Additive Friction Stir Deposition (AFSD) is an emerging metal additive manufacturing process gaining increased attention due to its solid-state processing nature, resulting in desirable mechanical properties and the capability of fabricating large parts. To understand the multi-physics nature of the process, modeling of thermal distribution and material velocity during processing is imperative, which can be addressed by solving the related partial differential equations (PDEs). However, complex coupled PDEs entail inaccuracies and slowness in utilizing the existing numerical methods, such as the Finite Volume Method (FVM). This work proposes a neural network-based method, coupled Physics Informed Neural Network (coPINN), as a fast and accurate mesh-free solution for predicting the thermomechanical states during the AFSD process. In the case studies, the proposed coPINN accurately predicts the velocity and temperature distribution for a steady-state non-linear coupled fluid dynamics-heat transfer model of the AFSD process faster than the existing FVM. Comprehensive comparisons with residual errors show the proposed method’s trustworthiness in handling coupled physics, non-linearity, and mixed boundary conditions. In addition, the solutions of our approach show better conformance with the actual experimental data than the FVM-based solver. Overall, this work shows more efficient neural network-based solutions of a PDEs-based model than a traditional numerical method.

This paper proposes a Geometry-Aware Physics Informed PointNet (GeoPIPN), a deep learning-based fast prediction model for solving Partial Differential Equations (PDE) acrossunseenadditive manufacturing (AM) geometries. Thermal distributions in AM parts are crucial and governed by diffusion models in the form of PDEs, which dictate their spatial and temporal behavior. Proper thermal distribution control ensures desired material properties, uniform microstructure, and dimensional accuracy – key to AM part quality and reliability. Existing thermal modeling primarily utilizes numerical solvers, namely, the Finite Element Methods (FEM), which are computationally costly for the fast-paced, iterative geometric design phase. Furthermore, these methods are tailored to a specific part geometry, requiring rerunning for each new geometry, making them inefficient for designs that change frequently. Our GeoPIPN considers various part geometries, represented as point clouds, as input, and is trained in a physics-informed setting where the learning objective combines PDE residual loss, boundary condition loss, and measurement-based loss from sparse sensor locations. During the training, both geometry-aware global and boundary features, as well as point-wise local features, are learned. The resulting GeoPIPN can predict the thermal distribution for a new AM part geometry at inference time without rerunning the FEM simulations. In the case studies, we demonstrate the superior performance of our GeoPIPN over benchmark methods on 2D AM test problems. Consequently, GeoPIPN offers an efficient solution for fast and accurate AM design iteration in 2D settings and provides a foundation for future extensions to fully 3D AM problems.

Sensor technology developments provide a basis for effective fault diagnosis in manufacturing systems. However, the limited number of sensors due to physical constraints or undue costs hinders the accurate diagnosis in the actual process. In addition, time-varying operational conditions that generate nonstationary process faults and the correlation information in the process require to consider for accurate fault diagnosis in the manufacturing systems. This article proposes a novel fault diagnosis method: clustering spatially correlated sparse Bayesian learning (CSSBL), and explicitly demonstrates its applicability in a multistation assembly system that is vulnerable to the above challenges. Specifically, the method is based on a practical assumption that it will likely have a few process faults (sparse). In addition, the hierarchical structure of CSSBL has several parameterized prior distributions to address the above challenges. As posterior distributions of process faults do not have closed form, this paper derives approximate posterior distributions through Variational Bayes inference. The proposed method’s efficacy is provided through numerical and real-world case studies utilizing an actual autobody assembly system. The generalizability of the proposed method allows the technique to be applied in fault diagnosis in other domains, including communication and healthcare systems. Note to Practitioners—This article proposes a new process fault diagnosis method: clustering spatially correlated sparse Bayesian learning. This method effectively diagnoses time-varying defects by leveraging the correlation structures in the process when sensor measurements are insufficient. The actual autobody assembly process is utilized to show the proposed method’s effectiveness. The proposed method performs superior to the benchmark methods in fault detection capability. In addition, the proposed method accurately estimates the severity of the process faults, providing significant information to the practitioners for their decision-making in the maintenance schedule. Specifically, the error between the estimation from the proposed method and the actual severity of the process faults achieves less than 10% of error of all the benchmark methods when there exists a high correlation between the variations of the fixture locators in the autobody assembly system.

