数字孪生 增材制造

Digital Twin -- a virtual replica of a physical system enabling real-time monitoring, model updating, prediction, and decision-making -- combined with recent advances in machine learning, offers new opportunities for proactive control strategies in autonomous manufacturing. However, achieving real-time decision-making with Digital Twins requires efficient optimization driven by accurate predictions of highly nonlinear manufacturing systems. This paper presents a simultaneous multi-step Model Predictive Control (MPC) framework for real-time decision-making, using a multivariate deep neural network, named Time-Series Dense Encoder (TiDE), as the surrogate model. Unlike conventional MPC models which only provide one-step ahead prediction, TiDE is capable of predicting future states within the prediction horizon in one shot (multi-step), significantly accelerating the MPC. Using Directed Energy Deposition (DED) additive manufacturing as a case study, we demonstrate the effectiveness of the proposed MPC in achieving melt pool temperature tracking to ensure part quality, while reducing porosity defects by regulating laser power to maintain melt pool depth constraints. In this work, we first show that TiDE is capable of accurately predicting melt pool temperature and depth. Second, we demonstrate that the proposed MPC achieves precise temperature tracking while satisfying melt pool depth constraints within a targeted dilution range (10\%-30\%), reducing potential porosity defects. Compared to PID controller, the MPC results in smoother and less fluctuating laser power profiles with competitive or superior melt pool temperature control performance. This demonstrates the MPC's proactive control capabilities, leveraging time-series prediction and real-time optimization, positioning it as a powerful tool for future Digital Twin applications and real-time process optimization in manufacturing.

No abstract available

Laser-directed-energy deposition (DED) offers advantages in additive manufacturing (AM) for creating intricate geometries and material grading. Yet, challenges like material inconsistency and part variability remain, mainly due to its layer-wise fabrication. A key issue is heat accumulation during DED, which affects the material microstructure and properties. While closed-loop control methods for heat management are common in DED research, few integrate real-time monitoring, physics-based modeling, and control in a unified framework. Our work presents a digital twin (DT) framework for real-time predictive control of DED process parameters to meet specific design objectives. We develop a surrogate model using Long Short-Term Memory (LSTM)-based machine learning with Bayesian Inference to predict temperatures in DED parts. This model predicts future temperature states in real time. We also introduce Bayesian Optimization (BO) for Time Series Process Optimization (BOTSPO), based on traditional BO but featuring a unique time series process profile generator with reduced dimensions. BOTSPO dynamically optimizes processes, identifying optimal laser power profiles to attain desired mechanical properties. The established process trajectory guides online optimizations, aiming to enhance performance. This paper outlines the digital twin framework's components, promoting its integration into a comprehensive system for AM.

No abstract available

The increasing demand for smaller batch sizes and mass customisation in production poses considerable challenges to logistics and manufacturing efficiency. Conventional methodologies are unable to address the need for expeditious, cost-effective distribution of premium-quality products tailored to individual specifications. Additionally, the reliability and resilience of global logistics chains are increasingly under pressure. Additive manufacturing is regarded as a potentially viable solution to these problems, as it enables on-demand, on-site production, with reduced resource usage in production. Nevertheless, there are still significant challenges to be addressed, including the assurance of product quality and the optimisation of production processes with respect to time and resource efficiency. This article examines the potential of integrating digital twin methodologies to establish a fully digital and efficient process chain for on-site additive manufacturing. This study focuses on wire arc additive manufacturing (WAAM), a technology that has been successfully implemented in the on-site production of naval ship propellers and excavator parts. The proposed approach aims to enhance process planning efficiency, reduce material and energy consumption, and minimise the expertise required for operational deployment by leveraging digital twin methodologies. The present paper details the current state of research in this domain and outlines a vision for a fully virtualised process chain, highlighting the transformative potential of digital twin technologies in advancing on-site additive manufacturing. In this context, various aspects and components of a digital twin framework for wire arc additive manufacturing are examined regarding their necessity and applicability. The overarching objective of this paper is to conduct a preliminary investigation for the implementation and further development of a comprehensive DT framework for WAAM. Utilising a real-world sample, current already available process steps are validated and actual missing technical solutions are pointed out.

With the rapid development of additive manufacturing technology, its potential in complex structural manufacturing has gained widespread attention. However, there are many challenges in the additive manufacturing process, such as printing accuracy control and process parameter optimization. These challenges directly affect the performance and reliability of the final product. In this paper, we combined reinforcement learning with digital twin, and proposed an additive manufacturing optimization method to enhance the intelligence and adaptability of the manufacturing process. Firstly, we established a virtual simulation platform for additive manufacturing processes with digital twin technology based on physical models and real-time sensor data fusion. The manufacturing process was monitored and predicted through realtime simulation and data-driven methods. Next, we use Deep Q-Network (DQN) to adaptively adjust key process parameters such as printing speed, layer thickness, temperature, etc., to optimize the printing quality and efficiency. The experimental results show that this method can significantly improve the stability of the manufacturing process and product quality.

No abstract available

Digital Twins (DTs) are transforming manufacturing by bridging the physical and digital worlds, enabling real-time insights, predictive analytics, and enhanced decision making. In Industry 4.0, DTs facilitate automation and data integration, while Industry 5.0 emphasizes human-centric, resilient, and sustainable production. However, implementing DTs in robotic metal additive manufacturing (AM) remains challenging because of the complexity of the wire-based laser metal deposition (LMD) process, the need for real-time monitoring, and the demand for advanced defect detection to ensure high-quality prints. This work proposes a structured DT architecture for a robotic wire-based LMD cell, following a standard framework. Three DT implementations were developed. First, a real-time 3D simulation in RoboDK, integrated with a 2D Node-RED dashboard, enabled motion validation and live process monitoring via MQTT (message queuing telemetry transport) telemetry, minimizing toolpath errors and collisions. Second, an Industrial IoT-based system using KUKA iiQoT (Industrial Internet of Things Quality of Things) facilitated predictive maintenance by analyzing motor loads, joint temperatures, and energy consumption, allowing early anomaly detection and reducing unplanned downtime. Third, the Meltio dashboard provided real-time insights into the laser temperature, wire tension, and deposition accuracy, ensuring adaptive control based on live telemetry. Additionally, a prescriptive analytics layer leveraging historical data in FireStore was integrated to optimize the process performance, enabling data-driven decision making.

A digital twin (DT), with the components of a physics-based model, a data-driven model, and a machine learning (ML) enabled efficient surrogate, behaves as a virtual twin of the real-world physical process. In terms of Laser Powder Bed Fusion (L-PBF) based additive manufacturing (AM), a DT can predict the current and future states of the melt pool and the resulting defects corresponding to the input laser parameters, evolve itself by assimilating in-situ sensor data, and optimize the laser parameters to mitigate defect formation. In this paper, we present a deep neural operator enabled computational framework of the DT for closed-loop feedback control of the L-PBF process. This is accomplished by building a high-fidelity computational model to accurately represent the melt pool states, an efficient surrogate model to approximate the melt pool solution field, followed by an physics-based procedure to extract information from the computed melt pool simulation that can further be correlated to the defect quantities of interest (e.g., surface roughness). In particular, we leverage the data generated from the high-fidelity physics-based model and train a series of Fourier neural operator (FNO) based ML models to effectively learn the relation between the input laser parameters and the corresponding full temperature field of the melt pool. Subsequently, a set of physics-informed variables such as the melt pool dimensions and the peak temperature can be extracted to compute the resulting defects. An optimization algorithm is then exercised to control laser input and minimize defects. On the other hand, the constructed DT can also evolve with the physical twin via offline finetuning and online material calibration. Finally, a probabilistic framework is adopted for uncertainty quantification. The developed DT is envisioned to guide the AM process and facilitate high-quality manufacturing.

Metal Additive Manufacturing is increasingly used in heavy industry for fabricating large-scale structural components by sequentially layering metal material based on digital 3D models. However, deviations in process parameters can cause significant material and operational losses. To address these challenges, a two-stage additive manufacturing strategy is proposed, integrating an initial standardized production phase followed by targeted customizations using Wire Arc Additive Manufacturing, specifically Tungsten Inert Gas welding. Despite WAAM’s advantages like stable arcs and high thermal efficiency, real-time melt depth control is complicated by arc illumination, electromagnetic interference, and elevated temperatures. This paper introduces a digital twin framework integrating geometric reconstruction via Neural Radiance Fields with analytical temperature-field modeling, enabling accurate melt pool depth prediction and immediate process parameter adjustments. Experimental evaluations yielded a NeRF reconstruction SSIM of 93.80% and a PSNR of 36.09 dB, achieving 95.31% accuracy in weld penetration prediction during welding process. The developed approach presents a rapid, precise, and cost-effective solution, significantly enhancing the manufacturing adaptability and quality assurance of large-scale structural components in heavy industry applications.

