The paper presents preliminary investigations into a strategy for solving linear systems resulting from 3D finite element simulations, based on the algebraic multigrid (AMG) method, enhanced using artificial intelligence techniques. In particular, we adapt to 3D problems the algorithm presented in Luz et al. (2020) for using a graph neural network to create the prolongation and restriction operators in a way that will improve convergence. The process of training the network proceeds on the basis of a set of system matrices obtained for tasks much smaller in scale than the target problem to be solved. Learning is aimed at decreasing the spectral radius of the error propagation matrix after applying modified prolongation and restriction. We describe some implementation details of the solver developed based on the presented strategy and show several numerical examples of its application for medium-sized problems.
Bipolar plates are critical core components in proton exchange membrane fuel cells (PEMFCs). Titanium-based materials are highly favored due to their excellent corrosion resistance and high specific strength. However, the plates often experience severe local thinning and poor consistency in forming dimensions during the two-stage stamping process. Although traditional finite element method (FEM) optimization can mitigate these defects, it comes with high computational costs and time consumption. This study proposes a die design optimization framework based on the Sequential Physics-Informed Neural Network (S-PINN). Unlike traditional single-layer neural network models, S-PINN adopts a sequential architecture that effectively maps the two-stage forming process of the plates. This architecture can explicitly predict the evolution of forming quality from the pre-forming stage to the final stage. By embedding the core physical laws of plastic deformation into the network loss function, the S-PINN model effectively predicts the complex nonlinear relationship between mold geometry and forming quality, while ensuring physical consistency. Experimental and simulation results show that the S-PINN model’s prediction accuracy for dimensional consistency (DC) is 73.8% higher than that of the PINN model and 33.9% higher than that of the S-ANN model. Compared with traditional modeling methods, the S-PINN-optimized die design can reduce the thinning rate and improve channel dimensional consistency.
This study investigates the sensitivity of factors influencing the contact resistance and reliability assessment of bolted copper–aluminum electrical connectors. Through continuous energized dynamic testing, the temporal evolution of the contact resistance and temperature of the contact area were monitored throughout the experiment. In conjunction with observations of contact surface morphology and microstructural composition analysis, the intrinsic mechanisms underlying anomalous changes in contact resistance were explored in depth. The results indicated that, during continuous current flow, the contact resistance of aluminum busbar assemblies without plating exhibited an “initially low → rapid increase → gradual stabilization” pattern, whereas nickel-plated aluminum assemblies maintained stable resistance. Quantitatively, the contact resistance of the nickel-plated group remained at approximately 15 μΩ, while that of the bare aluminum group rose to 400–750 μΩ. The evolution of the contact interface was asymmetric, with the dynamic growth of the oxide layer on the aluminum surface being the dominant mechanism increasing resistance. Moreover, the aluminum coating significantly reduced contact resistance and enhanced connection reliability.
In this study, Artificial Neural Networks (ANN) were created to develop a Digital Twin (DT) for temperature prediction in the laser hardening process of NC10 steel. The ANN were trained to predict temperature on the top layer during the laser hardening process of NC10 steel samples with different thicknesses and with various laser power and laser scanning speeds. The prediction developed during the project work was based on a parametric numerical model of the laser hardening process for a sample of NC 10 steel, using the Finite Element Method (FEM) within the ADINA software. Numerical simulations enabled a detailed analysis of the temperature produced on the surface of each sample, as well as a visualization of the structural changes made to the sample according to the laser hardening process. It is crucial to create data that reflects reality as closely as possible to assess the best setting for each process. A well created DT allows to make automatically important changes along laser hardening process. To obtain a set of the most efficient parameters for the desired result, Genetic Algorithms (GA) were integrated with the developed ANN. As a result, the authors developed an effective and efficient tool to predict the temperature produced along the laser hardening process.