The training and testing data for deep-neural-network-based classifiers are usually assumed to be sampled from the same distribution. When part of the testing samples are drawn from a distribution that is sufficiently far away from that of the training samples (a.k.a. out-of-distribution (OOD) samples), the trained neural network has a tendency to make high-confidence predictions for these OOD samples. Detection of the OOD samples is critical when training a neural network used for image classification, object detection, etc. It can enhance the classifier's robustness to irrelevant inputs, and improve the system's resilience and security under different forms of attacks. Detection of OOD samples has three main challenges: (i) the proposed OOD detection method should be compatible with various architectures of classifiers (e.g., DenseNet, ResNet) without significantly increasing the model complexity and requirements on computational resources; (ii) the OOD samples may come from multiple distributions, whose class labels are commonly unavailable; (iii) a score function needs to be defined to effectively separate OOD samples from in-distribution (InD) samples. To overcome these challenges, we propose a Wasserstein-based out-of-distribution detection (WOOD) method. The basic idea is to define a Wasserstein-based score that evaluates the dissimilarity between a test sample and the distribution of InD samples. An optimization problem is then formulated and solved based on the proposed score function. The statistical learning bound of the proposed method is investigated to guarantee that the loss value achieved by the empirical optimizer approximates the global optimum. The comparison study results demonstrate that the proposed WOOD consistently outperforms other existing OOD detection methods.

Differential equations are fundamental in modeling numerous physical systems, including thermal, manufacturing, and meteorological systems. Traditionally, numerical methods often approximate the solutions of complex systems modeled by differential equations. With the advent of modern deep learning, Physics-informed Neural Networks (PINNs) are evolving as a new paradigm for solving differential equations with a pseudo-closed form solution. Unlike numerical methods, the PINNs can solve the differential equations mesh-free, integrate the experimental data, and resolve challenging inverse problems. However, one of the limitations of PINNs is the poor training caused by using the activation functions designed typically for purely data-driven problems. This work proposes a scalable tanh-based activation function for PINNs to improve learning the solutions of differential equations. The proposed Self-scalable tanh (Stan) function is smooth, non-saturating, and has a trainable parameter. It can allow an easy flow of gradients and enable systematic scaling of the input-output mapping during training. Various forward problems to solve differential equations and inverse problems to find the parameters of differential equations demonstrate that the Stan activation function can achieve better training and more accurate predictions than the existing activation functions for PINN in the literature.

