Machine learning and other data-driven methods have developed at a prolific rate for industrial applications due to the advent of industrial big data. However, industrial datasets may not be especially well-suited to supervised learning approaches that require extensive domain knowledge in the complete and accurate labeling of datasets. To address these challenges, a semi-supervised learning approach is proposed that makes use of partially labeled subsets. The proposed methodology is applied to high-dimensional in-process measurement data, utilizing a convolutional autoencoder (CAE) for unsupervised feature extraction. A multiclass extension for semi-supervised anomaly diagnosis is proposed that utilizes principal component analysis (PCA) as the basis for anomaly scoring, and the proposed approach intersects the results of targeted one-against-all phases on partially labeled sets to classify faults. Experiments in a case study on semiconductor manufacturing measurement data are performed to explore the relationship between latent features extracted and anomaly detection performance. The application of the proposed algorithm achieves a true positive detection rate of over 90% with false positive rate under 9% for both local and global anomaly types, with these results accomplished while reducing over 99% of the original input data dimensions. In addition, the approach also allows for positive samples to be identified that were previously undetected by human experts. These results are promising for the application of the proposed semi-supervised methodology in real industrial settings.

In the present research, a digital twin of coaxial one-side resistance spot welding (COS-RSW) was established for the real-time prediction of transient temperature field. A three-dimensional (3D) model of COS-RSW joint was developed based on the in-house finite element (FE) code JWRIAN-SPOT. The experimental verified FE model was employed to generate the big data of temperature of COS-RSW process. Multiple dimension interpolation was applied to process database and output prediction. The FE model can predict the thermal cycle on COS-RSW joints under different parameter couples. The interpolation effect of individual welding parameters was discussed, and a power weight judgement for welding time was essential to ensure accuracy. With the support of big data, the digital twin can provide visualization prediction of COS-RSW within 10 s, whereas numerical modeling needs at least 1 h. The proposed application of digital twin has potential to improve the efficiency of process optimization in engineering.

This article considers the influence of the size of processed workpieces during their strain hardening based on the impact of deformation waves on the processed surface. For the first time, the specificity of the hardened layer formation by wave strain hardening (WSH) makes it possible to draw attention to the fact that, with equal volumes of hardened samples and processing modes, different diagrams of microhardness distribution in the surface layer are observed. The aim of this work is to establish the relationship between processing parameters, workpiece dimensions, and the WSH nature for the first time. The experiments are carried out on a specially designed bench. It is established that WSH occurs not only from the front side (along which the processing was performed), but also from the backside, while leaving the original material hardness in the middle of the sample. The results obtained form a basis for the further development of WSH technology.

Nowadays, the tolerancing integration in computer-aided design (CAD) tools remains among the major goals of mechanical manufacturers. In the virtual product development, ideal and rigid models are used in the digital mockup (DMU). Hence, research works developed integrated CAD models for tolerance analysis, while considering manufacturing defects. However, the tolerance analysis in the case of composite positional tolerance for feature patterns, commonly used in the industry, becomes a difficult activity with the consideration of parts deformations. Thus, this paper presents a novel CAD model for the tolerance analysis considering composite positional defect of features set and nonrigid component deformations due to external mechanical loads. The modeling of rigid components with dimensional defects is established based on the numerical perturbation method. Indeed, the relationships between driving and driven dimensions are determined to obtain the configurations in maximum and least material of the CAD model. Thereafter, the geometrical deviations are modeled by face displacements. The modeling of composite positional errors is performed while respecting the feature relating position tolerance zone framework and the pattern-location tolerance zone framework constraints, as well as the maximum or least material condition. The deviations caused by nonrigid part deformations are considered by the integration of finite element results into the CAD model. The realistic configurations of the assembly are obtained after the updating of mating constraints between rigid and nonrigid parts with defects. The composite positional tolerance is analyzed with the simulation of relative motion between parts. A case study is proposed to evaluate the developed tolerancing method.

Additive manufacturing (AM) is a novel fabrication technique capable of producing highly complex parts. Nevertheless, a major challenge is the quality assurance of the AM fabricated parts. While there are several ways of approaching this problem, how to develop informative process signatures to detect part anomalies for quality control is still an open question. The objective of this study is to build a new layer-wise process signature model to characterize the thermal-defect relationship. Based on melt pool images, we propose novel layer-wise key process signatures, which are calculated using multilinear principal component analysis (MPCA) and are directly correlated with the layer-wise quality of the part. The resultant layer-wise quality features can be used to predict the overall defect distribution of a fabricated layer during the build. The proposed model is validated through a case study based on a direct laser deposition experiment, where the layer-wise quality of the part is predicted on the fly. The accuracy of prediction is calculated using three measures (i.e., recall, precision, and F-score), showing reasonable success of the proposed methodology in predicting layer-wise quality. The proposed quality prediction methodology enables online process correction to eliminate anomalies and to ultimately improve the quality of the fabricated parts.

