透過您的圖書館登入
IP:3.238.71.155
  • Journals

機器學習輔助設計雷射積層參數製造超硬合金

Machine Learning Assisted Design to Optimize Laser Additive Manufacturing Parameters for Cemented Carbide

Abstracts


此文使用擇區雷射熔融(Selective Laser Melting,SLM)製造WC-Co超硬合金,擇區雷射熔融是一種積層製造(Additive Manufacturing, AM)技術,能克服傳統液相燒結超硬合金過程中,因模具的使用所造成之幾何限制,得以進一步製作複雜度高的組件。然而快速冷卻的製程容易導致裂縫生成等缺陷,使其生產具有挑戰性。本文透過機器學習演算法協助設計SLM參數以減少缺陷,並在緻密度(Densification)、硬度(Hardness)、韌性(Toughness) 和表面粗糙度(Surface Roughness)達到更好的綜合表現。本文章主要描述研究過程利用機器學習演算法種類中的隨機森林模型,輔助設計雷射積層參數,所建立之模型可以提供準確的預測能力及通用性,實驗驗證列印超硬合金,可以達到98.59 %的緻密度、1700 HV1的硬度、8.66 MPa√m的破壞韌性和0.1 mm的粗糙度,揭示機器學習為可被利用於優化3D列印參數之實用的工具。

Parallel abstracts


This article describes how to utilized selective laser melting (SLM) to fabricate WC-Co cemented carbide. SLM is an additive manufacturing (AM) technique, which can be used to avoid traditional manufacturing constraints and offer freedom in complexity design for products. However, the fabrication of WC-Co cemented carbide by SLM is very challenging because of the material is brittle and is prone to form defects such as crack due to rapid cooling process. This article describes how to apply one of the machine learning methods-Random Forest Algorithm in design to optimize SLM parameters to reduce defects and achieve a better combination of properties including densification, hardness, toughness and surface roughness. In summary, random forest model is an effective tool in design to optimize SLM parameters, according to the experimental results, a balanced properties with 98.59 % in densification, 1700 HV1 hardness, 8.66 MPa√m fracture toughness and 0.1 mm roughness have been achieved. This article shows that the machine learning can be a practical tool to optimize process parameters for 3D printing.

References


E. Uhlmann, A. Bergmann, and W. Gridin, “Investigation on additive manufacturing of tungsten carbide-cobalt by selective laser melting,” Procedia CIRP 35, 8-15, 2015.
S. Grigoriev, T. Tarasova, A. Gusarov, R. Khmyrov, and S. Egorov, “Possibilities of manufacturing products from cermet compositions using nanoscale powders by additive manufacturing methods,” Materials 12 (20) 3425, 2019.
G. Goh, S. Sing, and W. Yeong, “A review on machine learning in 3D printing: Applications, potential, and challenges,” Artificial Intelligence Review 54(1), 63-94, 2021
M. Khanzadeh, S. Chowdhury, M. Marufuzzaman, M.A. Tschopp, and L. Bian, “Porosity prediction: supervised-learning of thermal history for direct laser deposition, “ Journal of Manufacturing Systems 47, 69-82, 2018
王传洋 , 姜宁 , 陈再良 , 陈瑶 , 董渠 , “Prediction of sintering strength for selective laser sintering of polystyrene using artificial neural network,” 《东华大学学报:英文版》, 第 5 期 , 825-830页 , 2015

Read-around