Dangers of "Bolt Torque"

Posted in: , on 20. Mar. 2009 - 03:36

Dangers of "Bolt Torque" : Mill Failure Averted

This case study is a clear example of the danger of ingrained and, unfortunately, blindly-accepted bolting methods and assumptions This particular project clearly highlights the importance of load verification and fastener "tuning".:

One of our teams recently returned from an overseas SAG Mill Girth Gear replacement project. The owner engaged our services to ensure that the new bolts were tightened to the necessary preload during installation of the new gear segments (This had nothing to do with "torque" but rather with actual clamp load)

A description of the process follows:

Hydraulic torque wrenches were used to "tighten" the fasteners (Girth Gear-to-Shell and Head-to-Shell). We monitored the clamp load independently of the applied torque (by measuring bolt elongation). Most bolts reached the necessary clamp load/elongation at a hydraulic pump pressure of between 4,000 and 4,900 psi (variations caused by frictional differences and bolt interactions). However, there were three critical fasteners which did not achieve even minimal clamp load although the torque wrench "maxed-out" at 10,000 psi. This was caused by the nuts unknowingly galling into the threaded portion of the bolts during the tightening procedure.

To help illustrate this significance of this, it must be realized that "Torque" is not a measurement of a bolt's "tightness". It's only a measurement of the resistance encountered when trying to turn a nut. Torque specifications, unless accompanied by relevant elongation tolerances (as most mill OEMs now provide and insist upon reviewing before warranty coverage is extended) are very misleading and could be quite dangerous. Furthermore, "calibrated" torque tools, insidiously contribute to this false sense of security.

Had we not monitored the actual bolt clamp load, these very serious risks to operation would have been missed; the gear installation contractor would have simply "torqued" each fastener to one specific torque value. This would have resulted in many bolts not being tightened as per the designer's intention with three of the bolts being seriously under-tightened. The latter would have likely failed during service with a consequent forced shutdown of the mill. In the extreme, gear and pinion failure could have also occurred.

J Albrecht [COLOR="DarkOrange"][B][url]www.heviitech.com[/url][/B][/COLOR] [IMG]http://www.heviitech.com/Images/Hvi/HeviiReflect2.gif[/IMG]

Re: Dangers Of "Bolt Torque"

Erstellt am 20. Mar. 2009 - 12:46

These are a step in the right direction but, still don't provide a definitive indication of load. The biggest issue stems from the fact that with either, one can't tell if the joint has been over-tightned. Furthermore, load indicating washers provide no means of monitoring bolt load migration. Thus, the bolts can't be "tuned".

Rotabolts and their siblings can never be relied upon when configured as studs. In as much as they are 'calibrated', the accuracy and effectiveness of calibration depends on the effective length of the fastener relative to the rota pin. This length can change based on the position the nuts at the pin end. Since elongation of a tightened fastener is often only a very small amount, a very small change in the relative effective length of the fastener has a huge impact on the accuracy of this type of device!

There has been some development of this technology in attempt to allow indication of over-tightness. Some bolts actually change colour when tightened to the load which they were calibrated to and then change again when tightend beyond this level. Obvioulsy, interpretion of the results is highly subjective. However, as simple "go, no-go" indicators to quickly determine gross changes in load, they do serve a purpose.

Whereas rotabolts, load indicating washers or even simple torquing may be quite suitable for less demanding applications, the residual bolt load of critical fasteners under significant dynamic stress such as those found on mills should always be verified by actual measurement of elongation.

J Albrecht [COLOR="DarkOrange"][B][url]www.heviitech.com[/url][/B][/COLOR] [IMG]http://www.heviitech.com/Images/Hvi/HeviiReflect2.gif[/IMG]