This article discusses feasible methods for detecting dents in car bodies caused by transportation damage, commuting collisions, and hail. The authors review existing approaches exploiting their limitations, including smartphone-based ML detection algorithms and drive-through tunnels. The paper details the setup for capturing dents using computer vision with industry-grade cameras and structured light projection, emphasizing optimized data acquisition and computer vision setup. A particular emphasis is placed on acquiring high-quality input data thanks to the proper calibration and alignment of cameras, structured light, and the synchronization between them. Challenges related to obtaining high-quality footage in real-life conditions, such as car speed, body color, and lighting conditions, are thoroughly discussed. The method covers algorithms for detecting car paint, optimizing camera parameters, and identifying dents. Data annotation methods are described in detail, ensuring robust training datasets. Validation of the method is based on comparing the results of an inspection by professional car appraisers with algorithm detection outcomes. The results demonstrate the effectiveness of the proposed methods. Additionally, the article explores future research opportunities, such as scratch detection, damage severity estimation, and integrating these systems into automated production lines. The potential for enhancing vehicle inspection processes through advanced computer vision and structured light techniques is also considered.
This paper explores the discrepancies observed between 2D and 3D crystal plasticity finite element (CPFE) simulations and mean-field approximations in terms of macroscopic flow curves. Two hypotheses are proposed to address the discrepancies: (1) the type of yield function in the mean-field approach (2) differences in stress states between the two methodologies. Based on the first hypothesis, the type of yield function may influence the stress-strain partitioning in the mean-field approach. Consequently, the von Mises criterion is replaced with the Hershey yield function. To test the second hypothesis, CPFE simulations are extended to 3D to achieve comparable stress states in both methods. This analysis reveals that the exact shape of the yield function has a marginal impact on the discrepancies, whereas the proper 3D stress distribution significantly reduces them. This comprehensive study also uncovers a limitation of the mean-field approach in terms of accuracy in the prediction of macroscopic material response and stress partitioning for a two-phase polycrystalline material.
The bound states and decay time in a certain quantum well structure (GaMnAs/GaAs) were analyzed and identified at the minimum decay time. Through the analysis of quantum mathematical equations, we derived specific formulas for energies that significantly amplify in the numerical solutions of equations throughout all dimensions of confinement. The quantification and barriers, alongside the well width, without altering the parameters utilized, were predominantly influenced by the spatial dimension parameters, such as the barrier height and well width. The principal bound state and lowest decay time were determined at a well width of 40Å and a barrier thickness of 46.27Å. This work revealed a novel characteristic known as interfacial tunnelling, which refers to the phenomenon where an electron establishes a tunnelling state between two interfaces. This tunnelling process is significantly influenced by the characteristics of the materials used, as well as the dimensions of the wells and barriers.
The mechanical behavior of 316H austenitic stainless steel is investigated in this study under cyclic strain-controlled loading with and without hold periods at elevated temperatures. Understanding the low-cycle fatigue (LCF) and fatigue-creep interaction (FCI) characteristics is essential for ensuring the structural performance and safety of reactor components, particularly under conditions typical of modular and generation IV reactors. The new generation of nuclear power plants require more resistant and durable materials as the operating environments impose significantly higher demands, including increased neutron irradiation levels and elevated operating temperatures, leading to accelerated material degradation. A combined isotropic-kinematic hardening model within a crystal plasticity framework is employed to capture the cyclic and time-dependent mechanical response of the material. Model parameters are calibrated by fitting cyclic loading simulation results to experimental data at 550°C using polycrystalline representative volume elements (RVE). Strain-controlled uniaxial loading simulations are performed to analyze peak stress evolution throughout cyclic loading and stress relaxation behavior during strain-hold periods. The RVE simulation results are in strong agreement with experiments under LCF loading. For the loading with strain-holds, stress relaxation during hold periods exhibits two distinct stages: an initial rapid decay followed by a steady decline, both of which are captured in simulations. Beyond the macroscopic response, analyses reveal the heterogeneous evolution of field variables at the microstructural level, as strain hardening during loading and stress relaxation during hold periods varied across grains due to their crystal orientations and interactions with neighboring grains. These findings enhance the understanding of high-temperature mechanical behavior at both macroscopic and microstructural scales, contributing to the efforts for the design, operation, and life extension of nuclear reactor components.
