This study presents a robust multiscale formulation for stress-constrained topology optimization aimed at designing lightweight and structurally resilient components. Unlike classical compliance-based methods, which may result in topologies unable to support applied loads, the proposed approach minimizes structural volume while rigorously enforcing local stress constraints. A dual-scale framework integrates macro-structural optimization with periodic micro-structural design, leveraging the Solid Isotropic Material with Penalization (SIMP) method; though it remains adaptable to other established topology optimization techniques. To address the computational challenges arising from numerous local stress constraints, we implement an Augmented Lagrangian strategy combined with polynomial vanishing constraints, eliminating the need for aggregation functions such as the p-norm or Kreisselmeier-Steinhauser functions. The resulting optimization algorithm is accurate and scalable, supported by a detailed sensitivity analysis and adjoint-based gradient computation. Numerical experiments in two dimensions validate the effectiveness of the method, demonstrating superior stress distribution and structural efficiency compared to classical formulations. This work contributes a comprehensive and scalable methodology for multiscale topology optimization under stress constraints, suitable for high-performance engineering applications.

High-fidelity numerical models are widely used in civil engineering design and assessment, but their computational cost hinders extensive parametric studies and optimisation. Machine learning-based surrogate models offer a promising alternative by learning fast input–output mappings from limited simulation or experimental data. This work presents a systematic comparative study of three widely used surrogate techniques—artificial neural networks (ANN), Kriging metamodels and support vector machines (SVM)—applied to three representative civil engineering problems: (i) structural optimisation of an overhead travelling crane, (ii) assessment of the liquefaction potential of sandy soils and (iii) back-calculation of flexible pavement layer responses under different loading levels. For each case, a designed set of simulations or measurements is used to train the surrogates, and their predictive performance is evaluated using error-based metrics and computational efficiency indicators. The results provide a clear quantitative picture of model performance: Kriging and SVM consistently achieve the best classification accuracies for small to moderate datasets, with SVM and Kriging reaching top classification accuracies of 99.0

This work presents the integrated application of topology optimization and additive manufacturing (AM) to develop lightweight structural components for space robotic systems, focusing on the deployment arm and rocker support of the Lunar Volatiles Mobile Instrumentation-Extended lunar rover. These components are subject to stringent performance requirements under reduced gravity, launch-induced vibrations, and terrain-induced loading. A level set-based (LSB) topology optimization approach is applied to minimize structural compliance while adhering to constraints on mass, stress, symmetry, and manufacturability. Six mission-specific load cases, including quasistatic surface operations and launch scenarios derived from Miles' equation, are used to drive the design. The optimization achieves mass reductions of approximate to 50% for the deployment arm and the rocker support, while preserving mechanical integrity and functional interfaces. Post-optimization validation through finite element analysis confirms that the optimized designs meet all structural performance criteria. Prototypes are fabricated via fused deposition modeling to assess manufacturability and assembly integration, paving the way for future metal AM using aerospace-grade alloys. The results demonstrate a simulation-driven workflow that applies LSB topology optimization with additive manufacturing constraints to mission-specific load cases, integrating European Cooperation for Space Standardization compliant verification and manufacturability to develop structurally efficient rover suspension components.

Topology optimization has become an essential tool in structural analysis, allowing for the design of efficient, lightweight structures that meet specific performance requirements. This study focuses on the application of general optimization methods, particularly simulated annealing, in the topology optimization of trusses or other discrete structures. In recent years optimization functions are readily available in packages such as Matlab/Octave, and there is no lack of structural analysis software, often with no optimization capabilities. Simulated annealing offers the advantage of exploring the design space thoroughly, increasing the likelihood of finding a global minimum solution. This is critical for ensuring that optimized structures can withstand arbitrary loading conditions, such as seismic or dynamic loads. Additional constraints can be imposed, such as limiting the number of distinct cross-sections of the structural elements. These constraints add further complexity to the optimization process but also enhance the practicality of the solutions in real-world applications. With modern computers, topology optimization has become increasingly feasible for structures composed of discrete elements (trusses, frames), where the the elements (bars, beams, columns) are prefabricated or even 3D-printed. This research demonstrates how readily available global optimization techniques can lead to better structural designs by efficiently balancing load-bearing capacity and material use, while adhering to complex constraints.

