The uncertainty and lack of repeatability of the additive manufacturing (AM) processes have brought significant issues hindering the wider industrial adoption of these novel fabrication techniques. The process-induced defects have compromised the structural integrity of the specimens where the mechanical properties have failed to reach the required industrial qualifications. The uncertainty within the AM processes has caused issues such as variation in part quality. Therefore, it is crucial to develop a robust model to address the issue of uncertainty. Our objective in this study is to pave the way toward a better understanding of the uncertainty in the process-defect-structures relationship using an inverse robust design exploration method. The method involves two steps. In the first step, mathematical models are developed to characterize and model the forward flow of information in the intended additive manufacturing process. In the second step, inverse robust design exploration is carried out to investigate satisfying design solutions that meet multiple AM goals. We test the utility of the method for an AM problem; in the first step, a two-phase methodology is developed to predict the fatigue life from initial process parameters through intermediate process-induced defect properties such as maximum defect size. In the second step, we carry out robust inverse design exploration to maximize the predefined fatigue-related goals while managing the uncertainty in the design variables. The model is implemented for fatigue experiments of Ti-6AL-4 V material, and the results of satisfying design solutions are reported.

Additive manufacturing (AM) is a customizable layer-wise production process challenged by inherent process complexity, often leading to structural defects such as geometric deformation and distortion. A common cause of structural defects is uneven material expansion and contraction during the rapid heating and cooling phases. Various data types can monitor the process for defects, but many AM processes lack a unified in-process data collection and synchronization. Moreover, there is a gap in developing methods to utilize multimodal data to learn from previous experiments to save on the cost of resources. This work develops a thermal physics-informed PointNet methodology that addresses the lack of in situ data-driven distortion characterization models constrained by physical function by leveraging multimodal position, weld parameters, and thermal data. First, this research utilizes a digital environment developed in robot operating system 2 to merge and synchronize multimodal sensor data-including infrared thermal images, system joint states, and weld parameters-to collect in situ thermal point cloud data. Second, using a PointNet deep learning model accounts for various data types and sensor sources in point cloud data format and maps the relationship to distortion characterization. Finally, a physics-informed model training loss function is created with thermal expansion and contraction constraints using thermal expansion coefficients. The integration of these processes marks a significant step toward harnessing the power of physics-informed machine learning to better understand and predict distortions in AM, paving the way for more reliable and accurate manufacturing outcomes.

In collaborative additive manufacturing (AM), sharing process data across multiple users can provide small- to medium-sized manufacturers (SMMs) with enlarged training data for part certification, facilitating accelerated adoption of metal-based AM technologies. The aggregated data can be used to develop a process-defect model that is more precise, reliable, and adaptable. However, the AM process data often contains printing path trajectory information that can significantly jeopardize intellectual property (IP) protection when shared among different users. In this study, a new adaptive AM data de-identification method is proposed that aims to mask the printing trajectory information in the AM process data in the form of melt pool images. This approach integrates stochastic image augmentation (SIA) and adaptive surrogate image generation (ASIG) via tracking melt pool geometric changes to achieve a trade-off between AM process data privacy and utility. As a result, surrogate melt pool images are generated with perturbed printing directions. In addition, a convolutional neural network (CNN) classifier is used to evaluate the proposed method regarding privacy gain (i.e., changes in the accuracy of identifying printing orientations) and utility loss (i.e., changes in the ability to detect process anomalies). The proposed method is validated using data collected from two cylindrical specimens using the directed energy deposition (DED) process. The case study results show that the de-identified dataset significantly improved privacy preservation while sacrificing little data utility, once shared on the cloud-based AM system for collaborative process-defect modeling.