Sensor signals acquired during the manufacturing process contain rich information that can be used to facilitate effective monitoring of operational quality, early detection of system anomalies, and quick diagnosis of fault root causes. This paper develops a method for effective monitoring and diagnosis of multisensor heterogeneous profile data based on multilinear discriminant analysis. The proposed method operates directly on the multistream profiles and then extracts uncorrelated discriminative features through tensor-to-vector projection, and thus, preserving the interrelationship of different sensors. The extracted features are then fed into classifiers to detect faulty operations and recognize fault types. The developed method is demonstrated with both simulated and real data from ultrasonic metal welding.

The modern manufacturing industry faces increasing demands to customize products according to personal needs, thereby leading to the proliferation of complex designs. To cope with design complexity, manufacturing systems are increasingly equipped with advanced sensing and imaging capabilities. However, traditional statistical process control methods are not concerned with the stream of in-process imaging data. Also, very little has been done to investigate nonlinearity, irregularity, and inhomogeneity in the image stream collected from manufacturing processes. This paper presents the joint multifractal and lacunarity analysis to characterize irregular and inhomogeneous patterns of image profiles, as well as detect the hidden dynamics in the manufacturing process. Experimental studies show that the proposed method not only effectively characterizes surface finishes for quality control of ultraprecision machining but also provides an effective model to link process parameters with fractal characteristics of in-process images acquired from additive manufacturing. This, in turn, will allow a swift response to processes changes and consequently reduce the number of defective products. The proposed multifractal method shows strong potentials to be applied for process monitoring and control in a variety of domains such as ultraprecision machining and additive manufacturing.

Additive manufacturing (or three-dimensional (3D) printing) is constantly growing as an innovative process for the production of complex-shape components. Among the seven recognized 3D printing technologies, fused deposition modeling (FDM) covers a very important role, not only for producing representative 3D models, but, mainly due to the development of innovative material like Peek and Ultem, also for realizing structurally functional components. However, being FDM a production process involving high thermal gradients, non-negligible deformations and residual stresses may affect the 3D printed component. In this work we focus on meso/macroscopic simulations of the FDM process using abaqus software. After describing in detail the methodological process, we investigate the impact of several parameters and modeling choices (e.g., mesh size, material model, time-step size) on simulation outcomes and we validate the obtained results with experimental measurements.

Parts made via polymer extrusion are currently limited to a constant cross section. Additionally, the process is difficult to control, so desired final part dimensions are often achieved via a manual trial-and-error approach to parameter adjustment. This work seeks to increase the capability of polymer extrusion by using iterative learning control (ILC) to regulate the final width of a rectangular part through changing the width of a simple variable-geometry die. Simulation results determine the appropriateness of the learning algorithm and gains to be used in experiment. A prototype die on a production extruder was used to demonstrate the effectiveness of the approach. These experiments achieved automated control over both gross change in shape and final part dimension when the puller speed was held constant, which has not been seen previously in the literature.

In recent years, the requirements for high dimensional precision on Z-bent shaped parts have become increasingly stringent. To attain these requirements, the suitable selection of the Z-die bending type has to be considered much more strictly. In this research, two types of Z-bending processes, offset Z-die bending and wiping Z-die bending, were investigated using the finite element method (FEM) to identify the spring-back characteristics and dimensions of Z-bent shaped parts. In the case of offset Z-die bending, the spring-back characteristics on both bend angles were similar. In contrast, in the case of wiping Z-bending, the spring-back characteristics on both bend angles were different. In addition, the dimensions of the Z-bent shaped parts were investigated. It was found, in the case of wiping Z-bending, that web thinning was generated and the outer bend radius was out of tolerance. To validate the FEM simulation results, experiments were carried out. The FEM simulation results showed good agreement with the experimental results in terms of the bend angles and the overall geometry of the Z-bent shaped parts. To achieve precise Z-bent shaped parts, the suitable selection of Z-die bending type in the Z-die bending process is very important.

