Dataset for random uniform distributions of 2D circles and 3D spheres.

This dataset contains random uniform distributions for a large number of 2D and 3D balls, along with the description files. It provides the possibility for fast pick up of random, but repeatable, sets of smaller samples, with the guaranteed statistical properties such as random uniform distribution of balls, the predefined expected volume ratio of balls, and also the minimum distance between them. Samples are uniquely identified by the position coordinates in the provided large kernels. The sets of samples can be used in performing numerical predictions of different types for uniform ball distributions while keeping the numerical effort at a reasonable level. Specifically, this can be useful in computational homogenization of fiber and spherical particle reinforced composites, where the provided kernels can be viewed as representative volumes and the samples as the realizations of statistical volume elements. Some secondary results, like the numbers of samples of a given size assuring the required accuracy in expected ball volume ratio representation, are also provided. Data was created by means of the pseudo-random number generator using python scripting and can be loaded and used also in other programming environments.