Zhang Zhicheng, Fidan Ismail
Department of Mechanical Engineering, Tennessee Tech University, Cookeville, TN 38505, USA.
Department of Manufacturing and Engineering Technology, Tennessee Tech University, Cookeville, TN 38505, USA.
Materials (Basel). 2022 Jun 17;15(12):4292. doi: 10.3390/ma15124292.
Additive manufacturing (AM) is a widely used layer-by-layer manufacturing process. Material extrusion (ME) is one of the most popular AM techniques. Lately, low-cost metal material extrusion (LCMME) technology is developed to perform metal ME to produce metallic parts with the ME technology. This technique is used to fabricate metallic parts after sintering the metal infused additively manufactured parts. Both AM and sintering process parameters will affect the quality of the final parts. It is evident that the sintered parts do not have the same mechanical properties as the pure metal parts fabricated by the traditional manufacturing processes. In this research, several machine learning algorithms are used to predict the size of the internal voids of the final parts based on the collected data. Additionally, the results show that the neural network (NN) is more accurate than the support vector regression (SVR) on prediction.
增材制造（AM）是一种广泛使用的逐层制造工艺。材料挤出（ME）是最流行的增材制造技术之一。最近，开发了低成本金属材料挤出（LCMME）技术，以利用ME技术进行金属材料挤出，从而生产金属零件。该技术用于在对增材制造的注入金属的零件进行烧结后制造金属零件。增材制造和烧结工艺参数都会影响最终零件的质量。显然，烧结零件的机械性能与传统制造工艺制造的纯金属零件不同。在本研究中，使用了几种机器学习算法，根据收集的数据预测最终零件内部孔隙的尺寸。此外，结果表明，在预测方面，神经网络（NN）比支持向量回归（SVR）更准确。
