Manufacturing Process Research Articles

Heterogeneous graph neural network for modeling intelligent manufacturing systems

Abstract Currently, manufacturing systems have become more and more complex, often involving multiple machines, systems and processes to produce workpieces. In order to facilitate comprehensive analysis and control of manufacturing processes, the integration of connections between machine level, system level, and process level in manufacturing process modeling is needed. However, traditional graph deep learning models are unable to take into account the heterogeneity of different machines in the system when modeling manufacturing systems. To address this problem, this paper proposes a new approach: modeling manufacturing systems using the heterogeneous graph neural network sample and aggregate algorithm based on cutting-edge Bayesian neural networks and graph deep learning. This method considers the connection between different manufacturing system levels, treats machine operations as nodes, and connects different nodes through material flow and operational similarity. The effectiveness of the method is demonstrated through its application to model an aero-engine blade production line. Extensive numerical experiments show that the proposed graph modeling method is effective in expressing the heterogeneity of different machines and multi-level manufacturing processes. By integrating machine heterogeneity into the modeling of a manufacturing system, it not only facilitates comprehensive analysis and control of the manufacturing process, but also lays the foundation for cost savings and productivity improvements in the manufacturing system.

Ultra-high-rate Manufacturing of Thermoplastic Window Plug using Hybrid Overmolding

Highly loaded complex parts are machined from a solid block of metal by removing material through cutting, boring, drilling, and grinding to obtain the three-dimensional shape. Most cases, this involves removing as much as 80 percent of the material from a billet, which is time-consuming and inefficient considering the cost of the materials. Alternatively, these complex parts can be manufactured using hybrid overmolding process which involves injecting plastics over a reinforced substrate with continuous fiber-reinforcement eliminating subtractive machining or joining of multiple parts. The substrate is typically heated using an infrared (IR) oven and preformed over a solid mold to a three-dimensional shape using thermoforming. Injection overmolding material can also be reinforced with discontinuous fibers based on the design requirements. Both thermoforming and injection molding are matured automotive ultra-high-rate manufacturing processes. Adopting these technologies to aerospace requires advanced aerospace-grade thermoplastic material systems and developing the hybrid overmolding process that combines both thermoforming and injection molding processes. Material compatibility is one of the key enablers for the fusion across the overmolded interfaces. To demonstrate the overmolding technology for an aerospace application, Victrex AE250 low-melt polyaryletherketone (LM-PAEK) system reinforced with AS4 fibers and a 30% carbon fiber-reinforced polyetheretherketone (PEEK) were used to overmold aircraft window plugs to be used for cargo conversion of passenger aircraft. The fully automated overmolding process was demonstrated using KraussMaffei multi-robotic thermoforming and injection molding system. Thermal and mechanical test evaluations were employed to assess the material compatibility and overall part quality. Photomicrographs and flexure tests conducted on test samples extracted from the window plugs indicated satisfactory fusion across the overmolded interface and optimized holding of PEEK materials.

The Influence of Carbon Fiber on Crystallization and Morphology of Poly(Ether Ketone Ketone) Composites Towards High Rate Thermoplastic Manufacturing Processes

Recent interest in thermoplastic composites for aerospace applications is driven by the need for high-volume manufacturing of materials that display excellent mechanical, thermal, and solvent resistive properties. Specifically, a large component of the composite properties of thermoplastic composites are highly dependent on the crystalline fraction of the semi-crystalline polymer matrix of the reinforced system. This work investigates the melting behavior of poly (ether ketone ketone) (PEKK) and the influence of temperature and fiber on crystallization kinetics and morphology of (PEKK) composites utilizing, differential scanning calorimetry, Velaris-Sefaris modelling, and polarized optical microscopy (POM). Annealing temperatures of 380°C and above were required to fully erase crystalline ordering in the PEKK matrix. Furthermore, carbon fibers accelerated crystallization kinetics in PEKK at a range of temperatures observed via DSC, and Velaris-Sefaris modelling implied a shift from bulk to surface induced nucleation by the addition carbon fiber to the PEKK matrix, reducing activation energy of crystallization events. This nucleation behavior was visually confirmed through POM with observation of transcrystalline domains nucleated from the surface of a single carbon fiber.

