength. Considering that almost all bolts in service are pre-tensioned on installation why is it not standard practice to use as the applied tension load the highest of the pre-tension load or the actual applied tension load?

The specific question that was asked was the case of a bolt loaded near ultimate in shear, with a low applied axial load, where this bolt would be acceptable if the applied axial load was used in the interaction equation but considered to have failed if the pre-tension axial load was used in the interaction equation.

Thanks
Taz

 

[rb1957]

if the joint is gapping at ultimate load then you don't need to consider preload.

if it isn't then you should.

 

[Taz99]

Thanks for the reply, but accounting for pre-load doesn't seem to match with the actual stress analysis practices in industry.
I have looked at Bhrun, Flabel and HSB 20210 and none make any mention of requiring you to consider the bolt pre-load in the bolt shear-tension interaction calculation.
Bhrun page D1.3 mentions that the interaction curves were generated with finger tight nuts, but in his example on page D1.9 the applied tension force he uses for the bolt is 5800 lb tension, however the bolt has a tension allowable of 30000 lb ult and could reasonably be expected to have approx 19000 lb pre-load if torqued to 80% of 100 ksi yield stress, however the effect of pre-load is ignored. Maybe you could make an argument that this is a simple example or there was no pre-load on these bolts?
Additionally, when you look at HSB20210 it specifically states for Hi-Lok bolts the interaction curve Rt² + Rs² is applicable and for these types for bolts, but it is not possible to install these bolts without pre-loading them in tension, but again there is no stated requirement to consider pre-load in the analysis, and I am yet to see a Hi-Lock (or any other bolt) analyzed accounting for pre-load in any check stress work that has come across my desk.
So my thinking is still that it is industry practice to ignore pre-load for bolt interaction analysis.
Are people out there ignoring pre-load or always accounting for it?

 

[TVP]

The standard method is to use Rt² +Rs² < 1 as the failure criterion. Here are two references that you might find useful:

Aerospace Threaded Fastener Strength in Combined Shear and Tension Loading

http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20120003667_2012004005.pdf

Preloaded Joint Analysis Methodology for Space Flight Systems

http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19960012183_1996112183.pdf

 

[ESPcomposites]

It is standard practice to not include the preload in the shear calculation. You are also right about the interaction of tension and compression in a fastener. So what gives?

Consider this. If there is a preload in the fastener, then friction between the members will carry some of the shear load. Therefore, the applied shear load is not actually be transferred by the fastener. Conversely, if the entire shear load is being carried by the fastener, then there there would be not friction between the members (i.e. no preload). So it is not realistic to have the fastener carry both the entire preload and shear load.

In turn, the aerospace industry just uses the applied shear and neglects the preload (from what I observed and consistent with typical aircraft references). This is also typical practice by the aircraft companies.

Brian

http://www.espcomposites.com

 

[Taz99]

thanks, this makes sense to me "it is not realistic to have the fastener carry both the entire preload and shear load."

 

[rb1957]

ok, but i would include preload if the bolt wasn't gapped and shear (yes conservative, but i'd prefer to keep friction as an unknown).

if the joint is gapped (if the applied tension load exceeds the preload is a simple though not rigiously accurate) then the bolt reasonably experinces the applied tension and shear loads.

if the joint isn't gapped then the preload is being felt by the bolt and i would include the preload and the applied shear. Many places (OEMs) will use bolt stiffness models to include the effect of the tension load in the bolt slowly increasing as external tension is applied; as opposed to the simplistic assumption that the bolt reacts it's preload untill the preload is exceeded, and the joint gaps and the bolt carries the applied load.

using Rt^2+Rs^2 = 1 is a simple and conservative failure rule. you can squeeze some more out of the joint with other rules.