Glenn R McDowell, Wee L Lim and Andrew C Collop, School of Civil Engineering, University of Nottingham. This paper was first published in GE’s January 2003 edition.
There are two basic types of test used to determine ballast strength: the revolving drum type test and the quasi-static test. The Los Angeles Abrasion test, Mill Abrasion test, and Deval test each involve revolving ballast in a drum so that particles can continuously move and form new contacts with other particles (Selig and Waters, 1993).In the Los Angeles Abrasion test, steel balls are added to the ballast.
A degradation process, where particles can freely move around and form new contacts, is described as “unconstrained comminution” (Sammis et al, 1997). The use of such an approach assumes that a particle’s fracture probability is controlled mainly by its strength. Clearly, the mechanics of degradation in a revolving drum are completely different to the mechanics of degradation beneath the track, which can be described as “constrained comminution”.
In the constrained process, particles are relatively unable to move around. In this case, the number of contacts (“co-ordination number”) which a particle has, determines whether the induced tensile stress within it is large or small, and this plays a major role in determining the fracture probability of the particle.
In constrained comminution, a fractal distribution of particles tends to evolve (McDowell and Bolton, 1998), as large particles become protected by fines which become smaller and stronger. It is well known that unconstrained comminution models produce completely different particle size distributions to constrained models (Sammis et al, 1997).
The Aggregate Crushing Value (ACV) test is another constrained comminution process, and so should give a good representative measure of degradation beneath the track. The ACV test involves crushing an aggregate of 10mm-14mm ballast particles in a 154mm diameter and 134mm deep steel mould. The sample is compacted so as to be approximately 100mm thick, so the aspect ratio of the sample (ie H/D, where H is the sample height and D is the diameter) is 0.65.The applied uniaxial stress is increased to about 21MPa in 10minutes, and the ACV is calculated as the percentage by mass passing the 2.36mm sieve (Railtrack, 1998). The Railtrack line specification (Railtrack, 1998) specifies that the ACV should not exceed 22%.
However, the ACV test is flawed because the tensile strength of ballast particles is a function of size. The size effect is different for different ballasts, so the testing of an aggregate of 10mm-14mm particles will yield little information about the behaviour of 50mm particles beneath the track. The size effect on strength is discussed in detail in this paper. In addition, the ACV may introduce too much wall friction, because the aspect ratio of the sample is high. This will be demonstrated mathematically.
The use of the ACV is soon due to expire, and to be replaced in the UK by the Los Angeles Abrasion and Micro-Deval tests (CEN, 2001). Although these tests use ballast particles of the same size as those placed beneath the track, the mechanics are unconstrained, and therefore the ACV test is, at least in principle, much better.
It will be shown in this paper that to test an aggregate of 50mm particles in an ACV type test, the sample requires a sufficient thickness, acceptable aspect ratio and therefore a diameter much larger than 154mm. It will be shown that a diameter of 300mm might be considered acceptable. In this case the uniaxial stress of 21MPa can only be achieved in an oedometer test by a loading frame capable of applying 1.5MN. An Instron testing machine with a capacity of 2.5MN at the University of Nottingham has been used to perform oedometer tests on samples of 37.5mm-50mm and 10mm-14mm ballast particles. It will be shown that the aggregate yield stress reduces with a reduction in the average strength of the constituent particles, as determined by compression between flat platens, and that the ACV is clearly a function of particle size.