Binder Jetting (BJ) process is an additive manufacturing process in which powder materials are selectively joined by binder materials. Products can be manufactured layer by layer directly from 3D model data. It is not always easy for manufacturing engineers to choose proper BJ process parameters to meet the end-product quality and fabrication time requirements. This is because the quality properties of the products fabricated by BJ process are significantly affected by the process parameters. And the relationships between process parameters and quality properties are very complicated. In this paper, a process model is developed by Backward Propagation (BP) Neural Network (NN) algorithm based on 16 groups of orthogonal experiment designed by Taguchi Method to express the relationships between 4 key process parameters and 2 key quality properties. Based on the modeling results, an intelligent parameters recommendation system is developed to predict end-product quality properties and printing time, and to recommend process parameters selection based on the process requirements. It can be used as a guideline for selecting the proper printing parameters in BJ to achieve the desired properties and help to reduce the printing time.
A Hybrid Deep Learning Model for Layer-Wise Melt Pool Temperature Forecasting in Wire-Arc Additive Manufacturing Process
