Abstract
It is difficult for co-continuous metal/ceramic composite with complex spatial topology to establish three-dimensional (3D) solid model. A generation-based optimization method is proposed to reconstruct 3D solid model. Firstly, pre-processing of the composite SEM image is applied to obtain two-dimensional (2D) digital matrix using a maximum between-cluster variance method, metallic and ceramic phases are segmented accurately. 2D generation core field is obtained following the principles proposed in the paper. Secondly, an optimization method is proposed and applied for the randomly distributed 3D cores to match the core distribution characteristics. Thirdly, a modified real-time changing directional growth probability is applied to form a smooth interface between the two phases. Finally, volume fractions and 2-point correlation functions of the reconstructed model and a micro-CT model are compared to verify the accuracy. This method can be used to reconstruct the 3D model of other materials with different microstructures such as titanium alloy.
Original language | English |
---|---|
Pages (from-to) | 534-543 |
Number of pages | 10 |
Journal | Computational Materials Science |
Volume | 117 |
DOIs | |
Publication status | Published - May 2016 |
Keywords
- 2-point correlation function
- 3D reconstruction
- Co-continuous ceramic composites
- Model verification
- Volume fraction