XLVI Ibero-Latin American Congress on Computational Methods in Engineering

Structural Reliability Analysis of Reinforced Concrete: A Comparison of MCS, FORM, SORM, and Parallelized Subset Simulation in Julia

Kristian Yuiti Matsushita — 2025-12-01

In civil engineering, estimating the probability of rare events is essential for ensuring the safety and reliability of structural systems. Scenarios such as failures in reinforced concrete elements, progressive collapse, and performance-based design require accurate evaluation of low-probability events. Traditional methods like Monte Carlo Simulation (MCS), the First-Order Reliability Method (FORM), and the Second-Order Reliability Method (SORM) are widely used due to their theoretical foundation and practical simplicity. However, MCS becomes computationally prohibitive for rare events, while FORM and SORM may lose accuracy in nonlinear or non-Gaussian problems, common in structural analysis. Subset Simulation (SuS) overcomes these limitations by decomposing the rare event into a sequence of intermediate events, sampled efficiently through Markov Chain Monte Carlo (MCMC) techniques. To reduce computational time, SuS can be parallelized by distributing the sampling across multiple processors, this strategy enables faster analysis without compromising accuracy, making it suitable for large-scale models in civil engineering. This work explores the use of MCS, FORM, SORM, and SuS within the Julia programming environment, which offers high-performance computing capabilities. The methods are evaluated based on accuracy, computational cost, and robustness across a range of examples, including benchmark problems and structural models of reinforced concrete. Results indicate that SuS delivers accurate estimates with significantly fewer samples. When parallelized, it becomes a computationally efficient tool for reliability analysis in civil engineering, outperforming traditional methods in complex scenarios.

Optimal Design with Connectivity Constraints via Level-Set and Topological Derivatives Considering an Auxiliary Eigenvalue Problem

Giovanna Andrade — 2025-12-01

This work addresses the issue of imposing connectivity constraints in shape and topology optimization, with the goal of controlling the presence of void or solid islands in optimal designs. The approach relies on the spectrum of a two-phase differential operator with Dirichlet boundary conditions imposed on the external boundary. In contrast to previous works that adopt the same mathematical formulation for connectivity within the framework of density-based methods, we consider a level set representation of the domain combined with a topological-derivative-based optimization algorithm. Numerical experiments are presented to validate the strategy.

Numerical and theoretical evaluation of the influence of edge-to-hole distance on possible failure modes of sleeve connections in thin-walled square hollow sections

Gabriela Moura Azevedo — 2025-12-01

Tubular members in truss structures allow for spanning large distances with reduced self-weight. To facilitate manufacturing, transportation, and assembly, these members are subdivided and therefore require connections to join the segments. One example is the sleeve connection, which uses through bolts and eliminates the need for welded or flanged joints, making assembly quicker and more practical. The present study aimed to conduct a theoretical and numerical analysis of the influence of the edge-to-hole distance in sleeve connections with aligned bolts, applied to thin-walled square hollow sections. For this purpose, a numerical analysis was carried out using the finite element method through ANSYS software, with variations in edge-to-hole distance of 2, 2.5, 2.7, 3, and 3.5 times the bolt diameter, as well as variations in the number of bolts between 2 and 3. From this study, the possible failure modes of the connection were identified, with the dominant failure mode being the bearing failure of the bolt holes, also observed in the theoretical evaluation, occurring in the outer tubes. Furthermore, variations in the edge-to-hole distance and the number of bolts did not change this failure mode.

Modeling Approaches for Evaluating the Equivalent Mechanical Properties of Wind Turbine Blade Composites

Júlia Kunsch — 2025-12-01

The objective of this paper is to conduct a systematic and critical assessment of four models used to calculate the structural and inertial properties of wind turbine blades, with particular emphasis on those made from composite materials. Although traditional 3D Finite Element Analysis offers high accuracy, it is computationally expensive. Consequently, reduced-order modeling approaches, particularly one-dimensional geometrically exact beam theories, are commonly employed. However, the accuracy of these simplified models depends on the fidelity of the sectional properties provided as input.This study analyzes and compares four modeling approaches, each implemented through a dedicated computational tool: PreComp, VABS, CROSTAB, and GXBeam. These models are evaluated based on their theoretical foundations, computational capabilities, assumptions, and limitations. PreComp, while efficient due to its simplified assumptions, may lack accuracy when applied to complex geometries or heterogeneous structures. VABS (Variational Asymptotic Beam Section analysis) is distinguished by its rigorous mathematical framework, enabling accurate computation of both one-dimensional beam properties and three-dimensional field distributions. CROSTAB offers high computational efficiency but is constrained by assumptions regarding material homogeneity and in-plane wall behavior. GXBeam employs a finite element-based formulation that accurately determines the mass and stiffness properties of composite sections with arbitrary geometry and anisotropic materials. Importantly, it captures all relevant geometric and material-induced couplings, including bend–twist interactions. In this work, the structural and inertial properties are computed using GXBeam. The results are then compared with those previously obtained from PreComp, VABS, and CROSTAB, which serve as reference data for the analysis of sectional properties.To evaluate the performance of these models, the paper presents a case study based on a realistic wind turbine blade featuring two distinct cross-sections, both derived from the MH 104 airfoil. Each cross-section is modeled with five outer skin segments and two internal webs, located at 15% and 50% of the chord length. The blade’s material layup is highly detailed and representative of industry standards. The complexity of this configuration necessitates models capable of accurately capturing anisotropic behavior and elastic couplings.

Numerical investigation of side wall influence in RHS column–I-beam welded connections

Diego Souza — 2025-12-01

The evaluation of moment resistance, stiffness, and failure modes in welded connections between I-beams and rectangular hollow section (RHS) columns is essential for ensuring the structural integrity and performance of steel structures. However, current design codes and standard equations for estimating moment resistance do not account for all geometric parameters that influence connection behavior, such as the contribution of the side wall of the column. The equations provided by the EN 1993-1-8 and ISO 14346:2013 standards are based on the predominant failure mode of the connection. When the ratio between the beam width and the column width (β=b1/b0) is less than 0.85, failure occurs due to column frontal face plastification. For β equal to 1.0, failure is governed by column side wall yielding. Within the intermediate range of 0.85 lt; β lt; 1.0, the design standards suggest interpolating between these two failure modes. In all cases, the side wall of the column contributes to the moment resistance of the connection. In this context, the present study investigates the influence of this parameter on the structural behavior and moment resistance of uniplanar T-type welded connections. To achieve this, an investigation was conducted using finite element (FE) simulations, supported by a numerical model validated against experimental data. The analysis considered variations in the thickness (t0) and height (h0) of the hollow section side wall, as well as the thickness (t1), width (b1), and height (h1) of the beam, resulting in a total of 225 models. A comparison of the analytical formulations with the numerical results showed that the design equations tend to provide conservative estimates for the moment resistance of the connections. The findings indicate that higher values of t0 and ratios between the beam width and column width (β=b1/b0) lead to a greater influence of the column side wall on the connection behavior, resulting in increased moment resistance. Based on these observations, an equation was proposed for the case when β =1.0, which incorporates the contribution of the side walls of the column to the moment resistance of the connection, esulting in more consistent and accurate predictions.

Use of Physics-Informed Neural Networks to compute velocity and pressure fields around an airfoil in transonic flow

Vinícius Passeri Moraes de Souza — 2025-12-01

This work presents a novel application of Physics-Informed Neural Networks (PINNs) to simulate transonic (Mach 0.8–1.2) flow around an airfoil. The transonic regime features coexisting subsonic and supersonic zones, which generate shock waves and possibly boundary layer separation. This makes it a highly nonlinear and challenging flow condition for conventional Computational Fluid Dynamics (CFD) methods. Traditional finite-volume and finite-element solvers often require extremely fine shock-aligned meshes and significant computational cost to resolve transonic shocks and complex wave interactions. In contrast, our approach employs a PINN that embeds the compressible Navier–Stokes and continuity equations directly into the network’s loss function, enabling the flow field solution without a predefined mesh. PINN was implemented using DeepXDE library in a two-dimensional wind tunnel domain, with no-slip boundary conditions on the airfoil surface and far-field conditions on the outer boundaries. The mesh-free formulation avoids the grid generation burden and inherently sidesteps mesh-dependency in the solution, while working with residual-based adaptive collocation points sampling in the domain. We assess how well the PINN captures key transonic flow features—particularly the shock wave pattern and pressure field—by comparing its predictions to a high-fidelity finite-volume CFD solution. Applying PINNs to shock-dominated transonic flows remains largely uncharted, and previous attempts have struggled to resolve normal shock waves without special numerical treatments. To address this challenge by drawing on recent PINN advances, such as adaptive loss weighting and adding localized artificial viscosity at the shock was done. These help stabilize training and improve convergence for shock resolution. Preliminary results show that the PINN reproduces the flowfield accurately, capturing the shock location and overall pressure distribution in close agreement with the CFD benchmark. This study demonstrates the potential of PINNs as a promising, mesh-independent solver for complex aerodynamic simulations. It may offer advantages in regimes where traditional solvers encounter prohibitive meshing requirements or stability issues.

COMPARISON OF METHODS FOR HOMOGENIZING THE THERMAL PROPERTIES OF ULTRA-HIGH PERFORMANCE FIBER REINFORCED CONCRETE

Iury Fagundes — 2025-12-01

Ultra-high-performance fiber reinforced concrete (UHPFRC) stands out in the current context of civil construction as an alternative of great interest in the design of bold, slender and sophisticated structures. The performance under high thermal loads has been studied in order to understand the degradation of materials and a reduction in mechanical strength capacities, which indicates a loss of safety in structures. Computer simulations are a helpful tool for studying behavior patterns and identifying the mechanisms involved in the process of concrete damage when subjected to thermal stress. However, a challenge encountered in these models is the size difference between the fibers and the cementitious matrix in which they are embedded, which drives investment in homogenized models, as discrete modelling requires highly refined mesh generation and, consequently, high computational costs. In this context, this study aims to compare thermal models to represent the heating of synthetic UHPFRC specimens. Different models are discussed, including discrete fibers and homogenized systems, based on various methods in the literature. Analyses are carried out for concretes with fiber reinforcement rates ranging from 0 to 2% at temperatures of up to 600°C. The results show that the homogenized models have a good ability to approximate the thermal behavior observed in simulations with discrete fibers, with significant computational savings.