Metal-based powder-bed-fusion additive manufacturing (PBF-AM) is gaining increasing attention in modern industries, and is a promising direct manufacturing technology. Additive manufacturing (AM) does not require the tooling cost of conventional subtractive manufacturing processes, and is flexible to produce parts with complex geometries. Quality and repeatability of AM parts remain a challenging issue that persistently hampers wide applications of AM technology. Rapid advancements in sensing technology, especially imaging sensing systems, provide an opportunity to overcome such challenges. However, little has been done to fully utilize the image profiles acquired in the AM process and study the fractal patterns for the purpose of process monitoring, quality assessment, and control. This paper presents a new multifractal methodology for the characterization and detection of defects in PBF-AM parts. Both simulation and real-world case studies show that the proposed approach effectively detects and characterizes various defect patterns in AM images and has strong potential for quality control of AM processes.

Textured functional surfaces are finding applications in the fields of bioengineering, surface energy, hydrodynamics, lubrication, and optics. Electrical discharge machining (EDM), which is normally used to generate smoother surface finish on various automotive components and toolings, can also generate surfaces of rough finish, a desirable characteristic for texturing purposes. There is a lack of modeling efforts to predict the surface textures obtained under various EDM operating conditions. The aim of the current work is to capture the physics of the electrical discharge texturing (EDT) on a surface assuming random generation of multiple sparks with respect to (i) space, (ii) time, and (iii) energy. A uniform heat disk assumption is taken for each individual spark. The three-dimensional (3D) texture generated is utilized to evaluate a 3D roughness parameter namely arithmetic mean height, S a . Surface textures obtained from the model are validated against experimentally obtained ones by comparison of distribution of R a values taken along parallel sections along the surface. It was found that the distribution of simulated R a values agrees with that of experimental R a values.

Nerve conduits with topographical guidance have been recognized as the efficient repair of damaged peripheral nerves. In this study, polymeric hollow fiber membranes (HFMs) with grooved inner surface have been fabricated from a microstructured spinneret using a dry-jet wet spinning process for nerve regeneration studies. The effectiveness of HFM inner grooves has been demonstrated during an in vitro study of chick forebrain neuron outgrowth. It is of great importance that the groove geometry can be controllable to meet various needs in promoting nerve regeneration performance. While the overall groove geometry is determined by the spinneret design, fabrication conditions are also indispensable in fine-tuning the final groove geometry such as the groove height and width on the order of 10 μ m or less. It is found that the bore fluid flow rate can be utilized to effectively adjust the resulting groove height by at most 52% and groove width by at most 61%, respectively, without modifying the spinneret geometry. This enables a new approach to fabricate different grooved HFMs using the same spinneret. By comparing to the influences of bore fluid flow rate, the dope fluid flow rate is less effective in regulating the groove height and width when using the same microstructured spinneret. Both bore and dope fluid flow rates should be carefully selected for fine groove width tuning.

The studies about sustainable manufacturing engineering (SME) contain an increasing body of knowledge, motivated by the rising interest in the processes lifecycle sustainability. Its continuous improvement and optimization (including sustainability criteria) has become an emerging necessity. For this reason, new clean technologies and proposals of work methods are required; they have to integrate the ecological and social dimensions at an operational level in the manufacturing processes, maintaining the economic and technical feasibility attained up to this moment. However, a unified framework does not exist to orientate the lines of research in optimization when applied to sustainability. In this sense, the article reviews studies from scientific literature about sustainable machining developed in the last 15 years. The review has been carried out from the triple bottom-line (TBL) perspective, defined by the three general sustainability dimensions (economy, ecology, and equity). It contributes to the literature and current machining engineering knowledge, with its involvement in mitigating the metabolic rift. The results from the review have allowed to characterize the investigation effort, with regard to the optimization of the sustainable machining processes; even though numerous studies exist which optimize machining operations (with the aim to find the trade-off between different environmental and equity factors), in general, the technical and economic feasibilities are still the priority. The patterns defined through the analysis of the publications have established the current development trend; furthermore, as a consequence of the review results, we propose an outline of articulated lines of investigation with the aim to mitigate the metabolic rift through triple bottom-line, necessary so that machining engineering assumes the goal of finding the balance to achieve integral sustainability.

Increasing demands in profile manufacturing leads to a constant increase of the geometric complexity of the profile. Nowadays, lightweight design and load adaption are of huge importance and a need for new profile manufacturing technologies exists. In order to cope with this development, a new tube profile forming method was invented, the incremental profile forming (IPF). IPF allows the flexible manufacturing of profiles with varying cross section geometries along the longitudinal profile axis and offers therefore a high potential for the manufacturing of lightweight design parts. Due to the high degree of innovation, research work regarding the process fundamentals as well as the process limits is necessary. For this reason, the results of the first basic experimental and numerical investigations are presented. While in the experiments, the potential of the process is shown in feasibility studies, first analyses of the forming process, and the influence of different process parameters on the process were carried out in numerical investigations. In particular, it turned out that the tool infeed and tube dimensions have significant influence on the forming force, whereas the tool radius has a minor influence.