Aim: To present a Finite Element Model of the macula region of the eye’s posterior pole for structural analysis during ocular movements. Method: We used software ImageJ, Meshlab and Spaceclaim for image processing and meshing to create a patient specific geometry of the posterior pole from Ocular Coherent Tomography images. Then we used Ansys software to create a predominately linear FEM model to simulate the action of inferior oblique muscle on the posterior pole. We performed Static analysis and Eigenvalue buckle analysis in different biomechanical scenarios. Results: We recorded the stress, strains and volumetric changes on the posterior pole layers. We reproduced known clinical entities like partial detachment of vitreous cortex, macula rhexis and pigment epithelial detachments. We found that removal of cortex may have a protective effect on the retina by transferring stress and buckling events from the retina and its interfaces to the deeper layers of choroid and sclera. Conclusion: This is the first time that a FEM demonstrates the effects of ocular movements and particularly of inferior oblique muscle action, supporting our hypothesis that repetitive strain from ocular motility may be a common pathogenetic factor in various, potentially blinding, diseases of the macula.

Hydraulic infrastructures for flood management are typically designed to operate under a wide range of conditions with multiple roles. In this research, we present the capabilities of water flow modeling with physical maquettes created with photogrammetry using successive photos by drone. Digital terrain model can be 3D printed to acquire the maquette of the terrain, and to simulate how a rainfall event transforms into runoff on the maquette. Thus, we can acquire the experimental simulation of how water inundates along the river channel and the flood plain areas. This provides physical characteristics to the study of water flow and the necessary inputs for the designing process. By capturing water movement and physically observing the simulation of the model, it is easier to identify potential errors that may arise during the creation of the digital model, and which may not be easily detected otherwise.

Floods are catastrophic events that affect nearly every part of the globe, leading to substantial losses of life and causing extensive economic damage. During flood emergencies, one of the primary challenges faced by response teams is accurately identifying the flooded regions to establish access points and determine safe evacuation routes swiftly. As climate change intensifies, the frequency and severity of extreme weather events like floods are rising, making rapid flood mapping and response even more critical in minimizing damage and saving lives. This study compares several artificial intelligence models and spectral indexes for flood detection in Sentinel 2 multispectral images. The proposed methodology combines transfer learning with different models, such as neural networks, convolutional neural networks, and vision transformers. A variation of this model is the vision transformer (ViT), which can be applied to image classification tasks. Multispectral Instrument (MSI) images from Sentinel-2 contain images from different bands. By combining different zones of those bands different spectral indexes can be calculated. This study uses the Normalized Difference Water Index (NDWI), Modified Normalized Difference Water Index (MNDWI), Water Ratio Index (WRI), Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Automated Water Extraction Index (AWEI). By comparing different artificial intelligence models and different spectral indexes the best combination is determined, which is NDVI with VGG CNN model, and can be used for real-time flood detection using Sentinel 2 Multispectral Images.

Customized wrist splints, particularly for upper extremity orthoses like wrist support braces, are commonly used across numerous clinical scenarios. However, the traditional process for producing personalized wrist splints is largely manual and highly dependent on the expertise of orthopedic specialists. This experience-based approach often leads to suboptimal outcomes, necessitating further refinement of the designs. Recent advancements in Additive Manufacturing (AM) have brought significant innovation to various industries, including orthopedics. This study aims to present a comprehensive methodology that integrates advanced design tools ,like 3D Scanning, with digital manufacturing techniques to produce tailored wrist splints. The produced hand brace aims to offer enhanced mechanical performance and comfort by precisely fitting an individual’s anatomy while minimizing material usage and weight. To achieve optimal design efficiency, the study explores the application of a Topology Optimization (TO) approach for design, while the manufacturing process utilizes Fused Deposition Modeling (FDM), an evolving technology within the Additive Manufacturing (AM) sector.