Revisiting the fib blind competition: insights into the numerical modeling of steel-fiber reinforced concrete beams

Lucas Teotônio — 2025-12-01

The second blind simulation competition (BSC) organized by the fib WG 2.4.1. took place in 2021 proposing the numerical modeling of a hybrid fiber reinforced concrete (R/FRC) shallow beam to assess the predictive performance of FEM-based computational models on the design verification at serviceability and at ultimate limit state conditions (SLS and ULS, respectively). This work revisits this benchmark aiming to address two issues not detailed in the BSC. The commercial finite element software DIANA is the chosen numerical tool, in which a classical smeared crack model is used to simulate the FRC properties and bar elements to represent the longitudinal reinforcement. It is shown that the modeling of the steel/FRC bond plays as important role in the structural behavior at SLS. It also discusses the influence of the modeling of the FRC compressive properties on the structural behavior at ULS by applying three different curves. The better correlation between the numerical results and the 2nd BSC experimental tests are obtained by using the parabolic compression curve for the FRC and the no-failure bond-slip interface model for the steel/FRC connection.

Comparing 2D and 3D finite element simplified micro-models of unreinforced masonry

Rafael da Silva Rodrigues — 2025-12-01

Finite element models of masonry walls play an important role in the analysis of these structures, avoiding the need for costly physical experiments. With advances in technology, both 2D and 3D models have become viable for analyzing walls with in-plane loads, but there is no work dedicated to comparatively showing the efficiency of each approach. The objective of this study is to compare 2D and 3D simplified micro-models of unreinforced masonry to make possible an informed choice of which model to use. First, an experimental study is selected to be simulated, then the modeling methods and data are defined. Both models use the extended finite element method (XFEM) to simulate cracks, the bilinear cohesive law, the Benzeggagh-Kenane law, and the quadratic nominal stress criterion to model damage, and the Drucker Prager plastic model to characterize the nonlinear behavior of the quasi-brittle materials. The results are discussed and compared, showing the advantages and disadvantages of using the 2D model. In most cases, using the 3D model is unjustifiable, because processing time is twenty-eight times longer for only subtle differences in the physical aspects compared to the 2D model.

Probabilistic Approach based on Design Response Spectra to Evaluate the Human Comfort of Pedestrian Footbridges

Paula de Oliveira Bezerra Diniz — 2025-12-01

Nowadays, the structural project of pedestrian footbridges has been conceived based on the use of lightweight structures, with low natural frequencies and structural damping ratios. These facts have generated slender footbridges, sensitive to human dynamic excitations, and consequently modified the design serviceability limit states. Thus, current design codes and technical guides recommend the use of deterministic models to evaluate the footbridges dynamic structural behaviour. On the other hand, pedestrian walking is related to a stochastic phenomenon and the dynamic force generated at each step depends on the weight, frequency and length of each pedestrian’s step. This way, this research work aims to propose a probabilistic approach to assess the dynamic behaviour of pedestrian footbridges (steel; steel-concrete; and concrete footbridges), based on the use of design response spectra. The proposed analysis methodology considered the stochastic nature of pedestrian walking aiming to evaluate the structural response taking into account excessive vibrations that can induce human discomfort. Based on the use of a probabilistic approach, it is possible to determine the probability of the peak acceleration values of the footbridge exceeding or not the human comfort criteria. Therefore, it was possible to conclude that the 95% percentile values of accelerations calculated using the proposed design response spectra are lower when compared to the experimental tests peak accelerations. On the other hand, the results obtained in this study indicated that the peak acceleration values calculated through the deterministic methods can be overestimated in project situations.

Treatment of numerical instabilities for topology optimization using the LSM/IGABEM coupling

Mário Sérgio Oliveira César Filho — 2025-12-01

The volumetric reduction characteristic of distinct topology optimization techniques is attractive to both academia and industry. This interest stems from financial reasons associated with the reduction in material consumption and environmental aspect, which involves the reduction of natural resources and pollutant emissions. This study addresses a topology optimization formulation for structural elements based on the coupling between the Level Set Method (LSM) and the isogeometric formulation of the Boundary Element Method (IGABEM). The former method describes the boundary evolution whereas the latter determines the mechanical fields. As in IGABEM, isogeometric descriptions have been used in LSM, which provide accuracy, robustness, solution generality, and direct communication between both methods. The optimization problem can be defined according to the augmented Lagrangian method. Thus, the coupling between the methods has been written by defining the normal velocity to the reference level set curve through the shape sensitivity of the augmented Lagrangian function. A heuristic topology modification addresses the deficiency of LSM in generating holes. This study clarifies the possibility of encountering numerical instability problems when using the studied coupling formulation, which might also occur if another penalty method had been adopted for driving the optimization. An effective treatment is proposed for this issue, showing its capability to reduce the dependence on the parameters of the heuristic topology modification criterion for the process success. The results verify its robustness by comparing them with those from other coupling formulations between LSM and IGABEM where results from the Solid Isotropic Material with Penalization (SIMP) are the reference.

Advanced modeling of particle-laden gravity currents over flat and wavy surfaces using a residual-based variational multi-scale approach

Gabriel Guerra — 2026-01-01

In this study, we introduce a numerical framework for simulating dilute particle-laden gravity currents at high Reynolds numbers (Re) using a residual-based variational multi-scale (RBVMS) approach. Our formulation effectively represents the coupling between velocity fine scales and the residual of the density concentration equation, ensuring a more accurate and robust representation of the underlying physical processes.We simulate lock-exchange particle-laden gravity currents across different Reynolds numbers to evaluate the proposed methodology. First, we validate the formulation by modelling gravity currents over flat terrain. Subsequently, we extend our analysis to particle-laden gravity currents over wavy terrains and varying wave heights to explore their influence on flow. Results indicate that floor roughness reduces the front velocity while significantly enhancing instabilities, leading to the earlier emergence of secondary instabilities. Additionally, roughness serves as an additional vorticity generation mechanism. Findings demonstrate that the (RBVMS) framework achieves high accuracy while requiring significantly low mesh resolution.

Development and Validation of Optimized Multiscale Structures in 3D-Printed Polymers: An Integrated Simulation and Experimental Approach

Jailson Moraes Bastos Júnior — 2025-12-01

This study proposes an integrated workflow that combines topology optimization, numerical simulation, and additive manufacturing for the development and validation of polymer structures subjected to bending. The methodology employs the SIMP method to redistribute material within structural models, with the results converted into lattice geometries using a Diamond-type unit cell. Finite Element Analysis is used to simulate the mechanical behavior of the optimized geometries, while fabrication is carried out through Fused Deposition Modeling using PLA filament. The specimens are then experimentally evaluated through three-point bending tests, allowing for validation of the computational model. The results show that the optimized structure exhibits higher specific stiffness and mechanical behavior influenced by the geometric characteristics of the lattice. The good correlation between simulated and experimental data in the elastic region validates the adopted approach, while the discrepancies observed in the post-yield phase reflect the effects of printing-induced anisotropy, geometric imperfections, and simplified constitutive modeling. This work highlights the potential of topology optimization combined with additive manufacturing as an effective strategy for designing lightweight components, while also emphasizing the challenges related to accurately predicting the mechanical behavior of complex structures.

Dynamic Catenary: A Simulation Model for Heterogeneous Mooring Lines

Milton Mateus Guimarães dos Santos — 2025-12-01

This study presents Dynamic Catenary (DynCat), an extended methodology for simulating elastic mooring lines in free-catenary configurations connected to floating offshore units. Traditional analyses rely either on catenary models, which are computationally efficient but dynamically oversimplified, or on finite element models (FEM), which provide higher fidelity at a greater computational cost. DynCat bridges this gap by combining the low computational cost of catenary models with enhanced physical accuracy, grounded in principles of deformable-solid mechanics. The formulation revisits deformable cable elements within the FEM framework, assembles multi-material line segments into a unified model, and incorporates environmental loading effects, resulting in numerically stable simulations that provide instantaneous line geometry and tension profiles at each time step. Validation studies covering shallow, intermediate, and deep waters under various fairlead motion regimes demonstrate that DynCat maintains good accuracy under low-to-moderate motion conditions. Although its performance deteriorates under highly dynamic deep-water conditions, DynCat remains accurate and robust within the operational envelope of shallow-to-intermediate water depths and low-to-moderate motion intensities, with simulation time reductions of 60–70% compared to conventional FEM approaches. This makes DynCat a promising tool for preliminary design and time-constrained analyses.

Content-based macroscopic microbial image retrieval: Preliminary results

Angela Mestas Muñante — 2025-12-01

As image databases grow and evolve, the need for effective retrieval engines becomes increasingly critical. While Content-Based Image Retrieval (CBIR) has advanced considerably across various domains, its application in microbiology remains limited; particularly from a macroscopic perspective, where pure culture images often appear visually similar and are difficult to distinguish. This study pioneers a preliminary investigation into the development of a CBIR framework for accessing images of microorganisms grown on solid media. Our proposal leverages deep neural networks to extract image features and applies Uniform Manifold Approximation and Projection (UMAP) to reduce the dimensionality of feature vectors without sacrificing performance. To enable fast and efficient retrieval, we implement a K-dimensional tree structure in combination with the K-nearest neighbors algorithm. We evaluate the framework’s performance using a leave-one-out strategy, considering precision and recall, measured by mean average precision, along with execution time. Experiments on two original microbial image sets yield promising results, with vision transformers outperforming other pre-trained models. Our findings highlight the potential of the developed framework for accurately retrieving culture images and demonstrate its applicability in both clinical and microbiological research.

