Metal additive manufacturing molten pool shape prediction method based on dimensional analysis

Abstract

The invention discloses a metal additive manufacturing molten pool shape prediction method based on dimensional analysis. A quantitative physical model between additive manufacturing process parameters and a molten pool shape is established. The method comprises the following steps: firstly, determining an additive manufacturing process parameter space, performing dimensional analysis on the manufacturing process parameter space, establishing a function relationship among non-dimensional parameters, training the function relationship among the non-dimensional parameters by utilizing a machine learning method, and finally obtaining a quantitative relationship function model between the additive manufacturing process parameters and the shape of a molten pool. According to the method, the dependence on experimental samples can be reduced, and the generalization ability of the model is enhanced.

Description

Technical field [0001] The present invention relates to the field of additive manufacturing techniques, and more particularly to a metal additive method for producing a molten pool shape prediction method based on dimensional analysis. Background technique [0002] Metal additive manufacturing process is a multi-physique, multi-scale process, including microscopic scale material - laser interaction, photoconductive pond kinetics and macroscopic dimensions. Researchers hope to achieve control of the molten pool and the deposition process by research and optimization of forming parameters, and then improve the mechanical properties of the forming parts. In order to be described in clearer, more accurate ways, researchers have tried to build various physical models to find links between adding fabrication process parameters and the shape parameters of the molten pool. People often use parsing methods and numerical simulation methods. However, the disadvantage of analytical methods i...

AI Technical Summary

Problems solved by technology

They will automatically learn the relationship between input features and output targets based on previous data. Machine learning algorithms are migrating to traditional manufacturing fields including additive manufacturing. However, the accuracy of machine learning algorithms depends heavily on the quality of training data sets. lots of experiments

Images