IoT and Machine learning for in-situ process control using Laser Based Additive Manufacturing (LBAM) case study

AbstractAdditive Manufacturing (AM) has a large potential to revolutionize the manufacturing industry, yet the printing parameters and part design have a profound impact on the robustness of the printing process as well as the resulting quality of the manufactured components. To control the printing process, a substantial number of parameters is measured while printing and used primarily to control and adjust the printing process in-situ. The question raised in this paper is how to benefit from these data being gathered to gain insight into the print process stability. The case study performed included the analysis of data gathered during printing 22 components. The analysis was performed with a widely used Random Forest Classifier. The study revealed that the data did contain some detectable patterns that can be used further in assessing the quality of the printed component, however, they were distinct enough so that in case the test and train sets were comprised of separate components the predictions’ result was very poor. The study gives a good understanding of what is necessary to do a meaningful analytics study of manufacturing data from a design perspective.

Download Full-text

Toward in-situ flaw detection in laser powder bed fusion additive manufacturing through layerwise imagery and machine learning

Journal of Manufacturing Systems ◽

10.1016/j.jmsy.2021.01.008 ◽

2021 ◽

Vol 59 ◽

pp. 12-26

Author(s):

Zackary Snow ◽

Brett Diehl ◽

Edward W. Reutzel ◽

Abdalla Nassar

Keyword(s):

Machine Learning ◽

Additive Manufacturing ◽

Flaw Detection ◽

Powder Bed Fusion ◽

Laser Powder Bed Fusion ◽

Powder Bed

Download Full-text

Utility of optical heterodyne displacement sensing and laser ultrasonics as in situ process control diagnostic for additive manufacturing

Optical Engineering ◽

10.1117/1.oe.57.4.041415 ◽

2018 ◽

Vol 57 (04) ◽

pp. 1 ◽

Cited By ~ 3

Author(s):

Anthony J. Manzo ◽

Henry Helvajian

Keyword(s):

Additive Manufacturing ◽

Process Control ◽

Laser Ultrasonics ◽

Optical Heterodyne

Download Full-text

Using machine learning to identify in-situ melt pool signatures indicative of flaw formation in a laser powder bed fusion additive manufacturing process

Additive Manufacturing ◽

10.1016/j.addma.2018.11.010 ◽

2019 ◽

Vol 25 ◽

pp. 151-165 ◽

Cited By ~ 17

Author(s):

Luke Scime ◽

Jack Beuth

Keyword(s):

Machine Learning ◽

Additive Manufacturing ◽

Manufacturing Process ◽

Powder Bed Fusion ◽

Laser Powder Bed Fusion ◽

Powder Bed ◽

Melt Pool

Download Full-text

Multi-Extrusion Additive Manufacturing for Error Compensation: A Demonstrative Case Study

Volume 2: Manufacturing Processes; Manufacturing Systems; Nano/Micro/Meso Manufacturing; Quality and Reliability ◽

10.1115/msec2020-8497 ◽

2020 ◽

Author(s):

Chen Kan ◽

Zehao Ye ◽

Yiran Yang ◽

Lei Di ◽

Deep Shah ◽

...

Keyword(s):

Additive Manufacturing ◽

Error Compensation ◽

Manufacturing Industry ◽

Dimensional Accuracy ◽

Production Method ◽

Geometric Error ◽

Layer By Layer ◽

Subtractive Manufacturing

Abstract The global additive manufacturing industry has been rapidly increasing, owing to its unique layer-by-layer production method. While additive manufacturing has superior capabilities compared to traditional subtractive manufacturing, limitations still exist, which significantly hinder the larger-scale implementations of additive manufacturing. Some challenging issues include unsatisfactory dimensional accuracy, surface quality, etc. In the literature, extensive research efforts have dedicated to detecting, predicting, and compensating process errors using various methodologies. In this work, a new approach is proposed for error compensation using multi-extrusion additive manufacturing process. Three demonstrative case studies are conducted, i.e., multicolor and/or multimaterial printing, geometric error compensation, and rough surface compensation. Experimental results have shown that the proposed approach is effective in utilizing the multi-nozzle capability in additive manufacturing quality control. Notably, the proposed approach has remarkable potentials to be extended for in-situ error compensation. Our future forays will focus on integrating the proposed approach with in-situ process monitoring approaches for layer-wise defection and compensation of process anomalies.

Download Full-text

In-Situ Laser-Based Process Monitoring and In-Plane Surface Anomaly Identification for Additive Manufacturing Using Point Cloud and Machine Learning

10.1115/detc2021-69436 ◽

2021 ◽

Author(s):

Jiaqi Lyu ◽

Javid Akhavan Taheri Boroujeni ◽

Souran Manoochehri

Keyword(s):

Machine Learning ◽

Additive Manufacturing ◽

Data Processing ◽

Process Monitoring ◽

Point Cloud ◽

Plane Surface ◽

Laser Scanner ◽

Classification Model ◽

Depth Image

Abstract Additive Manufacturing (AM) is a trending technology with great potential in manufacturing. In-situ process monitoring is a critical part of quality assurance for AM process. Anomalies need to be identified early to avoid further deterioration of the part quality. This paper presents an in-situ laser-based process monitoring and anomaly identification system to assure fabrication quality of Fused Filament Fabrication (FFF) machine. The proposed data processing and communication architecture of the monitoring system establishes the data transformation between workstation, FFF machine, and laser scanner control system. The data processing performs calibration, filtering, and segmentation for the point cloud of each layer acquired from a 3D laser scanner during the fabrication process. The point cloud dataset with in-plane surface depth information is converted into a 2D depth image. Each depth image is discretized into 100 equal regions of interest and then labeled accordingly. Using the image dataset, four Machine Learning (ML) classification models are trained and compared, namely Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Convolutional Neural Network (CNN), and Hybrid Convolution AutoEncoder (HCAE). The HCAE classification model shows the best performance based on F-scores to effectively classify the in-plane anomalies into four categories, namely empty region, normal region, bulge region, and dent region.

Download Full-text