Understanding the local evolution of phase transformations in steels, particularly the γ (austenite) → α (ferrite) transformation, is crucial for controlling the microstructure and properties of steel components. Over recent decades, significant progress has been made in the numerical modeling of this complex phenomenon. This development has been driven by both scientific curiosity and industrial needs, especially in processes such as hot rolling, forging, thermal treatment, etc. The developed models have evolved from simple solutions based on local equilibrium to more complex approaches that consider local heterogeneities. Modern computational approaches, such as Phase-Field (PF), Level-Set (LS), Cellular Automata (CA), Monte Carlo (MC) or Vertex based simulations, allow for the precise reproduction of microstructural evolution considering local instabilities. They also enable the analysis of phase boundary motion in an explicit manner. These techniques also allow for direct integration with thermodynamic data and mechanical models, thereby better capturing the physical mechanisms of phase transformations, such as chemical composition, diffusion resistance, or the influence of deformation. An overview of the state of the art in this area is presented within the paper. The model’s concepts, assumptions, fundamental equations, advantages, limitations, and potential practical applications are summarized. Special attention is given to modeling the γ → α transformation by the Cellular Automata method. The importance of incorporating phenomena such as diffusion, nucleation, and growth is emphasized. The need for consistency between experimental results and simulations is also highlighted.
The evolutionary deep learning algorithm EvoDN2 is an emerging strategy for data-driven intelligent learning and many-objective optimisation capable of handling a large volume of noisy and non-linear data. This article provides the essential details of this algorithm and highlights a number of its recent applications.
In the presented study, plastometric tests using channel die compression were employed to define the boundary conditions for numerical simulations of the deformation processes of heterogeneous multi-layer systems composed of microalloyed steel, titanium, or magnesium. Various configurations, conditions, and deformation schemes were applied, which were then replicated in numerical simulations. Rheological models were used in the studies which, through computer simulations, enabled the modeling of interactions between the incoherent components of the microstructure. The primary outcome of the conducted experimental studies and numerical simulations is the ability to assess the heterogeneity of the studied multi-layer systems in terms of their mechanical states and influence on microstructural changes. This heterogeneity additionally arises from the diverse microstructural and rheological characteristics of the investigated materials (BCC vs. HCP), which, in turn, affect the strengthening mechanisms, primarily strain hardening. The results obtained from channel die compression tests were then used in simulations of multi-stage wire drawing, supporting both the design phase and the analysis of the resulting microstructural effects in the studied heterogeneous systems. It was observed that one of the key criteria for designing heterostructured wires from the examined materials is the proper selection of the volume fraction of the components, as well as the deformation history during multi-stage wire drawing, considering interpass heat treatment.
The paper presents the application of numerical simulations based on the Finite Element Method (FEM) for analyzing and optimizing the extrusion processes of aluminum and lead. These processes are efficient methods for manufacturing critical machine parts and metal components, ensuring excellent mechanical properties. A detailed analysis was conducted on the numerical modeling of the impact of die taper angles on strain distribution and forming forces during co-extrusion. The study found that a 45-degree angle provides optimal deformation conditions, minimizing extrusion forces and reducing the formation of dead zones compared to a 90-degree angle. Numerical simulations, supplemented by technological trials under semi-industrial conditions and image analysis involving the deformation of the coordinate grid, provided key insights into a material flow, strain distribution, and force parameters. The results emphasize the importance of validating numerical models with semi-industrial experiments to ensure accuracy and reliability, as assuming constant tribological conditions may not reflect actual process conditions, including the formation of dead zones for angles greater than 45°. It was only through a thorough analysis of the actual process and the introduction of variable friction coefficients for individual tools that a dead zone was achieved in the modelling. The findings from this research can serve as the foundation for further optimization and adaptation of technological processes, aiming to further enhance extrusion processes through the use of numerical simulations.