Light-curing additive manufacturing is extensively employed in high-precision industries due to its capability to generate products with exceptional surface quality. Nonetheless, given the susceptibility of the surface of light-curing additive manufacturing models to various factors, the current model is constructed based on design parameters, which may not precisely replicate the actual model surface. This constraint hampers the analytical work across various stages. Moreover, distinct analysis stages may necessitate varied physical models. While it is feasible to produce corresponding models for each stage, this approach may result in time wastage and reduced efficiency. To address this issue, this paper introduces a digital twin system for light-curing additive manufacturing, incorporating an integrated algorithm specifically designed for constructing rough surfaces. The algorithm employs the Fast Fourier Transform (FFT), Johnson Transformation System, and autocorrelation Function to generate the surface topography of the model. Additionally, the system can monitor the printer’s stability throughout the printing process. To validate the feasibility of this system, a DT system was implemented to monitor the printer’s stability throughout the printing process and construct the surface topography of the printed model. The surface of the physical model was measured using a 3D surface profiler and perform statistical analysis of the model surface data. Finally compared with the rough surface constructed by the DT system. The results indicate that the characteristic parameter errors of the model surface are all below 5%, providing evidence that the rough surface constructed by the system fulfills the specified requirements.

Large Format Additive Manufacturing (LFAM) enables the fabrication of large, complex structures but presents challenges in thermal management, particularly in determining the optimal layer time to ensure interlayer bonding and structural integrity. Digital Twin (DT) technology has emerged as a key solution for predicting temperature distributions and optimizing process parameters. However, existing Physics-Based and Data-Driven DT models provide static, one-time predictions, lacking the adaptability to dynamically update thermal profile predictions based on real-time parameter adjustments. To address this limitation, we propose an adaptive Digital Twin framework based on the Video Diffusion Transformer (VDT). Unlike traditional DT models, our approach leverages Generative AI to dynamically simulate future temperature distributions when layer time or other printing parameters change. This method ensures that adjustments in printing strategy are immediately reflected in updated temperature predictions, leading to enhanced efficiency, improved print quality, and greater adaptability in LFAM workflows. Experimental results demonstrate that our approach is highly effective, generating realistic future frames that accurately reflect the temperature distribution. This work represents a significant step forward in Digital Twin technology, highlighting the potential of Generative AI in manufacturing.

Emerging paradigms such as cloud-edge continuum, Software-Defined Networking (SDN), and Digital Twin (DT) technologies are transforming smart manufacturing systems by enabling intelligent automation, real-time monitoring, and scalable orchestration. These capabilities are particularly critical in additive manufacturing (AM), where latency-sensitive control and predictive maintenance are essential. However, current architectures often require dynamic network programmability, synchronized twin management, and secure telemetry pipelines. This paper presents an SDN-Integrated Cloud-Edge DT Framework tailored for real-time AM monitoring. The framework integrates KubeEdge’s DeviceTwin module for edge-local twin representation, a telemetry agent for structured data streaming, and SDN-controlled Open vSwitch (OVS) for adaptive traffic control. IoT-enabled AM devices, including 3D printers, CNC machines, and robotic arms, interface with the edge node for local state caching, while KubeEdge’s CloudCore aggregates device states for analytics, visualization, and policy enforcement. Experimental validation on a Kubernetes cluster demonstrates sub-100-ms twin synchronization, SDN enforcement under 312 ms, and streaming latency breakdowns across MQTT-Kafka stages. This work establishes a scalable, resilient, and programmable foundation for next-generation, Industry 5.0 manufacturing ecosystems.

No abstract available

No abstract available

No abstract available

The input data for LSTM training is categorized into two types: virtualized manufacturing data and semi-virtualized manufacturing data. The virtualized manufacturing dataset includes the position and speed information of the XYZBC axis motors, with labels of 0 and 1, where 0 denotes a normal state and 1 indicates an abnormal condition. The semi-virtualized manufacturing dataset encompasses a broader range of sensor signals, including the position and speed of the physical XYZBC axis motors, vibration angles, velocities, displacements of the extruder along the XYZ axes, motor temperatures

Material extrusion (MEX) continues to be a pivotal additive manufacturing (AM) process, involving the selective heating and layer-by-layer deposition of material. However, conventional finite-element models (finite-element analysis) face limitations in accurately simulating the MEX process, highlighting the need for experimental validation. This paper highlights an advanced material modeling technology that streamlines the development of composite parts using large format AM (LFAM). It specifically focuses on a thermoplastic acrylonitrile butadiene styrene (ABS) matrix composite material enriched with 20% short carbon fibers. The study employs integrated computational materials engineering, integrating (i) the manufacturing process, (ii) the material’s microstructure, (iii) homogenization techniques, and (iv) the performance of the final part. The development of a digital twin for pellet extrusion is proposed, emphasizing the importance of micro-structure characterization to account for warpage and residual stresses that lead to part distortion. The demonstrator manufactured for this study is a wind turbine mold of a blade section. Experimental tests revealed an elastic modulus of 5.5 GPa and a hardening modulus of 2.4 GPa for the composite. The numerical microscopic model showed a 16% error in the elastic modulus compared to experimental results. The study concludes that the homogenization techniques are effective in predicting the elastic properties but lack accuracy in the plastic region. The application of the model to the LFAM process of pellet extrusion is demonstrated, with simulation results showing a maximum deformation close to the center of the wind turbine blade mold.

No abstract available

No abstract available

No abstract available

No abstract available

Recent developments in the field of Additive Manufacturing have been improving the capabilities of the technique not only to be able to build complex geometry parts layer by layer with different materials, but also including the so-called Industry 4.0 technologies, namely Internet of Things (IoT), big data (BD) and Digital Twins (DT). The combination of these technologies with Additive Manufacturing allows online process monitoring and simulation, along with the cloud storage of the process and geometry data collected during the material deposition. The analysis of such data allows online and post-deposition identification of eventual process instabilities that can lead to quality problems. Considering the above-mentioned concepts, this work presents a DT architecture based on the ISO 23247-Digital Twin Framework for Manufacturing standard. In this sense, an approach of a Digital Twin framework for metal additive manufacturing process in a robotic cell composed of a robotic arm, positioning table and welding machine is presented and validated, focusing on the collection and cloud storage of both geometrical and process data along with near real-time process simulation.

A digital twin is a virtual representation that accurately replicates its physical counterpart, fostering bi-directional real-time data exchange throughout the entire process lifecycle. For Laser Directed Energy Deposition of Wire (DED-LB/w) additive manufacturing processes, digital twins may help to control the residual stress design in build parts. This study focuses on providing faster-than-real-time and highly accurate surrogate models for the formation of residual stresses by employing neural ordinary differential equations. The approach enables accurate prediction of temperatures and altered structural properties like stress tensor components. The developed surrogates can ultimately facilitate on-the-fly re-optimization of the ongoing manufacturing process to achieve desired structural outcomes. Consequently, this building block contributes significantly to realizing digital twins and the first-time-right paradigm in additive manufacturing.

The key differentiation of digital twins from existing models-based engineering approaches lies in the continuous synchronization between the physical and digital twins through data exchange. The success of digital twins, whether operated automatically or with humans in the loop, hinges on the quality of data, models, and computations, which influences the digital twins’ usability and effectiveness. This paper provides a framework for tracking digital twin performance in Additive Manufacturing (AM) applications and developing quality assessment tools to enhance the development and applications of AM digital twins. This framework is founded on a digital twin development activity model, which identifies the digital objects through the digital twin evolution life cycle, delineating their quality measures and the transfer of any quality-related problems. Quality metrics are defined for each type of digital objects in the framework and computation methods are outlined to track how quality issues in earlier stages affect subsequent activities and digital objects. The data characteristics linked to digital twin effectiveness, as identified by the framework, could serve as key performance indicators for AM data management. Furthermore, understanding the challenges in the uncertainty and quality transition between digital objects can lead to a strategic research agenda for AM digital twins.