Reservoir computing for chaotic time series prediction

Vitor de Barros Jr — 2025-12-01

Reservoir computing is a machine learning paradigm inspired by the dynamics of recurrent systems, where a fixed nonlinear dynamical system—the reservoir—transforms input signals into high-dimensional representations. These representations allow a simple linear combination at the readout to approximate complex temporal relationships with minimal training effort. Since its emergence in the early 2000s, this approach has shown great potential in predicting chaotic behavior, particularly by exploiting the underlying structure of nonlinear dynamical systems. However, key challenges remain in making reservoir computing more robust, interpretable, and broadly effective across diverse application domains. This work investigates the predictive capabilities of a reservoir computing architecture based on untrained recurrent dynamics combined with a trained linear readout. The method is applied to two benchmark chaotic systems: the Duffing oscillator and the Rössler attractor. A validation strategy specifically designed for chaotic time series is employed to optimize key model hyperparameters. The results demonstrate accurate short-term predictions and highlight the method’s sensitivity to initialization, underscoring the importance of proper validation and hyperparameter tuning.

Isogeometric Analysis vs. NURBS-Mapped Finite Elements: A Study on Selected Examples

Bianca Esteves Pirani — 2025-12-01

Isogeometric Analysis (IGA) is a numerical method that integrates finite element analysis (FEA) with CAD geometry representations using NURBS-based basis functions. In this work, we compare numerical simulations using IGA with a finite element approximation, where the geometric mapping is performed using NURBS, but the approximation space is constructed using traditional hierarchical finite element shape functions. The comparison focuses on problems with manufactured solutions involving both linear and curved geometries. Error analyses are conducted with respect to the number of degrees of freedom to assess the relative performance and accuracy of each approach.

Automatic Detection and Classification of Relevant Events in Oil Wells Using Numerical Derivatives and Rule-Based Segmentation

Andressa Silva — 2025-12-01

Anomaly detection in oil wells is a complex task, especially when identifying critical events that affect well operations. In this context, the oil industry has increasingly relied on computational techniques to enhance operational safety, leveraging time series data recorded by pressure and temperature sensors along wells. Significant variations in these signals over time may indicate relevant events. While all anomalous events are relevant, the converse is not necessarily true — not all relevant events are anomalies. However, focusing anomaly detection exclusively on relevant events simplifies the process. This work proposes an automated approach for detecting relevant events in production/injection oil wells, using a rule-based decision algorithm combined with statistical analysis tools. The objective is to identify the start and end points of relevant events, segmenting them into well-defined intervals. The technique applies numerical differentiation and Gaussian smoothing to sensor data, in order to reduce noise and highlight peaks caused by abrupt variations. Decision rules are then used to group these variations into coherent intervals. Each detected event is represented by a compact set of attributes, such as absolute, relative, and temporal differences between its start and end points. The main advantage of this approach lies in its ability to filter out noise and focus on relevant segments throughout the wells’ lifecycle, excluding periods of low variability that add little value and may introduce bias into machine learning processes. This focus on high-variability segments results in a cleaner, more representative dataset for modeling purposes. This initial detection serves as a starting point for distinguishing operational events from anomalies and preparing the data for more complex models. Initial results indicate that the statistical features extracted from the segmented intervals are promising for detecting anomalies in Inflow Control Valves (ICVs). Experiments using real well data, under different operational conditions and with recorded ICV failures, enabled the evaluation of the supervised classifier K-Nearest Neighbors (KNN) in distinguishing between normal and anomalous relevant events. The results indicate that the proposed method is effective in automating the identification and classification of anomalies, contributing to the continuous monitoring of oil well operations.

Synthetic Facial Aging: A Comparative Study of Stable Diffusion and StyleGAN

Bruno de Barros Oliveira — 2025-12-01

This study investigates the use of generative models for identity-preserving facial aging, focusing on the progression of real facial images over time. Traditional facial recognition systems often struggle with long-term age progression, as individuals undergo gradual yet significant changes in appearance.To address this challenge, we use real images of young adults from the FFHQ, BUPT-CBFace, and LFW datasets, ensuring balanced representation across ethnicities and genders. Facial aging is simulated at five target ages (30, 40, 50, 60, and 70 years) using two generative models: Stable Diffusion and StyleGAN. The outputs are evaluated in terms of identity preservation and perceptual quality.This approach enables a controlled comparison of age progression techniques. The results contribute to the advancement of AI-driven facial aging methods, with potential applications in forensic investigations, identity verification, and digital security.

Graph RAG vs. Vector RAG: A Performance Comparison in Response Generation

Cecília de Freitas Vieira Couto — 2025-12-01

The Natural Language Processing (NLP) area has been gaining prominence as one of the most promising fields in artificial intelligence, driven by recent advances in Large Scale Language Models (LSMs). These models have undergone significant development in a short period of time, becoming more robust and reaching levels of fluency and coherence increasingly close to that of human language. However, with the significant growth of LLMs, some relevant concerns have arisen, such as the occurrence of hallucinations, the importance of the knowledge used in training the models, and the environmental costs resulting from the high computational demand. Aiming to overcome these limitations, some alternatives have emerged, such as RAG (Retrieval-Augmented Generation) models. This approach combines text generation with external information retrieval, allowing the model to access dynamic databases during text generation, which can significantly reduce hallucinations. In conventional RAG models, this information is stored as vectors in a semantic space, allowing similarity searches. Despite its efficiency, this method may fail to represent more complex relationships between data. To overcome this limitation, an alternative to the traditional model is Graph RAG, a version that proposes organizing information in graph structures. This representation facilitates the mapping of explicit connections between entities, which tends to improve the quality and contextualization of the generated responses. To evaluate the impact on the performance of the models due to the data storage method, this work proposes the comparison between two models: a RAG with vector recovery and another based on graphs. For this, both will be trained using the same data set. After that, the models will be evaluated based on the responses provided to a standardized questionnaire, considering the quality and the time required to generate the responses.

Influence of Thermal Variations on the Behavior of Mooring Lines Using a Heterogeneous Extensible Catenary Model

Milton Mateus Guimarães dos Santos — 2025-12-01

This study investigates the mechanical behavior of mooring lines in floating offshore units subjected to thermal variations along the water column, employing a static numerical model based on a heterogeneous extensible catenary formulation. The growing global energy demand and the expansion of deepwater operations have increased the importance of mooring systems for platform stability and station-keeping. As key structural elements, mooring lines provide essential horizontal stiffness, govern restoring forces, and ensure structural integrity. However, conventional design models frequently neglect thermal effects, particularly the influence of depth-dependent temperature variations on deformed configuration of the line and its overall performance. To address this limitation, the proposed model incorporates both multimaterial line segments and vertical oceanic temperature profiles to evaluate their combined impact on horizontal stiffness and restoring capability. Results demonstrate that these thermal variations substantially alter the mechanical response of the line, directly affecting its interaction with the platform. This methodology enhances the reliability of mooring system design and contributes to a better understanding of thermo-mechanical behavior of mooring lines in offshore engineering.

Structural Optimization of Offshore Wind Turbine Towers under Stochastic Environmental Loads Integrating Finite Element Modelling and Reliability Analysis

André Victor da Silva Castilho — 2025-12-01

Based on national and international reports related the wind energy, it is evident that the Brazilian wind power sector is expected to experience exponential growth in the coming years. In this context, the structural optimization of offshore wind turbine towers emerges as a viable approach to reducing the costs associated with the commissioning of offshore wind farms currently waiting environmental licensing by IBAMA. Offshore wind towers are subjected to continuous loads resulting from wind and ocean actions. Frequently, design codes and technical literature treat these loads as deterministic, neglecting the inherently stochastic nature of such phenomena. Therefore, this research work proposes an analysis methodology for the structural optimization of offshore wind turbine towers through the integration of the Finite Element Method (FEM) and reliability analysis, aiming to contribute to the development of the wind energy sector in Brazil. Initially, a finite element model associated to a 5MW steel monopile offshore wind turbine (OWT) with 77.6m high was developed taking into account the effect of the soil-structure interaction, the non-deterministic dynamic behaviour of the wind and ocean loadings, and the aerodynamic forces related to the rotor. The proposed analysis methodology aims to establish an optimization framework that combines FEM and reliability analysis, incorporating the intrinsic uncertainties involved in the modelling of offshore wind turbine towers.

IMPACT OF DIRECTIONAL WELL INCLINATION ON THE CALCULATION OF ANNULAR PRESSURE BUILDUP

Maria Luiza Nunes Lopes — 2025-12-01

This study presents a preliminary analysis investigating Annular Pressure Buildup (APB) behavior as a function of the inclination angle in directional oil wells. APB refers to the pressure increase caused by the tendency of confined fluids to expand thermally within annular spaces. This phenomenon is critical in deep reservoirs, where temperature variations compromise casing integrity due to differential pressure. Despite its relevance, the influence of wellbore inclination on APB prediction remains underexplored. In this context, the present work preliminarily investigates the impact of several well inclinations on APB behavior. The adopted methodology consists of five main stages: a) a literature review on APB computational modeling in oil wells; b) definition of a reference well scenario; c) selection of well trajectories (inclinations angles) for analysis; d) simulation of the reference well with the selected trajectories; and e) analysis and discussion of the simulation results. The findings of this study are expected to provide a more accurate understanding of the impact of well inclination on APB behavior, highlighting the limitations associated to the use of vertical well models for its predictions.

Elastic stability and post-buckling behavior of circular rings partially supported on rigid saddles

Ricardo Silveira — 2025-12-01

This paper presents a numerical study on the nonlinear elastic response of a circular ring in bilateralcontact with a rigid cradle. The analysis focuses on geometric nonlinearities associated with large displacementswhile assuming linear elastic material behavior. The formulation employs an incremental-iterative approach basedon the corotational description, implemented using the general-purpose software MASTAN2 and the in-housecode CS-ASA, enabling cross-validation of results. Several geometric and loading parameters, including ringslenderness, saddle geometry, and support stiffness, are examined to characterize the load-displacement behaviorand the evolution of the contact region. The results highlight the influence of boundary conditions on the stabilityand post-buckling response, emphasizing the changes in stiffness due to the contact constraint. The close agreementbetween both computational tools confirms the reliability of the adopted modeling strategy. The study contributesto a better understanding of nonlinear elastic interactions between curved members and rigid supports, providinginsights relevant to structural and mechanical systems involving ring- or arch-like components.