The Astonishing Accomplishment of Biological Drug Delivery using Lipid Nanoparticles: An Ubiquitous Review.

Biotech drugs, including proteins, hormones, enzymes, DNA/RNA therapies, and cell-based treatments, are gaining popularity due to their effectiveness. However, effective delivery systems are needed to overcome administration challenges. Lipid nanoparticles (LNPs) have emerged as promising carriers for various therapies. LNPs are biocompatible, less likely to cause adverse reactions, and can stabilize delicate biological drugs, enhancing their stability and solubility. Scalable and cost-effective manufacturing processes make LNPs suitable for largescale production. Despite recent research efforts, challenges in stability, toxicity, and regulatory concerns have limited the commercial availability of LNP-based products. This review explores the applications, administration routes, challenges, and future directions of LNPs in delivering biopharmaceuticals.

The Influence of Carbon Fiber on Crystallization and Morphology of Poly(Ether Ketone Ketone) Composites Towards High Rate Thermoplastic Manufacturing Processes

Recent interest in thermoplastic composites for aerospace applications is driven by the need for high-volume manufacturing of materials that display excellent mechanical, thermal, and solvent resistive properties. Specifically, a large component of the composite properties of thermoplastic composites are highly dependent on the crystalline fraction of the semi-crystalline polymer matrix of the reinforced system. This work investigates the melting behavior of poly (ether ketone ketone) (PEKK) and the influence of temperature and fiber on crystallization kinetics and morphology of (PEKK) composites utilizing, differential scanning calorimetry, Velaris-Sefaris modelling, and polarized optical microscopy (POM). Annealing temperatures of 380°C and above were required to fully erase crystalline ordering in the PEKK matrix. Furthermore, carbon fibers accelerated crystallization kinetics in PEKK at a range of temperatures observed via DSC, and Velaris-Sefaris modelling implied a shift from bulk to surface induced nucleation by the addition carbon fiber to the PEKK matrix, reducing activation energy of crystallization events. This nucleation behavior was visually confirmed through POM with observation of transcrystalline domains nucleated from the surface of a single carbon fiber.

Probing an LSTM-PPO-Based reinforcement learning algorithm to solve dynamic job shop scheduling problem

With the growth of personalized demand and the continuous improvement in social productivity, the large-scale and few-variety centralized production model is gradually transitioning towards a personalized model of small batches and multiple varieties, which makes the manufacturing process of the job shop increasingly complex. Furthermore, disruptive events such as machinery failures and rush orders in the job shop increase the uncertainty and variability of the production environment. Traditional scheduling methods are usually based on fixed rules and heuristic algorithms, which are difficult to adapt to constantly changing production environments and demands. This may lead to inaccurate scheduling decisions and hinder the optimal allocation of job shop resources. To solve the dynamic job shop scheduling problem (JSP) more effectively, this paper proposes a Reinforcement Learning (RL) optimization algorithm integrating long short-term memory (LSTM) neural network and proximal policy optimization (PPO). It can dynamically adjust scheduling strategies according to the changing production environment, achieving comprehensive status awareness of the job shop environment to make optimal scheduling decisions. First, a state-aware network framework based on LSTM-PPO is proposed to achieve real-time perception of job shop state changes. Then, the state and action space of the job shop are described within the context of the state-aware network framework. Finally, an experimental environment is established to verify the algorithm’s effectiveness. Training the LSTM-PPO algorithm makes it feasible to achieve better performance than other scheduling methods. By comparing the initial planning time with the actual completion time of the rescheduling decision under different dynamic disturbances, the efficiency of the proposed algorithm is verified for the dynamic JSP

