Feature-based support generation in fused deposition modeling (FDM) machine

Abstract

Feature-based Support Generation is a technique that has been proposed in Fused Deposition Modeling (FDM) machine. This technique can provide information of volume and amount of support structure which are closely related to orientations of part deposition. There are two types of support features in FDM part model development, which are Self-Supported Features (SSF) and External-Supported Features (ESF). The SSF requires no support material while ESF involves the use of additional support material in their fabrication. Currently, various techniques have been suggested to identify features are limited to a specific manufacturing process. In other aspect, the LM process planning is not fully automatic and lead to part quality degradation and increases the possibility of making errors. Furthermore, many errors are occurred due to the involvement of human in this crucial process. Other issue is that the Stereolithography (STL) file format representation is used to transfer the CAD data to the LM process planning resulting to the loss of design and functional feature information. Determining the OPDO was found to be difficult and consumed longer build times that influenced by the speed and the change of nozzle's tip during material deposition. The main objective of this work is to integrate between Computer Aided Design (CAD) and Computer Aided Manufacturing (CAM) using a feature-based technique. This will help in automation of FDM process planning prior to the manufacturing of part model with less human error. In this work, the minimum volume and amount of support structure are selected in order to determine the optimum part deposition orientation. This work also focuses on the improvement of the non-contact surface area between the support structure and part model. The accuracy of the network is determined through five MLP structures (Structures 1 to 5). The accuracies for all MLP structures at specific hidden nodes are analysed.