scholarly journals A Convolutional Neural Network for Prediction of Laser Power Using Melt-Pool Images in Laser Powder Bed Fusion

IEEE Access ◽  
2020 ◽  
Vol 8 ◽  
pp. 23255-23263 ◽  
Author(s):  
Ohyung Kwon ◽  
Hyung Giun Kim ◽  
Wonrae Kim ◽  
Gun-Hee Kim ◽  
Kangil Kim
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Laser Power ◽  
Powder Bed Fusion ◽  
Laser Powder Bed Fusion ◽  
Powder Bed ◽  
Melt Pool

Measurement of the Melt Pool Length During Single Scan Tracks in a Commercial Laser Powder Bed Fusion Process

2017 ◽  
Author(s):  
J. C. Heigel ◽  
B. M. Lane
Keyword(s):  
Laser Power ◽  
High Speed ◽  
Infrared Camera ◽  
Powder Bed Fusion ◽  
Single Track ◽  
Laser Powder Bed Fusion ◽  
Powder Bed ◽  
Melt Pool ◽  
Significant Difference ◽  
Scan Speed

This work presents high speed thermographic measurements of the melt pool length during single track laser scans on nickel alloy 625 substrates. Scans are made using a commercial laser powder bed fusion machine while measurements of the radiation from the surface are made using a high speed (1800 frames per second) infrared camera. The melt pool length measurement is based on the detection of the liquidus-solidus transition that is evident in the temperature profile. Seven different combinations of programmed laser power (49 W to 195 W) and scan speed (200 mm/s to 800 mm/s) are investigated and numerous replications using a variety of scan lengths (4 mm to 12 mm) are performed. Results show that the melt pool length reaches steady state within 2 mm of the start of each scan. Melt pool length increases with laser power, but its relationship with scan speed is less obvious because there is no significant difference between cases performed at the highest laser power of 195 W. Although keyholing appears to affect the anticipated trends in melt pool length, further research is required.

Design Rules and In-Situ Quality Monitoring of Thin-Wall Features Made Using Laser Powder Bed Fusion

2019 ◽  
Author(s):  
Aniruddha Gaikwad ◽  
Farhad Imani ◽  
Prahalad Rao ◽  
Hui Yang ◽  
Edward Reutzel
Keyword(s):  
Neural Network ◽  
Aspect Ratio ◽  
Convolutional Neural Network ◽  
Thin Wall ◽  
Powder Bed Fusion ◽  
Laser Powder Bed Fusion ◽  
Design Rules ◽  
Powder Bed

Abstract The goal of this work is to quantify the link between the design features (geometry), in-situ process sensor signatures, and build quality of parts made using laser powder bed fusion (LPBF) additive manufacturing (AM) process. This knowledge is critical for establishing design rules for AM parts, and to detecting impending build failures using in-process sensor data. As a step towards this goal, the objectives of this work are two-fold: 1) Quantify the effect of the geometry and orientation on the build quality of thin-wall features. To explain further, the geometry-related factor is the ratio of the length of a thin-wall (l) to its thickness (t) defined as the aspect ratio (length-to-thickness ratio, l/t), and the angular orientation (θ) of the part, which is defined as the angle of the part in the X-Y plane relative to the re-coater blade of the LPBF machine. 2) Assess the thin-wall build quality by analyzing images of the part obtained at each layer from an in-situ optical camera using a convolutional neural network. To realize these objectives, we designed a test part with a set of thin-wall features (fins) with varying aspect ratio from Titanium alloy (Ti-6Al-4V) material — the aspect ratio l/t of the thin-walls ranges from 36 to 183 (11 mm long (constant), and 0.06 mm to 0.3 mm in thickness). These thin-wall test parts were built under three angular orientations of 0°, 60°, and 90°. Further, the parts were examined offline using X-ray computed tomography (XCT). Through the offline XCT data, the build quality of the thin-wall features in terms of their geometric integrity is quantified as a function of the aspect ratio and orientation angle, which suggests a set of design guidelines for building thin-wall structures with LPBF. To monitor the quality of the thin-wall, in-process images of the top surface of the powder bed were acquired at each layer during the build process. The optical images are correlated with the post build quantitative measurements of the thin-wall through a deep learning convolutional neural network (CNN). The statistical correlation (Pearson coefficient, ρ) between the offline XCT measured thin-wall quality, and CNN predicted measurement ranges from 80% to 98%. Consequently, the impending poor quality of a thin-wall is captured from in-situ process data.