INFLUENCE OF FILLER SIZE ON THE STRUCTURE AND PROPERTIES OF THE POLYMER MATRIX

Purpose. The aim of the study is to investigate the influence of aluminum oxide (Al2O3) particle size on the mechanical properties of a polymer composite with added chopped glass fiber, as well as to explore the interaction between aluminum oxide particles and chopped glass fiber. Research methods. Specimens were tested for tensile strength according to DSTU EN ISO 527-5:2018. Testing was conducted using an MTS Criterion Model 43 universal testing machine with a maximum load of 50 kN. Metallographic analysis was performed using a KEYENCE VHX microscope at magnifications of 50× and 500×. The microstructure of the polymer matrix was determined on unetched samples. Scanning electron microscopy was carried out using a JEOL JSM-5510LV microscope. Results. The influence of introducing different sizes of aluminum oxide fraction on the polymer matrix was studied. It was found that with an increase in the fraction size, the strength of the polymer composition decreases. Additionally, the interaction between chopped glass fiber and various fractions of aluminum oxide on the mechanical characteristics and morphology of the connection with the polymer composition was investigated. Scientific novelty. The interaction of chopped fiber with aluminum oxide fractions may affect the mechanical properties of the composite, such as strength, stiffness, and elasticity. Investigation of the morphology of the connection between chopped fiber and aluminum oxide fractions can help understand how they interact within the composite structure. This includes analyzing the adhesion between components, the structure of junctions, and possible defects that may arise during the manufacturing process. Studying this interaction opens up opportunities for the development of new composite materials with improved properties and diverse applications in industry, construction, aviation, automotive, and other fields. Practical value. The practical significance lies in refining manufacturing technologies and implementing new materials in industry. The ability to control the mechanical properties of composites by changing the size of aluminum oxide fractions can contribute to the creation of more efficient materials for the production of automobiles, aircraft, building constructions, and more. Additionally, understanding the interaction between chopped glass fiber and aluminum oxide fractions may open up new possibilities for developing composites with unique mechanical characteristics that meet the requirements of modern technologies and industrial production.

An EPQ Model with Varying Demand for Uncertain Quality Products

Objectives: This article delves into a production inventory model centered on non-deteriorating items amidst fluctuating demand. Methods: To minimize the total cost using the sine function as a demand function. Findings: In the manufacturing process, there's a chance of product damage, resulting in the final products being either flawless or flawed. Quality assessments occur during production to distinguish the flawless ones for immediate sale, while the flawed ones are marked down for later sale. “In this article, we use a sine function for demand which will increase first, and after a certain point it starts decreasing”. Novelty: The existing method has discussed only increasing or decreasing demand, but the current study deals with both increasing as well as decreasing at the same time. Applications: Seasonal methods like umbrellas, Christmas decorative items, crackers, etc., have increasing demand for the particular season. Keywords: Production Model, Fluctuating Demand, Sine function, minimizing Total Cost, Flawless items

Digital Manufacturing System of Custom Wigs with Measurable 3D Images

Wigs are major important means for improving life quality of the increasing people suffered from hair loss through their image enhancements. For the production of custom made wigs, there requires a mold that reproduces the wearer’s head dimension and the shape of hair loss. The conventional method for this is to make a resin film that adheres to the wig wearer’s head using vinyl tape, and to make a wig pattern with drawing the shape of hair loss onto the film, and to produce a head mold with this pattern, and then to plant hair in a wig cap made based on this mold and wig pattern. This method requires highly skilled manual work, and generates a large amount of chemical waste during the manufacturing process, and takes a long manufacturing period. However, the digital automation of measuring head dimension and the shape of hair loss can replace or omit the production of wig patterns. In this study, a 3D head image capable of actual measurement was implemented by using a 3D polygon head shape measurement device and a camera, which makes a digital production system possible, which can replace the manual wig pattern production process for custom wigs.

Advancements in Biofiller-Reinforced PLA Composites for 3D Printing: A Review

Additive manufacturing is key in realizing highly complex and high-performance composite materials. Among all the various kinds of functionality that could be added to a composite component, continuous fiber-reinforced composites have been drawing much attention because of their attractive combination of properties. The comprehensive spectrum of knowledge that ranges from the basics concerned with structure, morphology, synthesis, physical, and chemical properties finally reaches to this analytical study of advanced composites. Most generally, such a composite is structured on a thermoplastic or thermoset polymer matrix that embeds the load-carrying reinforcing fibers; probably best known are carbon, glass, and aramid fibers. These composites, which involve fibers continuously reinforcing, have high strength relative to their weight. They are also anisotropic, meaning they have qualities that vary from one direction to another – something which might be customisable for specific load cases. They warp and shrink less during their life in an outside environment than conventionally made bioplastics due to a fact that short fibers can provide reinforcement to them. This 3-D printing object is manufactured by special additive manufacture technology during the manufacturing process after compounding and extruding the fiber-reinforced composite filament. The final properties of 3D printed parts could be tailorable through changing variables such a fiber type, fiber length, volume fraction, polymer matrix material, orientation of fibers, and printing process parameters. These continuous fiber-reinforced 3D printed composites could achieve tensile strengths up to one GPa, have stiffness values reaching even a rating of countless Gpa, and have ad thicknesses in the range from about 1.4–1.8 g/cm³. This finding could be applied in basically all major industries, from aerospace and the automotive industry to sports gear. Such conclusions from the analysis may be useful in the selection of appropriate materials and techniques while even guiding the development of new applications with respect to such advanced composite materials.