Capillary condensation in porous media considering a phase transition multiphase model

Marcelo Adriano Fogiatto — 2025-12-01

Capillary condensation in building porous materials and the correct understanding of its underlying physics, leading to the enhancement of vapor transfer at low moisture contents, is a challenging research theme. In porous constrictions, the adsorbed moisture onto the solid surfaces of a porous material can grow around these surfaces, leading to what we know as capillary condensation: these constrictions are filled with liquid and, in consequence, form bridges, enhancing vapor diffusion through the porous material. Capillary condensation, thus, results from the intermolecular attraction forces between the vapor and the liquid molecules that are adsorbed onto the solid surfaces around constrictions. Vapor sorption has been modeled for materials presenting high hygroscopicity by considering the thermodynamic stability conditions of two-phase systems in single cavities of simple geometry, such as cylinders or spheres (Philippi et al., 1994), (Dutra et al., 2017). The continuous development of more rigorous molecular-based models of fluid systems, together with advances in computational resources, has resulted in an improved understanding of the underlying physics of these systems. In this study, an isothermal multiphase lattice-Boltzmann model with phase transition (Siebert et al., 2014) is used to simulate vapor condensation on the surface of a pore constriction represented by the spacing between two spheres (or cylinders in a two-dimensional simulation). The spacing between the solid bodies dictates if the liquid layers adhered to the walls give rise to capillary condensation. The results tend to agree with the Young-Laplace law. References Dutra, L., Mendes, N., & Philippi, P. (2017). On the characterization of pore size distribution of building materials. Journal of Building Physics. https://doi.org/10.1177/1744259117698515 Philippi, P. C., Yunes, P. R., Fernandes, C. P., & Magnani, F. S. (1994). The microstructure of porous building materials: study of a cement and lime mortar. Transport in Porous Media, 14(3), 219–245. https://doi.org/10.1007/BF00631003Siebert, D. N., Philippi, P. C., & Mattila, K. K. (2014). Consistent lattice Boltzmann equations for phase transitions. Physical Review E, 90(5), 053310. https://doi.org/10.1103/PhysRevE.90.053310

Design of Rectangular Concrete-Filled Steel Tube (RCFT) Columns of Noncompact and Slender Section

Glaubert Miranda — 2025-12-01

Concrete-filled steel tubular (CFT) columns have gained prominence in structural projects due to their high strength, stiffness and combined bending and compression performance. In Brazil, NBR 8800:2024 introduced provisions for the combined bending and compression design of noncompact and slender sections through Calculation Models I and III. This study presents a computational tool developed in MATLAB to verify the combined bending and compression resistance of rectangular composite filled columns, based on the formulations of the NBR 8800:2024. According to the normative stress diagrams, the neutral axis was determined by force equilibrium, applying the secant method for noncompact sections and the quadratic formula for slender sections, while the resistant moments were obtained by moment equilibrium. The normative procedures were evaluated by comparison with experimental results reported in the literature. The findings showed reasonable agreement with the experimental data, with a conservative trend in the assessed cases. However, the scarcity of tests involving noncompact and slender sections still limits a broader assessment of the normative models, highlighting the need for further experimental studies.

Computational Modeling of Fiber-Reinforced Cementitious Matrices

Lucca Rosa Parreira Borges — 2025-12-01

Various studies have been conducted to expand knowledge of the structural behavior of fiber-reinforced concrete (FRC), making it as a potential alternative to conventional reinforced concrete in specific applications. The addition of fibers to the concrete matrix primarily enhances the tensile behavior of plain concrete, improving its resistance to bending and shear forces, as well as its ductility and toughness. These characteristics are directly related to the fact that once fibers are embedded and bonded to the cementitious matrix, they hinder the initiation and propagation of cracks. A common strategy for investigating the behavior of FRC under bending and shear forces is numerical simulation using the Finite Element Method (FEM). The approach used here explicitly models the fibers as finite elements randomly dispersed within the continuous domain of the concrete. The primary objective is to evaluate the influence of the fiber-matrix interaction on the composite’s response by numerically simulating of a series of experimental tests carried out using the commercial FEM software DIANA FEA®. The results show that the response of the numerical model depends on the properties of the fiber-matrix interface used, as well as on the way the fibers are generated.

Application of Random Field Finite Element Analysis to Earthfill Dams

Albert Luiz Follmann — 2025-12-01

The safety assessment of earthfill dams has traditionally relied on limit equilibrium methods, which require the predefinition of a failure surface and provide limited insight into stress-strain behavior and material variability. This study presents a simplified application of Random Field Finite Element Method (RFEM) to model the spatial variability of geomechanical parameters in an earthfill dam. A two-dimensional stress-strain simulation of a clay-core rockfill dam section is performed, incorporating random fields to represent material heterogeneity. The model also implements the shear strength reduction method to estimate reliability metrics. The results are directly compared to those obtained using limit equilibrium analysis to evaluate the observed failure mechanisms and reduction factors. Additionally, a probabilistic analysis considering a homogeneous section was carried out for comparison, allowing the evaluation of the influence of spatial variability on the dam's safety assessment.

Many-objective truss structural optimization considering stability and vibrational aspects

Afonso Celso de Castro Lemonge — 2025-12-01

Many-objective structural optimization problems are well-known and have been discussed in the literature. Concerning framed structures, most formulations consider two objective functions: minimizing weight and minimizing the maximum nodal displacement. This paper aims to expand on and discuss many-objective truss structural optimization problems (MaOSOPs) with seven objectives. The objective functions are the weight, the number of different cross-section types, the maximum nodal displacement, the first natural frequency of vibration, the first load factor concerning the structure’s global stability, the sum of the absolute values of the first vibration mode, and the sum of the absolute values of the first stability mode. The sizing design variables are the cross-sectional areas of the bars, and the constraints are the nodal displacements and axial stress at the bars. The experiments are traditional benchmark problems in the literature. The algorithms used are the multi-objective meta-heuristic with iterative parameter distribution estimation (MM-IPDE), the success history–based adaptive multi-objective differential evolution (SHAMODE), the success history–based adaptive multi-objective differential evolution with whale optimization (SHAMODE-WO), and the third non-dominated sorting genetic algorithm (NSGA-III). Pareto fronts are obtained by providing the decision-maker with a set of non-dominated solutions that can be chosen according to their preferences.

CFD Analysis of Urban Airflow and Particulate Dispersion in a Mining-Influenced Neighborhood Using Real Topography Data

Eduardo Pescada — 2025-12-01

Urban areas located near ore management or mining sites may be affected by air pollution caused by the emission of particulate matter resulting from activities associated with the mineral production chain. Besides the deposition of dust in roads, plants and houses everywhere along the wind stream, suspended particulate matter has negative effects on the population’s health, from aggravating to being the cause of cardiovascular and respiratory diseases. Accordingly, the present study aims to evaluate the airflow through the streets and between residential blocks located in a neighborhood within the urban area of the municipality of Congonhas, in the state of Minas Gerais, situated near mining activities, using computational fluid dynamics (CFD). For this purpose, the software Ansys Fluent was employed to carry out the computational simulations, in which the actual topography of the terrain related to the studied case was considered. Recent works in particulate matter carried by wind uses simplified geometry of the terrain to model the basis of the air flow in urban regions. In contrast, the present analysis uses the official terrain scan made available by the city hall which excludes buildings and vegetation, but considers the detailed topography. Also, the input data of wind velocity profile is from a mining site, to better suit the local context and use an already fully developed airflow. Based on the analysis of different wind directions, it is possible to identify areas with low and high wind velocities, in which occurs deposition and emission of particulate matter. Furthermore, the integrated use of these tools and datasets enables a quantitative assessment of the performance of existing strategies in controlling particulate matter emissions. Moreover, the methodology supports the identification of optimal technologies and spatial deployment strategies for future interventions aimed at achieving effective dust pollution reduction.

Body-fitted mesh approach for structural topology optimization considering stress constraints

Lucas Oliveira Siqueira — 2025-12-01

Effective control of stress distribution in structural topology optimization is critical to avoid structural failure and ensure efficient material utilization. Based on that, a stress-based topology optimization framework is constructed to include stress constraints in the structural design through binary design variables. The solid structure is modeled using the Navier-Cauchy equation, considering linear elasticity. The governing equations are solved using the finite element method. The optimization problem is compliance minimization subjected to volume and stress constraints. The optimization is performed using the Topology Optimization of Binary Structures with Geometry Trimming (TOBS-GT) method, which employs Sequential Integer Linear Programming (SILP). The TOBS-GT method decouples the optimization and Finite Element Analysis (FEA) meshes. This decoupling offers several advantages, such as the flexibility to employ different sets of tools for geometry treatment, mesh building, FEA analysis, updating design variables, and post-processing results. Two classical benchmarking examples are explored: the L beam and the T pier. The proposed framework achieves the volume and stress constraints in both examples.

An optimized adaptive semi-explicit/explicit time-marching formulation for elastodynamic analyses

Caio Amaral — 2025-12-01

This work presents an optimized adaptive semi-explicit/explicit time integration strategy to analyse elastodynamic models. The discussed methodology is based on two time-integration parameters, which are allowed to assume different values for each element of the adopted spatial discretization. The computation of the first parameter is designed to improve accuracy and to ensure the stability of the analysis, and it defines the so-called semi-explicit and explicit elements of the hybrid model. The evaluation of the second parameter, on the other hand, focuses on enabling an effective numerical dissipative algorithm, and it defines the so-called dissipative and non-dissipative elements of the model, which are relabelled at each time step of the analysis. These adaptive time-integration parameters are computed taking into account the physical and geometrical properties of the elements of the spatial discretization, the adopted time-step value, and local previous time step results. The resulting time-marching formulation is very accurate and highly efficient, providing a very attractive solution procedure. To further improve the effectiveness of the discussed technique, optimal time-step values are also here automatically evaluated, taking into account a particle swarm optimization algorithm, which allows to establish the most efficient distribution of semi-explicit and explicit elements along the model for solution. As a consequence, the proposed formulation becomes entirely automated, requiring no expertise, decision nor effort from the user to be applied, further improving its usability. Numerical results are presented at the end of the paper, illustrating the good performance of the optimized hybrid approach.