This work presents a physics-informed fusion methodology for deformation detection using multi-sensor thermal data. A challenge with additive manufacturing (AM) is that abnormalities commonly occur due to rapid changes in the thermal gradient. Different non-destructive in-situ thermal sensors capture parts of the thermal history but are limited by the visible temperature spectrum and sensor field of view of the fabrication process. Various sensors mitigate problems with the loss of thermal history information; however, it brings forth challenges with integrating different data streams and the need to interpolate the internal thermal history. This study develops a thermal data-informed heat flux methodology that fills the gap in fusing numerical temperature approximation with data-driven knowledge of the surface of additive manufactured components. First, this study fuses infrared (IR) thermal data complexities during the AM process with the Goldak double ellipsoidal heat flux to model the energy input into the component. Second, a thermal physics-informed model input (PIMI) is created with thermal data-informed heat flux to capture internal thermal history. Lastly, a regression convolutional neural network (CNN) captures the relationship between the three-dimensional thermal gradient and the resulting surface deformation. The rapid thermal gradient formation and identification of deformation is a key step toward using thermal history data and machine learning to improve quality control in AM. The proposed surface deformation detection model achieved an mean squared error of 1.14 mm and an R-2 of 0.89 in the case study when fabricating thin-walled structures.

Background This study aimed to demonstrate both the potential and development progress in the identification of extracapsular nodal extension in head and neck cancer patients prior to surgery. Methodology A deep learning model has been developed utilizing multilayer gradient mapping-guided explainable network architecture involving a volume extractor. In addition, the gradient-weighted class activation mapping approach has been appropriated to generate a heatmap of anatomic regions indicating why the algorithm predicted extension or not. Results The prediction model shows excellent performance on the testing dataset with high values of accuracy, the area under the curve, sensitivity, and specificity of 0.926, 0.945, 0.924, and 0.930, respectively. The heatmap results show potential usefulness for some select patients but indicate the need for further training as the results may be misleading for other patients. Conclusions This work demonstrates continued progress in the identification of extracapsular nodal extension in diagnostic computed tomography prior to surgery. Continued progress stands to see the obvious potential realized where not only can unnecessary multimodality therapy be avoided but necessary therapy can be guided on a patient-specific level with information that currently is not available until postoperative pathology is complete.

Defect prevention and detection are very crucial for the quality improvement of additive manufacturing (AM) processes. Timely identification of imperfections and flaws in the manufactured products will allow effective and early implementation of corrective actions and thus impede the spread of defects to the whole industrial value chain. In laser metal deposition (LMD) AM processes, defects are imperfections which typically include "porosity" and "cracks formation". This study presents novel data-driven anomaly detection techniques that use both transfer learning and online learning to assess the quality of melt pool images taken during the LMD process of a part. The proposed methods incorporate the characteristics and dynamics of the manufactured part during the printing process along with previous knowledge acquired from historical data. Continual online learning models (K-means and self-organizing maps) are developed whose parameters are adapted to the incoming data collected in real-time as a new part is being manufactured. The proposed models significantly outperformed the performance of their batch-learning counterparts in detecting anomalous melt pool images in an additively manufactured Ti-6Al-4V thin-walled part. Both models required an average of similar to 0.07 s to process each incoming melt pool image, update their parameters and give a prediction on the image's health. This shows the potential of continual online learning for real-time anomaly detection.

We developed a deep fusion methodology of nondestructive in-situ thermal and ex-situ ultrasonic images for porosity detection in laser-based additive manufacturing (LBAM). A core challenge with the LBAM is the lack of fusion between successive layers of printed metal. Ultrasonic imaging can capture structural abnormalities by passing waves through successive layers. Alternatively, in-situ thermal images track the thermal history during fabrication. The proposed sensor fusion U-Net methodology fills the gap in fusing in-situ images with ex-situ images by employing a two-branch convolutional neural network (CNN) for feature extraction and segmentation to produce a 2D image of porosity. We modify the U-Net framework with the inception and long short term memory (LSTM) blocks. We validate the models by comparing our single modality models and fusion models with ground truth X-ray computed tomography (XCT) images. The inception U-Net fusion model achieved the highest mean intersection over union score of 0.93.