Open-die forging is a key process for manufacturing large components such as generator shafts and crankshafts for ship engines. Despite its industrial relevance, the process remains dependent on manual labour and operator expertise, leading to challenges in process stability, reproducibility, and efficiency. Traditional automation approaches are impractical due to the high variability and low production volumes typical of open-die forging. At the Institute of Metal Forming (IMF) at the TU Bergakademie Freiberg, a novel concept for autonomous open-die forging has been developed and tested. The system combines conventional forging equipment with advanced technologies, including industrial robotics, 3D laser scanning, thermal imaging, and modular control software. Central to the concept is a robot cell operating as a distributed system, where sensor data is used to create a digital twin of the workpiece. This enables adaptive process planning and real-time autonomous operative adjustments. A process planning tool generates pass sequences and commands for manipulator movements, while an electromechanical interface allows indirect control of the forging press. The modular software architecture, coordinated by a central core-module, ensures flexibility and facilitates integration into different production environments. Initial trials demonstrate the system’s potential to improve process stability and quality while reducing dependency on manual operation. Ongoing work focuses on refining the concept to meet industrial requirements and support advanced material applications.
In recent years, considerable progress has been made in the field of artificial intelligence and machine learning. This progress allows us to solve increasingly complex problems, but it also requires providing appropriate explanations to understand the actions taken by AI. For this purpose, research into the development of Explainable Artificial Intelligence has been initiated and interest in this topic is constantly growing. This review of XAI methods includes a justification for the need to introduce solutions to explain artificial intelligence models, describes the differences between various methods and presents example method/s that work in different cases. The purpose of this paper is to solve a real problem occurring in heavy industry. The third chapter describes the challenges to be faced, the solution developed and the results of the work. The entire study concludes with a summary of the research findings.
The paper presents an analysis of the possibility of measuring the residual stresses of metal sheets with the application of the so-called Barkhausen effect. The aim of the research was to compare the residual stress levels measured by two methods of multiphase steel sheets (ferritic-martensitic-bainitic) in grade HCT980C after flattening on a roller leveller in industrial conditions. The measurements were carried out using two methods: the Barkhausen effect method and the X-ray method. The paper describes in detail the methodology used for testing the measurement of residual stresses. The residual stress testing of sheets made of the CP1000 steel group was supplemented with tests of chemical composition, microstructure and mechanical properties (Re, Rm, A80, HRC hardness). In the analysis of the research results, elements of statistics were also used, in the form of ordinary correlation. The research results showed that in the case of sheets after flattening on a roller leveller in industrial conditions, it is possible to replace the commonly used and recognized, but labour-intensive X-ray method, with a simple, innovative and cheap to use method using the Barkhausen effect. Stress measurement using the Barkhausen effect has already been found to be applicable in the diagnostics of tracking changes in the stress value in the material in industrial pipelines, where access to the other measurement methods is difficult or even impossible. Currently, the measurement of stress in sheets by the magnetic method is introduced on the transverse cutting line when cutting the sheet metal from coils to metal sheets. The measurement of stresses in the production of steel sheets is important because the difference in stress between the top and bottom sides of the sheet has a significant effect on the flatness of manufactured metal sheets.
In the pursuit of lighter designs, many industries are shifting from conventional fasteners to adhesive joints, which offer a better strength-to-weight ratio and facilitate the use of fiber-reinforced polymers. However, modeling adhesive joints presents major challenges due to the complex behavior of polymeric adhesives and the microstructural changes induced by the substrates. As a result, various simulation methods have been developed to capture the behavior of adhesive joints across different scales. Molecular dynamics studies address atomistic and nanoscale phenomena, while continuum approaches — such as the finite element method, cohesive zone models, and peridynamics — focus on meso and macro scales. Additionally, multiscale methods combine particle and continuum approaches to provide a more comprehensive understanding of the adhesive bond behavior. This review paper offers an overview of the most relevant numerical methods employed to examine the mechanical behavior of adhesive joints and illustrates the simulations’ applications through examples from the literature.
