Towards Vision-based Closed-loop Additive Manufacturing: A Review

Moltumyr, Andreas Hanssen; Arbo, Mathias Hauan; Gravdahl, Jan Tommy

doi:10.1109/sims49386.2020.9121578

Cited by 6 publications

(2 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Image-assisted closed-loop control has been used for laser-based additive manufacturing 46,47 and their quality control 48 . In fact, the application of image-based control in 3D printing and additive manufacturing in general for layer height control, over and underfill correction, temperature and layer time control, print failure alert, standoff distance control, and process parameter control has been extensively discussed in recent reviews 49,50 . In the proposed method, for the first time, image-based feedback is used for precise control of projected patterns, high-speed triggering, and flow control.…”

Section: Introductionmentioning

confidence: 99%

Rapid Prototyping of High-resolution Large Format Microfluidic Device Through Maskless Image Guided In-situ Photopolymerization

Paul

Zhao

Coster

et al. 2022

Preprint

View full text Add to dashboard Cite

Microfluidic devices have been widely used in mechanical, biomedical, chemical, and materials research. As a result, it is becoming increasingly important to have a cheap, fast, and reliable method for rapid microfabrication. However, prototyping of microfluidic devices typically suffers from the high initial cost, low resolution, rough surface finish, and long turn-around time. Here we present a strategic approach to closed-loop control of deterministic fabrication process based on in-situ image analysis called image-guided in-situ maskless lithography (IGIs-ML). Using the closed-loop control along with flush-flow functionality and leveraging the swelling behavior of the photocurable polymer, we demonstrate the fabrication of sub-micron high aspect ratio channels (800 nm width and >10 µm height) close to the light diffraction limit. This outperforms any reported rapid prototyping platforms which can typically reach tens of µm in channel width. Such dimensional capability is even comparable to some of the most advanced fabrication methods currently available. Dynamic image analysis simultaneously provides superior repeatability of densely packed patterns across a large area and multiple devices. A general and robust approach is established to fabricate a wide variety of microfluidic devices. This resolves the critical over-curing issues and size limitations in rapid prototyping of microfluidic devices, enabling affordable and reliable exploration of design space beyond the resolution of traditional photolithography.

show abstract

Section: Introductionmentioning

confidence: 99%

Rapid Prototyping of High-resolution Large Format Microfluidic Device Through Maskless Image Guided In-situ Photopolymerization

Paul

Zhao

Coster

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Currently, research predominantly focuses on some form of in-situ imaging data for localized defect detection and property prediction as well as closed-loop control (Everton et al, 2016) (Moltumyr et al, 2020). However, other, less utilized in-situ data-specifically temporal sensor data-is also being investigated for process quality monitoring.…”

Section: Introductionmentioning

confidence: 99%

Localized Defect Detection from Spatially Mapped, In-Situ Process Data With Machine Learning

et al. 2021

View full text Add to dashboard Cite

In powder bed fusion additive manufacturing, machines are often equipped with in-situ sensors to monitor the build environment as well as machine actuators and subsystems. The data from these sensors offer rich information about the consistency of the fabrication process within a build and across builds. This information may be used for process monitoring and defect detection; however, little has been done to leverage this data from the machines for more than just coarse-grained process monitoring. In this work we demonstrate how these inherently temporal data may be mapped spatially by leveraging scan path information. We then train a XGBoost machine learning model to predict localized defects—specifically soot–using only the mapped process data of builds from a laser powder bed fusion process as input features. The XGBoost model offers a feature importance metric that will help to elucidate possible relationships between the process data and observed defects. Finally, we analyze the model performance spatially and rationalize areas of greater and lesser performance.

show abstract