Background/foreground separation is one of the most fundamental tasks in computer vision, especially for video data. Robust PCA (RPCA) and its tensor extension, namely, Robust Tensor PCA (RTPCA), provide an effective framework for background/foreground separation by decomposing the data into low-rank and sparse components, which contain the background and the foreground (moving objects), respectively. However, in real-world applications, the video data is contaminated with noise. For example, in metal additive manufacturing (AM), the processed X-ray video to study melt pool dynamics is very noisy. RPCA and RTPCA are not able to separate the background, foreground, and noise simultaneously. As a result, the noise will contaminate the background or the foreground or both. There is a need to remove the noise from the background and foreground. To achieve the three components decomposition, a smooth sparse Robust Tensor Decomposition (SS-RTD) model is proposed to decompose the data into static background, smooth foreground, and noise, respectively. Specifically, the static background is modeled by the low-rank tucker decomposition, the smooth foreground (moving objects) is modeled by the spatio-temporal continuity, which is enforced by the total variation regularization, and the noise is modeled by the sparsity, which is enforced by the$\ell _{1}$norm. An efficient algorithm based on alternating direction method of multipliers (ADMM) is implemented to solve the proposed model. Extensive experiments on both simulated and real data demonstrate that the proposed method significantly outperforms the state-of-the-art approaches for background/foreground separation in noisy cases. Note to Practitioners—This work is motivated by melt pool detection in metal additive manufacturing where the processed X-ray video from the monitoring system is very noisy. The objective is to recover the background with porosity defects and the foreground with melt pool in the presence of noise. Existing methods fail to separate the noise from the background and foreground since RPCA and RTPCA have only two components, which cannot explain the three components in the data. This paper puts forward a smooth sparse Robust Tensor Decomposition by decomposing the tensor data into low-rank, smooth, and sparse components, respectively. It is a highly effective method for background/foreground separation in noisy case. In the case studies on simulated video and X-ray data, the proposed method can handle non-additive noise, and even the case of high noise-ratio. In the proposed algorithm, there is only one tuning parameter$\lambda $. Based on the case studies, our method achieves satisfying performance by taking any$\lambda \in [{0.2,1}]$with anisotropic total variation regularization. With this observation, practitioners can apply the proposed method without extensive parameter tuning work. Furthermore, the proposed method is also applicable to other popular industrial applications. Practitioners can use the proposed SS-RTD for degradation processes monitoring, where the degradation image contains the static background, anomaly, and random disturbance, respectively.

Robust PCA (RPCA) and its tensor extension, namely, Robust Tensor PCA (RTPCA), provide an effective framework for background/foreground separation by decomposing the data into low-rank and sparse components, which contain the background and the foreground (moving objects), respectively. However, in real-world applications, the presence of missing pixels is a very common and challenging issue due to errors in the acquisition process or manufacturer defects. RPCA and RTPCA are not able to recover the background and foreground simultaneously with missing pixels. This study aims to address the problem of background/foreground separation with missing pixels by combining video recovery and background/foreground separation into a single framework. To achieve this goal, a smooth robust tensor completion (SRTC) model is proposed to recover the data and decompose it into the static background and smooth foreground, respectively. An efficient algorithm based on tensor proximal alternating minimization (tenPAM) is implemented to solve the proposed model with a global convergence guarantee under very mild conditions. Extensive experiments on actual data demonstrate that the proposed method significantly outperforms the state-of-the-art approaches for background/foreground separation with missing pixels.

Supervised machine learning techniques, such as classification models, have been widely applied to online process anomaly detection in advanced manufacturing. However, since abnormal process states rarely occur in regular manufacturing settings, the data collected for model training may be highly imbalanced, which may result in significant training bias for supervised learning and, thus, further deteriorate the anomaly detection accuracy. To reduce the training bias, a natural idea is to incorporate data augmentation techniques to generate effective artificial sample data for the abnormal process states. However, most of the existing data augmentation methods do not effectively consider the temporal orders of the sensor signals, and they also usually require large amounts of actual samples to ensure satisfactory augmentation performance. To address these limitations, this article developed a novel data-driven methodology termed augmented time regularized generative adversarial network (ATR-GAN). By incorporating a proposed augmented generator, ATR-GAN is capable of generating more effective artificial samples for training supervised learning models. The novelty of this augmented generator in the proposed methodology can be summarized into three aspects: 1) an augmented filter layer is introduced in the augmented generator to identify the high-quality artificial samples; 2) in the augmented filter layer, a new distance metric termed time-regularized Hausdorff (TRH) distance is developed to accurately measure the similarity between actual samples and the generated artificial samples; and 3) batching techniques are also employed in the proposed augmented generator to further increase the diversity of the artificial data and fully utilize the relatively limited training data. In addition, the effectiveness of the proposed ATR-GAN is also validated by both numerical simulation and a real-world case study in additive manufacturing. Note to Practitioners—Online process anomaly detection currently plays a significant role in advanced manufacturing since unexpected anomalies may damage product quality and even result in catastrophic loss. In practice, processes are mostly under normal conditions, and anomalies rarely occur. Therefore, the data collected under abnormal conditions are very limited compared to normal conditions, which causes the data imbalanced issue, leading to deterioration in detection accuracy. Many existing data augmentation methods, such as generative adversarial network (GAN), cannot synthesize diversified high-quality artificial samples only using relatively limited actual samples. There is an urgent need in developing an effective data augmentation methodology to address the data imbalanced issue in process anomaly detection. This article developed a novel approach called augmented time regularized GAN (ATR-GAN) for online sensor data augmentation. With this new approach, the applications in additive manufacturing demonstrate that the performance of data augmentation can be improved effectively, and thereafter, the anomaly detection accuracy is also increased significantly. Moreover, the developed methodology is inherently integrated into a generic framework. Thus, it can be further transformed for applications in many other areas that need data augmentation.

