Analysis and prediction of printable bridge length in fused deposition modelling based on back propagation neural network. Issue 3 (3rd July 2019)

• Record Type: Journal Article
• Authors: Jiang, Jingchao; Hu, Guobiao; Li, Xiao; Xu, Xun; Zheng, Pai; Stringer, Jonathan
• Is Part Of: Virtual and physical prototyping. Volume 14:Issue 3(2019)
• Page Start: 253
• Page End: 266
• Publication Date: 2019-07-03
• Subjects: Additive manufacturing -- support -- printable bridge length -- back propagation neural network; Rapid prototyping -- Periodicals; Virtual computer systems -- Periodicals; Prototypes, Engineering -- Periodicals; Production engineering -- Periodicals; Design, Industrial -- Periodicals; 670.285
• DOI: 10.1080/17452759.2019.1576010
• Languages: English
• ISSNs: 1745-2759

Abstract

ABSTRACT: In recent years, additive manufacturing has been developing rapidly mainly due to the ease of fabricating complex components. However, complex structures with overhangs inevitably require support materials to prevent collapse and reduce warping of the part. In this paper, the effects of process parameters on printable bridge length (PBL) are investigated. An optimisation is conducted to maximise the distance between support points, thus minimising the support usage. The orthogonal design method is employed for designing the experiments. The samples are then used to train a neural network for predicting the nonlinear relationships between PBL and process parameters. The results show that the established neural network can correctly predict the longest PBL which can be integrated into support generation process in additive manufacturing for maximising the distance between support points, thus reducing support usage. A framework for integrating the findings of this paper into support generation process is proposed.