Phase selection prediction and component determination of multiple-principal amorphous alloy composites based on artificial neural network model
(1. School of Materials Science and Engineering, Shaanxi University of Technology, Hanzhong 723001, China;
2. School of Materials and Energy, Guangdong University of Technology, Guangzhou 510006, China;
3. School of Aeronautics, Northwestern Polytechnical University, Xi’an 710071, China;
4. School of Mechanical Engineering, Shaanxi University of Technology, Hanzhong 723001, China;
5. School of Mathematics and Computer Science, Shaanxi University of Technology, Hanzhong 723001, China)
2. School of Materials and Energy, Guangdong University of Technology, Guangzhou 510006, China;
3. School of Aeronautics, Northwestern Polytechnical University, Xi’an 710071, China;
4. School of Mechanical Engineering, Shaanxi University of Technology, Hanzhong 723001, China;
5. School of Mathematics and Computer Science, Shaanxi University of Technology, Hanzhong 723001, China)
Abstract: The probability of phase formation was predicted using k-nearest neighbor algorithm (KNN) and artificial neural network algorithm (ANN). Additionally, the composition ranges of Ti, Cu, Ni, and Hf in 40 unknown amorphous alloy composites (AACs) were predicted using ANN. The predicted alloys were then experimentally verified through X-ray diffraction (XRD) and high-resolution transmission electron microscopy (HRTEM). The prediction accuracies of the ANN for AM and IM phases are 93.12% and 85.16%, respectively, while the prediction accuracies of KNN for AM and IM phases are 93% and 84%, respectively. It is observed that when the contents of Ti, Cu, Ni, and Hf fall within the ranges of 32.7-34.5 at.%, 16.4-17.3 at.%, 30.9-32.7 at.%, and 17.3-18.3 at.%, respectively, it is more likely to form AACs. Based on the results of XRD and HRTEM, the Ti34Cu17Ni31.36Hf17.64 and Ti36Cu18Ni29.44Hf16.56 alloys are identified as good AACs, which are in closely consistent with the predicted amorphous alloy compositions.
Key words: multiple-principal amorphous alloy composites; Ti-Cu-Ni-Hf alloy; phase selection; artificial neural network; machine learning