Deep neural networks (DNNs) have emerged as a powerful tool for solving regression problems, offering flexible architectures that can be tailored to diverse applications. In earthquake engineering, accurately predicting the seismic response of buildings remains a critical challenge. While traditional simplified models of single- or multi- degree-of-freedom systems, are widely used due to their computational efficiency and ability to facilitate rapid simulations, these approaches often fall short in capturing complex nonlinear structural behavior and spatial variability in ground motions. This research integrates earthquake time-histories with ambient vibration oscillators’ data to develop an advanced neural network-based predictive model. Presented through image representations, this framework is meticulously designed to enhance the accuracy of forecasting structural seismic responses taking also into account stiffness non-linearity. The goal is to accurately forecast a building’s seismic response. A dataset comprising 1197 MDOF 2D models was utilized, producing a total of 32,319 training samples for the model. The proposed framework is evaluated using except the loss function of a network model training, also with a mean absolute percentage error (MAPE). By combining AV and EQ response data into a neural network-based approach, the proposed network demonstrates a new direction developing tools using neural networks for predicting the seismic response of structures. Such advancements could not only enhance the understanding of building behavior during earthquakes but also support the development of resilient building designs, contributing to safer engineering practices and improved earthquake mitigation strategies.

The use of spinal braces is the most common solution to address adolescent idiopathic scoliosis by limiting the progression of spinal inclination. From a mechanics point of view, spinal braces are shell-type structures, relying on the fundamental engineering principles of three-point bending and inversion forces. A research project is ongoing to develop a procedure comprising laser scanning, advanced numerical simulation, topology optimization and additive manufacturing, to design and 3D print lightweight, personalized braces. In the present paper recent advances in the numerical simulation procedure are reported. Finite element simulation of the brace and its interaction with the patient’s body is employed, to evaluate the developing deformations and stresses during the brace’s use. For that purpose, the brace is modeled with shell finite elements and the body with a mirror surface of appropriate geometry, considering deformability and detachment by means of contact elements with an appropriate pressure – overclosure relation, employing realistic values of human body stiffness at its different parts. The mechanical properties of the brace material employed in the simulation are adopted from experimental tensile tests on 3D-printed Polylactic Acid (PLA) specimens. From the numerical simulation, von Mises stresses and displacements of the spinal brace are obtained and are used to interpret its structural behavior. To understand how the patient’s body is affected by the use of the brace, contact pressures are presented, reflecting the interaction between the brace and the body.

The objective of this research was to evaluate the influence of some parameters of 3D printing on the rough surface and flatness of the parts. For the study, all samples were produced from polylactic acid (PLA), following the Taguchi experimental design method. The printing parameter settings that optimised the roughness and flatness were firstly determined. Layer height, print speed and nozzle temperature were the key printing parameters used to devise the experiments. The experiments were defined by varying these according to the Taguchi L9 orthogonal array. Surface roughness in different orientations (X, Y, Z) and angles (15°, 30°, 45°, 60°, 75°) was then measured whereas flatness was evaluated based on the standard deviation from the value of a reference plane. This study potentially opens a new door to manufacturers to enhance surface finish and dimensional accuracy for various 3D printed components.

This study investigates the thermal conductivity of the Jordanian travertine rock, an important construction material, using the Hot Disk Transient Plane Source (TPS) 2200 technique. Thermal conductivity was correlated with the physical and engineering properties of the travertine through simple regression, multiple regression, and Artificial Neural Network (ANN) model to develop predictive relationships. To perform this study, 61 cylindrical core samples were extracted from different locations along the margins of the Dead Sea in Jordan and prepared with a diameter of 10 cm and height of 20 cm. The results indicated that the thermal conductivity ranged between 2.678 (W/(m*k)) to 3.407 (W/(m*k)) with an average approximately equal to 3 (W/(m*k)). The results showed that the thermal conductivity increases with increasing the density, hardness, rock strength, and P-wave velocity. On the other hand, the thermal conductivity decreases with porosity and water-absorbed capacity. The results showed that the ANN model outperformed other models in prediction accuracy. The developed models provide a foundation for estimating thermal conductivity based on basic physical properties, offering a cost-effective alternative to direct measurements for selecting suitable construction stones .

