AlloyManufacturingNet for discovery and design of hardness-elongation synergy in multi-principal element alloys

Description

1. Models saved after training the neural networks:

a. hardness_saved_models.zip

b. ductility_saved_models.zip

2. prediction_data_for_casting_process.zip: Alloy types and their composition variants are provided in the file multicomponent_alloys_variants_compositions.csv. The hardness prediction for casting process with alloys  are given in hardness_prediction_alloys_CAT-A.csv file whereas elongation prediction for the alloys for the same process (manufacturing route) are provided in elongation_prediction_alloys_CAT-A.csv. The composition sets D1_{x}D2_{y}(ZrHfNb)_{1-x-y} are referred to as Alloy A, whereas D1_{x}D2_{y}(VNbTa)_{1-x-y} are denoted as Alloy B in the columns of the csv files. The numeric value after alloy type denotes the specific pairs of the dopants [D1,D2] . For example, the alloy variants Ti_{x}Ta_{y}(ZrHfNb)_{1-x-y}, W_{x}Ta_{y}(ZrHfNb)_{1-x-y}, Mo_{x}Ta_{y}(ZrHfNb)_{1-x-y}, and Cr_{x}W_{y}(ZrHfNb)_{1-x-y} are referred to as Alloy A1, Alloy A2, Alloy A3 and Alloy A4 respectively. Similarly, Alloy A1, Alloy A2, Alloy A3 and Alloy A4 respectively denote Cr_{x}W_{y}(VNbTa)_{1-x-y}, Zr_{x}W_{y}(VNbTa)_{1-x-y}, Hf_{x}W_{y}(VNbTa)_{1-x-y}, and Mo_{x}Ti_{y}(VNbTa)_{1-x-y}.  Hence, if a column is represented as HV_A2, then it is the hardness prediction for the alloy systems W_{x}Ta_{y}(ZrHfNb)_{1-x-y}, and if the column header is EL_B3, then the  elongation of Hf_{x}W_{y}(VNbTa)_{1-x-y} alloy systems is estimated.