Voxel-wise segmentation for porosity investigation of additive manufactured parts with 3D unsupervised and (deeply) supervised neural networks

AbstractAdditive Manufacturing (AM) has emerged as a manufacturing process that allows the direct production of samples from digital models. To ensure that quality standards are met in all samples of a batch, X-ray computed tomography (X-CT) is often used in combination with automated anomaly detection. For the latter, deep learning (DL) anomaly detection techniques are increasingly used, as they can be trained to be robust to the material being analysed and resilient to poor image quality. Unfortunately, most recent and popular DL models have been developed for 2D image processing, thereby disregarding valuable volumetric information. Additionally, there is a notable absence of comparisons between supervised and unsupervised models for voxel-wise pore segmentation tasks. This study revisits recent supervised (UNet, UNet++, UNet 3+, MSS-UNet, ACC-UNet) and unsupervised (VAE, ceVAE, gmVAE, vqVAE, RV-VAE) DL models for porosity analysis of AM samples from X-CT images and extends them to accept 3D input data with a 3D-patch approach for lower computational requirements, improved efficiency and generalisability. The supervised models were trained using the Focal Tversky loss to address class imbalance that arises from the low porosity in the training datasets. The output of the unsupervised models was post-processed to reduce misclassifications caused by their inability to adequately represent the object surface. The findings were cross-validated in a 5-fold fashion and include: a performance benchmark of the DL models, an evaluation of the post-processing algorithm, an evaluation of the effect of training supervised models with the output of unsupervised models. In a final performance benchmark on a test set with poor image quality, the best performing supervised model was UNet++ with an average precision of 0.751 ± 0.030, while the best unsupervised model was the post-processed ceVAE with 0.830 ± 0.003. Notably, the ceVAE model, with its post-processing technique, exhibited superior capabilities, endorsing unsupervised learning as the preferred approach for the voxel-wise pore segmentation task.

Explaining the Anomaly Detection in Additive Manufacturing via Boosting Models and Frequency Analysis

Anomaly detection is an important feature in modern additive manufacturing (AM) systems to ensure quality of the produced components. Although this topic is well discussed in the literature, current methods rely on black-box approaches, limiting our understanding of why anomalies occur, making complex the root cause identification and the consequent decision support about the action to take to mitigate them. This work addresses these limitations by proposing a structured workflow designed to enhance the explainability of anomaly detection models. Using the wire arc additive manufacturing (WAAM) process as a case study, we examined 14 wall structures printed with INVAR36 alloy under varying process parameters, producing both defect-free and defective parts. These parts were classified based on surface appearance and welding camera images. We collected welding current and voltage data at a 5 kHz sampling rate and extracted features from both time and frequency domains using a knowledge-based approach. Isolation Forest, k-Nearest Neighbor, Artificial Neural Network, XGBoost, and LGBM models were trained on these features, and the results shown best performance of boosting models, achieving F1 scores of 0.927 and 0.945, respectively. These models presented higher performance compared to other models like k-Nearest Neighbor, whereas Isolation Forest and Artificial Neural Network posses lower performance due to overfitting, with an F1 score of 0.507 and 0.56, respectively. Then, by leveraging the feature importance capabilities of these models, we identified key signal characteristics that distinguish between normal and anomalous behavior, improving the explainability of the detection process and in general about the process physics.