The recent increase in applications of high-dimensional data poses a severe challenge to data analytics, such as supervised classification, particularly for online applications. To tackle this challenge, efficient and effective methods for feature extraction are critical to the performance of classification analysis. The objective of this work is to develop a new supervised feature extraction method for high-dimensional data. It is achieved by developing a clustered discriminant regression (CDR) to extract informative and discriminant features for high-dimensional data. In CDR, the variables are clustered into different groups or subspaces, within which feature extraction is performed separately. The CDR algorithm, which is a greedy approach, is implemented to obtain the solution toward optimal feature extraction. One numerical study is performed to demonstrate the performance of the proposed method for variable selection. Three case studies using healthcare and additive manufacturing data sets are accomplished to demonstrate the classification performance of the proposed methods for real-world applications. The results clearly show that the proposed method is superior over the existing method for high-dimensional data feature extraction.Note to Practitioners—This article forwards a new supervised feature extraction method termed clustered discriminant regression. This method is highly effective for classification analysis of high-dimensional data, such as images or videos, where the number of variables is much larger than the number of samples. In our case studies on healthcare and additive manufacturing, the performance of classification analysis based on our method is superior over the existing feature extraction methods, which is confirmed by using various popular classification algorithms. For image classification, our method with elaborately selected classification algorithms can outperform a convolutional neural network. In addition, the computation efficiency of the proposed method is also promising, which enables its online applications, such as advanced manufacturing process monitoring and control.

Sensor-based online process monitoring has extensive applications, such as in manufacturing and service industries. In real environments, though, sensor data are often contaminated with noise, leading to severe challenges in accurate data analysis. In the existing literature, noise is generally modeled as Gaussian to analyze sensor data for various applications, for example in fault detection and diagnostics. However, in some applications, such as due to challenging field conditions, sensor data may be disturbed by high levels of outliers such that the Gaussian assumption of sensor noise is inadequate, thus leading to large estimation errors. This paper focuses on online classification applications. A robust sparse representation classification method is proposed, which considers non-Gaussian noise, and thus can effectively analyze sensor data with higher levels of outliers. Case studies were completed, based on both numerically simulated sensor data and actual wearable sensor data from occupational manual material handling process monitoring. The proposed classification method could effectively analyze sensor data with non-Gaussian noise, and outperformed commonly used methods in the literature. Thus, this new method may be advantageous for solving classification problems in challenging field conditions, to address the difficulties of high levels of sensor outliers. Note to Practitioners - This paper proposes a fast, robust classification method for online sensor data classification. The proposed method is designed to cope with high levels of sensor outliers. The robustness of the method and its computational efficiency make it particularly appealing for online sensor data classification in challenging field conditions in which the presence of sensor outliers causes practical difficulties for most existing classification algorithms.