On the two-dimensional analysis of reinforced materials using the BEM/1D-BEM coupling

Mário Sérgio Oliveira César Filho — 2025-12-01

The design of structures composed of different materials—aiming to harness their individual advantages in the final product—has become increasingly common due to the demands for economic feasibility and high performance in new projects. Among the numerical alternatives for evaluating the mechanical quantities of such structures, the coupling of the Boundary Element Method (BEM) with its one-dimensional version, the 1D-BEM, has shown excellent results when compared to commercial software. Moreover, this coupling proves to be numerically more stable in comparison to the classical one with reinforcements represented by the Finite Element Method. The BEM/1D-BEM formulation presented herein accounts perfect adherence between the matrix domain and the reinforcements. To build the global system of equations, this coupling scheme requires the integral equation for assessing the displacements in internal points, which may be computationally costly. This study proposes a more efficient discretization methodology for the reinforcement domain, reducing the number of internal points required for evaluating the aforementioned integral. Additionally, methods are proposed to avoid accuracy loss due to near singularities resulting from the proximity of sources in a reinforcement domain to other domains. This issue is more common in two-dimensional analyses than in three-dimensional ones and is rarely avoided without proper treatment when analyzing randomly distributed reinforcements. Examples are presented using both the classical BEM formulation, with Lagrangian basis functions for boundary discretization, and the isogeometric one (IGABEM), with discretization based on Non-Uniform Rational B-Splines (NURBS) curves.

Simulating a Flexible Manipulator with Machine Learning: A Study Using Neural Networks

Fábio Félix — 2025-12-01

Parallel Kinematic Machines (PKMs) have been extensively studied in modern robotics due to their structural advantages, such as high stiffness, greater load capacity, and superior precision compared to serial manipulators. However, these advantages come with challenges, particularly regarding the complexity of modeling and dynamic control, due to the presence of multiple closed kinematic chains and strong coupling between links In this work, the development of a computational model is proposed to simulate a planar parallel manipulator of the 3RRR type, whose prototype was developed and built at EESC-USP (FAPESP 2018/21336-0). A relevant aspect of this study is the consideration of link flexibility, a feature that more accurately reflects the physical reality of the system but, on the other hand, makes analytical modeling even more challenging. Flexibility introduces nonlinear dynamic effects and elastic-structural couplings, requiring advanced numerical methods or alternative data-driven approaches for simulation and control. To address these difficulties, Artificial Neural Networks (ANNs), specifically Multi-Layer Perceptrons (MLPs), were adopted to build a predictive model of the manipulator. Eight experiments were conducted on the prototype, in which motor angles and strain signals obtained from strain gauges installed on the flexible links were recorded. These six signals were used as inputs to the MLP, while the end-effector pose data constituted the output targets. The MLP architecture was iteratively optimized by evaluating different combinations of the number of neurons, hidden layers, and learning rates. The training process incorporated early stopping techniques to prevent overfitting. The obtained results show that the MLP was able to successfully learn the relationship between input signals and the manipulator's pose, even in the presence of noisy data from the experimental environment. The neural network's predictions closely followed the actual trajectories, indicating the feasibility of using ANNs for modeling complex flexible manipulators. Future work includes exploring more sophisticated architectures such as recurrent or temporal convolutional networks, as well as integrating hybrid physics-informed models (Physics-Informed Neural Networks – PINNs), which can embed the system's physical equations to further enhance the model's generalization capability.

Automated Detection of Exposed Rebar in Civil Structures Using UAV-Based Image Processing: Case Study in a Real Infrastructure

Vinícius Mota — 2025-12-01

This study proposes and evaluates deep learning models for the segmentation of exposed steel rebar in concrete structures — a critical damage type often linked to concrete delamination. The motivation emerged from technical discussions with the Municipal Infrastructure Secretaria(SMI) of Rio de Janeiro, which identified this pathology as a priority in the inspection of urban bridges. The dataset used includes approximately 100 images, combining real UAV-captured frames and publicly available databases, with precise binary masks manually annotated. Three segmentation architectures were evaluated: U-Net with ResNet-34 encoder, lightweight U²-NetP, and a hybrid model combining Swin Transformer and DeepLabV3. Training employed k-fold cross-validation, data augmentation, and grid search with BCE+Dice+Focal loss functions. Only U-Net and Swin+DeepLabV3 were tested in real UAV inspection footage. Results show that Swin+DeepLabV3 achieved higher accuracy in detectingrebar under diverse lighting and textures, outperforming U-Net especially in multi-instance scenarios. The study highlights the potential of modern semantic segmentation models for field-ready structural diagnostics and contributes to closing a gap in the literature regarding real-world applications focused on rebar exposure.

INFLUENCE OF CASING DECENTRALIZATION ON THE THERMO-POROELASTIC INTEGRITY ASSESSMENT OF CEMENT SHEATHS IN OIL WELLS

Maria Clara Medeiros — 2025-12-01

This study preliminarily investigates the influence of casing decentralization on the structural integrity of the cement sheath in oil wells under thermo-poroelastic conditions. The cement sheath, located between two casings or the casing and surrounding rock formation, performs essential functions, including ensuring casing support, providing hydraulic isolation between producing zones, and maintaining the structural stability of the well. Traditionally, its design is based on thermoelastic models with circular cross-sections and concentric geometry. However, geometric imperfections, such as casing decentralization, can significantly modify stress distribution and compromise the structural integrity of the cement sheath. In addition, the porous behavior of hardened cement, often disregarded, can meaningfully influence its mechanical response. In this context, the methodology comprises: (i) definition of the numerical model; (ii) establishment of material properties and boundary conditions; (iii) verification of the eccentric model; (iv) definition of the case studies; and (v) structural integrity assessment. The results show that decentralization intensifies circumferential stresses in the region of reduced thickness, increasing susceptibility to failure. The inclusion of thermo-poroelastic behavior, although reducing absolute stress levels, amplifies the effects of eccentricity. Thus, this study integrates both factors, providing initial insights for safer and more realistic designs in the oil and gas industry.

Microscale Computational Modeling of the Infiltration of Suppressors into the Porous Surface of Iron Ore Pellets

João Victor Melo Costa — 2025-12-01

The mining industry faces significant challenges with dust emissions during the processing of iron ore pellets. This study employs microscale computational modeling using COMSOL Multiphysics software to simulate the biphasic flow of dust suppressants into the porous surface of pellets. The research aims to enhance the understanding of the interaction between suppressants and the porous matrix of the pellets, analyzing how this interaction affects infiltration efficiency and dust reduction effectiveness. The results demonstrate how certain key properties, such as the internal pore pressure field and the volumetric fraction of solution absorbed by porous channels, are influenced by changes in parameters related to the spraying process and the interaction between the suppressant and the pellet. The study contributes to the theoretical foundation necessary for future research and highlights the importance of understanding the factors that influence the process of mineral dust suppression, in alignment with environmental regulations and sustainable practices. The developed model serves as a valuable tool for engineers and researchers focused on improving the sustainability of mining operations.

Multi-Objective Optimization of Steel and Concrete Tubular Composite Columns

Luis Arthur Pereira Corrêa — 2025-12-01

With the increasing trend of building verticalization, the demand for more efficient cross-sections of structural elements has become essential. Within this context, the use of concrete-filled steel tubular composite columns emerges as a promising alternative, as they provide reduced cross-sectional dimensions while offering high load-bearing capacity, and they eliminate the need for formwork compared to reinforced concrete sections or concrete-encased composite sections. The objective of the present study is to develop a multi-objective optimization formulation for concrete-filled composite columns, considering both the minimization of CO₂ emissions and the maximization of the column’s load-bearing capacity across different cross-sectional geometries. The constraints adopted in the optimization process comply with the specifications of the Brazilian standard for tubular structural profiles. The optimization problem was solved using the Multi-Objective Particle Swarm Optimization (MOPSO) algorithm. The resulting Pareto fronts are compared with those from single-objective optimization problems found in the literature, where CO₂ emissions were considered as the objective function to be minimized.

Automatic Design of Steel and Composite Steel-Concrete Beams According to NBR 8800 Using Genetic Algorithms

Paula Bastos — 2025-12-01

Technological advancements have driven the use of new computational techniques and tools in structural engineering, enabling the accurate simulation of the behavior of increasingly slender and complex structures. Making use of such tools, a Python routine was developed for the design of steel beams with non-slender webs, covering welded, rolled profiles, and mixed steel-concrete beams, in accordance with the NBR 8800:2008 standard. The routine incorporates a genetic algorithm (PyGad) for discrete parametric optimization, selecting the most suitable profile among commercial options. The weighting function considers parameters such as inertia, mass, modulus of elasticity, radius of gyration, and bending strength, with weights adjusted by the algorithm. The verification of serviceability limit states and vibration frequencies is also part of the process, offering a comprehensive approach to design and optimization. After ensuring compliance with the NBR 8800:2008 standard requirements, it was possible to perform several additional analyses with the selected beam, such as cost, weight/performance ratio, boundary conditions, degree of utilization, as well as comparisons between different profile types and beam models. The case study involved the analysis of distinct beam models, allowing an effective comparison between different types of profiles and their characteristics. The results highlighted the program’s ability to provide optimized solutions, considering criteria such as price, weight, boundary conditions and utilization.

Reliability Analysis of Litzka Castellated Beams

Washington Batista Vieira — 2025-12-01

This paper presents a safety evaluation of Litzka castellated beams, based on calculation models proposed in the technical literature. It is noteworthy that there are currently no applicable Brazilian standards or official design criteria for this type of structure. The safety assessment is carried out through structural reliability analysis of the beams and advanced finite element-based analysis of their actual resistance. The effects of geometric imperfections and residual stresses on the resistance of the beams are considered. Statistical parameters related to model error, material strength, geometric properties, and permanent and variable loads are taken into account in the reliability evaluation. Reliability indices are obtained for load combinations according to Brazilian, American, and European standards. These indices reflect the safety level of the beams designed according to the analyzed models. The results are compared with target reliability indices recommended by ASCE and Eurocode.