The Synergistic Impact of Human-AI Collaboration: A Multi-Domain Analysis

This comprehensive article explores the transformative impact of human-AI collaboration across healthcare, manufacturing, and scientific research. The article examines how the synergistic integration of artificial intelligence with human expertise is revolutionizing key industries, enhancing efficiency, accuracy, and innovation. In healthcare, we discuss the implementation and impact of AI-assisted diagnostics and AI-augmented surgery, highlighting improvements in diagnostic accuracy and surgical outcomes. The manufacturing sector is analyzed through the lens of AI-driven quality control systems and the introduction of collaborative robots (cobots), demonstrating significant gains in productivity and worker safety. In scientific research, we explore AI's role in data-driven research and hypothesis testing, particularly in areas such as drug discovery and genomics. The article also provides a critical analysis of the complementary strengths of humans and AI, addresses ethical considerations in AI implementation, and speculates on the future potential of this collaboration. Drawing on recent research and case studies, this article presents a holistic view of how human-AI synergy is driving progress across multiple domains, while also considering the challenges and ethical implications of this technological integration. The findings suggest that the future of human-AI collaboration holds immense potential for accelerating scientific discovery, improving healthcare outcomes, and revolutionizing manufacturing processes, ultimately benefiting society as a whole.

Tailorable Fluorescent Perovskite Quantum Dots for Multiform Manufacturing via Two-Step Thiol-Ene Click Chemistry.

In practical applications, fluorescent perovskite quantum dots (PQDs) must exhibit high efficiency, stability, and processibility. So far, it remains a challenge to synthesize PQDs with stable dispersibility in tailorable monomers both before and after photocuring. In this work, a novel strategy of UV-induced two-step thiol-ene "click chemistry" is proposed to prepare PQDs with these attributes. The first step aims to epitaxially grow a shell around the PQD core to ensure stable dispersibility in a thiol-ene monomer solution. The second step is to achieve stable dispersibility in the photocured thiol-ene matrixes for multiform manufacturing processes. The tailorable PQDs (T-PQDs) not only have the highest photoluminescence quantum yield (PLQY) to ≈100% for green emission and over 96% for red emission, but also exhibit remarkable stability under severe conditions, including "double 85" aging, water exposure, and mechanical stress. Moreover, their exceptional processability allows for various processing techniques, including slot-die coating, inkjet printing, direct-laser writing, UV-light 3D printing, nanoimprinting, and spin coating. The high efficiency and stability of T-PQDs facilitate their multiform manufacturing for a wide range of applications.

3D-Printed Multi-Axis Alignment Airgap Dielectric Layer for Flexible Capacitive Pressure Sensor

Flexible pressure sensors are increasingly recognized for their potential use in wearable electronic devices, attributed to their sensitivity and broad pressure response range. Introducing surface microstructures can notably enhance sensitivity; however, the pressure response range remains constrained by the limited volume of the compressible structure. To overcome this limitation, this study implements an aligned airgap structure fabricated using 3D printing technology. This structure, designed with a precisely aligned triaxial airgap configuration, offers high deformability under pressure, substantially broadening the pressure response range and improving sensitivity. This study analyzes the key structural parameters—the number of axes and pore size—that influence the compressibility and stability of the dielectric material. The results indicate that the capacitive pressure sensor with an aligned airgap structure, manufactured via 3D printing, exhibits a wide operating pressure range (50 Pa to 500 kPa), rapid response time (100 ms), wide limit of detection (50 Pa), and approximately 21 times enhancement in sensitivity (~0.019 kPa−1 within 100 kPa) compared with conventional bulk structures. Furthermore, foot pressure monitoring trials for wearable sensor applications demonstrated exceptional performance, indicating the sensor’s suitability as a wearable device for detecting plantar pressure. These findings advocate for the potential of 3D printing technology to supplant traditional sensor manufacturing processes.

Modelling and Fabrication of Tri-Head Gantry Crane Model

During the manufacturing process, various raw materials and finished goods have to be handled in manufacturing industries. As the shapes and material properties of each raw material / finished good to be handled are different, it gets difficult for a single-head crane to handle diverse materials. At present forklifts and hook-headed cranes are being used for this purpose, but it requires a lot of labor and time to handle the materials and finished goods due to miscellaneous work associated with this equipment. Looking at the present day’s industrial requirements, it becomes inevitable for every manufacturing company to use the least manpower to offer competitive prices to customers without compromising the timely completion of projects. Tri head gantry crane is one of the best solutions to handle various raw materials and finished goods with the least manpower & time. The tri-headed gantry crane head consists of electromagnets to hold flat surfaces of magnetic materials, where the grab helps in holding the cylindrical, square, and rectangular materials finally if the material or a finished good is not in the shape that can be handled by using grab or magnetic heads, then we can use hook arrangement. The project involves study of various types of handling equipment being used in manufacturing industries, modeling and fabrication of tri-headed gantry crane model and testing of mechanisms of the model for handling diverse shaped materials [3].

