Bolt Failure 1

[dik]

I've attached an interesting picture of a Grade 5 bolt failure. Failure was one of 12 and was catastrophic.

We were trying to press a pin from a hydraulic assembly and I designed a frame so that the bolts would fail before the steel web of the cross beam would 'tear out'. The assembly was designed for 190K and the contractor took the load up to 190 tons (by intent, not accident).

The bolts were loaded in nearly pure shear with the transfer of shear to the head and nut being via tension. As the bolt failed, initially in shear and then in tension, you can see the shear 'crescent' and the grain boundry failure of the bolt in tension. The energy stored in the loading frame caused the catastrophic failure of all 12 fasteners (failure appeared instantaneous). Slightly developed 'Luders' bands are visable under magnification in the crescent area.

Dik

 

[fegenbush]

Dik,

Nice photo. I love it when the photo is large enough to see the failure mechanism.

What size was the bolt and in what pattern were they arranged?

 

[dik]

7/8" dia bolts and in two rows of 6. Failure load was nearly twice the limit load.

Dik

 

[MikeHalloran]

Nice picture.

Now, how you gonna get the pin out?



Mike Halloran
Pembroke Pines, FL, USA

 

[dik]

We were trying to extract it intact. It was assembled 40 years ago as a sliding fit... and a little corrosion later... The pin, 8"x3" wide, has been frozen in time <G>. The original faces had been machined down 1/2" in the interior to provide less surface...

We tried heating, penetrating oil, vibration, and 190T and didn't budge it... I suspect it may have 'slipped' due to loading and created a 'ridge' within. The force may be pushing against the ridge. The earlier pin... almost identical condition was pushed out with 19K...

Dik

 

[TXStructural]

Nice job on choosing your failure mode and matching materials to your intent. And also on (apparently) not letting anyone be exposed to the danger when it did go.

 

[dik]

Isn't that what hard hats are for?

I had written up a procedure for safety... I didn't think the contractor would take the load up as much as he did... With hydraulic load cells, when the deformation starts quickly, the load drops off. I was concerned with the energy stored in the loading frame. It had a hydraulic cell pressing against the pin and on the opposite end, there was an HSS tube that surrounded the pin and pressed against a 2" steel plate that the pin passed through.

Dik

 

[Ron]

dik...I'm confused!

The bolts were loaded in nearly pure shear with the transfer of shear to the head and nut being via tension.

The excellent photo shows tension as the dominant failure mode. The "clean" failure surface also indicates a brittle fracture as does your description. Any chance the bolts were not Grade 5?

 

[dik]

Ron... that's why I posted the pix!

On the top of the assembly, six bolts were loaded in single shear bolting a steel plate to the web of a 24" deep S section. On the bottom of the assembly, approximately 3' away, the same condition was mirrored. The two plates were connected with a W beam between and the beam resisted the hydraulic load cell placed at mid point of the beam, putting both plates in nearly identical tension. This tension was transmitted to the six bolts via (6 bolts top and 6 bolts bottom) shear and put the S section web in tension. A similar restraint was constructed on the opposite end of the S section except a brace for the material surrounding the pin was constructed.

The load cell put a force of 190 tons on the beam, pressing against the pin on the opposite end of the cell. There was a lot of energy stored in the system at failure. The overall length of the assembly was about 12'.

As the bolts were loaded in shear, because of the relatively thin plate and web, the bolts resisted the shear by developing a tension stress between the nut and the bolt head. What I suspect happened is the bolt(s) started to fail in shear (all 12 had nearly identical markings) as shown in the small crescent shaped smooth area. Because of the large amount of energy in the system, as they failed, as soon as the first slipped (sheared a bit), they all failed, with the induced tension providing the brittle fracture shown over the balance of the bolt. The failure surface (brittle looking part) for all bolts is pristine with no damage because they elastically shortened fast enough to pull away from the shear plane. The assembly went from load carrying to broken in a split-second.

The bolts were grade 5; I had spec'd 7/8"dia A307 (the holes in the S beam web were for 7/8" dia bolts). On short notice the A307's were not available and grade 5 was used as an alternative.

Dik

 

[dik]

Shearing deformation of bolt and bolt head...

Dik

 

[dik]

ditto... difficult to get a close shot of deformation... too many planes...

Dik

 

[Ron]

dik...were the bolts pre-tensioned? If so, to what level?

 

[dwallace1971]

Why would you design joints to be shear-critical?

 

[ToadJones]

still don't understand the tension failure mechanism.
It looks like the bolts began to shear and maybe started to bend as the plate pulled away.

Putting the bolts in double may have provided a more reliable failure load.

 

[dwallace1971]

Keep in mind that a torqued bolt can be subject to both tensile and shear loads.

 

[dik]

As the bolts started to bend, they went into tension, I suspect. From the failure surface, it's the best WAG that I could come up with for the 'brittle' failure mode. I don't know what the effects of the very rapid failure are with the added effect of the large amount of stored energy released.

Putting them in double shear would have required a weaker bolt because I didn't want the web of the S beam to tear out (A36 material); I wanted the bolts to be the weak link. A307's were about as soft as I could use (and then they couldn't get them.)

It failed as it was supposed to, albeit at twice the load. The holes in the S beam web deformed a bit and it was likely the higher ultimate strength of the A36 material that prevented the web failure.

As interesting as the failure is, I'm not overly happy that the contractor overloaded the system in this fashion.

Dik

 

[dwallace1971]

dik, what was the purpose behind designing the bolts as the "weak link"?

 

[dik]

Ron:
The bolts were 'snug-tight' only... no significant pre-load.

Dik

 

[dik]

dwallace...
The bolts were designed as the weak link as not to damage the S beam assembly. They were loaded in shear (with threads excluded) because this was the most reliable failure mode and the manner of loading lent itself to this type of connection.

The S beam assembly consists of two 24" S beams separated by two built-up C sections with a 2" thick web. The plates are separated by about 16". The 16" dia hydraulic ram shaft fits between between the plates and is secured in place by the 3"x8" pin we were trying to extract. It was a matter of extracting the pin without damaging the S beam assembly.

Dik