Powder bed fusion (PBF) is an emerging metal additive manufacturing (AM) technology that enables rapid fabrication of complex geometries. However, defects such as pores and balling may occur and lead to structural unconformities, thus compromising the mechanical performance of the part. This has become a critical challenge for quality assurance as the nature of some defects is stochastic during the process and invisible from the exterior. To address this issue, digital twin (DT) using machine learning (ML)-based modeling can be deployed for AM process monitoring and control. Melt pool is one of the most commonly observed physical phenomena for process monitoring, usually by high-speed cameras. Once labeled and preprocessed, the melt pool images are used to train ML-based models for DT applications such as process anomaly detection and print quality evaluation. Nonetheless, the reusability of DTs is restricted due to the wide variability of AM settings, including AM machines and monitoring instruments. The performance of the ML models trained using the dataset collected from one setting is usually compromised when applied to other settings. This paper proposes a knowledge transfer pipeline between different AM settings to enhance the reusability of AM DTs. The source and target datasets are collected from the National Institute of Standards and Technology and National Cheng Kung University with different cameras, materials, AM machines, and process parameters. The proposed pipeline consists of four steps: data preprocessing, data augmentation, domain alignment, and decision alignment. Compared with the model trained only using the source dataset, this pipeline increased the melt pool anomaly detection accuracy by 31% without any labeled training data from the target dataset.

Large Format Additive Manufacturing (LFAM) has gained prominence in the aerospace and automotive industries, where topology optimization has become crucial. LFAM facilitates the layer-by-layer production of sizeable industrial components in carbon fiber (CF) reinforced polymers, however 3D printing at large scales results in warpage generation. Printed components are deformed as residual stresses generated due to thermal gradients between adjacent layers. This paper tackles the problem at two different scales: the micro and macroscale. Initially, the microstructure characterization of the thermoplastic ABS matrix composite material enriched with 20% short CF is used in the development of numerical models to understand the mechanical behavior of the studied material. Numerical modeling is performed simultaneously by means of Mean-Field (MF) homogenization methods and Finite Element Analysis (FEA). Outcomes validated with corrected experimental mechanical testing results show a discrepancy in the elastic modulus of 7.8% with respect to FE multi-layer analysis. Micro-level results are coupled with the a macroscopic approach to reproduce the LFAM process, demonstrating the feasibility of the tool in the development of a Digital Twin (DT).

Additive manufacturing (AM), such as Laser Powder Bed Fusion (LPBF), enables intricate fabrication of custom components via layer-by-layer deposition. This study focuses on improving precision in additive manufacturing (AM), specifically Laser Powder Bed Fusion (LPBF), by using regression machine learning, including deep neural networks (DNNs), to predict and optimize melt pool geometry (depth, width, length). Leveraging process parameters and material properties, the goal is to refine the geometry to meet design and performance criteria. Integrating digital twin and IoT technologies enables real-time monitoring and dynamic control. The use of NSGA-II optimization further enhances melt pool behavior analysis. This research aims to advance AM quality, reliability, and efficiency in digital manufacturing.

In robotic multi-axis additive manufacturing, achieving high precision and consistency is crucial. This paper presents a comprehensive calibration procedure and optimization strategy for process parameters in a filament-based extruder mounted on an industrial robot. A digital twin of the manufacturing setup is utilized to simulate and validate the process, enabling precise adjustments and automation of code generation for robotic operations. Experimental calibration artifacts are produced and measured to verify the effectiveness of the proposed approach, highlighting the impact of variables such as hot-end temperature, layer height, and tool center point velocity on print quality. A laser tracker is employed to ensure precise measurement of calibration artifacts, enhancing the accuracy of the calibration process. The results highlight the potential for advancements in the precision and reliability of multi-axis additive manufacturing processes through systematic calibration and optimization, enabling the stable production of complex 3D-printed structures and reduced production waste.

ABSTRACT Additive manufacturing (AM) is expanding for fabricating complex, high-value, and large-size and low-volume parts. While AM holds significant potential for producing a variety of components, it is essential to conduct quality assessment of AM components to ensure their safe and reliable operation. Low-volume fabrication of parts imposes limitations on the number of inspection samples for quality verification. Limited availability of inspection samples can be addressed by nondestructive testing (NDT) of fabricated samples, and the development of digital twin models. In this study, the Process-Structure-Property (P-S-P) of stainless-steel AM parts fabricated via laser powder bed fusion technique are evaluated using a limited number of inspection samples by various material characterization and NDT. The correlation between process parameters and part properties has been assessed. Results indicate the P-S-P relationship where apart from the ultimate strength values, all experimental results display bimodal distributions, featuring two distinct modes. The correlation analysis reveals that higher laser power is associated with increased RF Z-score and Modulus. Utilizing the experimental findings, a FEM serving as a digital twin of the parts is constructed. This model enables the extension of quality assessment for the parts across a broader spectrum and facilitates the examination of P-S-P relations.

No abstract available

Abstract Early detection and correction of defects are critical in additive manufacturing (AM) to avoid build failures. In this paper, we present a multisensor fusion-based digital twin for in-situ quality monitoring and defect correction in a robotic laser-directed energy deposition process. Multisensor fusion sources consist of an acoustic sensor, an infrared thermal camera, a coaxial vision camera, and a laser line scanner. The key novelty and contribution of this work are to develop a spatiotemporal data fusion method that synchronizes and registers the multisensor features within the part's 3D volume. The fused dataset can be used to predict location-specific quality using machine learning. On-the-fly identification of regions requiring material addition or removal is feasible. Robot toolpath and auto-tuned process parameters are generated for defect correction. In contrast to traditional single-sensor-based monitoring, multisensor fusion allows for a more in-depth understanding of underlying process physics, such as pore formation and laser-material interactions. The proposed methods pave the way for self-adaptation AM with higher efficiency, less waste, and cleaner production.

Additive manufacturing is a promising manufacturing process with diverse applications, but ensuring the quality and reliability of the manufactured products are key challenges. The digital twin has emerged as a technology solution to address these challenge, allowing real-time monitoring and control of the manufacturing process. This paper proposes a digital twin system framework for additive manufacturing that integrates machine learning models, employing Unity, OctoPrint, and Raspberry Pi for real-time control and monitoring. Particularly, the system utilizes machine learning models for defect detection, achieving an Average Precision (AP) score of 92%, with specific performance metrics of 91% for defected objects and 94% for non-defected objects, demonstrating high efficiency. The Unity client user interface is also developed for control and visualization, facilitating easy additive manufacturing process monitoring. This research article presents a detailed description of the proposed digital twin framework and its workflow for implementation, the machine learning models, and the Unity client user interface. It also demonstrates the effectiveness of the integrated system through case studies and experimental results. The main findings show that the proposed digital twin system met its functional requirements and effectively detects defects and provides real-time control and monitoring of the additive manufacturing process. This paper contributes to the growing field of digital twin technology and additive manufacturing, providing a promising solution for enhancing the quality and reliability in the field of additive manufacturing.

No abstract available

Modern agricultural intelligent manufacturing faces critical challenges including low automation levels, safety hazards in high-temperature processing, and insufficient production data integration. Digital twin technology and 3D printing offer promising solutions through real-time virtual–physical synchronization and customized equipment manufacturing, respectively. However, existing research exhibits significant limitations: inadequate real-time synchronization mechanisms causing delayed response, poor environmental adaptability in unstructured agricultural settings, and limited human–machine collaboration capabilities. To address these deficiencies, this study develops a digital twin-driven intelligent sorting system for 3D-printed agricultural tools, integrating an Articulated Robot Arm, 16 industrial-grade 3D printers, and the Unity3D 2024.x platform to establish a complete “printing–sorting–warehousing” digitalized production loop. Unlike existing approaches, our system achieves millisecond-level bidirectional physical–virtual synchronization, implements an adaptive grasping algorithm combining force control and thermal sensing for safe high-temperature handling, employs improved RRT-Connect path planning with ellipsoidal constraint sampling, and features AR/VR/MR-based multimodal interaction. Validation testing in real agricultural production environments demonstrates a 98.7% grasping success rate, a 99% reduction in burn accidents, and a 191% sorting efficiency improvement compared to traditional methods, providing breakthrough solutions for sustainable agricultural development and smart farming ecosystem construction.

