TY - GEN
T1 - In-situ monitoring of laser powder bed fusion process anomalies via a comprehensive analysis of off-axis camera data
AU - Prasad Vallabh, Chaitanya Krishna
AU - Xiong, Yubo
AU - Zhao, Xiayun
N1 - Publisher Copyright:
Copyright © 2020 ASME.
PY - 2020
Y1 - 2020
N2 - In-situ monitoring of a Laser Powder-Bed Fusion (LPBF) additive manufacturing process is crucial in enhancing the process efficiency and ensuring the built part integrity. In this work, we present an in-situ monitoring method using an off-axis camera for monitoring layer-wise process anomalies. The in-situ monitoring is performed with a spatial resolution of 512 × 512 pixels, with each pixel representing 250 × 250 µm and a relatively high data acquisition rate of 500 Hz. An experimental study is conducted by using the developed in-situ off-axis method for monitoring the build process for a standard tensile bar. Real-time video data is acquired for each printed layer. Data analytics methods are developed to identify layer-wise anomalies, observe powder bed characteristics, reconstruct 3D part structure, and track the spatter dynamics. A deep neural network architecture is trained using the acquired layer-wise images and tested by images embedded with artificial anomalies. The real-time video data is also used to perform a preliminary spatter analysis along the laser scan path. The developed methodology is aimed to extract as much information as possible from a single set of camera video data. It will provide the AM community with an efficient and capable process monitoring tool for process control and quality assurance while using LPBF to produce high-standard components in industrial (such as, aerospace and biomedical industries) applications.
AB - In-situ monitoring of a Laser Powder-Bed Fusion (LPBF) additive manufacturing process is crucial in enhancing the process efficiency and ensuring the built part integrity. In this work, we present an in-situ monitoring method using an off-axis camera for monitoring layer-wise process anomalies. The in-situ monitoring is performed with a spatial resolution of 512 × 512 pixels, with each pixel representing 250 × 250 µm and a relatively high data acquisition rate of 500 Hz. An experimental study is conducted by using the developed in-situ off-axis method for monitoring the build process for a standard tensile bar. Real-time video data is acquired for each printed layer. Data analytics methods are developed to identify layer-wise anomalies, observe powder bed characteristics, reconstruct 3D part structure, and track the spatter dynamics. A deep neural network architecture is trained using the acquired layer-wise images and tested by images embedded with artificial anomalies. The real-time video data is also used to perform a preliminary spatter analysis along the laser scan path. The developed methodology is aimed to extract as much information as possible from a single set of camera video data. It will provide the AM community with an efficient and capable process monitoring tool for process control and quality assurance while using LPBF to produce high-standard components in industrial (such as, aerospace and biomedical industries) applications.
KW - Anomaly detection
KW - Deep learning
KW - In-situ process monitoring
KW - Spatter analysis
UR - http://www.scopus.com/inward/record.url?scp=85100945969&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85100945969&partnerID=8YFLogxK
U2 - 10.1115/MSEC2020-8300
DO - 10.1115/MSEC2020-8300
M3 - Conference contribution
AN - SCOPUS:85100945969
T3 - ASME 2020 15th International Manufacturing Science and Engineering Conference, MSEC 2020
BT - Additive Manufacturing; Advanced Materials Manufacturing; Biomanufacturing; Life Cycle Engineering; Manufacturing Equipment and Automation
T2 - ASME 2020 15th International Manufacturing Science and Engineering Conference, MSEC 2020
Y2 - 3 September 2020
ER -