Please use this identifier to cite or link to this item:
|Title:||Investigation of geometrical characteristics and porosity in direct energy deposition for process parameters optimization||Authors:||Kho, Jin Yao||Keywords:||Engineering::Mechanical engineering||Issue Date:||2021||Publisher:||Nanyang Technological University||Source:||Kho, J. Y. (2021). Investigation of geometrical characteristics and porosity in direct energy deposition for process parameters optimization. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/150953||Project:||B221||Abstract:||Additive Manufacturing (AM) is gaining popularity around the world as a result of its enormous potential and benefits. Across various industries, it is important that the AM fabricated parts have the most favorable geometrical characteristics and porosity because this can impact the mechanical properties. Furthermore, process parameters are significant in AM systems as they can influence the geometrical characteristics and porosity of the deposition. In this paper, direct energy deposition of stainless steel 316L powder was used to fabricate single-track specimens. The design of experiment was used to design the set of process parameters in particular laser power, scanning speed, and powder mass flow rate. Analysis of variance (ANOVA) and machine learning are used to analyze the correlation and to build prediction models between the process parameters and experiment results. Furthermore, the optimum model will be used for process parameter optimization. The correlation result from both ANOVA and machine learning shows a relatively similar trend as previous studies. However, the result of powder mass flow rate in ANOVA shows a contradiction in the machine learning correlation. In addition, among all the various models, the best prediction model obtained was Extreme Gradient Boosting (XGB). The XGB prediction model can achieve a percentage error of 0.037% to 4% for the geometrical characteristic. The accurate XGB prediction model has the potential to be extended to multi-layer deposition, which is more similar to real-world component fabrication.||URI:||https://hdl.handle.net/10356/150953||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||MAE Student Reports (FYP/IA/PA/PI)|
Updated on Sep 27, 2021
Updated on Sep 27, 2021
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.