There is an urgent need for developing collaborative process-defect modeling in metalbased additive manufacturing (AM). This mainly stems from the high volume of training data needed to develop reliable machine learning models for in-situ anomaly detection. The requirements for large data are especially challenging for small-to-medium manufacturers (SMMs), for whom collecting copious amounts of data is usually cost prohibitive. The objective of this research is to develop a secured data sharing mechanism for directed energy deposition (DED) based AM without disclosing product design information, facilitating secured data aggregation for collaborative modeling. However, one major obstacle is the privacy concerns that arise from data sharing, since AM process data contain confidential design information, such as the printing path. The proposed adaptive design deidentification for additive manufacturing (ADDAM) methodology integrates AM process knowledge into an adaptive de-identification procedure to mask the printing trajectory information in metal-based AM thermal history, which otherwise discloses substantial printing path information. This adaptive approach applies a flexible data privacy level to each thermal image based on its similarity with the other images, facilitating better data utility preservation while protecting data privacy. A real-world case study was used to validate the proposed method based on the fabrication of two cylindrical parts using a DED process. These results are expressed as a Pareto optimal solution, demonstrating significant improvements in privacy gain and minimal utility loss. The proposed method can facilitate privacy improvements of up to 30% with as little as 0% losses in dataset utility after de-identification. [DOI: 10.1115/1.4056488]

BACKGROUND Diagnosis and treatment management for head and neck squamous cell carcinoma (HNSCC) is guided by routine diagnostic head and neck computed tomography (CT) scans to identify tumor and lymph node features. The extracapsular extension (ECE) is a strong predictor of patients' survival outcomes with HNSCC. It is essential to detect the occurrence of ECE as it changes staging and treatment planning for patients. Current clinical ECE detection relies on visual identification and pathologic confirmation conducted by clinicians. However, manual annotation of the lymph node region is a required data preprocessing step in most of the current machine learning-based ECE diagnosis studies. PURPOSE In this paper, we propose a Gradient Mapping Guided Explainable Network (GMGENet) framework to perform ECE identification automatically without requiring annotated lymph node region information. METHODS The gradient-weighted class activation mapping (Grad-CAM) technique is applied to guide the deep learning algorithm to focus on the regions that are highly related to ECE. The proposed framework includes an extractor and a classifier. In a joint training process, informative volumes of interest (VOIs) are extracted by the extractor without labeled lymph node region information, and the classifier learns the pattern to classify the extracted VOIs into ECE positive and negative. RESULTS In evaluation, the proposed methods are well-trained and tested using cross-validation. GMGENet achieved test accuracy and area under the curve (AUC) of 92.2% and 89.3%, respectively. GMGENetV2 achieved 90.3% accuracy and 91.7% AUC in the test. The results were compared with different existing models and further confirmed and explained by generating ECE probability heatmaps via a Grad-CAM technique. The presence or absence of ECE has been analyzed and correlated with ground truth histopathological findings. CONCLUSIONS The proposed deep network can learn meaningful patterns to identify ECE without providing lymph node contours. The introduced ECE heatmaps will contribute to the clinical implementations of the proposed model and reveal unknown features to radiologists. The outcome of this study is expected to promote the implementation of explainable artificial intelligence-assiste ECE detection.

Manufacturing-as-a-Service (MaaS) can accelerate additive manufacturing (AM) process-defect modeling by augmenting training data to all collaborating users via a data sharing network. However, sharing process data may disclose product design information. This paper aims to evaluate design information disclosure of various thermal history-based feature extraction methods for metal-based AM anomaly detection. This is accomplished by evaluating the design information (i.e., printing orientation) retained, and the overall data usability (i.e., anomaly detection) preserved in the extracted features for various state-of-the-art feature extraction methods. The evaluation results indicate that there are urgent needs in privacy preserving data sharing for additive MaaS (AMaaS).

