Listening out for printing defects

Researchers at the Ecole Polytechnique Fédérale de Lausanne (EPFL) in Switzerland have developed a pioneering approach to detecting defects during the laser additive manufacturing of metallic objects. They report this novel approach in a paper in Nature Communications.

The progression of laser additive manufacturing – the 3D printing of metallic objects using powders and lasers – has often been hindered by unexpected defects. Traditional monitoring methods, such as thermal imaging and machine learning algorithms, have shown significant limitations. They often either overlook defects or misinterpret them, making precision manufacturing elusive and barring the technique from essential industries like aeronautics and automotive manufacturing.

But what if it were possible to detect defects in real time based on the differences in the sound the printer makes during a flawless print and one with irregularities? Up until now, the prospect of detecting defects this way was deemed unreliable. However, researchers at the Laboratory of Thermomechanical Metallurgy (LMTM) at EPFL's School of Engineering have successfully challenged this assumption.

"There's been an ongoing debate regarding the viability and effectiveness of acoustic monitoring for laser-based additive manufacturing,” said Roland Logé, head of the LMTM. “Our research not only confirms its relevance but also underscores its advantage over traditional methods."

This research is of paramount importance to the industrial sector as it introduces a groundbreaking yet cost-effective method for monitoring and improving the quality of products made through Laser Powder Bed Fusion (LPBF). "The synergy of synchrotron X-ray imaging with acoustic recording provides real-time insight into the LPBF process, facilitating the detection of defects that could jeopardize product integrity," said Milad Hamidi Nasab, lead researcher of the study.

LPBF is a cutting-edge technique that's reshaping metal manufacturing. Essentially, it uses a high-intensity laser to meticulously melt minuscule metal powders, building up detailed 3D metallic constructs layer-by-layer. It can be thought of as the metallic version of conventional 3D printing, but with an added degree of sophistication.

Rather than melted plastic, LPBF employs a fine layer of microscopic metal powder, which can vary in size from the thickness of a human hair to a fine grain of salt (15–100 μm). The laser moves across this layer, melting specific patterns based on a digital blueprint. This technique allows the crafting of bespoke, complex parts like lattice structures or distinct geometries, with minimal excess. Nevertheless, this promising method isn't devoid of challenges.

When the laser interacts with the metal powder, creating what is known as a melt pool, the metal fluctuates between liquid, vapor and solid phases. Occasionally, due to variables such as the laser's angle or the presence of specific geometrical attributes of the powder or of the part, this process might falter. Such ‘inter-regime instabilities’ can sometimes prompt shifts between two melting regimes, known as ‘conduction’ and ‘keyhole’.

Unstable keyhole regimes, where the molten powder pool delves deeper than intended, can create pockets of porosity, culminating in structural flaws in the end product. To help measure the width and depth of the melt pool in X-ray images, the Image Analysis Hub at the EPFL Center for Imaging developed an approach for visualizing small changes associated with the liquid metal and a tool for annotating the melt pool geometry.

In a joint venture with the Paul Scherrer Institute (PSI) and the Swiss Federal Laboratories for Materials Science and Technology (Empa), also in Switzerland, the EPFL team has now formulated an experimental design that melds operando X-ray imaging experiments with acoustic emission measurements. Using a miniaturized LPBF printer developed in the group of Steven Van Petegem, the team conducted its experiments at the TOMCAT beamline of the Swiss Light Source at PSI.

By positioning an ultra-sensitive microphone inside the printing chamber, the team was able to pinpoint distinct shifts in the acoustic signal during regime transitions, thereby directly identifying defects during manufacturing.

A pivotal moment in the research came with the introduction of an adaptive filtering technique by signal processing expert Giulio Masinelli from Empa. "This filtering approach," Masinelli explained, "allows us to discern, with unparalleled clarity, the relationship between defects and the accompanying acoustic signature."

Unlike typical machine-learning algorithms, which excel at extracting patterns from statistical data but are often tailored to specific scenarios, this approach provides broader insights into the physics of melting regimes, while offering superior temporal and spatial precision.

Reinforcing Switzerland's reputation for meticulous craftsmanship and manufacturing accuracy, this study underscores the need for consistent manufacturing techniques. Furthermore, it suggests the potential for early detection and correction of defects, enhancing product quality.

"This research paves the way for a better understanding and refinement of the manufacturing process and will ultimately lead to higher product reliability in the long term," Logé concludes.

This story is adapted from material from EPFL, with editorial changes made by Materials Today. The views expressed in this article do not necessarily represent those of Elsevier. Link to original source.