A Ritz approach for two-dimensional piezoelectric media

Francisco Monteiro — 2025-12-01

In this work, a Ritz procedure is presented for the analysis of two-dimensional piezoelectric media subjected to either applied mechanical loads or applied voltages. The solution domain is thought of as an assembly of disjoint subdomains to properly construct Ritz bases that account for discontinuities in the field variables. Specifically, the procedure is employed in a rectangular bimorph strip divided into two subdomains along the thickness, with convergence achieved by progressively increasing the number of Ritz bases. Only the required C⁰ continuity of displacement and electric potential is strictly enforced between subdomains. The parabolic distribution of the electric potential across the thickness of the strip, a common assumption in beam theories, is critically examined through numerical examples.

The Neighbor Effect Factor: Quantifying the Influence of Adjacent Buildings on Drag Force

GREGORIO VIEIRA — 2025-12-01

Urban areas have been increasingly expanding their built environment. Initially, buildings may be constructed in isolation; however, over time, new structures are inevitably erected nearby, potentially altering wind action in a non-negligible manner—conditions often not accounted for in the original design. This study evaluates the Neighborhood Factor (NF)—defined as the ratio between the drag force on a building in the presence of an adjacent structure and the drag force on the same building in isolation—based on experimental wind tunnel tests. The neighboring building, with the same dimensions as the main building, was positioned in four different locations, all aligned with the studied building. Three wind directions were considered: 0°, 45°, and 90°. For each configuration, mean NF values were computed from time series of drag force measurements obtained experimentally. Results indicate that the presence of neighboring buildings can significantly increase drag force, reaching values above 18% compared to the isolated case, depending on their relative position and distance. These findings highlight the importance of accounting for the aerodynamic influence of surrounding structures in wind load assessments for buildings.

Multi-objective structural optimization of shallow domes considering geometric nonlinearity and grouping of bars

João Marcos de Paula Vieira — 2025-12-01

The formulation and solution of structural optimization problems in trusses is broadly discussed in the literature, commonly involving objective functions such as minimizing the structure's weight and nodal displacements. In most of these works, the structures present elastic and geometrically linear behavior. In this paper, a multi-objective structural optimization problem (MOSOP) is proposed, involving objective functions such as minimizing the weight of the structure, maximizing the first natural frequency of vibration, maximizing the critical load factor related to the global stability and minimizing the number of distinct cross-sectional areas of the bars, simplifying, for instance, the manufacturing, assembling, checking, welding, etc.. The constraints are related to the maximum nodal displacements. The analyzed structure is a shallow dome with geometric nonlinearity. Therefore, a geometrically nonlinear analysis is performed to determine the deformed configuration of the structure, using the cylindrical arc-length method. This analysis allows the designer to obtain more accurate values for the objective functions and constraints. Three evolutionary algorithms are applied to solve the MOSOP, and their performances and solutions are compared. The Pareto front obtained in the proposed problem is presented, for example, illustrating how the growth of the dome's weight leads to increases in the first critical load factor and the first natural frequency of vibration. It is also possible to note how the structural weight can be reduced by grouping bars with a higher number of distinct cross-sectional areas. Finally, optimized solutions are extracted from the Pareto front based on the decision-maker's preferences.

An Artificial Neural Network Application for Efficient System Reliability Analysis

Bruno Gustavo Dos Santos — 2025-12-01

Artificial neural networks (ANNs) have emerged as a powerful tool for structural reliability analysis due to their capacity to effectively capture intricate, nonlinear relationships between input and output variables. In this domain, ANN models are typically trained on independent variables, such as material and geometric properties, to establish the connection between these inputs and the output variable, usually a probability of failure. However, despite their potential, ANNs are often disregarded in favor of well-established and computationally tractable methods like Polynomial Chaos Expansions and Kriging. Nevertheless, ANNs offer a significant advantage in addressing system reliability problems with multiple outputs. Their inherent ability to generate multiple outputs simultaneously eliminates the need for resorting to numerous surrogate models or intricate formulations, thereby leading to reduced computational demands. This paper proposes an efficient ANN-based surrogate modelling approach for system realibility analysis. A meta-heuristic optimization algorithm is employed to select the best architecture, considering three possible types of metamodels: multilayer perceptron (MLP), radial basis function (RBF) and long short-term memory networks (LSTM). A single optimal ANN is then used as a surrogate model for system reliability problems, whose several limit states are represented by the neurons on the last layer of the optimal network. Results are compared with classical surrogate models. Five examples are addressed, showing that multiple output ANNs can be a viable approach in this context.

Numerical Benchmarking of Stabilized Finite Element Methods Applied to Population Balance Equations for Modeling Crystal Size Distributions

Andrés Mauricio Nieves Chacón — 2025-12-01

Crystallization fouling is a prevalent phenomenon observed in heat exchangers, involving the accumulation of crystals on the heat transfer surfaces. It degrades equipment performance and increases operational costs in industry. Crystals are transported by diffusion from higher to lower concentration regions, with a reaction rate governing their incorporation into the crystal lattice formed on the heat exchanger walls. Its removal is promoted by the action of hydrodynamic forces and the effect of shear stresses. Mathematical models are developed to reproduce both the deposition and removal processes. This paper presents a preliminary study intended to support future research in which a new mathematical formulation for the removal term will be proposed. Removal is directly affected by crystal size, as its variation influences mechanical properties such as material strength, Young’s modulus, and hardness, as well as thermophysical properties, including thermal conductivity and specific mass. Thus, characterizing crystal size distribution within the fouling layer is crucial. Population balance equations (PBEs) are typically used to model size distribution dynamics, where different mathematical methods, such as the Method of Characteristics or Finite Volume Schemes, have been proposed for solving these equations. In this paper, we investigate the performance of several Stabilized Finite Element Methods (SFEMs) applied to PBEs for modeling the growth, aggregation, and nucleation processes involved in the crystallization fouling process. The stabilized formulations examined in this work include the Streamline Upwind Petrov-Galerkin (SUPG) method with various stabilization parameters, the Unusual Stabilized Finite Element Method (USFEM), and the Rothe’s method, employed here as a less conventional discretization technique for transient transport equations. The performance of these SFEMs is assessed based on their ability to reproduce expected physical behavior, providing insights for future model development.

Full-Waveform Reconstruction of Microseismic Sources From Hydroacoustic Observations

Welerson Kneipp — 2025-12-01

This work presents a full-waveform reconstruction strategy to identify microseismic events in solid–fluid systems using multi-frequency acoustic pressure measurements collected in the fluid. The goal is to recover both the locations and seismic moment tensors of the events, motivated by applications such as monitoring reaction-induced fracturing in subsurface carbon storage formations.Microseismic sources are modeled as dipole and double-couple point forces embedded in the solid domain. The reconstruction is carried out via a topological derivative approach, in which the leading-order perturbation of a misfit functional is driven by the elastic energy released through the creation of an infinitesimal fracture at a trial location.The forward problem is formulated in the frequency domain, coupling elastodynamics in the solid and acoustics in the fluid. The inversion algorithm combines a discrete grid search, adaptive refinement strategies, and multi-frequency data assimilation. Numerical simulations in two dimensions demonstrate the ability to reconstruct multiple sources using a modest number of hydrophones. The results also highlight the interplay between spatial and spectral aperture, showing that limitations in one can, to a degree, be compensated by abundance in the other.

Flexible job shop scheduling problem with non-linear routes, energy awareness and position-based learning effect

Débora Pretti Ronconi — 2025-12-01

The flexible job shop scheduling problem is notable for its many practical applications, such as in the on-demand printing industry, glass industry, steel production planning, and automotive repair shops. Additionally, sustainability has become one of the main objectives across all human activities, particularly in production environments. This work we consider the flexible job shop scheduling problem with the objective of minimizing energy consumption. As it is known that a considerable part of the energy consumption occurs when the machines are on and idle, the addressed problem includes the possibility of turning the machines off and on between processing operations. To bring the problem closer to the large variety of real-world problems it encompasses, we include two relevant factors: nonlinear routes and position-based learning effect. The treated problem is formally described through a mixed integer linear programming model. We propose constructive heuristics and two types of neighborhood with which we construct local search schemes. The local search strategies both modify solutions by removing and reinserting operations, but in different ways. The first, called SRRN (Single-Operation Removal and Reinsertion Neighborhood), removes and reinserts a single operation in a position that avoids forming cycles. The second, called SRDRRN (Single-Operation Removal, Destruction, Reinsertion, and Reconstruction Neighborhood), removes an operation along with its successors. If it is reinserted elsewhere, the new position’s operation and its successors are also removed. The resulting partial solution is then rebuilt using the proposed constructive heuristic. In addition, three meta-heuristics – namely, variable neighborhood search, greedy adaptive search procedure, and simulated annealing – have been tailored for the problem at hand. We conduct a large number of experiments to evaluate the performance of the introduced methods, on small-sized and large-sized instances. In the large-sized instances, the general variable neighborhood search, that combines the two neighborhoods into a single method, is particularly effective. In the small-sized instances with known optimal solution, the greedy randomized adaptive search procedure finds solutions that, on average, are within 0.22% of the optimal solution. The instances evaluated in this work are all available at http://www.ime.usp.br/∼egbirgin/ for future comparisons.

CFD-Based Study of Boat Tail Length Effects on 155 mm Projectile Drag

André Luiz Tenório Rezende — 2025-12-01

The study of projectile aerodynamics plays a crucial role in the development and optimization of ammunition, particularly in enhancing range and accuracy. In this context, the present work aims to investigate the effects of varying the Boat Tail length of a 155 mm artillery projectile on its aerodynamic drag coefficient. The methodology involved computational fluid dynamics (CFD) simulations in which several geometrical configurations of the projectile’s Boat Tail were analyzed. These simulations allowed for a detailed evaluation of the pressure and velocity fields around the projectile's surface, leading to the determination of aerodynamic coefficients for each configuration. The numerical results were compared to reference data obtained from the Projectile Design and Analysis System (PRODAS), serving as a benchmark to validate the computational findings. The analysis revealed which Boat Tail lengths are most effective in reducing aerodynamic drag. It was observed that specific geometric proportions result in a noticeable decrease in the drag coefficient, thus confirming the efficiency of the Boat Tail design technique in aerodynamic optimization. Moreover, the study underscores the relevance of rear-end geometry in the overall aerodynamic performance of artillery projectiles. The results provide valuable technical insights for the design and improvement of future ammunition, contributing to the broader field of external ballistics. The findings also highlight the importance of computational modeling as a powerful tool for evaluating structural modifications and guiding design decisions aimed at reducing aerodynamic resistance and improving projectile efficiency.