In this work we develop and validate neural-network (NN) constitutive models for structural steels under monotonic and cyclic loading, using internal-variable inputs to encode path dependence. For monotonic tension at 200-800^∘C (Kirby-Preston dataset), a feed-forward NN ( 1-10-10-1 , sigmoid activation) reproduces experimental stress-strain curves to high fidelity; across temperature-specific curves extracted from the experimental data, the median normalized Chamfer distance is 0.355% of the plot height (IQR: 0.026-0.361%) . For cyclic compression of stainless steel 316 (Chaboche dataset), the internal-variable NN ( 5-30-30-1 ) aligns with the experimental hysteresis loops with median normalized Chamfer distance 0.032% , indicating near-pixel-level curve overlap. We provide new quantitative validation tables and sensitivity plots; results indicate weak dependence on hidden-node count and epochs within practical ranges. The approach is computationally lightweight at inference and amenable to generalization via internal-state encoding. Limitations at very high temperatures ( >750^∘C ) and low-strain transitions are discussed, along with integration paths for physics-guided constraints.

Prior research has demonstrated that morphological characteristics and cell division timings can indicate the quality and developmental potential of an embryo. However, conventional evaluation techniques miss minute quantitative changes that can be identified by sophisticated computer vision analysis. Time-lapse data and embryo classifications from 2170 embryos grown for at least 110 h at a single IVF centre were examined in this retrospective cohort study. An extra 326 embryos were set aside for blind testing so that the model’s performance could be compared to that of an embryologist. The metrics were extracted by training an Attention U-Net model to segment the embryos in time-lapse images, enabling a robust analysis of the morphological patterns. The final machine learning mode was trained using extracted data from embryo surface metrics time series to perform a binary classification task, in order to discriminate between embryos that reached blastocyst stage and those who did not. The machine learning model was able to predict if an embryo will reach the blastocyst stage with an AUC of 0.85 [95

Objective of the present study is the simulation of an experimental process for the design of 3D-printed joints of a scissor-based deployable shelter for emergency response [1]. The test specimen consists of a 3D-printed joint made of Polylactic Acid (PLA) and four aluminum bars with rectangular hollow sections hinged to the joint using steel bolts. At their other ends the aluminum bars are hinged to a rigid steel base, and the load is applied by a tension rod system to the center of the joint. Geometry and material nonlinear analyses (GMNA) were carried out in Abaqus CAE software Version 2021 [2], where all members of the specimen were simulated by 3D solid finite elements, with the aim of predicting the evolution of the experimental process but also determining the deviations between numerical and experimental findings. Nonlinear material laws were adopted for the joints, the bars and the bolts, representing PLA [3], aluminum, and steel, respectively. In addition, all possible contacts of the members were simulated through contact elements. The analysis results demonstrated a high concentration of stresses and material yielding around the bearings of the bars and the joint, already at relatively low load levels, while high stresses develop also inside the joint at higher loads. Significant margins for optimization of the joint topology were established, as inactive areas of the joint were observed for the applied loads.

This project explores the dynamic intersection of traditional sculpture techniques and state-of-the-art 3D scanning and 3D printing technology. Two intricate clay sculptures were crafted by G.-Fivos Sargentis, serving as the initial embodiment of the creative vision. Following this traditional sculpting phase, each model underwent a transformative process, marking a synthesis of classical artistry and modern innovation. To bridge the analogue and digital realms, the clay sculptures were 3D scanned using mobile apps for smartphones, capturing the fine details and nuances of the original handcrafted forms. This digital replication paved the way for further digital sculpture processing and the utilization of 3D printing technology, as the scanned models were recreated in tangible, three-dimensional form. The printed sculptures became a canvas for the artist’s continued exploration, fostering a unique dialogue between the physical and the virtual. The final models were printed in various materials, including plastic, resin (for casting) and wax (for casting). Some were then cast in bronze using the lost-wax casting method.

