| abstract |
A method for predicting high temperature mechanical properties of heat-resistant alloys based on deep learning, comprising the following steps: S1, forming an original experimental database; S2, performing data preprocessing on microstructure photos in the original experimental database; S3, according to the high temperature of the heat-resistant alloys The distribution of experimental values of mechanical properties, divide the original experimental database into consecutive N categories, and use the divided category values as the category labels of the corresponding images; group the labeled image data; S4, and separate all groups of image data. Perform digital tensorization processing; S5, build a deep learning model, configure the model structure and model parameters, and optimize the prediction effect of the deep learning model; S6, use an optimized deep learning model to predict its high-temperature mechanical properties according to the microstructure pictures of heat-resistant alloys. The invention can realize the direct prediction of the heat-resistant alloy from the microstructure to the high-temperature mechanical properties, improve the high-temperature performance detection efficiency of the heat-resistant alloy, and save the high-temperature detection cost of the heat-resistant alloy. |