Improvements in Injection Moulds Cooling and Manufacturing Efficiency Achieved by Wire Arc Additive Manufacturing Using Conformal Cooling Concept

The plastic injection moulding industry is a constantly developing industrial field. This industrial process requires the manufacturing of metal moulds using complex heating and cooling systems. The purpose of this research is to optimize both the plastic injection moulding process and the mould manufacturing process itself by combining practices in this industry with current additive manufacturing technologies, specifically Wire Arc Additive Manufacturing (WAAM) technology. A mould punch was manufactured by using both WAAM technology, whose internal cooling system has been designed under the concept of Conformal Cooling, and conventional cooling channel designs and manufacturing techniques in order to carry out a comparative analysis. Theoretical results obtained by CAE methods showed an improvement in heat extraction in the WAAM mould. In addition, the WAAM mould was able to achieve better temperature homogeneity in the final part, minimizing deformations in the final part after extraction. Finally, the WAAM manufacturing process was proven to be more efficient in terms of material consumption than the conventional mould, reducing the buy-to-fly ratio of the part by 5.11.

An integrated approach of zero defect manufacturing and process mining to avoid defect occurrence in production and improve sustainability

Purpose This paper aims to integrate zero defect manufacturing (ZDM) with process mining (PM) to avoid defect occurrence during production and improve sustainability. Design/methodology/approach The method is developed based on literature review in ZDM and PM. It uses PM for process discovery as an initial strategy in priority to predict-prevent strategies of ZDM. Findings It presents the applicability of the proposed approach in observing manufacturing process behavior, identifying dynamic causes of defects during production, predicting the time of defect occurrence and preventing defective products. It also identifies, explains and measures criteria for environmental, social and economic pillars of sustainability affected by defects and presents the impacts of the proposed approach on sustainability improvement. Originality/value The extended view of this research, as well as its analytical approach, helps practitioners to develop their ZDM and PM approaches more holistically. The practical application of this research is illustrated through implementing it in a real-life manufacturing case, where the outcomes prove its applicability in avoiding defect occurrence and improving all three pillars of sustainability.

A Short Review on the Wire-Based Directed Energy Deposition of Metals: Mechanical and Microstructural Properties and Quality Enhancement

Wire-based directed energy deposition (WDED) is an emerging additive manufacturing process garnering significant attention due to its potential for fabricating metal components with tailored mechanical and microstructural properties. This study reviews the WDED process, focusing on fabrication techniques, mechanical behaviors, microstructural characteristics, and quality enhancement methods. Utilizing data from the Web of Science, the study identifies leading countries in WDED research and highlights a growing interest in the field, particularly in materials engineering. Stainless steel, titanium, aluminum, and copper-based alloys are prominent materials for WDED applications. Furthermore, the study explores post-processing techniques such as machining, heat treatment, and surface finishing as integral steps for quality enhancement in WDED components.

Optimizing ironing parameters in material extrusion (MEX) technology: enhancing efficiency and performance

This study explores the optimization of ironing parameters in Material Extrusion (MEX) technology, specifically targeting improvements in the mechanical properties of polylactic acid (PLA) printed objects, with a focus on enhancing Ultimate Tensile Strength (UTS) and minimizing printing time. To achieve this, we employed a systematic experimental approach that varied flow rate, path speed, and spacing between passes, followed by a comprehensive statistical analysis using Analysis of Variance (ANOVA). Our findings indicate that path speed and spacing between passes significantly influence both UTS and printing time, while flow rate has a minor effect within the tested range. Optimal parameters were determined to be a flow rate of 14.34%, path speed of 30 mm/s, and spacing between passes of 0.3 mm, achieving a composite desirability of 0.970537. Validation using the response optimizer feature in Minitab software confirmed a strong correlation between predicted and experimental results. This research provides practical insights for optimizing MEX parameters, contributing to enhanced mechanical properties and efficiency in additive manufacturing processes.