Additive manufacturing (AM) enables flexible part geometry and functionality, and reduces product development life cycle by direct layer-wise fabrication from CAD files. In the last decade, great achievements are made on AM materials, machines, processes, etc. However, the quality of the AM parts is still questionable for industrial specifications. On the one hand, AM part quality variables can be either quantitative, such as dimensional accuracy, or qualitative, such as binary indicators for voids, missing features, or surface roughness. On the other hand, both offline process setting variables and functional in situ process variables can be measured and modeled with both quantitative and qualitative (QQ) quality response variables. In this paper, the QQ quality response variables are modeled by offline process setting variables and in situ process variables via functional QQ models. The modeling of these in situ process variables provides the basis for real-time monitoring and control for AM processes. Simulation studies and experimental data from a fused deposition modeling process are performed to demonstrate the effectiveness of the proposed method.

The objective of this paper is to monitor complex process dynamics manifest in multivariate (multidimensional) time series data using a spectral (algebraic) graph theoretic approach. We test the hypothesis that the spectral graph-based topological invariants detect incipient process drifts earlier [lower average run length (ARL1)] and with higher fidelity (consistency of detection) when compared with the conventional statistics-based approaches. The presented approach maps a multidimensional sensor data stream XN×d(visualize N as time and d as the number of sensors) as an unweighted and undirected network graph G(V, E), indexed by its vertices V and edges E, i.e., X → G(V, E). The rationale is that the graph-based topological invariants are surrogate representatives of the system state. We compare the monitoring performance of spectral graph theoretic invariants with conventional statistical features in an exponentially weighted moving average control chart setting. The practical utility of the approach is substantiated in the context of process monitoring in two advanced manufacturing scenarios, namely, ultraprecision machining (UPM) and semiconductor chemical mechanical planarization. These studies corroborate the hypothesis that graph theoretic invariants, when used as monitoring statistics, lead to lower ARL1and more consistent detections in contrast to conventional statistical features. For instance, in the UPM case, the fault detection delay using graph theoretic invariants is less than 160 ms, compared with over 8 s of delay with statistical features.

The goal of this work is to use sensor data for online detection and identification of process anomalies (faults). In pursuit of this goal, we propose Dirichlet process Gaussian mixture (DPGM) models. The proposed DPGM models have two novel outcomes: 1) DP-based statistical process control (SPC) chart for anomaly detection and 2) unsupervised recurrent hierarchical DP clustering model for identification of specific process anomalies. The presented DPGM models are validated using numerical simulation studies as well as wireless vibration signals acquired from an experimental semiconductor chemical mechanical planarization (CMP) test bed. Through these numerically simulated and experimental sensor data, we test the hypotheses that DPGM models have significantly lower detection delays compared with SPC charts in terms of the average run length (ARL1) and higher defect identification accuracies (F-score) than popular clustering techniques, such as mean shift. For instance, the DP-based SPC chart detects pad wear anomaly in CMP within 50 ms, as opposed to over 140 ms with conventional control charts. Likewise, DPGM models are able to classify different anomalies in CMP. Note to Practitioners-This paper forwards novel Dirichlet process Gaussian mixture (DPGM) models for online process quality monitoring. The practical outcome is that the deleterious impact of process drifts on product quality is identified in their early stages using the presented DPGM models. For instance, sensor signal patterns from contemporary advanced manufacturing processes rarely follow distribution symmetry or normality assumptions endemic to traditional statistical process control (SPC) methods. These assumptions limit the effectiveness of traditional SPC methods for detection of process anomalies from complex heterogeneous sensor data. In comparison, the proposed DP-based SPC is capable of detecting process changes in the sensor data notwithstanding the characteristics of the underlying distributions. Moreover, we show that the recurrent hierarchical DP clustering model identifies process anomalies with higher fidelity compared with traditional methods, such as mean shift clustering.

