Powder Bed Defect Extraction of Laser Powder Bed Fusion Additive Manufacturing with Tensor Robust Principal Component Analysis

被引:1
作者
Jiang, Hao [1 ]
Zhang, Xingwu [1 ]
Zhao, Zhibin [1 ]
Wang, Chenxi [1 ]
Miao, Huihui [1 ]
Chen, Xuefeng [1 ]
机构
[1] Xi An Jiao Tong Univ, Sch Mech Engn, Xian, Peoples R China
来源
2024 IEEE INTERNATIONAL INSTRUMENTATION AND MEASUREMENT TECHNOLOGY CONFERENCE, I2MTC 2024 | 2024年
基金
中国博士后科学基金; 国家重点研发计划; 中国国家自然科学基金;
关键词
Laser powder-bed fusion; additive manufacturing; powder bed defects; tensor robust principal component analysis; ANOMALY DETECTION; CLASSIFICATION;
D O I
10.1109/I2MTC60896.2024.10560776
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
The process monitoring and quality control of metal additive manufacturing has been a research hotspot in recent years. AI-driven laser powder bed fusion process monitoring is currently the most popular research idea. However, machine learning methods have high requirements for data sets and require a lot of cost. Therefore, this paper attempts to use a new technology, namely tensor robust principal component analysis, to directly analyze and process powder bed images, to realize the extraction of powder bed defects. The main steps include: 1) synthesizing a large number of powder bed images into a tensor; 2) separating the tensor into low-rank components and sparse components. By analyzing and processing the powder bed images collected during the two printing processes and the powder bed images with serious defects collected during a long period, the problems and suitable use scenarios of tensor robust principal component analysis in dealing with powder bed defects are discussed. It is found that it has the best effect in dealing with continuous powder bed images with variable defects.
引用
收藏
页数:6
相关论文
共 12 条
[1]   Study of the spatter distribution on the powder bed during selective laser melting [J].
Bin Anwar, Ahmad ;
Quang-Cuong Pham .
ADDITIVE MANUFACTURING, 2018, 22 :86-97
[2]   Part defects identification in selective laser melting via digital image processing of powder bed anomalies [J].
Boschetto, A. ;
Bottini, L. ;
Vatanparast, S. ;
Veniali, F. .
PRODUCTION ENGINEERING-RESEARCH AND DEVELOPMENT, 2022, 16 (05) :691-704
[3]   Distributed optimization and statistical learning via the alternating direction method of multipliers [J].
Boyd S. ;
Parikh N. ;
Chu E. ;
Peleato B. ;
Eckstein J. .
Foundations and Trends in Machine Learning, 2010, 3 (01) :1-122
[4]   In Situ Monitoring of Additive Manufacturing Using Digital Image Correlation: A Review [J].
Cunha, Filipa G. ;
Santos, Telmo G. ;
Xavier, Jose .
MATERIALS, 2021, 14 (06)
[5]   Tensor robust principal component analysis based on Bayesian Tucker decomposition for thermographic inspection [J].
Hu, Yue ;
Cui, Fangsen ;
Zhao, Yifan ;
Li, Fucai ;
Cao, Shuai ;
Xuan, Fu-zhen .
MECHANICAL SYSTEMS AND SIGNAL PROCESSING, 2023, 204
[6]   Tensor Robust Principal Component Analysis with a New Tensor Nuclear Norm [J].
Lu, Canyi ;
Feng, Jiashi ;
Chen, Yudong ;
Liu, Wei ;
Lin, Zhouchen ;
Yan, Shuicheng .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2020, 42 (04) :925-938
[7]   Layer-wise anomaly detection and classification for powder bed additive manufacturing processes: A machine-agnostic algorithm for real-time pixel-wise semantic segmentation [J].
Scime, Luke ;
Siddel, Derek ;
Baird, Seth ;
Paquit, Vincent .
ADDITIVE MANUFACTURING, 2020, 36
[8]   Using machine learning to identify in-situ melt pool signatures indicative of flaw formation in a laser powder bed fusion additive manufacturing process [J].
Scime, Luke ;
Beuth, Jack .
ADDITIVE MANUFACTURING, 2019, 25 :151-165
[9]   A multi-scale convolutional neural network for autonomous anomaly detection and classification in a laser powder bed fusion additive manufacturing process [J].
Scime, Luke ;
Beuth, Jack .
ADDITIVE MANUFACTURING, 2018, 24 :273-286
[10]   Anomaly detection and classification in a laser powder bed additive manufacturing process using a trained computer vision algorithm [J].
Scime, Luke ;
Beuth, Jack .
ADDITIVE MANUFACTURING, 2018, 19 :114-126