Application of Machine Learning Based on Big Data in Metal 3D Printing

Currently, metal 3D printing technology is widely used in high-end manufacturing fields such as aerospace and medical devices. However, due to the complex parameters, difficulty in defect prediction and insufficient quality control during the printing process, the accuracy and performance of the finished products cannot meet industrial needs. This study aims to optimize the metal 3D printing process parameters, improve the printing quality and reduce the defect rate through machine learning technology based on big data. This paper collects a large amount of process parameter data during metal 3D printing, uses sensors and monitoring equipment to record the temperature field, molten pool morphology and finished product quality indicators in real time during the printing process. Then the data is cleaned and preprocessed to remove outliers and redundant data, and principal component analysis (PCA) is used to reduce the dimension and extract key features. Then a variety of machine learning algorithms are used to establish a prediction model, and the effects of defect prediction, strength prediction and porosity control are trained and verified respectively. After the test, the defect prediction accuracy of the random forest model reaches 97%. The printed product of the SVM model with optimized parameter settings reaches 70MPa, and the porosity of the XGBoost model is as low as 0.1%. Machine learning technology based on big data can effectively improve the quality control capability of metal 3D printing and provide reliable technical support for the manufacturing of high-precision and high-performance metal parts.