Dataset for Machine learning assisted prediction of defects for laser powder bed fusion of nickel-based superalloy

Description

This document contains partial defect image processed by semantic segmentation model , deep learning code, binary model code, machine learning regression model code, and AM parameters with different defect datasets

Steps to reproduce

Artificially calibrate a certain amount of defect image data, in order to train the semantic segmentation model of deep learning, and finally realize the classification and calibration of different defects, use the trained model for the recognition and statistics of other defect images, and establish a database of additive process parameters and defects. Based on the database, a binary classification method is used to determine the formability of AM superalloy, and prefered machine learning regression model is used to realize the prediction of defects.

Institutions

Categories

Funding

Licence