Close

November 30, 2018

Case study – Bolt Failures

Strand7 model

If the design forces on a bolt group are less than their code-calculated capacity it doesn’t always mean a satisfactory design for the extent of the design life.

That’s what our client discovered when the bolts on their vibratory screen kept failing, even after replacing them with bigger ones, so they came to Yenem for help. We’re known in our industry for doing the hard jobs. This one sure got us thinking!

 

Failure study - bolts that should have worked

The image shows the location of the connection within the red box.

A quick hand calc suggested M30 bolts were okay. But clearly they weren’t due to a number of failures so we needed to do something else in order to find the cause of the failure.

In order to accurately simulate the bolt connection under working conditions, our graduate engineer Shawn Cao built a solid Strand7 brick model to include part of the exciter box and the bolt connection with all clamped parts. Strand7 is our FEA software, FEA being Finite Element Analysis, which in its simplest form means a method for calculating stresses and deflections in objects of any shape and any material. We use it often, whenever a structure can’t be modelled properly using basic beam/column software like Space Gass and Microstran.

The Strand7 Model

Failure study - bolts that should have worked

 

The connection block mounts the exciter shaft onto the drive beam, which is fitted between the walls of the screen.

As shown in the image, only a quarter of the whole connection was modelled. This is because of the symmetry of this connection. Even this quarter model has more than 10,000 brick elements, comprising the plates, and the bolts, which are pre-tensioned to clamp the components together.

 

The connection block mounts the exciter shaft onto the drive beam, which is fitted between the walls of the screen.

As shown in the image, only a quarter of the whole connection was modelled. This is because of the symmetry of this connection. Even this quarter model has more than 10,000 brick elements, comprising the plates, and the bolts, which are pre-tensioned to clamp the components together.

What we found

Failure study - bolts that should have worked

  • The modelling showed us that due to the connection arrangement, and the pre-tension in the bolts, the bolt stresses were different – higher – than the hand calculation provided.
  • 8 off M30 bolts were not enough, and a hand calculation wasn’t good enough.
  • We revised the model to use M36 bolts and changed the amount of pre-tension force – the result came out positively. Bolt stresses and clamped parts were within the allowable stress range.
  • Changing the pre-stress changes the distribution of forces in the bolts and resulting stresses in the bolts and the plates.
  • Insufficient pre-stress results in cyclical stresses in the bolts that can lead to early failure due to fatigue.
  • Sufficient pre-stress eliminates the possibility of bolt fatigue in dynamic structures.

The primary possible causes of loss of the pretension

  • The bolts are incorrectly tensioned.
  • Embedding of the contact plates.
  • Surface coating – creep can become an issue when the coating is generating an added total deformable thickness.
  • Rotational self-loosening caused by axial loading and vibration, torsional loading and transverse vibration.

Recommendations – applies to all bolted connections

  • Use bolts that are compliant with AS 1252 Parts 1 & 2. This is a new standard that ensures bolts sourced locally and overseas meet the material, ductility and strength requirements expected.
  • Ensure bolts are correctly tensioned. This includes ensuring mating surfaces are free from lubricants, anti-seize compounds etc, and following a certified and compliant procedure and validation method.
  • Pre-tension bolts to 80% of proof-load. As a higher preload can raise the decompression point, it is better to pretension a fastener close to its limit rather than only partially pretension it.
  • Don’t over-tension a bolt. There’s no added benefit from a pre-tension greater than 80% of proof-load.
  • All clamping parts should be configured symmetrically. Asymmetric design can cause bending of the bolts due to uneven pressure on the mating surfaces. An FEA analysis can assist with this.
  • Self-loosing cannot be avoided. Inspect dynamic connections at 6 monthly intervals. If self-loosing is detected, replace and re-tension the bolts. Locking nuts may assist.

Like this article?

Subscribe to our newsletter to receive a weekly email from Yenem’s director, Dave Meney.

Subscribe To Our Newsletter

Read More Articles

Leave a Reply

Your email address will not be published. Required fields are marked *