Manufacturability of topology optimized models is a key element when it comes to the coupling of topology optimization (TO) and additive manufacturing. However, most topology optimization techniques generate models incorporating geometric inconsistencies that require additional post-processing. In this article, Taubin smoothing method is applied to an optimized Messerschmitt–Bölkow–Blohm beam using Solid Isotropic Microstructure with Penalization method of stainless steel 316L and produced using the selective laser melting (SLM) process. Furthermore, a performance-based analysis is realized to measure the mesh quality, using a set of quality metrics of the smooth mesh. Taubin method demonstrated a high capacity in preserving the overall volume while smoothing voxel elements engendering greater formability of the structure. The implemented conditioning of Taubin smoothing alongside SLM’s printing parameters produced high-quality surfaces with reasonable roughness. Numerical and experimental three-point bending tests are set to investigate both the stiffness performance and surface quality of the designed parts. The obtained results showed that the smooth manufactured parts are less stiff than the original TO model. Potential contributors are discussed, including the formation of an anisotropic microstructure of stainless steel 316L.

Data-driven porosity prediction in Laser Metal Deposition (LMD) is mainly done with supervised machine learning methods. These methods require labeled thermal signatures for model training, with the "labels" being post-process evaluations of porosity. In practice, acquiring porosity records for newly printed parts is expensive and time-consuming; matching thermal signatures from the printing process with the porosity records is subject to data registration errors. To enable convenient porosity prediction for new part geometry, this study proposes a "knowledge transfer" method to transfer prior statistical knowledge about printed parts to new printing processes. The prior knowledge is leveraged to evaluate the statistical property of new thermal signatures and assign them labels. Supervised machine learning methods can be readily trained with labeled data. The effort for postprocess porosity inspection is therefore saved, and the efficiency of data-driven porosity prediction is significantly improved. The proposed method is validated with datasets from an LMD machine, specifically an OPTOMEC LENS 750 system. The statistical inference knowledge about a Ti-6Al-4V thin wall is transferred to two different Ti-6Al-4V cylinders, respectively, to label their thermal signatures and train Convolutional Neural Networks (CNNs) for in-situ porosity prediction. The case study results demonstrate the effectiveness of the proposed method.

Society 5.0 refers to an advanced society based on big data, artificial intelligence, sensors, and robots to improve many aspects of life in a smart city. The role of sensors in Society 5.0 is critical. Sensors and the Internet of Things can be considered to work as a service system. Specifically, sensors can track millions of objects to support city security. This article considers the competitive sensor networks in terms of Quality of Service, which can be quantified by service price. The sensors are modeled as an $M/M/1/n$ queueing service system with a location on a secured grid area. A competitive admission fee is considered for tracking orders from society, which makes the arriving tracking demands price-sensitive. Moreover, a bi-level nonlinear program is developed wherein the first level maximizes the sensors’ revenue and social benefit revenue while also minimizing the wait time of tracking orders from society; in the second level, the expected damage cost is minimized from a disruptive scenario toward valuable infrastructure. Moreover, the second-level model is linearized, and it solves the problem by a branch and bound and enumeration algorithm. Finally, an illustrative example of the proposed model is presented.

Multi-channel sensor fusion can be challenging for real-time machinery fault identification and diagnosis when a substantial amount of missing data exists. Usually, some (or even all) sensors may not function correctly during real-time data acquisition due to sensor malfunction or transmission issues. Additionally, multi-channel sensor fusion yields a large volume of data. Imputation of missing entries can also be challenging with a large volume of data, which can predominantly affect the accuracy of machinery fault diagnosis. However, how to impute a substantial amount of missing data for machinery fault identification is an open research question. In light of the above challenges, this paper proposes constructing time-domain tensors based on heterogeneous sensor signals. Subsequently, the fully Bayesian CANDECOMP/PARAFAC (FBCP) factorization method is adopted for missing data imputation of diverse bearing faults signals. To validate the effectiveness of this proposed method, a machinery fault simulator was used to collect diverse bearing fault signals by incorporating both acoustics and vibration sensors. A varying percentage of continuous missing signal scenarios are introduced at the random locations among different acoustics and vibration channels to construct incomplete tensors. Subsequently, the FBCP method was leveraged to complete the incomplete tensors and calculate estimated tensors. To evaluate the performance of continuous missing data imputation, relative standard errors are computed based on the estimated and actual time-domain tensors. Experimental results show that this proposed method can effectively impute a substantial portion of continuous missing data from diverse bearing fault scenarios.