A Framework for 2D Mesh Generation from Computed Tomography of Rock Samples

Victor Villegas — 2025-12-01

Generating meshes is essential in fields like numerical simulations, computer graphics, and finite element analysis. CT scan technology is now widely used in different areas beyond medicine, including geosciences. The use of digital representations of rocks enables capturing different physical properties of real samples without altering the original rock samples. This study introduces a structured approach to digitally reconstruct rock samples from raw CT scan data and create 2D finite element meshes on Gmsh, incorporating both external libraries and self-developed algorithms. Tests are conducted on diverse rock types, including sandpack and carbonate. The level of mesh coarseness is regulated by specifying an area tolerance parameter between the original connected regions and its discretized representations. The resulting meshes are applied in finite element simulations using the NeoPZ environment.

A Technique for Synthesizing Spectrum-Compatible Artificial Ground Motions

Daniel Barbosa Mapurunga Matos — 2025-12-01

In seismic engineering, nonlinear dynamic analyses have become vital tools for accurately assessing the safety of critical infrastructure, in which estimating collapse probability and verifying serviceability levels are priorities. However, pseudo-acceleration response spectra prescribed by standards such as Eurocode 8, NSR10 or NBR 15421 lack temporal information, which prevents the adequate representation of seismic load evolution. This limitation hinders the ability to capture phenomena such as stiffness degradation, damage accumulation, or stress redistribution, all of which are fundamental to nonlinear analyses. Consequently, the creation of artificial seismic records stands out as a feasible approach for studies in the time domain. Classical methods such as the Kanai-Tajimi model (KANAI, 1961; TAJIMI, 1960), are widely used due to their simplicity but show poor spectral fidelity when compared to target design spectra. To overcome this challenge, this study introduces a methodology inspired by Clough and Penzien (1995). The proposed methodology facilitates the generation of synthetic accelerograms whose spectral response aligns with high fidelity to the normative pseudo-acceleration spectrum, ensuring their effective application in nonlinear dynamic analyses. An additional application of this methodology is the reduction of real signals to optimize nonlinear structural analysis. It is common in the literature to use trimmed segments of real signals to reduce processing time. However, such trimming can alter the original statistical properties of the signal. The methodology proposed in this study enables the generation of artificial signals that preserve the statistical characteristics of real ones, but with shorter duration, maintaining data integrity for structural analysis. The main objective of this study is to assess the method’s effectiveness in generating artificial records from normative spectra and real signals, verifying both its spectral accuracy and applicability in advanced seismic design and structural assessment contexts. Preliminary results show strong agreement with design spectra and stable frequency content distribution. In conclusion, the proposed method constitutes a precise, robust, and conceptually consistent tool for generating synthetic seismic signals from normative spectra and real signals, fulfilling the spectral and temporal requirements necessary for their use in nonlinear structural response studies.

MODELING AND SIMULATION OF PULL-OUT OF FLEXIBLE LAZY-WAVE RISERS

Luis Carlos do Nascimento Filho — 2025-12-01

The decommissioning of offshore structures is a challenge due to the numerous factors influencing the decision-making process regarding the methodology to be adopted. Activity planning involves a multidisciplinary study that considers regulatory, logistical, economic, environmental, and operational issues. One of the main stages is the pull-out, which consists of disconnecting and retrieving flexible pipes. This stage is a time-consuming and costly procedure due to the need for support vessels. Consequently, there is a constant search for alternatives to optimize this activity. A simple and economical alternative is the surface cutting of the riser followed by its free fall to the seabed. Assessing the technical feasibility of this alternative requires computational simulations to determine the geometry and internal forces during the fall. It is important to note that cutting the riser leads to the sudden release of a large amount of energy, in addition to generating a compression wave that travels through the riser, which can cause the phenomenon of dynamic buckling when another compression wave returns. Thus, this work aims to develop a computational model able to simulate the surface cutting of a flexible lazy-wave riser, considering the dynamic and nonlinear aspects involved. The studied parameters include mesh discretization, the time increment of the integration algorithm, and the effect of the drag coefficient. Initially, the riser’s static configuration will be obtained to verify the loads before cutting. Subsequently, the surface cutting of the structure will be simulated, followed by a study of the riser’s configuration during the fall, as well as the envelopes of velocity, acceleration, effective tension and compression, and curvature radius, which are all essential for assessing the operation's feasibility.

ECG Heartbeat Classification using Artificial Neural Networks: a comparative study of PyTorch and TensorFlow

Ítalo Flexa Di Paolo — 2025-12-01

The automatic classification of arrhythmias from ECG signals transformed into images represents a major advance in computer vision for cardiology, enabling faster and more accurate diagnoses. This study presents a comparative analysis of the deep learning frameworks PyTorch and TensorFlow for arrhythmia classification, using the publicly available PTB Diagnostic ECG Database. ECG signals were segmented and labeled into two categories (Abnormal and Normal), generating 14,552 vectors with 187 positions each at 125 Hz, divided into training and testing sets (80/20). Different signal representations were evaluated, including one-dimensional (1D) and two-dimensional (2D) formats, with Recurrence Plot used for image transformation. Implemented models included multilayer Perceptron’s and convolutional neural networks (ResNet-18 and AlexNet). Performance was evaluated using various metrics, while computational efficiency was measured by training time and hardware utilization. The results showed that AlexNet in PyTorch achieved the best performance (99.55% accuracy, 99.49% precision, 99.38% recall, 99.44% F1 score, 0.9889 MCC), outperforming TensorFlow (99.17%, 98.97%, 98.97%, and 98.97%, 0.9784). The comparative analysis showed that PyTorch proved more efficient and accurate under identical conditions.

Non-Intrusive Data-driven Surrogate Models for Predicting Turbidity Currents Deposition from 3D Simulations

Roberto Machado Velho — 2026-01-01

Non-intrusive data-driven methodologies, such as POD-DL, offer a powerful approach for constructing surrogate models to address complex parametric problems. POD-DL integrates deep neural networks and follows a multi-step dimensionality reduction process, beginning with a linear reduction via Proper Orthogonal Decomposition (POD) and followed by a nonlinear reduction using a deep autoencoder. A subsequent nonlinear regression, implemented with a forward neural network, accounts for temporal and parametric coefficients.The framework involves numerous hyperparameters, including the number of POD basis modes, network architecture specifications, and typical neural network parameters like learning rate and batch size, all of which influence both training time and model accuracy. In a previous work, we applied this scheme to study gravity currents under the 2D lock-exchange configuration [3] for multiple angles of the initial configuration [4]. The angle of the channel served as a parameter, i.e., different angles generated different dynamics that were learned by the surrogate model. The regression neural network could then predict the dynamics for unseen angles. Now, we extend the methodology to more realistic scenarios, a 3D channel-basin configuration for gravity currents, adapted from [5], analyzing variations in inlet velocities and sediment concentrations. Synthetic data were generated through large-scale parallel finite element simulations, allowing POD-DL to predict deposition maps for unseen parameter values. [1] M. Cracco et al., “Deep learning-based reduced-order methods for fast transient dynamics”, Arxiv Preprint 2212.07737, 2022. [2] S. Fresca and A. Manzoni, “POD-DL-ROM: Enhancing deep learning-based reduced order models for nonlinear parametrized PDEs by proper orthogonal decomposition”, Comput. Methods Appl. Mech. Engrg., 2022. [3] V. K. Birman, B. A. Battandier, E. Meiburg, and P. F. Linden, “Lock-exchange flows in sloping channels”, Journal of Fluid Mechanics, 577:53–77, 2007. [4] R. M. Velho, A. M. Cortes, G. F. Barros, F. A. Rochinha, and A. L. G. A. Coutinho, Advances in Data-Driven Reduced Order Models Using Two-Stage Dimension Reduction for Coupled Viscous Flow and Transport. Finite Elements in Analysis and Design, vol. 248, 2025.[5] T. Spychala, J. T. Eggenhuisen, M. Tilston and F. Pohl,The influence of basin setting and turbidity current properties on the dimensions of submarine lobe elements, SEDIMENTOLOGY, 2020.

FAST ModDef: An interative software for pre-processing Finite Element and Isogeometric Models of Composite Stuctures

Elias Barroso — 2025-12-01

The finite element method (FEM) is a numerical method widely used for structuralanalaysis, specially in the case of structures with complex material, loading and supportconditions. In FEM based simulation systems, the pre-processing stage involves theconstruction of geometric model using a interactive CAD module, the construction ofanalysis model using a mesh generation module, and definition and application ofanalysis attributes (simulation parameters, material properties, loads and supports). Incontext of the Isogeometric Analysis approach, the underlying mathetical representationsadopted in geometric modeling are also used in the simulation formulation, as RationalBézier Surfaces and NURBS, eliminating the geometry approximation error presented inclassical FEM and providing superior refinement capabilities (hpk-refinements). Having aproper interactive graphic system is crucial for pre-processing finite element models. Thispaper presents FAST ModDef, an academic pre-processing FEM tool for strcuturalanalysis of isotropic and composite structures, for FAST (Finite element AnalySis Tool)created at Laboratório de Mecânica Computacional and Visualização (LMCV). FASTModDef focus in the modification of existing analysis data, or preliminar analysis model(FEM mesh). It focus in definition and modification of analysis attributes, and also providesbasic refinement routines. It support both 2D models (e.g. plante stress, plate, plane frameand truss), and 3D models (solid, surface shells and 3D frames and space truss) aresupported. FAST ModDef is written in C++ using Qt 6 framework for user interfaceinteraction and OpenGL 4 for rendering. The program data structure (i.e. topology,geometry and simulation data) is detailed, as well the routines used for rendering andmanipulation of the analysis model. Moreover, the impact of the software in manyacademic activities (graduation and post-gradutaion courses) at LMCV is discussed.

