Kansas State University researchers develop AI system for 3D printing process monitoring

Researchers from Kansas State University’s Department of Industrial and Manufacturing Systems Engineering (IMSE) have developed a new quality monitoring system for the 3D printing process.

With integrated supervised machine learning, a camera, and image processing software, the researchers created a production quality monitoring system for assessing 3D printed parts in real-time. IMSE researchers Ugandhar Delli and Dr. Shing Chang stated:

“Conventionally, the quality of 3D printed parts is being checked after the printing is done. Detecting defects during the printing process not only help to eliminate waste of material and time but also prevent the need to reprint the whole part.”

Step-by-step process monitoring

Part qualification, particularly within metal 3D printed components, can be a lengthy process for manufacturers that can potentially withhold a product from its intended market. However, this study demonstrates that with step-by-step quality checks, production-scale 3D printing operations can be improved with cost and time efficiency.

With a LulzBot Mini 3D printer, Delli and Dr. Chang printed several parts, demonstrating no defects, by stopping production during ‘critical stages’ where the geometry of the part significantly changes. The study states “consider a complex part which involves different stages of printing like skirt/base, body and the top. These stages could be considered as the desired checkpoints to inspect quality at.”

The researchers proposed a three-step quality monitoring system to identify and assess the stages of the 3D printing process. Firstly, the researchers established checkpoints for a 3D printed part according to its geometry. Secondly, with the help of a mounted camera, images were taken of the semi-finished parts at each checkpoint.

Finally, part quality was automatically assessed through image processing and a supervised machine learning algorithm called a support vector machine (SVM).

The process monitoring setup. Image via Kansas State University/Delli and Dr. Chang.

Defective or effective?

Concluding the research, Delli and Dr. Chang observed that their method can detect both completion failure defects, due to filament running out, or geometrical defects, caused by unwanted stoppages. Nevertheless, this three-step quality monitoring system had its drawbacks.

According to the study, “[this] method might not be able to detect the defects on the vertical plane which cannot be seen in the top view image. This gives us a direction for future research to incorporate cameras on the sides of the printer as well to detect defects on both the horizontal and vertical planes.”

Devices such as Dyze Design’s filament sensors, and 3DGence’s high-quality extruders that prevent clogging, have also been seen to remedy potential defects within the 3D printing process. Now, Delli and Dr. Chang will work to incorporate a multi-camera system on 3D printers to capture part production errors.

A PLA printed part stopped and assessed at a checkpoint. Photo via the Kansas State University/Delli and Dr. Chang.

The research paper “Automated Process Monitoring in 3D Printing Using Supervised Machine Learning” is co-authored by Dr. Shing Chang and Ugandhar Delli.

Keep up with the latest news in 3D printing by subscribing to the 3D Printing Industry newsletter. Also, follow us on Twitter, and like us on Facebook.

Seeking a new position in the additive manufacturing industry? Search and post 3D Printing Jobs for opportunities and new talent across engineering, marketing, sales and more.

Featured image shows the Lulzbot Mini 2 3D printer. Photo via Lulzbot.