This paper proposes a digital twin-based 3D printer and model monitoring method, aiming to address the shortcomings of the current digital twin technology, which can only perform one-dimensional monitoring. The core of the research lies in constructing a comprehensive 3D printer digital twin model covering geometric, physical, and data models, to realize an all-round mapping of 3D printers. In addition, this paper proposes a monitoring method based on the improved YOLOv5 and Long Short-Term Memory (LSTM) network, which is capable of tracking and analyzing the status of 3D printing models and devices in real-time to achieve multi-dimensional monitoring of the printing process. Through experimental verification, the proposed method shows good feasibility and effectiveness, and can significantly improve the monitoring capability and response speed of the 3D printing process. The research results provide new ideas and solutions for the future development of intelligent manufacturing and promote the in-depth application of digital twin technology in the field of 3D printing.

No abstract available

No abstract available

No abstract available

No abstract available

Additive manufacturing (AM, also known as 3D printing) is a bottom–up process where variations in process conditions can significantly influence the quality and performance of the printed parts. Digital twin (DT) technology can measure process parameters and printed part characteristics in real-time, achieving online monitoring, analysis, and optimization of the AM process. Existing DT research on AM focuses on simulating the printing process and lacks real-time defect detection and twinning of actual printed objects, which hinders the timely detection and correction of defects. This study developed a DT system for fused deposition modeling (FDM) AM technology that not only accurately simulates the printing process but also performs real-time quality monitoring of the printed parts. A laser profilometer and industrial camera were integrated into the printer to detect and collect real-time morphological data on the printed object. The custom-developed DT software could convert the morphological data of the printed parts into a DT model. By comparing the DT model of the printed object with its three-dimensional model, defect detection of the printed parts was achieved, where the quality of the printed parts was evaluated using a defect percentage index. This study combines DT and AM to achieve process quality monitoring, demonstrating the potential of DT technology in reducing printing defects and improving the quality of printed parts.

Aerosol Jet (AJ) printing is a versatile additive manufacturing technique capable of producing high-resolution interconnects on both 2D and 3D substrates. The AJ process is complex and dynamic with many hidden and unobservable states that influence the machine performance, including aerosol particle diameter, aerosol carrier density, vial level, and ink deposition in the tube and nozzle. Despite its promising potential, the widespread adoption of AJ printing is limited by inconsistencies in print quality that often stem from variability in these hidden states. To address these challenges, we develop a digital twin model of the AJ process that offers real-time insights into the machine's operations. The digital twin is built around a physics-based macro-model created through simulation and experimentation. The states and parameters of the digital model are continuously updated using probabilistic sequential estimation techniques to closely align with real-time measurements extracted from the AJ system's sensor and video data. The result is a digital model of the AJ process that continuously evolves over a physical machine's lifecycle. The digital twin enables accurate monitoring of unobservable physical characteristics, detects and predicts anomalous behavior, and forecasts the effect of control adjustments. This work presents a comprehensive end-to-end digital twin framework that integrates customized computer vision techniques, physics-based macro-modeling, and advanced probabilistic estimation methods to construct an evolving digital representation of the AJ equipment and process. While the methodologies are customized for aerosol jet printing, the process for constructing the digital twin can be applied for other advanced manufacturing techniques.

This paper presents the development and experimental validation of a digital twin for a selective laser melting (SLM) system designed for operation under constrained technical conditions. The proposed design includes a custom-built powder feeding and leveling mechanism, along with a telemetry module enabling realtime acquisition and analysis of thermal data. The digital twin is implemented as an integrated model that combines thermomechanical analysis, G-code execution, and corrective control of printing parameters. The mathematical foundation is based on the transient heat conduction equation with a volumetric heat source and a Gaussian distribution of laser power density.An experimental setup employing pyrometers and thermocouples in the melt zone was used to validate the model. The comparison of simulated and measured data showed a mean absolute error of less than 2.5 °C. The application of the digital twin resulted in a 12–15% reduction in residual stresses, as confirmed by X-ray diffraction analysis. The developed system demonstrates high efficiency in predictive quality control and can be integrated into adaptive control loops. The approach aligns with the Industry 4.0 concept, offering increased stability, repeatability, and reliability in SLM processes

In the digital age, the digital twin eliminates physical barriers and risks, facilitating seamless activities in both real and virtual worlds. In the context of additive manufacturing, testing 3D printers can be resource-intensive and prone to printing issues. This research introduces a digital twin-based system that employs the innovative ensemble 3D-AmplifAI algorithm for fault monitoring in 3D printers. The system continuously monitors real-time temperature values and detects faults to prevent potential damage to the printer. Through an ensemble method, the 3D-AmplifAI algorithm combines multiple machine learning models to enhance fault detection in 3D printers. The digital twin environment, developed using Unity, serves as the bridge connecting the physical printer to the virtual world. Comparative evaluations against state-of-the-art algorithms, including Ridge Regression, XGBoost, InceptionTime, Time Series Transformer (TST), Rocket Ridge, Logistic Regression, Rocket XGBoost, ResNet, and Rocket Ridge Regression, demonstrate the superior performance of the 3D-AmplifAI algorithm in terms of accuracy, precision, recall, and F1-score.

In the age of digitalization, virtually everything around us is on the cloud. The metaverse allows people to perform the things they do in the real and virtual worlds. In the virtual world of the metaverse, everything from work to communication is on demand and connected. One of the advantages of the metaverse is the lack of physical barriers in development. Performing tests on expensive equipment are no longer expensive because companies do not have to purchase physical hardware. In addition, the risks in performing tests are virtually eliminated. Hardware errors may cause damage to a physical system, but in a metaverse, all one needs to do is reboot the virtual machine and have it reloaded with an operating system. In additive manufacturing, testing a 3D printer may take time and resources. Furthermore, miscalculations and unfavorable environmental conditions may lead to printing issues. Developing a simulation tool to test a 3D printer in variable environments is explored to address possible problems with additive manufacturing. The behavior of the digital twin will be diagnosed using various deep learning algorithms implemented in various studies. CNN-BiLSTM has proven to be the most accurate of the deep learning algorithms explored.

This work presents a novel predictive model and a virtual metrology for Flexible and Printable Electronic (FHE) device fabrication using deep learning techniques. Our method leverages the Deep Neural Network (DNN) based model Pix2Pix to predict outcomes of 3D inkjet printing from layout images. We prepared an experimental FHE dataset that captures printed electronics’s intricate patterns and printing process variability. By training the Pix2Pix model on this dataset, we successfully demonstrated its ability to learn and accurately predict the outcome of inkjet printing of structures using a conductive ink. The best mean absolute errors (MAE) for width and gap prediction are 5.07 microns and 7.99 microns. This approach enables improvement of efficiency and accuracy of printed electronics, paving the way for advancements in FHE manufacturing technology and workflows.

Digital twin and additive layer manufacturing plays a vital role of the fourth industrial revolution. Digital twin is the ideal solution for data-driven optimisation of additive manufacturing challenges. It is helpful in understating, analysing, and improving 3D printing machining process variables and consequently reducing the number of trial-and-error and component’s non-conformance and shorten product development lead time. Furthermore, the development of genuine digital twin still requires more research efforts to develop a thorough understanding of its concept, data management framework, and development techniques. Therefore, this paper aims to capture important data-driven digital twin requirements for additive layer manufacturing through a systematic approach by identifying the requirements, analysing technologies and processes for digital twin development. The main novelty of this research is applying a holistic approach to build digital twin of additive manufacturing process by capturing the requirements from both literature review and world-class aerospace industrial experts. Overall, the captured requirements will not only serve industries as a basis for implementing digital twin for additive manufacturing and modernize existing data management systems but also opens new research areas in the digital twin domain.