The aim of this paper is to detect the incipient anomalies in a ultraprecision machining (UPM) process by integrating multiple in situ sensor signals. To realize this aim we forward a Bayesian non-parametric Dirichlet Process (DP) decision-making approach for real-time monitoring of UPM process using the data gathered from multiple, heterogeneous sensors. The sensor signals are acquired under different experimental conditions from a UPM setup instrumented with a heterogeneous sensing array consisting of miniature tri-axis force, tri-axis vibration, and acoustic emission (AE) sensors mounted in close proximity to the cutting tool. We track the prominent nonlinear and non-Gaussian signal patterns evident in the experimentally acquired sensor data using an adaptive non-parametric DP modeling technique. A cohesive decision concerning the UPM process condition is made by developing a new supervised learning method, which integrates the DP-model state estimates with an evidence theoretic sensor data fusion method. Using this combined DP-evidence theoretic approach, UPM process drifts and anomalies, such as sudden changes in the depth of cut, feed rate, and spindle speed that deleteriously affect surface finish, and hence cause high yield losses, are detected and classified with over 90% accuracy (with${<} 5\hbox{\%}$standard deviation). We compared these results with popular classification techniques, e.g., naïve Bayes, self-organizing map, and support vector machine; these conventional techniques had classification accuracy in the range of 83%–88%. Consequently, this research makes the following practically relevant contributions: 1) real-time identification of the incipient UPM process anomalies from multiple sensors and 2) prescribing the optimal subset of sensors signals contingent to particular process anomalies.

Occupational jobs often involve different types of manual material handling (MMH) tasks. Performing such tasks can be physically demanding, and which may put workers at an increased risk of work-related musculoskeletal disorders (WMSDs). To control and prevent WMSDs, there has been a growing interest in online posture monitoring using wearable sensors. In this paper, we developed an online, supervised, task classification algorithm for monitoring and evaluation of MMH activities. The classification algorithm is based on a fast sparse estimation methodology, which makes it computationally efficient for online decision making. We further propose an optimization approach to improve classification performance, by differentially weighting sensors, thereby representing the relative influence of a sensor in classification performance. Optimizing these weights enables us to determine the most relevant sensors for classification. A case study using 37 sensors with 111 channels of data was completed to validate performance of the proposed method. With only 30 optimally selected sensor channels, our method provides high classification accuracy (>84%) and outperforms several benchmark methods, including support vector machine, quadratic discriminant analysis, and neural network.

Dimensional integrity has a significant impact on the quality of the final products in multistation assembly processes. A large body of research work in fault diagnosis has been proposed to identify the root causes of the large dimensional variations on products. These methods are based on a linear relationship between the dimensional measurements of the products and the possible process errors, and assume that the number of measurements is greater than that of process errors. However, in practice, the number of measurements is often less than that of process errors due to economical considerations. This brings a substantial challenge to the fault diagnosis in multistation assembly processes since the problem becomes solving an underdetermined system. In order to tackle this challenge, a fault diagnosis methodology is proposed by integrating the state space model with the enhanced relevance vector machine (RVM) to identify the process faults through the sparse estimate of the variance change of the process errors. The results of case studies demonstrate that the proposed methodology can identify process faults successfully.

Yield-based sensitivity analysis methods and algorithms are developed for process capability evaluation in this paper. Yield, the conformity to product specifications, is subject to critical design parameters such as dimensions, tolerances, and specification limits. The uncertainties in determining these parameters in design and manufacturing affect the process capability of producing high-quality products. Yield is a transparent and thus a desirable index, especially for multivariate process capability evaluation. Thus, yield sensitivity with respect to these design parameters indicates key contributors to final product quality, providing valuable information in design for quality control. Yield is formulated as a high dimension probability integral over a specification region. In multistage assembly processes, an assembly variation model links the quality characteristics to design parameters. This model is adopted in yield sensitivity analysis. Derivatives of yield with respect to the design parameters are developed using matrix calculus. Three sensitivity analysis algorithms, i.e., finite difference, yield derivative, and regression modeling, are implemented. Monte Carlo simulation is used for yield estimation in the three algorithms. A case study using floor pan assembly in automotive body manufacturing is presented for the validation of the proposed methodology.