Simulating the particle size distribution of rockfill materials based on its statistical regularity

1. Introduction Rockfill has excellent engineering characteristics, such as good compactness, better seepage ability, high density, low sedimentation, high shear strength and so on. So it is widely used in dams, bed of railway and deep foundations[1-2]. Two types of rockfill are commonly used in constructions, that is inartificial rockfill such as gravel and scree, and blasted rockfill, e.g. fragmental stone, macadam and moraine soil. Both types are hetero-geneous mixture composed of numerous …

Simulating the particle size distribution of rockfill materials based on its statistical regularity

The particle size distribution of rockfill is studied by using granular mechanics, mesomechanics and probability statistics to reveal the relationship of the distribution of particle size to that of the potential energy intensity before fragmentation, which finds out that the potential energy density has a linear relation to the logarithm of particle size and deduces that the distribution of the logarithm of particle size conforms to normal distribution because the distribution of the potential energy density does so. Based on this finding and by including the energy principle of rock fragmentation, the logarithm distribution model of particle size is formulated, which uncovers the natural characteristics of particle sizes on statistical distribution. Exploring the properties of the average value, the expectation, and the unbiased variance of particle size indicates that the expectation does notequal to the average value, but increases with increasing particle size and its ununiformity, and is always larger than the average value, and the unbiased variance increases as the ununiformity and geometric average value increase. A case study proves that the simulated results by the proposed logarithm distribution model accord with the actual data. It is concluded that the logarithm distribution model and Kuz-Ram model can be used to forecast the particle-size distribution of inartificial rockfill while for blasted rockfill, Kuz-Ram model is an option, and in combined application of the two models, it is necessary to do field tests to adjust some parameters of the model.