As additive manufacturing by fused filament fabrication has gained popularity, computational analysis has become fundamental in predicting the mechanical behavior of 3D models. This paper proposes the development of a method for the finite element (FE) simulation of 3D-printed parts, implementing model design reverse engineering using G-code to obtain their digital twins (DTs). Samples were printed under the ASTM D638 standard with different nozzle diameters and layer heights, which allowed them to be mechanically characterized by tensile tests. The tensile tests determined that the diameter of the nozzles used (between 0.2 mm and 1.0 mm) influences the material’s tensile strength. The greater the diameter, the greater the stiffness, which translates into a change in the Young’s modulus, as well as greater tensile strength and thus a reduction of the deformation, for which a value of 2.66 ± 0.6 % was obtained, i.e., the filament diameter did not influence this aspect. After carrying out the reverse engineering process of the samples to obtain DTs of the physical models, the printing G-code was used with the help of a Python script for their conversion to trajectories. These trajectories were introduced into Rhinoceros software with the Grasshopper add-on to obtain the reconstructed 3D models. The deposited filament profile used to reconstruct the DT was obtained by microscopy of the section of the physical samples. The predominant profile observed was that of a flattened oval. FE simulation was then carried out, obtaining a similarity of 90% between the simulated and mechanical tests, which validated the proposed method of predicting mechanical stresses in printed 3D elements.

More and more mission-critical industrial systems are deploying hardware (e.g. FPGA) based controls these days for increased cyber-security and reliability eliminating cyber-risks which is higher for software based controls. These industrial systems include but not limited to healthcare devices, power supplies for 3D metal printing, wind turbine controls etc. Most of their controllers are based on proportional-integral-derivative-feedforward (PIDF) algorithms. This paper presents methodology for modeling, designing, and developing field programmable gate arrays (FPGA) based PIDF controller for a representative industrial system (referred to as a physical twin). A digital twin of the system is identified to optimally design and develop the controller for implementing on a resource-constraint FPGA. We also validate the digital electronic circuit of the controller on the physical twin and observe successful tracking of the target by that physical system. The observed error between the physical and digital twin control effort is ±0.2%. That error is caused due to the variance in sampling time of their control efforts. The physical twin control effort is logged using a JTAG cable. Its sampling time is found to be variable while the sampling time used for digital twin simulation is fixed causing the mismatch between digital and physical twin control efforts.

Metal Additive manufacturing is increasingly employed in the production of bus components due to its flexibility in producing complex, lightweight structures. It enables prototype fabrication without costly tooling, supports bionic designs for weight reduction, and shortens lead times. However, its high energy consumption makes it more expensive than conventional manufacturing. A key opportunity for energy consumption reduction lies in optimizing preheating, cooling, and build plate temperature control, as these parameters directly affect the printing process's efficiency, print quality, and overall energy costs. This Work-in-Progress Paper proposes an assisted simulation approach for the TruPrint 3000 metal 3D printer, leveraging real-time machine state monitoring and simulation models to evaluate NC-code execution. This approach acts as a fundamental building block for the development of a digital twin, enabling the analysis and experimentation of new parameter configurations that reduce electrical power consumption. The study aims to provide insights into optimizing thermal management strategies and enhancing energy efficiency in metal additive manufacturing.

This work introduces a novel architecture of federated learning (FL)-enabled digital twin (DT) for the smart additive manufacturing industry, especially 3D printing. The proposed architecture tackles the previous limitation of the centralized approach that requires a large number of communication costs by efficiently updating the fault detection model on each server with distributed learning methods. A CNN-based model is also proposed to efficiently learn sensory data from a 3D printer for a fast and reliable fault detection model. To provide a robust system in intelligent manufacturing, a DT platform is also designed for seamless monitoring and control purposes. The proposed DT platform is able to initiate, monitor, and terminate the 3D printing process of physical assets via a virtual environment. Based on the simulation results, the FL process demonstrates that the proposed CNN-based model is superior to other DL models with 8% accuracy enlargement while maintaining the low training period. Furthermore, experimental work is conducted to evaluate the proposed architecture with real-world devices. Finally, the findings indicate that the overall latency given by the proposed system is relatively low, with an average of 1026.16 ms from the physical 3D printer to the DT platform.

The paper discusses the method of obtaining a digital passport of the material and the development of a digital twin of the product at various stages of its manufacture. The object of research is a conical mesh structure. The subject of research is the processes occurring in the product at the manufacturing stages. The following main stages of creating a mesh structure were considered in the work: 3D printing of a workpiece tooling, laying out a carbon unidirectional material, heating and impregnation of the preform with a binder, polymerization of the binder, and warpage of the product geometry. An algorithm for determining the properties of materials and their calibration using modern software and hardware and universal equipment was described. Modeling of the tooling 3D printing process was carried out in the Ansys software package. Step-by-step technological modeling of the transfer molding process was carried out using the ESI PAM-COMPOSITE software package. The result of the simulation is the optimal technological manufacturing parameters and the geometry of the shaping tooling with anticipation of warping.

The present work aims to develop a digital twin system for a small-scale robot workstation for intelligent manufacturing, based on ROS and Unity 3D. Such digital twin system can be used to remotely visualize, monitor and control the manufacturing process, which is of great significance in the development of industrial automation and intelligent manufacturing. In the present work, the system is preliminarily developed for a pick-and-place task. To extend this framework enabling it to be more intelligent, we have considered integrating, in our framework, the 3D vision perception system with deep learning based vision algorithms, especially for perception of complex objects. The purpose is primarily for real time monitoring of dynamic manufacturing processes such as detecting moving workpiece, 3D formation of complex workpieces in 3D printing, etc., the data of which cannot be obtained from controllers of manufacturing stations. The 3D vision system and the developed algorithm are based on point cloud perception.

In the 3D printing process, various error factors can affect the accuracy of the final printing quality. However, current 3D printing error compensation methods have limited effects and usually cannot work in real-time. The 3D printing error compensation process is modeled as a Markov decision process (MDP) in this paper, and Deep Reinforcement Learning (DRL) is applied for dynamic error compensation. This method learns autonomously through trial and error by interacting with the printing environment, which makes it adaptable to various types of 3D printers without specific training. Then, we simulate the digital light processing (DLP) 3D printing. Due to the huge state and action space of sliced images, applying the DRL algorithm to DLP is challenging. We propose an error compensation method based on morphological image operation and use Autoencoder to extract error features to reduce the state space. We then implement our method using a Twin Delayed Deep Deterministic policy gradient algorithm (TD3). The results demonstrate the effects of our method in compensating 3D printing errors.

3D-inkjet-printing is just beginning to take off in the optical field. Advantages of this technique include its fast and cost-efficient fabrication without tooling costs. However, there are still obstacles preventing 3D inkjet-printing from a broad usage in optics, e.g., insufficient form fidelity. In this article, we present the formulation of a digital twin by the enhancement of an optical model by integrating geometrical measurement data. This approach strengthens the high-precision 3D printing process to fulfil optical precision requirements. A process flow between the design of freeform components, fabrication by inkjet printing, the geometrical measurement of the fabricated optical surface, and the feedback of the measurement data into the simulation model was developed, and its interfaces were defined. The evaluation of the measurements allowed for the adaptation of the printing process to compensate for process errors and tolerances. Furthermore, the performance of the manufactured component was simulated and compared with the nominal performance, and the enhanced model could be used for sensitivity analysis. The method was applied to a highly complex helical surface that allowed for the adjustment of the optical power by rotation. We show that sensitivity analysis could be used to define acceptable tolerance budgets of the process.

This paper presents IoT based real-time monitoring system for 3D printing. However, additive manufacturing or 3D printing allows complex solutions and innovations due to its superior advantages over conventional material removal processes. Monitoring and quality control of the printing process are still challenging to implement. There is a need for research efforts in creating IoT-based monitoring, control systems, smart manufacturing systems, and digital twin for the additive manufacturing process. The paper provides a real-time monitoring system for 3D printing based on the data receiving an approach from different embedded sensors in a real-time regime. The sensors are composed of thermocouples, accelerometers, thermistors, and cameras. Matlab-Arduino support package tools were employed for visualization data from sensors. The advantages of our proposed monitoring system are the simplicity of design and the availability of embedded sensors which can track actual data from the 3D printer machine and interface it through remote monitoring systems.

No abstract available

Background: The growth of digital technologies, such as IoT and Cyber-Physical Production Systems (CPPS), is driving the Fourth Industrial Revolution. Additive Manufacturing (AM) plays a key role in this revolution but raises concerns regarding intellectual property (IP) theft and quality control due to the lack of physical interaction with printers. Methodology: This paper proposes a cloud-integrated cyber-physical system (CPS) for defect detection in AM using deep learning models. Two models, AlexNet and InceptionV3, were trained on a dataset of $1,5573 \mathrm{D}$ printer images. The models were deployed in a cloud environment for real-time monitoring and adaptive retraining. Results: Performance evaluation showed that AlexNet achieved 99% classification accuracy, while InceptionV3 achieved 97 %, significantly outperforming baseline models such as ResNet101 (86%), VGG16 (95%), and DNN (82%). The proposed models demonstrated superior defect detection and quality control capabilities in AM. Conclusion: The proposed CPS framework integrates Industry 4.0 technologies for automated and remote defect detection in AM.

