An official website of the United States government
Abstract Machine learning (ML) models are used for in-situ monitoring in additive manufacturing (AM) for defect detection. However, sensitive information stored in ML models, such as part designs, is at risk of data leakage due to unauthorized access. To address this, differential privacy (DP) introduces noise into ML, outperforming cryptography, which is slow, and data anonymization, which does not guarantee privacy. While DP enhances privacy, it reduces the precision of defect detection. This paper proposes combining DP with Hyperdimensional Computing (HDC), a brain-inspired model that memorizes training sample information in a large hyperspace, to optimize real-time monitoring in AM while protecting privacy. Adding DP noise to the HDC model protects sensitive information without compromising defect detection accuracy. Our studies demonstrate the effectiveness of this approach in monitoring anomalies, such as overhangs, using high-speed melt pool data analysis. With a privacy budget set at 1, our model achieved an F-score of 94.30%, surpassing traditional models like ResNet50, DenseNet201, EfficientNet B2, and AlexNet, which have performance up to 66%. Thus, the intersection of DP and HDC promises accurate defect detection and protection of sensitive information in AM. The proposed method can also be extended to other AM processes, such as fused filament fabrication.
more »
« less
Identifying build orientation of 3D ‐printed materials using convolutional neural networks
Abstract The advent of additive manufacturing (AM) processes brought with it intense research into various materials and manufacturing processes. At the same time, the need for validation of material properties, as well as study and forecasting of aging, has arisen. Modern imaging techniques, like X‐ray computed tomography (XCT), are a convenient vehicle for such studies; however, the large datasets they produce require novel analysis techniques to efficiently extract critical information. In this paper, we present our work on developing a 3D extension of the ResNet architecture to distinguish between two build orientations of tensile bars produced by AM. Using only information from XCT, our method achieves a 99.3% correct classification at a misidentification of 1%.
more »
« less
- Award ID(s):
- 1633216
- PAR ID:
- 10360129
- Publisher / Repository:
- Wiley Blackwell (John Wiley & Sons)
- Date Published:
- Journal Name:
- Statistical Analysis and Data Mining: The ASA Data Science Journal
- Volume:
- 14
- Issue:
- 6
- ISSN:
- 1932-1864
- Page Range / eLocation ID:
- p. 575-582
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
The goal of this work to mitigate flaws in metal parts produced from laser powder bed fusion (LPBF) additive manufacturing (AM) process. As a step towards this goal, the objective of this work is to predict the build quality of a part as it is being printed via deep learning of in-situ layer-wise images obtained from an optical camera instrumented in the LPBF machine. To realize this objective, we designed a set of thin-wall features (fins) from Titanium alloy (Ti-6Al-4V) material with varying length-to-thickness ratio. These thin-wall test parts were printed under three different build orientations and in-situ images of their top surface were acquired during the process. The parts were examined offline using X-ray computed tomography (XCT), and their build quality was quantified in terms of statistical features, such as the thickness and consistency of its edges. Subsequently, a deep learning convolutional neural network (CNN) was trained to predict the XCT-derived statistical quality features using the layer-wise optical images of the thin-wall part as inputs. The statistical correlation between CNN-based predictions and XCT-observed quality measurements exceeds 85%. This work has two outcomes consequential to the sustainability of additive manufacturing: (1) It provides practitioners with a guideline for building thin-wall features with minimal defects, and (2) the high correlation between the offline XCT measurements and in-situ sensor-based quality metrics substantiates the potential for applying deep learning approaches for the real-time prediction of build flaws in LPBF.more » « less
-
