Find in Library

A fractal dimension index for aggregate shape and an area index for aggregate size were used in this paper to construct fractal aggregate cross-sections which could characterize real aggregate contours. Firstly, images of the real section of asphalt concrete specimens were processed for the extraction of independent aggregate contours and the recognition of their fractal dimensions. Secondly, with square as the datum figure, the midpoint displacement method and the fractal dimensions were used to generate fractal aggregate cross-sections. Thirdly, numerical specimens were individually generated by square, circular and fractal aggregate cross-sections under three load conditions of uni-axial compression, uni-axial stretch and horizontal shear. Research results show that: 1) based on roundness, SP value and girth indices, fractal aggregate cross-section is closer to the real aggregate cross-section in terms of angularity when compared with the circular or square cross-section; 2) shapes of aggregate cross-sections have obvious influences on the mechanical properties of asphalt concrete specimens. With the randomness and statistical features of aggregate cross-sections taken into account, the method of generating fractal aggregate cross-sections in the paper can construct reasonable numerical aggregates. Keywords: Concrete, Aggregate section, Fractal dimension, Aggregate angularity, Image processing, Microstructure model