A Densification prediction Model of Selective Laser Melting Based on BP Neural Network
Lu Pan, Bin Qian, Yaqin Wang, Xinyu Liu, Hua Jiang, Liang Wang- General Materials Science
Background:
During the process of Selective Laser Melting (SLM), there is a complex nonlinear relationship between forming quality (Densification, elongation, and mechanical properties) and laser process parameters, and improper laser process parameters will directly lead to forming defects, including holes, cracks and even printing failure.
Objective:
Forming quality is limited by a series of factors, such as raw material properties, equipment properties, laser process parameters, and the post-treatment process, etc.
Methods:
In this paper, the feasibility test and density data test (laser power 130-280 w, laser scanning speed 1200-1500 mm/s, laser scanning distance 0.01 mm, and thickness 0.03 mm) were carried out by experiments. And the mathematical model of the Zl205A densification prediction curve and the densification distribution cloud plot were obtained.
Results and Conclusion:
The BP neural network prediction system for ZL205A by SLM was developed with the help of the BP neural network toolbox. The prediction system was applied to ZL205A densification prediction with an error of less than 5%.