Fused Deposition Modeling Source Identification Using Smartphone Photographs and Deep Learning

Sustainable additive manufacturing supply chains with a plithogenic stakeholder analysis: Waste reduction through digital transformation

The flexibility offered by additive manufacturing (AM) technologies to fabricate complex geometries poses several challenges to non-destructive evaluation (NDE) and quality control (QC) techniques. Existing NDE and QC techniques are not optimized for AM processes, materials, or parts. Such lack of reliable means to verify and qualify AM parts is a significant barrier to further industrial adoption of AM technologies. Electromechanical impedance measurements have been recently introduced as an alternative solution to detect anomalies in AM parts. With this approach, piezoelectric wafers bonded to the part under test are utilized as collocated sensors and actuators. Due to the coupled electromechanical characteristics of piezoelectric materials, the measured electrical impedance of the piezoelectric wafer depends on the mechanical impedance of the part under test, allowing build defects to be detected. This paper investigates the effectiveness of impedance-based NDE approach to detect internal porosity in AM parts. This type of build defects is uniquely challenging as voids are normally embedded within the structure and filled with unhardened model or supporting material. The impact of internal voids on the electromechanical impedance of AM parts is studied at several frequency ranges.

Fused deposition modeling (FDM) is one of the most common additive manufacturing (AM) technologies for thermoplastic materials. With the development of carbon fiber-reinforced polymer (CFRP) filament for FDM, AM parts with improved strength and functionality can be realized. CFRP is anisotropic material and its mechanical properties have been well studied, however, AM printing strategy for CFRP parts has not been developed. This paper proposes a systematic optimization of the FDM 3D printing process for CFRP. Starting with standard coupon specimen tests to obtain mechanical properties of CFRP, finite element analyses (FEA) were conducted to find principal directions of the AM part and utilized to determine fiber orientations. A specific tool-path algorithm has been developed to distribute fibers with the desired orientations. To predict/assess the mechanical behavior of the AM part, the 3D printing process was simulated to obtain the anisotropic mechanical behavior induced by the customized tool-path printing. Bolt hole plate and spur gear were selected as case studies. FE simulations and associated experiments were conducted to assess their performance. CFRP parts printed by the optimized tool-path shows about 8% higher stiffness than those printed at regular infill patterns. In summary, assisted by FEA, a customized 3D printing tool-path for CFRP has been developed with case studies to verify the proposed AM design optimization methodology for FDM.

PurposeDue to the complexity of and variations in additive manufacturing (AM) processes, there is a level of uncertainty that creates critical issues in quality assurance (QA), which must be addressed by time-consuming and cost-intensive tasks. This deteriorates the process repeatability, reliability and part reproducibility. So far, many AM efforts have been performed in an isolated and scattered way over several decades. In this paper, a systematically integrated holistic view is proposed to achieve QA for AM.Design/methodology/approachA systematically integrated view is presented to ensure the predefined part properties before/during/after the AM process. It consists of four stages, namely, QA plan, prospective validation, concurrent validation and retrospective validation. As a foundation for QA planning, a functional workflow and the required information flows are proposed by using functional design models: Icam DEFinition for Function Modeling.FindingsThe functional design model of the QA plan provides the systematically integrated view that can be the basis for inspection of AM processes for the repeatability and qualification of AM parts for reproducibility.Research limitations/implicationsA powder bed fusion process was used to validate the feasibility of this QA plan. Feasibility was demonstrated under many assumptions; real validation is not included in this study.Social implicationsThis study provides an innovative and transformative methodology that can lead to greater productivity and improved quality of AM parts across industries. Furthermore, the QA guidelines and functional design models provide the foundation for the development of a QA architecture and management system.Originality/valueThis systematically integrated view and the corresponding QA plan can pose fundamental questions to the AM community and initiate new research efforts in the in-situ digital inspection of AM processes and parts.