Powder Bed Fusion with Laser Beam of Metals (PBF-LB/M) faces challenges in reproducibility and quality assurance, even for widely applied alloys like AlSi10Mg. This work introduces a digital provenance framework for PBF-LB/M, showcased through the EOS M 300–4 multi-laser machine. An Extract, Transform, Load (ETL) pipeline autonomously captures machine data, including scan vectors as well as process signals, and organizes them into a Digital Shadow (DS). The DS is further extended by external data sources, such as Melt Pool Monitoring (MPM), to enable comprehensive analysis and root cause identification. This approach ensures continuous data representation and facilitates the development of new quality metrics. Moreover, the framework enhances quality assurance and traceability, supports compliance with industry standards, and improves productivity. It also enables more precise cost calculations and predictive maintenance. By addressing these challenges, the framework is essential for advancing PBF-LB/M in industrial applications, achieving greater consistency and scalability in production.

Fused Filament Fabrication (FFF) is one of the most common additive manufacturing tools. With the aid of the open-source community, constant improvements in the mechanical build, electronic components, thermal design, and software algorithms allowed for increased reliability and part quality. However, issues persist in extrusion control, manifesting as deposition errors that degrade the visual and mechanical properties of the printed part. Current feedforward control techniques use a simplified model of the extrusion process as a first order system. However, such assumption fails to capture nonlinear extrusion dynamics. Therefore, to build a controller that can improve quality under varied extrusion velocities, a more descriptive model is required. In this work, we introduce a cyberphysical system with industrial grade motion stages and controllers that can record feedback information from the drive motors of the 3D printer such as position, velocity, acceleration, and motor current. Additionally, a temperature sensor near the nozzle exit records the surface temperature of the nozzle. The printed geometry is scanned using a laser line profilometer. Using this system, single bead lines were printed with sinusoidally varying bead width. To create a model for extrusion dynamics, a nonlinear auto regressive exogenous (NARX) neural net is employed to predict the bead area given the extrusion velocity, tangent velocity, temperature, and motor current. The resulting trained NARX net provided predictions with low mean absolute error values between 0.0923 and 0.0994 using testing sets that are independent from training sets.

Additive Manufacturing (AM) for Industry 4.0 requires a number of networking, integrated control and cloud technologies to enhance the connectivity and performance. There is now, however, a dearth of widely available solutions for Cloud-based AM. Regrettably, the repeatability and monitoring of quality in the production process are not sufficiently dependable to be used in mass production. Thus, quality monitoring can be used as an important tool in AM for defect detection to minimize material and time waste during printing. Therefore, this study is aimed to provide a cyber physical system-based AM framework of evaluating the reliability of the automatically printed components by including sensor to capture images and machine learning approach in an industry 4.0 environment. Images of semi-finished parts are taken while the extruder's vibration goes into the above threshold level vibration. The proposed system incorporates an accelerometer and camera module connected with raspberry pi attached with 3D Printer. Azure machine learning studio, connected to the Azure IoT hub, where a machine learning method, convent, is proposed to classify the parts into the ‘good’ or ‘defective’ category. Thus, experimental runs are reduced in Fused Deposition Modeling (FDM) AM printing through parametric optimization using Taguchi Design studies approach. Finally, the developed model has been validated using variance analysis (ANOVA) and signal-to-noise ratio (S/N), in an intelligent environment with Industry 4.0 for defect-free production. This Framework have the potential to successfully implementation of cloud based closed loop quality monitoring system for FDM based Additive manufacturing process in industry 4.0 environment.

This paper presents a model-based feed rate calibration methodology for robotic fused deposition modeling (FDM), developed to improve dimensional accuracy in thin-walled structures. In single-perimeter geometries, where deviations in material deposition directly influence wall thickness and surface quality, precise feed rate control is essential. The proposed computational approach establishes a volumetric relationship between the robot motion speed, layer height, and target extrusion width, enabling accurate determination of the material feed rate from measurable process parameters rather than empirical tuning. The system was implemented within a Cyber-Physical Systems (CPS) architecture integrating independent control of the robot and extrusion unit, allowing synchronized motion and material deposition. Experimental validation on cuboid and cylindrical specimens confirmed the model’s effectiveness, achieving minimum wall-thickness deviations of 0.027 mm for the cuboid and 0.065 mm for the cylindrical sample. The methodology provides a practical framework for feed rate calibration in robotic FDM systems, supporting consistent extrusion quality, reduced material variation, and faster process setup. Although the methodology was tested on a specific extrusion system, the approach is scalable to large-format robotic additive manufacturing (RAM) and adaptable to various extrusion materials and configurations.

Industry 5.0 is an emerging value-driven manufacturing model in which human–machine interface-oriented intelligent manufacturing is one of the core concepts. Based on the theoretical human–cyber–physical system (HCPS), a reference framework for human–machine collaborative additive manufacturing for Industry 5.0 is proposed. This framework establishes a three-level product–economy–ecology model and explains the basic concept of human–machine collaborative additive manufacturing by considering the intrinsic characteristics and functional evolution of additive manufacturing technology. Key enabling technologies for product development process design are discussed, including the Internet of Things (IoT), artificial intelligence (AI), digital twin (DT) technology, extended reality, and intelligent materials. Additionally, the typical applications of human–machine collaborative additive manufacturing in the product, economic, and ecological layers are discussed, including personalized product design, interactive manufacturing, human–machine interaction (HMI) technology for the process chain, collaborative design, distributed manufacturing, and energy conservation and emission reductions. By developing the theory of the HCPS, for the first time its core concepts, key technologies, and typical scenarios are systematically elaborated to promote the transformation of additive manufacturing towards the Industry 5.0 paradigm of human–machine collaboration and to better meet the personalized needs of users.

The material extrusion 3D printing process known as fused deposition modeling (FDM) has recently gained relevance in the additive manufacturing industry for large-scale part production. However, improving the real-time monitoring of the process in terms of its mechanical properties remains important to extend the lifespan of numerous critical applications. To enhance the monitoring of mechanical properties during printing, it is necessary to understand the relationship between temperature profiles and ultimate tensile strength (UTS). This study uses a cyber–physical production system (CPPS) to analyze the impact of four key thermal parameters on the tensile properties of polylactic acid (PLA). Layer thickness, printing speed, and extrusion temperature are the most influential factors, while bed temperature has less impact. The Taguchi L-9 array and the full factorial design of experiments were implemented along with the deposited line’s local fused temperature profile analysis. Furthermore, correlations between temperature profiles with the bonding strength during layer adhesion and part solidification can be stated. The results showed that layer thickness is the most important factor, followed by printing speed and extrusion temperature, with very close influence between each other. The lowest impact is attributed to bed temperature. In the experiments, the UTS values varied from 46.38 MPa to 56.19 MPa. This represents an increase in the UTS of around 17% from the same material and printing design conditions but different temperature profiles. Additionally, it was possible to observe that the influence of the parameter variations was not linear in terms of the UTS value or temperature profiles. For example, the increase in the UTS at the 0.6 mm layer thickness was around four times greater than the increase at 0.4 mm. Finally, even when it was found that an increase in the layer temperature led to an increase in the value of the UTS, for some of the parameters, it could be observed that it was not the main factor that caused the UTS to increase. From the monitoring conditions analyzed, it was concluded that the material requires an optimal thermal transition between deposition, adhesion, and layer solidification in order to result in part components with good mechanical properties. A tracking or monitoring system, such as the one designed, can serve as a potential tool for reducing the anisotropy in part production in 3D printing systems.