Physics-Informed Gaussian Process Based Optimal Control of Laser Powder Bed Fusion

2020 ◽  
Author(s):  
Yong Ren ◽  
Qian Wang
Keyword(s):  
Optimal Control ◽  
Optimal Control Problem ◽  
Control Problem ◽  
Gaussian Process ◽  
Laser Power ◽  
Reference Value ◽  
Powder Bed Fusion ◽  
Laser Powder Bed Fusion ◽  
Powder Bed ◽  
Melt Pool

Abstract Regulating the melt-pool size to a constant reference value during the build process is a challenging task in Laser Powder Bed Fusion additive manufacturing (LPBF-AM). This paper considers adjusting laser power to achieve a constant melt-pool volume during laser processing of a multi-track build under LPBF-AM. First, a Gaussian Process Regression (GPR) is applied to model the variation of the melt-pool volume along the deposition distance, with physics-informed input features. Then a constrained finite-horizon optimal control problem is formulated, with a quadratic cost function defined to minimize the difference between the melt-pool volume and a reference value. A projected gradient descent algorithm is applied to compute the sequence of laser power in the proposed optimal control problem. The GPR modeling of melt-pool dynamics is trained and tested using simulated data sets generated from a commercial finite-element based AM software, and the same commercial AM software is used to evaluate the control performance. Simulation results demonstrate the effectiveness of the proposed GPR modeling and optimal control in regulating melt-pool volume for building multi-track parts with LPBF-AM.

scholarly journals Hyperspectral imaging for prediction of surface roughness in laser powder bed fusion

2021 ◽  
Author(s):  
Niklas Gerdes ◽  
Christian Hoff ◽  
Jörg Hermsdorf ◽  
Stefan Kaierle ◽  
Ludger Overmeyer
Keyword(s):  
Neural Network ◽  
Surface Roughness ◽  
Quality Assurance ◽  
Convolutional Neural Network ◽  
Absolute Error ◽  
Hyperspectral Data ◽  
Future Research ◽  
Powder Bed Fusion ◽  
Laser Powder Bed Fusion ◽  
Powder Bed

AbstractThis article discusses the relevance of in situ quality assurance in metal additive manufacturing for cost-efficient product qualification. It presents an approach for monitoring the laser powder bed fusion (LPBF) process using an area-scan hyperspectral camera to predict the surface roughness Rz with the help of a convolutional neural network. These investigations were carried out during LPBF processing of the magnesium alloy WE43 that, due to its bioresorbability and compatibility, holds significant potential for biomedical implants. A data acquisition and processing methodology has been set up to enable efficient management of the hyperspectral data. The hyperspectral images obtained from the process were labeled with the surface roughness Rz as determined by a confocal microscope. The data was used to train a convolutional neural network whose hyperparameters were optimized in a hyperparameter tuning process. The resulting network was able to predict the surface roughness within a mean absolute error (MAE) of 4.1 μm over samples from three different parameter sets. Since this is significantly smaller than the spread of the actual roughness measured (MAE = 14.3 μm), it indicates that the network identified features in the hyperspectral data linking to the roughness. These results provide the basis for future research aiming to link hyperspectral process images to further part properties relevant for quality assurance.

Simulation Aided Process Development with Multi-Spot Strategies in Laser Powder-Bed Fusion

2021 ◽  
Vol 1161 ◽  
pp. 75-82
Author(s):  
Marcel Slodczyk ◽  
Alexander Ilin ◽  
Thomas Kiedrowski ◽  
Jens Schmiemann ◽  
Vasily Ploshikhin
Keyword(s):  
Laser Power ◽  
Process Development ◽  
Research Work ◽  
Spatial Arrangement ◽  
Quality Level ◽  
Powder Bed Fusion ◽  
Laser Powder Bed Fusion ◽  
Powder Bed ◽  
Single Spot ◽  
Melt Pool

A challenge in laser powder-bed fusion is to achieve high process speed while maintaining quality level of the melting tracks. One approach to increase productivity is to distribute available laser power over several laser spots, resulting in higher melting rate. Using multiple laser spots opens up new parameter spaces in comparison to the conventional single-spot exposure. In addition to classical process parameters, e.g. total laser power and scanning speed, the distribution of power to the specific spots and the respective spatial arrangement have an impact on resulting process quality and speed. Within the scope of this research work, a physically based model is presented to define multi-spot process strategies for the generation of desired melt pool dimensions. Diffractive optical elements are used in order to adjust power or spatial arrangement of multiple laser spots. Resulting melt pool has more width and less depth compared to single-spot generated melt pools. Simulations and experiments show an optimum in applied spot distance between laser spots to obtain higher melting rates.

