Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
1575183 | Materials Science and Engineering: A | 2014 | 8 Pages |
Abstract
An artificial neural network (ANN) was employed to investigate the fracture toughness of directionally solidified Nb-silicide in situ composites. The microstructures of the composites were quantified with a metallographic statistics method. Both microstructural features and composition of the constituent phases were used as the candidate inputs of the artificial neural network model while the fracture toughness of the composites was employed as the outputs. The effects of different inputs on the fracture toughness were investigated and evaluated by the trained network. When all of the candidate inputs were taken into account, outstanding performance of the neural network was achieved. A new alloy with optimized microstructure and fracture toughness was produced according to the prediction of the model. The fracture toughness of the new alloy reached 19.5 MPa m1/2, which was 25.5% higher than the best inputted alloy (15.5 MPa m1/2).
Related Topics
Physical Sciences and Engineering
Materials Science
Materials Science (General)
Authors
Kai Guan, Lina Jia, Xiaojun Chen, Junfei Weng, Fei Ding, Hu Zhang,