Additive Manufacturing (AM) has revolutionized industries by enabling the production of complex, customized products with unparalleled efficiency. However, the increasing reliance on AM in critical sectors such as aerospace, healthcare, and defense has exposed it to significant cybersecurity and reliability challenges, including intellectual property theft, process sabotage, and data tampering. These vulnerabilities as well as reliability issues can compromise product integrity, safety, and operational continuity, posing severe risks to both industry and national security. In this work, we propose an novel methodology for modeling the AM process chain as a Cyber-Physical System (CPS) using multi-modal data structured in a graph format. Our methodology leverages Graph Neural Networks (GNNs) to detect and localize anomalies across diverse data modalities, enabling precise identification of both the nature and source of attack/fault. By integrating data fusion, advanced anomaly classification, and localization techniques, our solution provides a robust methodology for enhancing the security and reliability of AM processes, ensuring their safe deployment in critical applications. Furthermore, the proposed technique is adaptable to other industrial systems, underscoring its potential for broader impact in securing critical infrastructure.

The paper considers the problems of technologies for creating products using 3D printing based on the application of a cyber-physical approach. An analysis of 3D printing management was conducted, problems in operation were identified that reduced quality or led to various defects in parts. Existing methods for eliminating the selected problems in 3D printer management were analyzed. The object of the study is the process of printing a part on a 3D printer. The subject of the study is a method and algorithm for operational management of 3D printing based on monitoring data. The relevance of the work lies in the fact that modern additive manufacturing is a complex process with electronics, software and network connection, requires constant monitoring of changes in operating parameters and adaptation of cyber-physical system management to improve performance. The goal is to reduce the defectiveness of products obtained on a 3D printer through operational management based on monitoring data. To solve this problem, a new method for reducing defects when using a 3D printer has been developed, based on an algorithm for collecting data from monitoring the external environment, analyzing the data obtained and controlling the 3D printer, a model based on graph theory has been built, allowing to predict the occurrence of defects by analyzing the monitoring data, an algorithm for controlling the 3D printer has been built using the analysis of temperature and camera monitoring data, a software product has been developed through the VisualStudio platform in the C# language, due to which it can be integrated into an automated monitoring system for a 3D printer, which has been successfully tested. The use of a 3D printer defect reduction system has proven the advantages of its use and reducing defects when printing on a 3D printer by analyzing the environmental monitoring data. The program has prospects for further development and implementation in systems for improving the quality and efficiency of production.

This paper explores the integration of machine-to-machine calibration within a cyber-physical system. As Industry 4.0 revolutionizes manufacturing processes, the focus shifts to interconnected machines for enhanced efficiency. This study addresses the critical role of cyber-physical systems in managing 3D printers for additive manufacturing and exploring automatic pressure advance calibration using machine vision. An instrumented 3D printer was used to calibrate an un-instrumented 3D printer, showing a systematic calibration process. Test beads were printed on the instrumented system and the optimal pressure advance time constant value was automatically found. The optimal pressure advance time constant value was then applied to a non-instrumented system to print a test part. A case study on ball valve production illustrates the effectiveness of our approach in mitigating defects and optimizing print quality. The results of the case study demonstrate the effectiveness of our automated calibration system.

The challenges in a manufacturing system are lack of timely, accurate, and lack of information to featured product prediction, shop floor resources, product flow, product inspection, product status to customer, product delivery status and factory adaption for customized product. The proposed idea is to design IoT visualization based Smart Factory for Additive Manufacturing System (ISFAMS) that creates a way towards progressively from traditional automation to a fully connected mass customization and flexible cyber-physical system. The ISFAMS utilize a consistent stream of information from associated tasks and creating frameworks to learn and adjust factory productions to new requests from the customer. The system utilizes the Industrial Controller to control the operation of individual systems and sequence of product flow in the Smart Factory setup. The wireless sensor network acquires real-time manufacturing information and information is stored, accessed and visualized using cloud computing. The vision system and automated platform enable the inspection of products shape and dimensions based on the machine learning approach and to transfer the product from section to section and separate the product for packaging section. This digitization of manufacturing system increases flexibility, reliability, smart sensing and control, resource wastage, easy access to manufacturing information and logistics management.

3D printing systems have expanded the access to low cost, rapid methods for attaining physical prototypes or products. However, a cyber attack, system error, or operator error on a 3D printing system may result in catastrophic situations, ranging from complete product failure, to small types of defects which weaken the structural integrity of the product, making it unreliable for its intended use. Such defects can be introduced early-on via solid models or through G-codes for printer movements at a later stage. Previous works have studied the use of image classifiers to predict defects in real-time as a print is in progress and also by studying the printed entity once the print is complete. However, a major restriction in the functionality of these methods is the availability of a dataset capturing diverse attacks on printed entities or the printing process. This paper introduces a visual inspection technique that analyzes the amplitude and phase variations of the print head platform arising through induced system manipulations. The method uses an image sequence of a 3D printing process captured via an off the shelf camera to perform an offline multi-scale, multi-orientation decomposition to amplify imperceptible system movements attributable to a change in system parameters. The authors hypothesize that a change in the amplitude envelope and instantaneous phase response as a result of a change in the end effector translational instructions, to be correlated with an AM system compromise. A case study is presented that tests the hypothesis and provides statistical validity in support of the method. The method has the potential to enhance the robustness of cyber-physical systems such as 3D printers that rely on secure, high quality hardware and software to perform optimally.

Smart Hybrid Manufacturing (SHM)—the coordinated integration of additive, subtractive, lean, and smart control capabilities—has emerged as a central pathway for enabling responsive, high-quality, and resource-efficient production in modern manufacturing systems. Yet, empirical evidence explaining how these multidimensional capability foundations contribute to productivity, quality, efficiency, and agility remains limited. This study addresses this gap by developing and testing a comprehensive capability–performance model using data from a heterogeneous sample of industrial plants and hybrid manufacturing cells operating across multiple sectors, complexity classes, and volatility conditions. The research examines how additive and subtractive process maturity, lean–smart integration governance, and cyber-physical smart control jointly shape agile production system performance (APSP), while also assessing the mediating role of smart control capability and the moderating role of lean maturity. Additional multi-group comparisons evaluate whether SHM performance impacts differ across hybridization architectures and environmental turbulence. Measurement model evaluation demonstrated strong psychometric robustness. All constructs achieved acceptable levels of internal consistency (Cronbach’s α ≥ 0.85; CR ≥ 0.90), convergent validity (AVE ≥ 0.70), and discriminant validity across Fornell–Larcker, HTMT, and CFA criteria. The overall measurement structure supported a multidimensional representation of SHM capability, as well as second-order factors for SHM Capability (SHMC) and Agile Production System Performance (APSP). Model fit indices (CFI = 0.957; TLI = 0.948; RMSEA = 0.045; SRMR = 0.037) indicated strong alignment between data and the hypothesized measurement structure. Structural equation modeling results provided strong support for the hypothesized relationships. SHM Capability exhibited a significant and substantial effect on APSP (β = 0.62), explaining more than half of the observed variance in agile system performance. Each SHM sub-dimension demonstrated its expected directional relationship with specific performance outcomes: additive maturity predicted agility, subtractive finishing predicted quality, and lean–smart integration predicted productivity and efficiency. Smart Control Maturity partially mediated the SHM → quality relationship, confirming that in-situ sensing and closed-loop correction serve as essential mechanisms for enabling conformance and reducing variability. Lean maturity significantly moderated the SHM → productivity and SHM → agility relationships, indicating that SHM benefits are amplified under strong pull-flow discipline and structured continuous improvement.

Cyber-Physical Production Systems (CPPS) will usher a new era of smart manufacturing. However, CPPS will be vulnerable to cross-domain attacks due to the interactions between the cyber and physical domains. To address the challenges of modeling cross-domain security in CPPS, we are proposing GAN-Sec, a novel conditional Generative Adversarial Network based modeling approach to abstract and estimate the relations between the cyber and physical domains. Using GAN-Sec, we are able to determine if various security requirements such as confidentiality, availability, and integrity are met. We provide a security analysis of an additive manufacturing system to demonstrate the applicability of GAN-Sec.

The article discusses an approach of creating an Industrial Cyber-Physical Platform in a research laboratory to support the application of Industrial Cyber-Physical Systems technologies in production processes. The research is based on the processes of small series production using injection molding with polymer forming parts made with the use of additive technology. The article considers the issues of improving physical production processes using information generated by identification of digitalized objects in the production system. The need for using process simulation systems and the possibility of processes improvement by collecting and analyzing digitalized data at all stages of technological manufacturing preparation are noted.