Measurement of the Melt Pool Length During Single Scan Tracks in a Commercial Laser Powder Bed Fusion Process

2018 ◽  
Vol 140 (5) ◽  
Author(s):  
J. C. Heigel ◽  
B. M. Lane
Keyword(s):  
Laser Power ◽  
High Speed ◽  
Infrared Camera ◽  
Powder Bed Fusion ◽  
Single Track ◽  
Laser Powder Bed Fusion ◽  
Powder Bed ◽  
Melt Pool ◽  
Significant Difference ◽  
Scan Speed

This work presents high-speed thermographic measurements of the melt pool length during single track laser scans on nickel alloy 625 substrates. Scans are made using a commercial laser powder bed fusion (PBF) machine while measurements of the radiation from the surface are made using a high speed (1800 frames per second) infrared camera. The melt pool length measurement is based on the detection of the liquidus–solidus transition that is evident in the temperature profile. Seven different combinations of programmed laser power (49–195 W) and scan speed (200–800 mm/s) are investigated, and numerous replications using a variety of scan lengths (4–12 mm) are performed. Results show that the melt pool length reaches steady-state within 2 mm of the start of each scan. Melt pool length increases with laser power, but its relationship with scan speed is less obvious because there is no significant difference between cases performed at the highest laser power of 195 W. Although keyholing appears to affect the anticipated trends in melt pool length, further research is required.

scholarly journals Melt Pool Features Analysis Using a High Speed Coaxial Monitoring System for Laser Powder Bed Fusion of Ti-6Al-4V Grade 23

2021 ◽  
Author(s):  
Aditi Thanki ◽  
Louca Goossens ◽  
Agusmian Partogi Ompusunggu ◽  
Mohamad Bayat ◽  
Abdellatif Bey-Temsamani ◽  
...  
Keyword(s):  
Energy Density ◽  
Laser Power ◽  
High Speed ◽  
Powder Bed Fusion ◽  
Laser Powder Bed Fusion ◽  
Porosity Formation ◽  
Powder Bed ◽  
Binary Images ◽  
Melt Pool ◽  
Scan Speed

Abstract In laser powder bed fusion (LPBF), defects such as pores or cracks can seriously affect the final part quality and lifetime. Keyhole porosity, being one type of porosity defects in LPBF, results from excessive energy density which may be due to changes in process parameters (laser power and scan speed) and/or result from the part’s geometry and/or hatching strategies. To study the possible occurrence of keyhole pores, experimental work as well as simulations were carried out for optimum and high volumetric energy density conditions in Ti-6Al-4V grade 23. By decreasing the scanning speed from 1000 mm/s to 500 mm/s for a fixed laser power of 170 W, keyhole porosities are formed and later observed by X-ray computed tomography. Melt pool images are recorded in real-time during the LPBF process by using a high speed coaxial Near-Infrared (NIR) camera monitoring system. The recorded images are then pre-processed using a set of image processing steps to generate binary images. From the binary images, geometrical features of the melt pool and features that characterize the spatter particles formation and ejection from the melt pool are calculated. The experimental data clearly show spatter patterns in case of keyhole porosity formation at low scan speed. A correlation between the number of pores and the amount of spatter is observed. Besides the experimental work, a previously developed, high fidelity finite volume numerical model was used to simulate the melt pool dynamics with similar process parameters as in the experiment. Simulation results illustrate and confirm the keyhole porosity formation by decreasing laser scan speed.

Real-time Image-based Feedback Control of Laser Powder Bed Fusion

2021 ◽  
pp. 1-12
Author(s):  
Aleksandr Shkoruta ◽  
Sandipan Mishra ◽  
Stephen Rock
Keyword(s):  
Real Time ◽  
Laser Power ◽  
Powder Bed Fusion ◽  
Single Track ◽  
Laser Powder Bed Fusion ◽  
Powder Bed ◽  
Melt Pool ◽  
Real Time Image ◽  
Time Image ◽  
Pool Image

Abstract This letter presents the design and experimental validation of a real-time image-based feedback control system for metal laser powder bed fusion (LPBF). A coaxial melt pool video stream is used to control laser power in real-time at 2 kHz. Modeling of the melt pool image response to changes in the input laser power is presented. Based on this identified model, a real-time feedback controller is implemented experimentally, on a single track and part scales. On a single-track scale, the controller successfully tracks a time-varying melt pool reference. On a part-level scale, the controller successfully regulates the melt pool image signature to the desired reference value, reducing layer-to-layer signal variation, and eliminating within-layer signal drift.

Sign in / Sign up

Export Citation Format

Share Document