High entropy alloys, characterized by their near-equimolar compositions of five or more elements, exhibit unique properties including high strength, thermal stability, and corrosion resistance, making them ideal candidates for demanding applications. Unfortunately, experimental research on their behavior under processing and in-use conditions is expensive and time-consuming. Therefore, the use of computer-aided technology design is required. However, reliable constitutive material models for these alloys are rarely available in the literature. Thus, this research aims to develop a constitutive material model of a Mo-Mn-Fe-Co-Ni high entropy alloy through a structured two-phase inverse analysis. First, a preliminary inverse analysis was conducted to recalculate load-displacement data measured during uniaxial compression tests at varied temperatures and strain rates to the required flow stress data. This first phase helps mitigate the impact of testing artifacts – such as friction and localized heating – that can introduce inhomogeneities in the material and affects the hardening behavior. Then, a full inverse analysis was performed to precisely calibrate the constitutive model parameters, ensuring an accurate representation of the alloy’s flow stress behavior under the tested conditions. This second phase optimizes the model to reflect the material’s inherent properties rather than external test-induced effects, thus improving the robustness and reliability of the flow stress data across a range of loading scenarios. As a result, a reliable form of the constitutive model, along with the identified parameters, was obtained and can be used during computer-aided technology design.
Controlling the microstructure of a diffusion welded interface is a critical point to ensure optimum mechanical properties and the homogeneity of the joint. Beyond the intimate contact formation between bonded parts studied in the literature, this article focuses on the grain boundary crossing of the interface during this process and its measurement. Following this perspective, a Level-Set method has been used for full-field microstructure simulations in 2D with various interface parameters. Two crossing measurement models have been formulated, tested and discussed over the simulations.
W artykule przedstawiono analizę stanu i jakości oleju hydraulicznego stosowanego w koparce gąsienicowej Doosan 380 LC-5 eksploatowanej w kopalni surowców skalnych. Celem badań była analiza zmian jakościowych oleju hydraulicznego Orlen Hydrol L-HV 46 w funkcji ilości przepracowanych motogodzin oraz jej ocena do określania stanu oleju hydraulicznego w układzie. Badaniom poddano pięć próbek oleju, w tym jedną zawierającą olej świeży. Badania przeprowadzono z wykorzystaniem mobilnego analizatora OPCom Portable Oil Lab, umożliwiającego ocenę czystości oleju metodą laserowo-optyczną zgodnie z normą PN-EN ISO 4406:2021. Uzyskane wyniki potwierdzają postępującą degradację oleju wraz z czasem eksploatacji oraz zasadność systematycznego monitoringu jego czystości.
W zaprezentowanym artykule przedstawiono wyniki badań podczas rozruchu przesiewacza wibracyjnego w zależności od nastawionego czasu rozruchu. Badania przeprowadzono, wykorzystując podwieszany przesiewacz wibracyjny o odcinkowej trajektorii drgań – w którym wymuszenie drgań realizowane jest przez dwa silniki wibracyjne mocowane bezpośrednio do rzeszota przesiewacza. Silniki wibracyjne były zasilane napięciem trójfazowym i połączone z siecią przez układ przemienników częstotliwości umożliwiających sterowanie prącą napędu zarówno w stanach ustalonych, jak i w trakcie rozruchu. Rejestrowano również moc chwilową pobieraną przez napęd przesiewacza i określono maksymalny prąd płynący w układzie zasilania dla dziewięciu przyjętych czasów rozruchu (od 0 s do 8 s). Badania te pozwoliły na określenie najkorzystniejszych parametrów rozruchu pozwalających na zmniejszenie zapotrzebowania mocy napędu niewpływające znacząco na parametry kinematyczne przesiewacza podczas jego rozruchu.
