Is there a means of testing rails for flaws before/after/cropping/welding (whatever) before they are placed in the track?It is routine to inspect every weld within a few days of the weld being performed using a hand held ultrasonic detector, but to do that along the whole length of every rail would be very slow, and tedious as said by The Meddling Monk. Rail Flaw Detectors are usually machines mounted on bespoke hi-rail vehicles. Even so, when they detect a possible defect, it is confirmed (or dismissed) with a hand held unit.
In completely relaying the Ararat - Maryborough section there was the opportunity to produce a first class job with 'proper' concrete sleepers and heavier rail but that opportunity has now been lost.
From the jaws of success we seem able to seize failure every time, although from previous experience elsewhere good 80 lb rail, properly welded, should have been adequate for the task.
Weld failure rates are usually part of the contract specifications. A failure rate of 2% is tolerable for alumnothermic welding, but I should think a lower failure rate would be expected for flashbutt welding.
There is a lot of assumption and misinformation regarding ultrasonic testing of rail on this thread. Some bits are right, some wrong, so careful.
Ultrasonic testing with machine based systems are very good (no way would someone contemplate hand testing outside of defect sizing). An aspect missing in the discussion is the rail was unused and rusty. It would not have been possible to test it prior to the job. Only with a few trains on it, would it be clean enough to test.
Then there is the issue of legacy. Who knows how it was joined in the scheme of things. There was nothing in the scope to address the integrity of the rail in any meaningful way. The scope was to throw it aside it aside, put it back, weld it up as-is. The have a wave of broken welds is not unusual in rail re-purposed and is likely fatigued.
I could go on all night about different failure modes of rail, however the two main aspects that condemn a rail is wear (easy to understand) and fatigue. Fatigue can be in the parent rail (crack growth around inclusions) and the rail join. It is however mainly in the join is most circumstances. I wont go into jointed rail too much other than to say to flashbutt a rail without removing the bolt holes and crippled ends, is a fools game. Fatigue at the weld is usually driven by the geometry at manufacture of the weld - hence I consistently advocate about weld quality obtaining the correct geometry. There can be issues with metallurgy resulting in the wrong structure, but this is yet again another branch of the topic and not that common.
For those needing the simple explanation of fatigue, its the development of flaws in the rail due to the number of stress loadings and the magnitude of the stress loading. Big stresses and big cycles equals short life. In the case of a weld the magnitude of the stress is exacerbated by poor geometry of the weld. To give you an idea, a 0.6mm dip in the weld can yield stresses in the rail 2.5 times that of the wheel rolling along a normal flat rail. To add a further dimension, the relationships between rail life and the number of cycles it will take before failure is not linear, indeed the diagrams usually represent the number of cycles logarithmically.
One can draw a link to my criticism of historic Victorian practices, as it never been known for reliably producing quality welding - the result is the rail has a shortened life. Unfortunately this realisation is many years after the responsible managers/engineers/failed quality processes have departed the scene.
I cant speculate too much on what is going on with Maryborough to Ararat, I have not seen it. I do know that rail that has questionable fatigue does not like to be assigned to a new duty. Different sleepers, new loading needs, new outcomes. There ends up being a wave of failures in the first bit. It will settle down I would think.
Another aspect of rail is the handling it would have seen while being relaid. The stressing of the rail section sideways is on the weaker axis and its easy to damaged welds by poor handling. The risk is defects and damage to the foot of the rail. One aspect of a notch or crack in the foot is its not detectable to normal ultrasonics. It is however a sure broken rail. Then there is the reality of this rail being left to sit in the grass for many years, a corroded foot will also add to the risk of a parent broken rail (and write the rail off)
In this instance the first task is to scrutinise the ultrasonic results and quantify what is there. With some trains on it and a clean head, this is now possible. Examine the broken rails, how they broke. What's causing it. There is a bit of information to gather, but only with this information can decisions be made. Its not panic stations yet.
The rail is tied together well with new sleepers, it will hold a broken rail reasonably well.