In-process thermal melt pool images and post-fabrication porosity labels are acquired for Ti-6Al-4V thin-walled structure fabricated with OPTOMEC Laser Engineered Net Shaping (LENS™) 750 system. The data is collected for nondestructive thermal characterization of direct laser deposition (DLD) build. More specifically, a Stratonics dual-wavelength pyrometer captures a top-down view of the melt pool of the deposition heat-affected zone (HAZ), which is above 1000∘C, and Nikon X-Ray Computed Tomography (XCT) XT H225 captures internal porosity reflective of lack of fusion during the fabrication process. The pyrometer images provided in Comma Separated Values (CSV) format are cropped to center the melt pool to temperatures above 1000℃, indicative of the shape and distribution of temperature values. Melt pool coordinates are determined using pyrometer specifications and thin wall build parameters. XCT porosity labels of sizes between 0.05 mm to 1.00 mm are registered within 0.5 mm of the melt pool image coordinate. An XCT porosity-labeled table provided in the Excel spreadsheet format contains time stamps, melt pool coordinates, melt pool eccentricity, peak temperature, peak temperature coordinates, pore size, and pore label. Thermal-porosity data utilization aids in generating data-driven quality control models for manufacturing parts anomaly detection.

This study aims to develop an intelligent, rapid porosity prediction methodology for additive manufacturing (AM) processes under varying process conditions by leveraging knowledge transfer from the existing process conditions. Conventional machine learning (ML) algorithms are extensively used in porosity prediction for AM processes. These approaches assume that the underline distribution of the source (training) and target (testing) is the same and that target labels are available for modeling purposes. However, the source and target sometimes follow different distributions in real-world manufacturing environments as the diversity of industrialization processes leads to heterogeneous data collection under different production conditions. This will reduce the ability of decision-making with conventional approaches. Transfer learning (TL) is one of the robust techniques that enables transferring learned knowledge between the target and source to establish a robust relationship while the target has fewer data. Therefore, this paper presents an unsupervised grouping-based transfer learning method to characterize the relationship between an unknown target and sources. The similarities between sources and targets are learned by forming a new mixed domain, which organizes data into identity groups. Then, a group-based learning process is designated to transfer knowledge to make target predictions. The effectiveness of the proposed method is evaluated by predicting porosity based on thermal images collected from the AM process under different process conditions, i.e., single-source and multi-source transfer to target porosity prediction. The performance comparison demonstrates that the in situ porosity prediction using the proposed method outperformed state-of-art classification models support vector machine (SVM), convolutional neural network (CNN), and different TL methods such as TL with NNs (TLNN), and TL with CNNs (TLCNN).

This work summarizes the state-of-the-art data-driven methods for prediction of the Remaining Useful Life (RUL). It discusses challenges and open problems faced in PdM. This study presents a discussion on the new problems that need to be considered towards the Industry 4.0 goals. We propose the future direction for each challenge discussed in this article.