PurposeTo conduct a head-to-head comparison between deep learning (DL) and radiomics models across institutions for predicting microvascular invasion (MVI) in hepatocellular carcinoma (HCC) and to investigate the model robustness and generalizability through rigorous internal and external validation.MethodsThis retrospective study included 2304 preoperative images of 576 HCC lesions from two centers, with MVI status determined by postoperative histopathology. We developed DL and radiomics models for predicting the presence of MVI using B-mode ultrasound, contrast-enhanced ultrasound (CEUS) at the arterial, portal, and delayed phases, and a combined modality (B + CEUS). For radiomics, we constructed models with enlarged vs. original regions of interest (ROIs). A cross-validation approach was performed by training models on one center’s dataset and validating the other, and vice versa. This allowed assessment of the validity of different ultrasound modalities and the cross-center robustness of the models. The optimal model combined with alpha-fetoprotein (AFP) was also validated. The head-to-head comparison was based on the area under the receiver operating characteristic curve (AUC).ResultsThirteen DL models and 25 radiomics models using different ultrasound modalities were constructed and compared. B + CEUS was the optimal modality for both DL and radiomics models. The DL model achieved AUCs of 0.802–0.818 internally and 0.667–0.688 externally across the two centers, whereas radiomics achieved AUCs of 0.749–0.869 internally and 0.646–0.697 externally. The radiomics models showed overall improvement with enlarged ROIs (P < 0.05 for both CEUS and B + CEUS modalities). The DL models showed good cross-institutional robustness (P > 0.05 for all modalities, 1.6–2.1% differences in AUC for the optimal modality), whereas the radiomics models had relatively limited robustness across the two centers (12% drop-off in AUC for the optimal modality). Adding AFP improved the DL models (P < 0.05 externally) and well maintained the robustness, but did not benefit the radiomics model (P > 0.05).ConclusionCross-institutional validation indicated that DL demonstrated better robustness than radiomics for preoperative MVI prediction in patients with HCC, representing a promising solution to non-standardized ultrasound examination procedures.

COVID-19 is highly infectious and causes acute respiratory disease. Machine learning (ML) and deep learning (DL) models are vital in detecting disease from computerized chest tomography (CT) scans. The DL models outperformed the ML models. For COVID-19 detection from CT scan images, DL models are used as end-to-end models. Thus, the performance of the model is evaluated for the quality of the extracted feature and classification accuracy. There are four contributions included in this work. First, this research is motivated by studying the quality of the extracted feature from the DL by feeding these extracted to an ML model. In other words, we proposed comparing the end-to-end DL model performance against the approach of using DL for feature extraction and ML for the classification of COVID-19 CT scan images. Second, we proposed studying the effect of fusing extracted features from image descriptors, e.g., Scale-Invariant Feature Transform (SIFT), with extracted features from DL models. Third, we proposed a new Convolutional Neural Network (CNN) to be trained from scratch and then compared to the deep transfer learning on the same classification problem. Finally, we studied the performance gap between classic ML models against ensemble learning models. The proposed framework is evaluated using a CT dataset, where the obtained results are evaluated using five different metrics The obtained results revealed that using the proposed CNN model is better than using the well-known DL model for the purpose of feature extraction. Moreover, using a DL model for feature extraction and an ML model for the classification task achieved better results in comparison to using an end-to-end DL model for detecting COVID-19 CT scan images. Of note, the accuracy rate of the former method improved by using ensemble learning models instead of the classic ML models. The proposed method achieved the best accuracy rate of 99.39%.

