TechArchive

Original article date: April 2000

We asked KEN FARROW, head of materials engineering services at Materials Engineering, to guide us through the problems that can cause materials failures

Things are not always exactly what they seem. When something breaks down, the immediate response is to want it to be repaired and back in service. And the more critical the item and the higher the capital investment, the sooner it is needed to be back in service.

A common cause of failure is a fracture of a bolt or other fastener. A simplistic solution would then be to put in another one, but, since it has failed, perhaps something stronger. Unfortunately, that solution may not work and can lead to a repeat failure, possibly even sooner and more dramatically.

Why? Mechanical overload of a bolt could develop owing to tensile, bending, torsional or shear stresses. But a bolt can fail as a result of several processes other than just overload. Even the overload failures can take more than one form. At the extremes, these would be ductile overload or brittle fracture. A ductile fracture would be preceded by plastic deformation – stretching or bending would tend to provide some warning that all was not well, possibly with progressive tearing taking place. A brittle fracture could take place, with negligible prior deformation occurring before a sudden unexpected failure took place.

The factors favouring brittle fracture include sharp stress concentrations, high rates of loading and low temperatures. Conventional steels undergo a definite ductile-to-brittle transition with changing temperature. Above the transition, the steel will behave in a ductile manner. Below the transition temperature, the steel can be extremely brittle. However, it is less well appreciated that the ductile/brittle transition temperature can even be above ambient, depending on heat treatment history.

There are other failure mechanisms: fatigue cracking is a common failure mode for bolts and develops as a result of repeated applications of loads, which can be well below those needed to produce yielding. Those loads could be cyclical, periodic or intermittent, but in general terms, the lower the load the greater the number of loadings required. Contrary to myth, the metal does not recrystallise, but instead develops an expanding crack, which eventually leads to an overload fracture of the remaining uncracked metal. Factors favouring the development of fatigue cracks include sharp notches and residual surface tensile stresses. The fatigue life is also reduced by increasingly corrosive environments.

But many fatigue failures of bolts are the result of incorrect tightening or the bolt becoming loose. Friction grip bolts are designed to transmit load through frictional forces between the parts they hold together, but, if loose, those bolts could see fatigue loads or even become mere shear pins. If bolts with machined ‘cut’ threads were supplied instead of rolled threads, then it would not be surprising if they suffered premature failures. For different reasons, bolts produced from ‘free machining’ steels can suffer early failure, for example where corrosion has developed around the inclusions in a free-machining stainless steel. Other mechanisms potentially causing failure include stress corrosion cracking (SCC) and hydrogen embrittlement (HE), both of which can have similarities. SCC needs stress and environment together and results in a development of cracking which can lead to a partially brittle fracture face, the rest being overload. Such cracking can develop with austenitic stainless steels, which can suffer SCC if stressed whilst exposed to hot chloride solutions, or with brasses exposed to ammonia compounds. Hydrogen embrittlement can develop due to hydrogen picked up during manufacture (electroplating), fabrication (electric arc welding) or in service, due to corrosion product hydrogen.

Hydrogen picked up during electroplating tends to diffuse out of steel bolts, but cadmium plating acts as a barrier. Baking is normally carried out to remove the hydrogen, but if that is not done then failure in service can develop. Since the hydrogen movement is diffusion controlled, that failure can take time. Similar delays can occur where welding results in hydrogen pick-up, such as when tack welding a fastener.

Whatever the source of the hydrogen, the hydrogen embrittlement of steels is more of a problem the higher the strength of the steel and is a relatively low risk with low strength steels. In general terms, SCC is also more of a problem with higher strength materials.

Depending on the cause of failure, the use of a higher strength material for replacements can be better or worse. But a variety of problems can cause failures, like fatigue failures due to inadequate torque or delayed failure due to hydrogen not baked out after a cadmium plate.

So things are not always what they seem. And hence the need to investigate…

• Materials Engineering

April 2000