Laser metal deposition (LMD) is an additive manufacturing method for metal parts by using focused thermal energy to fuse materials as they are deposited. During LMD, transient thermal signatures such as the in-situ thermal images of melt pool, contain rich information about process performance. Early prediction of such transient thermal signatures provides opportunities for process monitoring and defect prevention. While physics-based models of LMD have been conventionally used for thermal signature prediction, they have limitations and are computationally expensive for real-time prediction. A scalable, efficient data-science-based model is therefore needed. This paper develops a deep-learning-based surrogate model, called LMD-cGAN, to predict and emulate the transient thermal signatures in LMD. The model generates images for the thermal dynamics of melt pool conditionally on the deposition layer. It enables early prediction of future-layer thermal signatures for an in-process part based on its early-layer thermal signatures. To respect the physics in LMD, a physics-guided image selection (PGIS) mechanism is integrated with LMD-cGAN to calibrate the predictions against physical benchmarks of transient melt pool for the process. The effectiveness and efficiency of the proposed method are demonstrated in a case study on the LMD of Ti-4Al-6V thin-walled structures. Note to Practitioners—With online sensing, many LMD applications have real-time process data that convey valuable information about the process status and part quality. The proposed method leverages these data for thermal signature prediction. LMD-cGAN is a deep-learning-based surrogate model that learns the population profile of real thermal signatures and generates thermal signatures from there. The proposed PGIS mechanism in LMD-cGAN ensures the physical validity of these predictions by benchmarking them against physical insights about the process. LMD-cGAN can be applied to predict thermal signatures in future layers based on early-layer thermal signatures of an in-process part (an implicit assumption here is that the in-process part to be predicted for is the same type). LMD-cGAN can also be applied to emulate thermal signatures in specific layers. To generate thermal signatures for generic, non-defect parts, the training data should be selected with caution – the part where the data were collected should have no obvious defects, so the thermal signatures generated by LMD-cGAN show the regular thermal dynamics. Compared with pure physical models, the proposed method incorporates process uncertainties captured from the early-layer data, hence “on-the-fly” emulation of the melt pool, while characterizing the inherent relationship between the LMD process and thermal signatures.

Background: Universities are at risk for COVID-19 and Fall semester begins in August 2020 for most campuses in the United States. The Southern States, including Mississippi, are experiencing a high incidence of COVID-19. Aims: The objective of this study is to model the impact of face masks and hybrid learning on the COVID-19 epidemic on Mississippi State University (MSU) campus. Methods: We used an age structured deterministic mathematical model of COVID-19 transmission within the MSU campus population, accounting for asymptomatic transmission. We modeled facemasks for the campus population at varying proportions of mask use and effectiveness, and Hyflex model of partial online learning with reduction of people on campus. Results: Facemasks can substantially reduce cases and deaths, even with modest effectiveness. Even 20% uptake of masks will halve the epidemic size. Facemasks combined with Hyflex reduces epidemic size even more. Conclusions: Universal use of face masks and reducing the number of people on campus may allow safer universities reopening.

The process uncertainty induced quality issue remains the major challenge that hinders the wider adoption of additive manufacturing (AM) technology. The defects occurred significantly compromise structural integrity and mechanical properties of fabricated parts. Therefore, there is an urgent need in fast, yet reliable AM component certification. Most finite element analysis related methods characterize defects based on the thermomechanical relationships, which are computationally inefficient and cannot capture process uncertainty. In addition, there is a growing trend in data-driven approaches on characterizing the empirical relationships between thermal history and anomaly occurrences, which focus on modeling an individual image basis to identify local defects. Despite their effectiveness in local anomaly detection, these methods are quite cumbersome when applied to layer-wise anomaly detection. This paper proposes a novel in situ layer-wise anomaly detection method by analyzing the layer-by-layer morphological dynamics of melt pools and heat affected zones (HAZs). Specifically, the thermal images are first preprocessed based on the g-code to assure unified orientation. Subsequently, the melt pool and HAZ are segmented, and the global and morphological transition metrics are developed to characterize the morphological dynamics. New layer-wise features are extracted, and supervised machine learning methods are applied for layer-wise anomaly detection. The proposed method is validated using the directed energy deposition (DED) process, which demonstrates superior performance comparing with the benchmark methods. The average computational time is significantly shorter than the average build time, enabling in situ layer-wise certification and real-time process control.