Electricity load forecasting is a crucial undertaking within all the deregulated markets globally. Among the research challenges on a global scale, the investigation of deep transfer learning (DTL) in the field of electricity load forecasting represents a fundamental effort that can inform artificial intelligence applications in general. In this paper, a comprehensive study is reported regarding day-ahead electricity load forecasting. For this purpose, three sequence-to-sequence (Seq2seq) deep learning (DL) models are used, namely the multilayer perceptron (MLP), the convolutional neural network (CNN) and the ensemble learning model (ELM), which consists of the weighted combination of the outputs of MLP and CNN models. Also, the study focuses on the development of different forecasting strategies based on DTL, emphasizing the way the datasets are trained and fine-tuned for higher forecasting accuracy. In order to implement the forecasting strategies using deep learning models, load datasets from three Greek islands, Rhodes, Lesvos, and Chios, are used. The main purpose is to apply DTL for day-ahead predictions (1–24 h) for each month of the year for the Chios dataset after training and fine-tuning the models using the datasets of the three islands in various combinations. Four DTL strategies are illustrated. In the first strategy (DTL Case 1), each of the three DL models is trained using only the Lesvos dataset, while fine-tuning is performed on the dataset of Chios island, in order to create day-ahead predictions for the Chios load. In the second strategy (DTL Case 2), data from both Lesvos and Rhodes concurrently are used for the DL model training period, and fine-tuning is performed on the data from Chios. The third DTL strategy (DTL Case 3) involves the training of the DL models using the Lesvos dataset, and the testing period is performed directly on the Chios dataset without fine-tuning. The fourth strategy is a multi-task deep learning (MTDL) approach, which has been extensively studied in recent years. In MTDL, the three DL models are trained simultaneously on all three datasets and the final predictions are made on the unknown part of the dataset of Chios. The results obtained demonstrate that DTL can be applied with high efficiency for day-ahead load forecasting. Specifically, DTL Case 1 and 2 outperformed MTDL in terms of load prediction accuracy. Regarding the DL models, all three exhibit very high prediction accuracy, especially in the two cases with fine-tuning. The ELM excels compared to the single models. More specifically, for conducting day-ahead predictions, it is concluded that the MLP model presents the best monthly forecasts with MAPE values of 6.24% and 6.01% for the first two cases, the CNN model presents the best monthly forecasts with MAPE values of 5.57% and 5.60%, respectively, and the ELM model achieves the best monthly forecasts with MAPE values of 5.29% and 5.31%, respectively, indicating the very high accuracy it can achieve.

Two types of additive manufacturing (AM) techniques, fused filament fabrication (FFF) and selective laser sintering (SLS), were employed to fabricate specimens from three widely used AM polymers: polylactic acid (PLA), acrylonitrile butadiene styrene (ABS), and polyamide (PA): the first two using FFF and PA using SLS. The single‐edge notch bending (SENB) test was employed to measure the mode‐I fracture toughnesses of bulk and AM specimens in three directions relative to the printing direction. Unlike their bulk counterparts, AM polymers exhibited anisotropic fracture toughness. The degree of anisotropy in FFF‐PLA was much larger than FFF‐ABS and SLS‐PA. The smallest fracture toughness value for each AM polymer was achieved along the transverse printing direction. Different AM techniques affect fracture toughness differently. For instance, FFF is shown to increase PLA's plastic deformation, while SLS reduces PA's ductility. The decrease in fracture toughness is consistently observed when transitioning from bulk polymers to AM ones, reaching as much as 75%. This study provides insights into the fracture toughness of AM parts made via popular plastics, which are crucial for designing structural parts using FFF and SLS processes.

Much of the potential of industrial additive manufacturing (AM) is said to lie in the digital specification of components that can be transmitted seamlessly and unambiguously to partners fostering flexible outsourcing. In industry, we observe nuanced AM supply chain governance structures that result from make‐or‐buy decisions, with a tendency to implement AM in‐house. Thus, there is a discrepancy between what is discussed in the literature and implemented in practice. We apply a multiple‐case study approach to investigate why and how AM impacts the make‐or‐buy decision of manufacturing firms. We identify four decision profiles demonstrating the spectrum of specific governance structures and develop a framework to explain the underlying rationales. We find strong arguments for in‐house AM including firms’ perceived need to protect their digitally encapsulated intellectual property, reevaluation of their core competencies, commitment to internal learning, and senior management's enthusiasm for AM. By using transaction cost economics and the resource‐based view, we contribute to the understanding of how arguments of these general theories are modified by the digital and emerging traits of AM. We reveal contradicting guidance in the theories’ argumentation for the case of AM and provide managers a clear perspective on alternative strategies for their AM implementation process.

