We use cookies to enhance your experience. By continuing to browse this website, you agree to our use of cookies. More information.

The researchers devised a data-centric approach to generative modeling of 3D printed steel. The team believes that their model helps determine the quality of the design and materials before manufacturing. The research was published in Proceedings A of the Royal Society.

Research: Data-centric 3D printing steel generation modeling method. Image Credit: Nuttawut Uttamaharad/Shutterstock.com

The additive manufacturing (AM) process, commonly referred to as 3D printing, presents many exciting opportunities for many industries including engineering, aerospace, and automotive. This is because the development of additive manufacturing of metallic materials has promoted the production of complex and complex parts, such as jet engine fuel nozzles.

However, unlike traditional materials, metals manufactured and formed by the AM process show greater changes in their geometric and mechanical properties.

(a) The 3D printing protocol developed by MX3D uses a welding head connected to a robotic arm (picture from Joris Laarman, www.jorislaarman.com). (b) Pedestrian bridge made of three-dimensional printed steel. (c) A close-up of the change in the surface geometry of the material. Image source: Dodwell TJ et al., Proceedings of the Royal Society A

These changes are not widely understood, which hinders the establishment of safety standards for post-manufacturing testing, which means that manufacturers will encounter certain obstacles.

Among all the emerging technologies of AM, one of the most promising technologies for producing large parts is arc additive manufacturing (WAAM). This process is a variant of Direct Energy Deposition (DED) technology that uses an arc welding process to 3D print metal parts.

Unlike traditional metal powder additive manufacturing methods, WAAM uses an electric arc as a heat source on the metal substrate to melt the metal wire. When the wire melts and expresses in the form of beads on the substrate. When the beads adhere to each other, a layer of metallic material is created. Then repeat the process layer by layer until the metal part is completed.

However, due to uncertainties in the structure and mechanical properties and complex thermal deformation, the process still presents challenges. Therefore, a method is needed to support the effective material characterization of WAAM: “Specifically, we have developed a generative statistical model that can perform integration-based predictions on the performance of stainless steel WAAM components before they are manufactured,” Dodwell explained.

Laser scanning of 3D printed steel plates. (a) A photo of a handheld scanning device. (b) The orthographic projection of the scanned paper (part of it), representing a panel. The nominal thickness of the panel is 3.5 mm. (Observe the slight curvature in the nominal plate introduced by the residual stress). Image source: Dodwell TJ et al., Proceedings of the Royal Society A

In order to effectively and successfully characterize the mechanical properties of WAAM steel, the team also needed to develop a method to isolate the characterization of geometric changes. Then, a unified statistical model was generated by combining the generated statistical model of mechanical and geometric changes in WAAM steel. More importantly, this generative statistical model handles the two sources of variation independently.

"Combine the above statistical models into a unified statistical model, which can be instantiated in the finite element software ABAQUS, so that the nominal design of CHS can have real surface geometry and material properties based on this unified statistical data. Model," Dodwell said.

The team proved that obtaining a relatively small amount of training data can be used to successfully train the generative statistical model of AM steel. "Approximate prediction of the performance of hypothetical components is a prerequisite for efficient and low-cost design using WAAM," Dodwell said.

This paved the way for general predictions of the performance of various structural lengths, but the team stated that experimental testing is still essential when safety-critical certification is involved.

However, demonstrating the ability to generate statistical models for AM steel and WAAM can enhance the potential for future research, namely the use of stochastic simulations to determine key performance aspects of AM metallic materials.

(a) Samples from training the generative model. (b) Buckle cylinder, experimental production. (c) Based on the prediction of the generative model and the experimental load-displacement curve, introducing and not introducing macroscopic geometric defects, and a perfect elastic model. The two sd areas are shaded. Image source: Dodwell TJ et al., Proceedings of the Royal Society A

Dodwell and his team believe that their model will be very suitable for this task, "be able to quantify (and distinguish) cognitive uncertainty (due to limited training data) and random uncertainty (due to the inherent changes in the printing protocol)."

Stochastic simulation is important in the additive manufacturing process because it helps to identify areas related to structural and mechanical integrity. These simulations will play a key role in improving printing protocols and reducing security certification costs.

A key difference between the previous work on AM and the current work done by Dodwell and his research team is that their models are both generated and independent of any specific design. This means that limited training data for only one type of component may be sufficient to facilitate prediction of the performance of another component that has not yet been seen.

Dodwell TJ, Fleming LR, Buchanan C., Kyvelou P., Detommaso G., Gosling PD, Scheichl R., Kendall WS, Gardner L., Girolami MA, and Oates CJ 2021 Data-centric 3D printing generative modeling method SteelProc . R. Soc. A.4772021044420210444 https://royalsocietypublishing.org/doi/10.1098/rspa.2021.0444

Disclaimer: The views expressed here are those expressed by the author in a personal capacity, and do not necessarily represent the views of the owner and operator of this website AZoM.com Limited T/A AZoNetwork. This disclaimer forms part of the terms and conditions of use of this website.

David is an academic researcher and interdisciplinary artist. David's current research explores how science and technology, especially the Internet and artificial intelligence, can be put into practice to influence a new shift to utopianism and the reemergence of commons theory.

Please use one of the following formats to cite this article in your paper, essay, or report:

Cross, David. (2021, November 16). Simulate additive manufacturing of steel. AZoM. Retrieved from https://www.azom.com/news.aspx?newsID=57336 on November 18, 2021.

Cross, David. "Simulation of Additive Manufacturing Steel". AZoM. November 18, 2021. <https://www.azom.com/news.aspx?newsID=57336>.

Cross, David. "Simulation of Additive Manufacturing Steel". AZoM. https://www.azom.com/news.aspx?newsID=57336. (Accessed on November 18, 2021).

Cross, David. 2021. Simulate additive manufacturing of steel. AZoM, viewed on November 18, 2021, https://www.azom.com/news.aspx?newsID=57336.

Do you have any comments, updates, or anything you want to add to this news story?

AZoM talks with Dr. Robert Shepherd from Cornell University. In their research, Dr. Shepherd and his team produced a key component of a technology that can make inflatable braille that changes shape under the touch of the user a reality. Triggered by burning, Dr. Shepherd and his team created a hapt

Michael is part of a team of researchers at the Massachusetts Institute of Technology, which has developed a data-driven system that accelerates the process of discovering new 3D printing materials.

Coxem has developed a new automated large-area particle analyzer based on SEM (Scanning Electron Microscope), which can be used to analyze and classify particles by size and element.

The knife grinder GRINDOMIX GM 200 has two sharp, sturdy blades and a powerful 1000 W motor, making it an ideal instrument for grinding and homogenizing food and feed.

The Extrel VeraSpec atmospheric pressure ionization mass spectrometer (APIMS) is designed to provide reliable and reproducible low part-per-trillion detection limits for pollution control in ultra-high purity (UHP) gases used in semiconductor and other high-tech industrial applications.

For many years, rotovaps have been the standard for laboratories and industries that perform chemistry, such as laboratories in the pharmaceutical, chemistry, life sciences, materials, environment, and cannabis sectors.

New research in China Physics Letters investigated the superconductivity and charge density waves in single-layer materials grown on graphene substrates.

This article will explore a new method that makes it possible to design nanomaterials with an accuracy of less than 10 nm.

This article reports on the preparation of synthetic BCNTs by catalytic thermal chemical vapor deposition (CVD), resulting in rapid charge transfer between the electrode and the electrolyte.

AZoM.com-AZoNetwork website

Owned and operated by AZoNetwork, © 2000-2021