This study aims to discuss the state-of-the-art digital factory (DF) development combining digital twins (DTs), sensing devices, laser additive manufacturing (LAM) and subtractive manufacturing (SM) processes. The current shortcomings and outlook of the DF also have been highlighted. A DF is a state-of-the-art manufacturing facility that uses innovative technologies, including automation, artificial intelligence (AI), the Internet of Things, additive manufacturing (AM), SM, hybrid manufacturing (HM), sensors for real-time feedback and control, and a DT, to streamline and improve manufacturing operations. Design/methodology/approachThis study presents a novel perspective on DF development using laser-based AM, SM, sensors and DTs. Recent developments in laser-based AM, SM, sensors and DTs have been compiled. This study has been developed using systematic reviews and meta-analyses (PRISMA) guidelines, discussing literature on the DTs for laser-based AM, particularly laser powder bed fusion and direct energy deposition, in-situ monitoring and control equipment, SM and HM. The principal goal of this study is to highlight the aspects of DF and its development using existing techniques. FindingsA comprehensive literature review finds a substantial lack of complete techniques that incorporate cyber-physical systems, advanced data analytics, AI, standardized interoperability, human–machine cooperation and scalable adaptability. The suggested DF effectively fills this void by integrating cyber-physical system components, including DT, AM, SM and sensors into the manufacturing process. Using sophisticated data analytics and AI algorithms, the DF facilitates real-time data analysis, predictive maintenance, quality control and optimal resource allocation. In addition, the suggested DF ensures interoperability between diverse devices and systems by emphasizing standardized communication protocols and interfaces. The modular and adaptable architecture of the DF enables scalability and adaptation, allowing for rapid reaction to market conditions. Originality/valueBased on the need of DF, this review presents a comprehensive approach to DF development using DTs, sensing devices, LAM and SM processes and provides current progress in this domain.more » « less
-
null (Ed.)Abstract Additive manufacturing (AM) processes present designers with unique capabilities while imposing several process limitations. Designers must leverage the capabilities of AM — through opportunistic design for AM (DfAM) — and accommodate AM limitations — through restrictive DfAM — to successfully employ AM in engineering design. These opportunistic and restrictive DfAM techniques starkly contrast the traditional, limitation-based design for manufacturing techniques — the current standard for design for manufacturing (DfM). Therefore, designers must transition from a restrictive DfM mindset towards a ‘dual’ design mindset — using opportunistic and restrictive DfAM concepts. Designers’ prior experience, especially with a partial set of DfM and DfAM techniques could inhibit their ability to transition towards a dual DfAM approach. On the other hand, experienced designers’ auxiliary skills (e.g., with computer-aided design) could help them successfully use DfAM in their solutions. Researchers have investigated the influence of prior experience on designers’ use of DfAM tools in design; however, a majority of this work focuses on early-stage ideation. Little research has studied the influence of prior experience on designers’ DfAM use in the later design stages, especially in formal DfAM educational interventions, and we aim to explore this research gap. From our results, we see that experienced designers report higher baseline self-efficacy with restrictive DfAM but not with opportunistic DfAM. We also see that experienced designers demonstrate a greater use of certain DfAM concepts (e.g., part and assembly complexity) in their designs. These findings suggest that introducing designers to opportunistic DfAM early could help develop a dual design mindset; however, having more engineering experience might be necessary for them to implement this knowledge into their designs.more » « less
-
From wool to Kevlar, one-dimensional (1D) fiber has experienced the transition from clothing materials to structural applications in the past centuries. However, the recent advancements in tooling engineering and manufacturing processes have attracted much attention from both academia and industry to fabricate novel, versatile fibers with unique microstructures and unprecedented properties. This mini-review focuses on the fabrication techniques of porous, coaxial, layer-by-layer, and segmented fibers with continuous solution and melt fiber spinning methods. In each section of this review article, the unique structure-related applications, including intelligent devices, healthcare devices, energy storage systems, wearable electronics, and sustainable products, are discussed and evaluated. Finally, the combination of additive manufacturing (AM) for 1D fiber patterning in two-dimensional (2D) and three-dimensional (3D) devices, in addition to challenges in the reviewed fiber microstructures, is briefly introduced in the conclusion section.more » « less
