A novel method for generation of random aggregate structure and its application in soil-rock mixture

The generation of random aggregate structure (RAS) is encountered in many fields and has received wide attention. The random sequential addition (RSA) is a frequently-used approach that can generate RAS by dispersing particles randomly and sequentially. The classical RSA is not efficient due to the excessive overlap detections involved in every trial of placing new aggregates. In this work, a new version of RSA called correlative element method (CEM) is proposed based on the concept of "correlative element" on a background grid. The overlap detection is replaced by the checking of the occupied state of correlative elements. Furthermore, a multilayer CEM is also presented to guarantee the minimum gap between adjacent aggregates, leading to a more evenly distributed aggregate model. The proposed CEM can handle not only polygonal aggregates but also cir-cular and elliptical aggregates in a unified way. Other related issues, such as the aspect ratio and orientation of aggregates, the packing density, and the efficiency of the proposed CEM are also tested and discussed. The results show that the proposed CEM has advantages of easy implementation, high efficiency, and wide application range. Finally, the proposed CEM is applied to simulate the direct shear test of soil-rock mixture.