We developed a multimodal ultrasound (US) deep learning (DL) fusion model to automatically classify early fibrosis in patients with chronic kidney disease (CKD). This prospective study included patients with CKD who underwent continuous gray-scale US, superb microvascular imaging, and strain elastography from May to November 2022. According to the pathological tubular atrophy and interstitial fibrosis score, patients were divided into minimal and mild groups (affected area ≤10% and 11 − 25% of the total cortical volume, respectively). The dataset was divided into training (70%) and test (30%) sets. A DL model combining the features of the three US modes was developed to predict early fibrosis in patients with CKD. We compared these findings with the area under the receiver operating characteristic curve (AUC) of the clinical model by analyzing the receiver operating characteristic curve in the test set. The AUC of single-mode DL based on gray-scale US, superb microvascular imaging, and strain elastography was 0.682, 0.745, and 0.648, respectively, while that of the multimodal US DL model was 0.86. The accuracy, specificity, and sensitivity of the multimodal US DL model were 0.779, 0.767, and 0.796, respectively, and the negative and positive predictive values were 0.842 and 0.706, respectively. The AUC of the multimodal US DL model was significantly better than that of the single-mode DL and clinical models. The DL algorithm developed using multimodal US images can effectively predict early fibrosis in patients with CKD with significantly greater accuracy than single-mode DL or clinical models.

We aimed at comparing intratumoral and peritumoral deep learning, radiomics, and fusion models in predicting KRAS mutations in rectal cancer using endorectal ultrasound imaging. This study included 304 patients with rectal cancer from Fujian Medical University Union Hospital. The patients were randomly divided into a training group (213 patients) and a test group (91 patients) at a 7:3 ratio. Radiomics and deep learning models were established using primary tumor and peritumoral images. In the optimally performing regions-of-interest, two fusion strategies, a feature-based and a decision-based model, were employed to build the fusion models. The Shapley additive explanation (SHAP) method was used to evaluate the significance of features in the optimal radiomics, deep learning, and fusion models. The performance of each model was assessed using the area under the receiver operating characteristic curve (AUC) and decision curve analysis (DCA). In the test cohort, both the radiomics and deep learning models exhibited optimal performance with a 10-pixel patch extension, yielding AUC values of 0.824 and 0.856, respectively. The feature-based DLRexpand10_FB model attained the highest AUC (0.896) across all study sets. In addition, the DLRexpand10_FB model demonstrated excellent sensitivity, specificity, and DCA. SHAP analysis underscored the deep learning feature (DL_1) as the most significant factor in the hybrid model. The feature-based fusion model DLRexpand10_FB can be employed to predict KRAS gene mutations based on pretreatment endorectal ultrasound images of rectal cancer. The integration of peritumoral regions enhanced the predictive performance of both the radiomics and deep learning models.

Deep learning breast cancer risk models demonstrate improved accuracy compared with traditional risk models but have not been prospectively tested. We compared the accuracy of a deep learning risk score derived from the patient's prior mammogram to traditional risk scores to prospectively identify patients with cancer in a cohort due for screening. We collected data on 119 139 bilateral screening mammograms in 57 617 consecutive patients screened at 5 facilities between September 18, 2017, and February 1, 2021. Patient demographics were retrieved from electronic medical records, cancer outcomes determined through regional tumor registry linkage, and comparisons made across risk models using Wilcoxon and Pearson χ2 2-sided tests. Deep learning, Tyrer-Cuzick, and National Cancer Institute Breast Cancer Risk Assessment Tool (NCI BCRAT) risk models were compared with respect to performance metrics and area under the receiver operating characteristic curves. Cancers detected per thousand patients screened were higher in patients at increased risk by the deep learning model (8.6, 95% confidence interval [CI] = 7.9 to 9.4) compared with Tyrer-Cuzick (4.4, 95% CI = 3.9 to 4.9) and NCI BCRAT (3.8, 95% CI = 3.3 to 4.3) models (P &lt; .001). Area under the receiver operating characteristic curves of the deep learning model (0.68, 95% CI = 0.66 to 0.70) was higher compared with Tyrer-Cuzick (0.57, 95% CI = 0.54 to 0.60) and NCI BCRAT (0.57, 95% CI = 0.54 to 0.60) models. Simulated screening of the top 50th percentile risk by the deep learning model captured statistically significantly more patients with cancer compared with Tyrer-Cuzick and NCI BCRAT models (P &lt; .001). A deep learning model to assess breast cancer risk can support feasible and effective risk-based screening and is superior to traditional models to identify patients destined to develop cancer in large screening cohorts.

