October 09, 2005

Fatigue Fracture Pt.1 

During the early steam trains age, the owners of these trains (which sometimes were just sort of circus attractions, running on a small circular track) noticed a strange phenomenon: sometimes, steels rails - and even axles - just broke apart even if the load applied to them was well below the strength of the rails. Besides being curious in itself, this fact obviously had also serious implications: you cannot build a successful railway if your rails break apart for no apparent reason.

So, the engineers of the age began to study the situation, and built machines that could simulate the operative loads on rails and carriage axles, and discovered a very interesting fact: if a stress below the maximum design load of a mechanical part is applied and released repeatedly (thousands of times), fractures will develop in the part until its strength is so low that it will break even under absolutely normal operative loads.
And this is fatigue. If you want to try it at home, take a piece of iron or copper wire of 3 - 5 mm diameter, some 30 cm long and try to bend it back and forth in the same point: after a certain number of cycles (that depends mainly from the properties of the specific wire) the wire will break at the bending point, revealing a peculiarly ruvid surface, with tiny crystal faces in view.

Since then, fatigue has been studied extensively, and down to the microscopic level of how these fatigue cracks form and grow in the material's structure. Metals are the prime candidates for fatigue cracking - especially light alloys - but almost any material can suffer from it. Cracks generally originate where a stress concentration occurs: it can be something macroscopic such as a gas bubble or piece of slag remaining inside the metal from its melting and forging, or a tiny crystal defect, or a notch caused by mishandling of the part, or the stress concentration at a square corner.

If you take another wire like the one in the example above, and make a notch in it with a hacksaw, file or chisel and bend it back and forth again, it will break after a smaller number of cycles. Avoiding square corners is common design practice in many cases: when I designed a small pressure vessel, the senior engineer supervising the project told me to design the vessel with radiuses instead of square corners at the wall-bottom junction and at the flange-body junction, expressly to prevent stress concentration.

A fatigue crack is born as a tiny one, just microns deep. But the stresses will concentrate at the tip of the crack, and make it a bit deeper with each cycle - until eventually the piece fractures. A crack's growth can also stop, if it encounters a region of material with different characteristics - but this is an issue I do not know very well.

Fatigue is an issue in almost all metal structures, but it becomes of life and death (not to mention economic) importance in aeronautics - and railways in the second place.

These are both fields of high stresses: trains travel at less than 300 km/h in general, but they are massive things, with big steel bogies, wheels and axles. With a huge mass, even a small accelaration is enough to produce a considerable force, and the rolling action of a wheel on a rail is a very violent phenomenon, that leaves the top layer (a few micrometers) of the rail completely messed up. Fatigue cracks can generate in this layer, and grow vertically in the rail with the bending loads that a train wheel applies to the rail. Following one serious accident in England, investigators discovered that, due to fatigue cracks, a 1 m long section of rail disintegrated in 200-odd pieces when the train passed over it.

Ah, finally Blogger behaved. Stay tuned for the second part then.


Post a Comment

This page is powered by Blogger. Isn't yours?