No abstract available

No abstract available

No abstract available

The use of digital twin and shadow concepts for industrial material processes has introduced new approaches to bridge the gap between physical and cyber manufacturing processes. Consequently, many multidisciplinary areas, such as advanced sensor technologies, material science, data analytics, and machine learning algorithms, are employed to create these hybrid systems. Meanwhile, new additive manufacturing (AM) processes for metals and polymers, based on emerging technologies, have shown promise for the manufacturing of sophisticated parts with complex geometries. These processes are undergoing a major transformation with the advent of digital technology, hybrid physical-data-driven modeling, and fast-reduced models. This study presents a fresh perspective on hybrid physical-data-driven and reduced order modeling (ROM) techniques for the digitalization of AM processes within a digital twin concept. The main contribution of this study is to demonstrate the benefits of ROM and machine learning (ML) technologies for process data handling, optimization/control, and their integration into the real-time assessment of AM processes. Therefore, a novel combination of efficient data-solver technology and an architecturally designed neural network (NN) module is developed for transient manufacturing processes with high heating/cooling rates. Furthermore, a real-world case study is presented, showcasing the use of hybrid modeling with ROM and ML schemes for an industrial wire arc AM (WAAM) process.

Additive Manufacturing (AM) has been widely used in the industrial field. Failures in the manufacturing process may lead to a decrease in product quality and resource waste. Existing research has focused on predicting failures in additive manufacturing. But most of them lack accurate modeling of complex dynamic processes and fail to effectively integrate real-time data and simulation models. To address this challenge, we proposed a novel failure prediction method WTCAutoFormer for additive manufacturing with the additive manufacturing digital twin platform. Firstly, we utilize Wavelet Transform Convolution (WTC) to preprocess the multidimensional sensor data collected during the additive manufacturing process. WTC extracts high-frequency and low-frequency features with wavelet transform to more accurately capture abnormal fluctuations in sensor sequences. Subsequently, we adopt AutoFormer to extract hidden features in sensor sequences. The experimental results show that the proposed method has significant advantages in the accuracy of failure prediction. Compared to traditional prediction methods, WTC-AutoFormer can more accurately identify potential problems and provide warnings, thereby reducing risks and costs in the production process.

Digital twins (DTs) are an emerging capability in additive manufacturing (AM), set to revolutionize design optimization, inspection, in situ monitoring, and root cause analysis. AM DTs typically incorporate multimodal data streams, ranging from machine toolpaths and in-process imaging to X-ray CT scans and performance metrics. Despite the evolution of DT platforms, challenges remain in effectively inspecting them for actionable insights, either individually or in a multidisciplinary, geographically distributed team setting. Quality assurance, manufacturing departments, pilot labs, and plant operations must collaborate closely to reliably produce parts at scale. This is particularly crucial in AM where complex structures require a collaborative and multidisciplinary approach. Additionally, the large-scale data originating from different modalities and their inherent 3D nature pose significant hurdles for traditional 2D desktop-based inspection methods. To address these challenges and increase the value proposition of DTs, we introduce a novel virtual reality (VR) framework to facilitate collaborative and real-time inspection of DTs in AM. This framework includes advanced features for intuitive alignment and visualization of multimodal data, visual occlusion management, streaming large-scale volumetric data, and collaborative tools, substantially improving the inspection of AM components and processes to fully exploit the potential of DTs in AM.

No abstract available

No abstract available

Additive manufacturing is often used in rapid prototyping and manufacturing, allowing the creation of lighter, more complex designs that are difficult or too expensive to build using traditional manufacturing methods. This work considers the implementation of a novel digital twin ecosystem that can be used for testing, process monitoring, and remote management of an additive manufacturing–fused deposition modeling machine in a simulated virtual environment. The digital twin ecosystem is comprised of two approaches. One approach is data-driven by an open-source 3D printer web controller application that is used to capture its status and key parameters. The other approach is data-driven by externally mounted sensors to approximate the actual behavior of the 3D printer and achieve accurate synchronization between the physical and virtual 3D printers. We evaluate the sensor-data-driven approach against the web controller approach, which is considered to be the ground truth. We achieve near-real-time synchronization between the physical machine and its digital counterpart and have validated the digital twin in terms of position, temperature, and run duration. Our digital twin ecosystem is cost-efficient, reliable, replicable, and hence can be utilized to provide legacy equipment with digital twin capabilities, collect historical data, and generate analytics.

This study presents a digital twin system designed to monitor, predict, and optimize linear welding parameters in real time, with applications in Industry 4.0 and smart manufacturing. The system leverages Kubernetes to seamlessly integrate various sensors, enabling their dynamic allocation to different aggregates and ensuring scalability for adding new devices as needed. The selected computational models, including machine learning and fuzzy logic-based approaches, provided effective real-time feedback to the welding aggregate by dynamically adjusting welding power. GitOps practices and event-driven communication using message queues facilitated efficient deployment and management in this dynamic and distributed environment, facilitating continuous system updates and minimizing downtime. By providing real-time feedback to welding machine operators and functioning as a digital twin, this system enhances both simulation capabilities and physical process control, demonstrating potential for cross-domain interoperability and advanced decision-making frameworks.

This paper presents a developed model of a Digital Twin (DT) for a fused deposition modeling (FDM) printer, real-time defect detection, and proposed frameworks for preventing cyber-attacks in real-time. It also highlights a model predictive control (MPC) algorithm based on a real-time feedback system for controlling the material feed. The system is designed and developed based on DT, and MPC with integrated machine learning (ML) algorithms to establish real-time process control and enhance the safety and reliability of the physical plant. ML algorithm is used for anomaly detection based on the convolutional neural network (CNN) model. The developed system can be practically utilized in smart manufacturing industries as well as cyber-physical systems-based plants. The work is novel and original as this type of DT and cyber-physical systems (CPS) are very new to additive manufacturing (AM) industries. There are several conceptual models in the literature and there is a critical need for such implemented working systems.

No abstract available

This paper addresses the problem of a comprehensive quality assurance strategy for additively manufactured components with integrated in-situ inspection and artificial intelligence and machine learning (AIML) models. A custom test setup was created around a fused filament fabrication (FFF) 3D printing apparatus, incorporating sensors to capture real-time data concerning part quality. This image data was harnessed to formulate an AI/ML dataset for training a Convolutional Neural Network (CNN). A robust framework for predicting defects in real-time during the additive manufacturing process and validating the accuracy of AIML predictions has been presented. A test setup, generating a varied dataset, crafting AI/ML models, and optimizing the AI/ML model for precise defect prediction has been made. The proposed methodology’s practical applicability and potential to redefine quality assurance in additively manufactured parts have been presented. Results were compared between the Matlab AlexNet pre- trained model and the user-designed model on Google Colab, which has the capability of hyperparameter tuning. Performance parameters, including accuracy, loss, precision, and recall, were plotted over epochs to analyze the model’s merit after training. With the addition of hyperparameter tuning to the AlexNet, the best model was chosen as per the training accuracy. Much better accuracies were observed in the earlier epochs, with the initial loss being highly reduced compared to the non-hyperparameter-tuned model. The ResNet50 model, which did not have hyperparameter tuning capabilities, produced less accurate results. On the other hand, the loss was much lower with ResNet50 compared to both the AlexNet Matlab model and the AlexNet model on Google Colab without hyperparameter tuning. ResNet50, at the same time, gave accuracy comparable to that of the AlexNet Model with hyper-parameter tuning.

With the growing adoption of additive manufacturing (AM) technology across various industries, concerns regarding the possible release of hazardous volatile organic compound (VOC) emissions have surfaced, particularly in VAT photopolymerization (VPP) processes. This study investigates VOC emissions in VPP AM by implementing machine learning (ML) and advanced digital twins to monitor, predict, and mitigate VOC release. An Industrial Internet of Things (IIoT) sensor network, integrated with an Anycubic Mono X 6 K 3D printer, captured data on critical parameters, including layer thickness, exposure time, and light intensity. Subsequent ML model analysis identified exposure time as a principal factor influencing VOC emissions. A Unity-based digital twin was developed to support proactive process optimization, offering real-time visualization and predictive analytics of emission trends. The system aligns with Industry 4.0 objectives, showing considerable potential to enhance operational efficiency and environmental sustainability in VPP AM. This integrated approach significantly advances environmentally responsible AM practices in industrial settings.

No abstract available