• No results found

Factors Affecting Particle Breakage .1 Influence of State and Level of Stress



2.5.2 Factors Affecting Particle Breakage .1 Influence of State and Level of Stress

Lee and Seed (1967) and Vesic and Clough (1968) from tests on sand at very high stress levels observed that there is a stress (called breakage stress) beyond which dilatancy disappears, and the crushing of the grains becomes the only mechanism that, in addition to simple slip, governs the shear strength of sand. The level of breakage stress is a function of the crushing strength and angle of physical friction of the mineral. A lower value of breakage stress is associated with a low particle crushing strength and hence, a higher angle of friction of the mineral. An increase in the fracturing phenomenon was reported when the stress level was increased. It is noteworthy that a substantially greater proportion of particle degradation was found to be caused by the action of shear stress rather than the application of isotropic consolidation stresses.

Chapter 2: Critical review of granular media with special reference to railway ballast

Research on various types of rockfill (Fumagalli et al., 1970; Marachi et al., 1972;

Marsal, 1973; Frassoni et al., 1982) and gravels (Leslie, 1963; Lee and Farhoomand, 1967) reported similar behaviour. However, due to the coarser grains, the crushing of grains occurred at an even lower level of confining pressure. It was also shown that while the extent of particle crushing increased with the increase in level of confining pressure, the rate of particle breakage decreased with the increase in confining stress.

While the detrimental effect of grain breakage on the ballast deformation is recognized, very little is known about the breakage mechanism of railway ballast. Feng (1984) and Selig and Devullapaly (1991) reported from both field observation and ballast box tests that there is proportional relationship between the applied load and both breakage level and plastic deformation developed. The breakage (at the end of the test) was expressed as percentage by weight of grains smaller than 9.5 mm. In addition Selig and Devullapaly (1991) proposed a similar method as that described by Edwards (1986) for settlement, which enabled correlation of breakage produced by any cumulative load to a reference load. Ravitharan and Martin (1996) reached similar findings, reporting that independent of type of ballast, the degradation rate per million cycles of loading recorded during tests in a semi-confined cylindrical apparatus was highly dependent on the level of cycled stress.

Earlier research on railway ballast by Raymond et al. (1975, 1976) reported the crushing characteristics of a Coteau dolomite ballast with a particle size ranging from 4.75 mm to 38 mm. The CID triaxial tests were performed at various levels of confining pressure on specimens of different initial densities. To quantify the degree of particle fragmentation particles smaller than 4.75 mm were defined as fines.

Surprisingly, it was found that the stress level did not affect the extent of particle fragmentation as a value of only 0.6 % was reported. Also, Raymond (1977) concluded that the contact area rather than the contact forces controls ballast abrasion. However, Feng (1984) argued that the abrasion and breakage of ballast were related to the contact forces between individual grains. Influence of Material Properties

a. Grains Characteristics (size, shape and strength)

There is sufficient evidence in the literature to correlate the extent of particle degradation to grain characteristics, namely, mineral composition and structure, texture, friability, degree of weathering, brittleness, hardness, shape and size. From comparisons between research on sand and rockfill it was shown that coarser grains were generally associated with more particle crushing (Lee and Seed, 1967; Vesic and Clough, 1968; Lee and Farhoomand, 1967; Fumagalli et al., 1970). It has been shown that, at a similar level of stress and independent of type of material, gradation and degree of packing, the crushing of angular particles was higher than that of rounded grains (Lee and Seed, 1967; Robert and De Souza, 1958; Lee and Farhoomand, 1967).

Leslie (1963) indicated that the effect of oversized particles on the extent of grain breakage was insignificant. However, Marachi et al. (1972) showed that for different sizes of rockfill specimens, the amount of particle breakage increased as the maximum particle size increased. The increase was more distinct for tests with high confining stresses. This behaviour can be explained by the Griffiths (1921) crack theory, which states that the fracture strength of a particle decreases as its size increases due to the increased probability of inherent flaws and cracks in the particles.

Chapter 2: Critical review of granular media with special reference to railway ballast

Very few studies on road base and railway ballast were concerned with the effect of grains characteristics on the crushing behaviour of these materials. The published data showed however, that presence of small amount (< 25 %) of flaky grains in the road base specimens increased the extent of particle breakage (Gur et al., 1971), while Roner (1985) reported an increase in grain crushing when the particle size of quartzite railway ballast was increased.

b. Sample Characteristics (gradation and density)

Several types of sand prepared for a range of relative density and with different PSD curves were studied by Lee and Seed (1967) and Vesic and Clough (1968). The tests were performed at a high level of confining stress and collected data revealed that the extent of breakage increased as the coefficient of uniformity approached the limiting value 1.

Marsal and La Rosa (1976) performed triaxial tests on rockfill-soil (sand, silt and clay) mixture. They found that the grain breakage gradually reduced with an increase in percentage of the soil fraction. The highest breakage was recorded for the clean rockfill specimens. These findings confirmed earlier reports by Kjærnsli and Sande (1963) and Marsal (1973) that the crushing of uniform material was higher than that of well-graded material. It was also shown that grains exhibited higher fracturing as their porosity increased, i.e. lower initial dry density.

Earlier studies on railway ballast were not concerned with relating the gradation of material to the crushing mechanism. However, more recent research by Ravitharan and Martin (1996) on various types of railway ballast reported that, by eliminating fractions

finer than 19.1 mm and larger than 40 mm from a given gradation, a higher rate of degradation was observed for the same test conditions. The cycled tests were performed in a semi-confined cylindrical apparatus and the load was applied at a frequency of 6 Hz, though the type of ballast was reported. Effect of Particle Breakage on Strength and Deformability

The most important factor affecting both the shear strength and the compressibility of a granular material is the phenomenon of fragmentation (when subjected to changes in its state of stress during both isotropic compression and shear). Marsal (1973) presented the correlation of principal stress ratio at failure to particle breakage for different types of rockfill materials, as shown in Figure 32. This figure shows that the strength of the materials decreases as the particle breakage increases. Figure 33 shows the effects of particle breakage and applied axial stress on the one-dimensional compressibility of rockfill. The relationship between particle breakage and change in void ratio for both loose and dense material are S-shaped curves. This change in void ratio may be a result of the rearrangement of grains and the filling of the voids within the new particle distribution due to breakage.

Studies by Marachi et al. (1972) showed that the relationship between the angle of internal friction and the particle breakage was unique for each tested material, provided the initial void ratio was the same. This trend was independent of the maximum particle size of the sample. Leslie (1963) found that the decrease in angle of internal friction was proportional to the increase in particle breakage.

In contradiction with the findings from tests on sand and rockfill, Raymond and Davies

Chapter 2: Critical review of granular media with special reference to railway ballast

Figure 32. Stress ratio at failure versus particle breakage (Marsal, 1973)

Figure 33. Effect of particle breakage on compressibility from oedometer tests (Marsal, 1973)

(1978) concluded that the change in stress state on the specimens did not produce any effect on the extent of particle degradation. It should be mentioned that the findings were reported for only one type of ballast and that they are in contradiction with the general grain crushing characteristics of granular media. More recent studies by Selig and Devullapany (1991) and Ravitharan and Martin (1996) reached the conclusion that there is a correlation between stress state and the magnitude of particle degradation.

Consequently, an extensive investigation is necessary to clarify the breakage mechanism applicable to railway ballast.