Tsunamis are one of the most devastating natural hazards, with the potential to cause extensive loss of life, property damage and socioeconomic disruptions. Developing robust and accurate early warning systems is critical to mitigating these impacts. In this study, a neural network-based early warning system is proposed to predict tsunami wave heights nearshore, focusing on the Vancouver Island area on the western coast of Canada. The Vancouver Island region, which is extremely susceptible to tsunami hazards because of its closeness to the Cascadia Subduction Zone, is the area used to generate the synthetic data. In tsunami research, synthetic data are essential because they enable the investigation of a variety of possible earthquake and tsunami scenarios, including uncommon but highly consequential occurrences. The dataset, which contains 5000 simulation scenarios, used includes parameters such as fault slip parameters, bathymetry, hypocenter position, and earthquake magnitude, as well as the related tsunami wave heights at particular nearshore locations. The parameters used to train the model are the maximum wave heights off shore at different stations and the parameter that the model is trained to predict is the maximum wave height near shore in different depth zones (0 m, 5 m, 10 m, and 100 m). The neural network architecture was designed to model the nonlinear relationships between input parameters (maximum wave heights off shore at different stations) and resulting tsunami wave heights (near shore at different depths). By training, validating, and testing the neural network, the model demonstrated a high level of accuracy in predicting wave heights nearshore. The performance metrics, including mean absolute error and correlation coefficients, indicate that the neural network effectively captures the complexities of tsunami wave dynamics, making it suitable for early warning applications. According to the results, the neural network can accurately forecast tsunami heights close to shore, facilitating prompt evacuation preparation and disaster relief. This method is a major improvement over conventional physics-based models, which frequently demand a large amount of time and resources, by providing a computationally effective and scalable solution. Overall, this study demonstrates how machine learning, and in particular neural networks, might improve early warning systems for tsunamis.

The energy crisis presents a challenging concern for many international organizations in different areas around the world. Conventional construction methods allow thermal radiation to enter buildings in all its forms, which in turn consumes more energy. Structural buildings consume about 40

The purpose of the research paper in hand is to investigate the effect of axially applied and cyclic lateral loadings on the effectiveness of using carbon fiber-reinforced polymer (CFRP) material in preserving the structure’s behavior as well as governing the failure's mode of the different-in-strength-of-concrete CFT circular-shaped steel columns. For this purpose, the nonlinear finite element analysis (NLFEA) method has been employed. To begin with, a CFT column model was verified using the findings of previously published research; then, the experimental model was extended to include the influence of concrete’s strength in the search. Eighteen FEA CFT column samples were prepared and confined – at their ends – with different numbers of CFRP layers, representing the crucial position (with respect to the lateral loading capacity) and, also, to prevent the column samples from localized buckling outwardly. Thus, the samples would acquire more strength, greater net drifting, and higher dissipation of energy. The parameters of the experimental research were: i) the number of CFRP’s layers, and ii) the level of axial load. The research aimed to numerically explore, utilizing NLFEA, the influence of the research's parameters on the CFT FEA models’ behavior after the samples had been put to adequate calibration and validation as per credible experimental findings. The obtained findings indicated that when the columns were strengthened with CFRP, there was an enhancement in the cyclic behavior, represented by: more improved capacity of load, bigger horizontally oriented displacements, more displacement’s ductility, more dissipated energy, and less deterioration in secant stiffness Keywords: concrete strength, CFT columns, CFRP, lateral loading, cyclic axial loading, NLFEA

In this work, a holistic approach to optimizing driving behavior by exploiting acceleration-based metrics is proposed. By employing a Genetic Algorithm (GA), a search for smooth and safe acceleration profiles, particularly minimal “jerk,” which is the time derivative of acceleration, that also satisfy practical constraints such as maximum acceleration limits, traveling time, and specific target distance is executed. Contrasting conventional gradient-based optimizers, GA presents the ability to effectively explore a vast and potentially non-smooth solution area, thereby overcoming the risk of converging to local minima or infeasible solutions. A detailed formulation of the optimization problem, reflecting the constraints and objectives incorporated in a GA formulation is presented. Optimization results demonstrate that GA-based solutions identify smooth acceleration profiles, maximize passenger safety, and satisfy key performance metrics, such as comfort, and constraints.