Goal: This research identifies the optimal location for additive manufacturing (AM) facilities within the spare parts supply chain, addressing a critical yet underexplored dimension in the literature on strategic facility location for AM technology deployment.Design / Methodology / Approach: The study integrates agent-based discrete event simulation with scenario analysis to evaluate four facility locations for four distinct spare parts. The model aims to reduce order fulfillment losses and opportunity costs by supporting evidence-based location decisions.Results: The findings indicate that the Thane location is the most suitable for establishing an AM facility, offering a balanced trade-off between supply reliability and economic efficiency.Limitations of the investigation: The scope is confined to a specific case involving four locations and four spare parts, which may limit generalizability. The study’s scenario-based approach also relies on assumptions that warrant further validation.Practical implications: The results offer actionable insights for firms adopting AM in spare parts management. Strategic facility placement enhances service responsiveness, reduces downtime, and supports cost-effective ful-fillment strategies.Originality / Value: This research provides an incremental empirical contribution by proposing and demonstrat-ing a simulation-based AM facility location selection framework, supported by real-world case data. It addresses a less explored strategic decision of AM facility placement, considering dynamic operational uncertainties, thereby extending the existing literature on AM supply chain design.

Additive manufacturing (AM) processes, such as Fused Filament Fabrication (FFF), enable the production of lightweight parts with high stiffness-to-weight ratios, making them highly suitable for a wide range of engineering applications. However, ensuring the mechanical reliability of these components, particularly for load-bearing purposes, requires systematic mechanical testing of well-designed specimens to asses their suitability. While the tensile properties of additively manufactured materials have been extensively studied, the compressive behavior of components produced via AM, particularly those made from thermoplastic materials, remains comparatively underexplored and insufficiently characterized in the existing body of research. Among these materials, polylactic acid (PLA)—a biodegradable thermoplastic derived from renewable resources—has gained prominence in AM applications. Recent studies have investigated the compression properties of PLA in reinforced materials; however, the focus has primarily been on solid, semi-solid, or porous specimens. These investigations largely overlook thin-walled structures, which are integral to weight-saving designs and commonly feature in topology-optimized structures. Understanding the mechanical behavior of monolayers, the fundamental building blocks of most AM components, is essential for accurately predicting the overall performance of multilayer structures. Monolayers represent the smallest, most basic structural elements of AM parts, and their properties directly influence the behavior of the final, more complex assemblies. Establishing a methodology that correlates monolayer properties with those of multilayer components could significantly streamline testing procedures. By performing mechanical tests on monolayers, instead of on more intricate multilayer specimens, manufacturers could reduce testing complexity and cost while accelerating the development process. The current literature reveals a gap in the design and analysis of thin-walled AM specimens, especially monolayers, under compressive loads. Specifically, the design of monolayer or thin-walled AM compression specimens without infill has not been thoroughly explored. This article addresses this gap by investigating the design and testing of AM monolayer compression specimens produced using FFF of PLA. Three distinct specimen geometries are considered—circular, helicoidal, and S-shaped—to evaluate their potential for understanding and predicting the compressive behavior of AM monolayer structures.

The past few years have seen a surge of applying Deep Learning (DL) models for a wide array of tasks such as image classification, object detection, machine translation, etc. While DL models provide an opportunity to solve otherwise intractable tasks, their adoption relies on them being optimized to meet latency and resource requirements. Benchmarking is a key step in this process but has been hampered in part due to the lack of representative and up-to-date benchmarking suites. This is exacerbated by the fast-evolving pace of DL models. This paper proposes DLBricks, a composable benchmark generation design that reduces the effort of developing, maintaining, and running DL benchmarks on CPUs. DLBricks decomposes DL models into a set of unique runnable networks and constructs the original model's performance using the performance of the generated benchmarks. DLBricks leverages two key observations: DL layers are the performance building blocks of DL models and layers are extensively repeated within and across DL models. Since benchmarks are generated automatically and the benchmarking time is minimized, DLBricks can keep up-to-date with the latest proposed models, relieving the pressure of selecting representative DL models. Moreover, DLBricks allows users to represent proprietary models within benchmark suites. We evaluate DLBricks using $50$ MXNet models spanning $5$ DL tasks on $4$ representative CPU systems. We show that DLBricks provides an accurate performance estimate for the DL models and reduces the benchmarking time across systems (e.g. within $95\%$ accuracy and up to $4.4\times$ benchmarking time speedup on Amazon EC2 c5.xlarge).