Precision improvement in sheet metal stamping has been the concern that the stamping researchers have engaged in. In order to improve the forming precision of sheet metal in stamping, this paper devoted to establish the generalized holo-factors mathematical model of dimension-error and shape-error for sheet metal in stamping based on BP neural network. Factors influencing the forming precision of stamping sheet metal were divided, altogether ten factors, and the generalized holo-factors mathematical model of dimension-error and shape-error for sheet metal in stamping was established using the back-propagation algorithm of error based on BP neural network. The undetermined coefficients of the model previously established were soluble according to the simulation data of sheet punching combined with the specific shape based on the BP neural network. With this mathematical model, the forecast data compared with the validate data could be obtained, so as to verify the fine practicability that the previously established mathematical model had, and then, it was shown that the generalized holo-factors mathematical model of size error and shape-error had fine practicality and versatility. Based on the generalized holo-factors mathematical model of error exemplified by the cylindrical parts, a group of process parameters could be selected, in which forming thickness was between 0.713 mm and 1.335 mm, major strain was between 0.085 and 0.519, and minor strain was between −0.596 and 0.319 from the generalized holo-factors mathematical model prediction, at the same time, the forming thickness, the major strain, and the minor strain were in good condition.

The ease of entering a car is one of the important ergonomic factors that car manufacturers consider during the process of car design. This has motivated many researchers to investigate factors that affect discomfort during ingress. The patterns of motion during ingress may be related to discomfort, but the analysis of motion is challenging. In this paper, a modeling framework is proposed to use the motions of body landmarks to predict subjectively reported discomfort during ingress. Foot trajectories are used to identify a set of trials with a consistent right-leg-first strategy. The trajectories from 20 landmarks on the limbs and torso are parameterized using B-spline basis functions. Two group selection methods, group non-negative garrote (GNNG) and stepwise group selection (SGS), are used to filter and identify the trajectories that are important for prediction. Finally, a classification and prediction model is built using support vector machine (SVM). The performance of the proposed framework is then evaluated against simpler, more common prediction models.

The ability of additive manufacturing (AM) processes to produce components with virtually any geometry presents a unique challenge in terms of quantifying the dimensional quality of the part. In this paper, a novel spectral graph theory (SGT) approach is proposed for resolving the following critical quality assurance concern in the AM: how to quantify the relative deviation in dimensional integrity of complex AM components. Here, the SGT approach is demonstrated for classifying the dimensional integrity of standardized test components. The SGT-based topological invariant Fiedler number ( λ 2 ) was calculated from 3D point cloud coordinate measurements and used to quantify the dimensional integrity of test components. The Fiedler number was found to differ significantly for parts originating from different AM processes (statistical significance p-value <1%). By comparison, prevalent dimensional integrity assessment techniques, such as traditional statistical quantifiers (e.g., mean and standard deviation) and examination of specific facets/landmarks failed to capture part-to-part variations, proved incapable of ranking the quality of test AM components in a consistent manner. In contrast, the SGT approach was able to consistently rank the quality of the AM components with a high degree of statistical confidence independent of sampling technique used. Consequently, from a practical standpoint, the SGT approach can be a powerful tool for assessing the dimensional integrity of the AM components, and thus encourage wider adoption of the AM capabilities.

Growing environmental awareness leads production engineers to focus increasingly on energy and material efficiency of manufacturing processes. However, only a few holistic approaches have been applied on the manufacturing process level and they often disregard product quality. In this study, sustainability indicators for the discrete manufacturing process of grinding are defined and discussed. Various temporal and spatial boundaries for the sustainability analysis are evaluated with regard to their effect on the results. Selected indicators, here energy and waste intensity, are then used to evaluate different dressing strategies in a case study. This study highlights the challenges in setting the boundaries for a sustainability analysis and stresses the importance of clearly defining these in research papers.

This paper considers a way of measuring a process capability index in order to obtain the geometric tolerance of a pattern of position elements according to the ASME Y14.5 standard. The number of elements present in the pattern, as well as its material condition (least LMC or maximum MMC), are taken into consideration during the analysis. An explicit mathematical model will be developed to identify the distribution functions (PDF and CDF) of defects on the location and diameter. Using these distributions and the Hasofer–Lind index, we will arrive at a new definition of process capability—meaning the value of tolerances that can meet the threshold of x% compliance. Finally, our method is validated using a variety of typical case studies.