Multi-Objective Optimization of Tubular Composite Trusses: Analysis of CO2 Emission and Load Bearing Capacity

Carolina Oliveira — 2025-12-01

The use of steel-concrete composite systems has been increasing over the past decades. Among the existing systems, steel-concrete tubular composite trusses stand out for their ability to span longer distances with reduced mass compared to full-web beams. The aim of this study is to present a multi-objective formulation for the topological optimization of steel-concrete tubular trussed beams, considering the possibility of filling the upper chord with concrete. The problem was modeled using bar elements with two degrees of freedom per node. The objective functions considered were CO₂ emissions, energy of fabrication process, and the load-bearing capacity of the truss, while the constraints were based on the requirements of Brazilian standards for laminated and tubular profiles. To solve the optimization problem, Multi-Objective Particle Swarm Optimization (MOPSO) was employed to generate Pareto fronts. The formulation was validated and compared with examples from the literature in which single-objective optimization of CO₂ emissions in beams was performed. Preliminary results indicate that the algorithm was effective in determining solutions when compared to the single-objective approach and led to more efficient structural systems, with increased load-bearing capacity and minimal CO₂ emissions and energy

Numerical Analysis of External Diaphragm Connections Between I-Beams and Circular Hollow Section Column

Macksuel Soares de Azevedo — 2025-12-01

This article addresses the structural behavior of the connection between circular hollow steel tubular columns and I-section beams in civil construction. As a strategy to improve the stiffness of this connection, external diaphragms are used — flat components incorporated at the interface between the column and the beam. It is noteworthy that the Brazilian standard ABNT NBR 16239:2013 does not cover the design of this type of connection. The objective of this study is to numerically analyze the behavior of welded connections between a circular hollow section column and I-section beams, using external diaphragms in X-type and XX-type connections. These connections were subjected to compression and bending loads, and variations in the column diameter and wall thickness were also considered. Numerical modeling was performed using the finite element method and showed agreement with the numerical and experimental results obtained by Winkel (1998). The performance of the external diaphragm proved to be satisfactory, as it resulted in a considerable increase in stiffness, enhancing the connection’s load-bearing capacity, as well as the positive effects of increasing the column’s diameter and wall thickness.

Comparative Analysis of Concrete Strengths for High-Rise Column Design

Marlon Justiniano Pansiere — 2025-12-01

This resarch presents a case study aimed to evaluate the technical and economic feasibility of applying concrete with a characteristic compressive strength (fck) of 90 MPa in the columns of a 130 m high building. To this end, a comparative analysis was conducted with the fck 50 MPa concrete originally specified in the project. By maintaining the original geometric dimensions of the columns and utilizing the fck 90 MPa concrete, a reduction in the longitudinal reinforcement rate was observed in some columns at the lower floors, reaching the minimum reinforcement rate at the upper floors. The final stage of the study included a comparative analysis of the costs and the reduction in concrete volume between the evaluated models.

Assessment of High Buildings Human Comfort Subjected to Wind Actions based on Simplified Design Formulations and a Nondeterministic Dynamic Structural Analysis

George Lucas da Silva Quintanilha — 2025-12-01

The requirements for space in major urban centres have driven the development of structural designs for high-rise buildings. Consequently, these structures are more susceptible to wind-induced effects, which may lead to excessive vibrations and occupant discomfort. This way, this investigation aims to evaluate the human comfort levels in a 48-storey steel-concrete composite building with total height of 172.8 m, when subjected to wind loads. The mean and fluctuating wind responses were determined in accordance with the recommendations of the Brazilian standard NBR 6123, considering the wind velocity variations ranging from 5 m/s (18 km/h) to 50 m/s (180 km/h). The numerical model of the building was developed using the Finite Element Method (FEM), implemented through the ANSYS computational software. The assessment of the investigated building human comfort was based on the maximum accelerations values, in time and frequency domain, considering several design standards and simplified formulations, and also a nondeterministic dynamic structural analysis. The investigated scenario allows a general comparative assessment of the results calculated based on the use of simplified formulations and through a nondeterministic structural response. Finally, it must be emphasized that the structural engineers have to be alerted about the relevant quantitative differences associated to the peak accelerations values determined by simplified formulations and based on a traditional nondeterministic dynamic structural analysis.

Analysis of Prestressed Concrete Behavior under Partial Strand Failure

Isaac Ishigami Bastos Terra — 2025-12-01

This study presents a technical investigation into the structural behavior of prestressed concrete elements under partial failure scenarios involving strands rupture. The research combines a literature review of typical failure mechanisms—such as prestress losses, localized cracking, and the consequences of strand rupture—with a numerical simulation performed using the TQS, a structural analysis software. In the computational phase, a prestressed concrete slab is modeled with realistic loading conditions, material properties, and prestressing profiles. The model simulates the abrupt rupture of one or more strands in critical regions to evaluate the structural response. The results indicate that the failure of a single strand promotes internal force redistribution and discrete alterations in deflection behavior, without significant impacts. Conversely, the rupture of multiple strands in the same slab notably compromises the structural performance, with redistribution of efforts and sharper deflections. The numerical model proved effective in visualizing and quantifying these vulnerabilities.

On the effect of fractional damping on the nonlinear vibrations of structures

Zenon Jose Guzman del Prado — 2025-12-01

In this work the effect of the fractional-order damping on a discrete system described by the forced Duffing oscillator is studied. For this, numerical methods are implemented a parametrical study is performed to evaluate the influence of fractional order derivatives on the nonlinear dynamic response. Numerical simulations implemented in to study the nonlinear dynamic behavior of bifurcations, chaos, and chaotic attractors and the stability of fixed points is analyzed. It is possible to observe that by increasing the value of the fractional-order parameter, the resulting discrete response is stabilized.

PARAMETRIC ANALYSIS OF FLOW IN A VERTICAL POROUS CHANNEL WITH MULTIPLE FRACTURES

Lucas Dias — 2025-12-01

During the drilling stage of oil wells, the flow of drilling fluids through the annular space is influenced by discontinuities and high-permeability regions in the geological formation. This phenomenon, known as lost circulation, reduces operational efficiency and compromises well integrity. One of the preventive control methods for this issue is the selection of a drilling fluid whose formulation is best suited to the drilling phase and well characteristics. This process can be assisted through numerical simulations using CFD (Computational Fluid Dynamics), where the annular space is simplified as a vertical channel, fractures are modeled as transverse channels with constant thickness through which leakage flow occurs, and the high-permeability geological formation is treated as a porous medium. This characterization refers to a partially porous and fractured channel (PPFC), modeled as a bidisperse porous medium, where both the free-flow and porous regions are considered homogeneous. This study aims to analyze lost circulation in a PPFC with multiple discrete fractures. The flow is assumed to be steady, incompressible, and the fluid is modeled as a power-law fluid. The porous medium is considered isotropic. Flow in the free channel and fractures is governed by the Navier-Stokes equations, while flow in the high-permeability region is described by the generalized porous media equation proposed by Vafai and Tien (1981). Numerical simulations are performed using the finite volume method in OpenFOAM version 24.06, employing an adapted version of the porousSimpleFoam solver to represent non-Newtonian fluid flow in porous media. The simulations are set up with a porosity of ε = 0.7 and permeability of 10⁻⁶ m². The drilling fluid is idealized with a consistency index k = 0.205 Pa·sⁿ, a power-law index n = 0.568, a density ρ = 1012.60 kg/m³, and a Reynolds number Re = 1000, calculated based on the annular space. The results are obtained from combinations of 2 or 3 fractures with thicknesses eₓ = 4, 16, and 40 mm, spaced at distances h = 100, 200, and 400 mm. The analysis focuses on evaluating the velocity, pressure, and viscosity fields, the leakage rate, and the pressure drop along the channel.

The Influence of the Fluctuation Term on the Accuracy of Multiscale Analysis Results

Dayane Lima — 2025-12-01

This work investigates the influence of the fluctuation term in the formulation of the homogenized constitutive tensor within a multiscale approach applied to the analysis of heterogeneous media. A phase-field constitutive model is used, coupled with the multilevel finite element method, which allows for the simultaneous consideration of macro and micro scales. The phase-field model enables the diffuse representation of fractures, capturing the evolution of the displacement field and crack propagation without the need for an explicit description of the fracture geometry. This makes the model particularly useful for materials like concrete, which may exhibit complex cracking patterns.At the macroscopic scale, the material's constitutive behavior is not directly defined; it must be obtained from solving the microscopic scale, where the local material properties, such as variations in stress and strain, are better represented. The connection between the scales requires defining homogenized parameters that adequately represents the material’s behavior. Literature suggests that this transition cannot be done through a simple average of the local constitutive tensors, as this approach disregards important fluctuations in the displacement field.To ensure theoretical consistency, the inclusion of a fluctuation term is required, which depends on the inverse of the stiffness matrix of the microscale problem. However, calculating this term results in a significant computational cost. With the multiscale implementation in place, comparative tests were conducted between simulations with and without the fluctuation tensor, evaluating both the accuracy of the results and the impact on processing time. All implementations were carried out using INSANE, an open-source platform developed by the Department of Structural Engineering (DEES) at the Federal University of Minas Gerais (UFMG).

A comparative study of contact models for determining geometric properties of the wheel/rail interaction

Thiago Leister Sa — 2025-12-01

Creep forces are the primary mechanism for steering a wheelset on a rail track. Therefore, their evaluation is essential for modeling the train lateral dynamics, wear and rolling contact fatigue on rails and wheels. These forces are highly dependent on geometric properties such as contact position and wheel rolling radius. This study presents a comparative analysis of obtaining such contact properties using three different 2D contact models: rigid (geometric) contact, elastic penalty contact and elastic barrier contact. The case study is a scenario of a wheelset with new wheels in contact with a track with new rails. The three contact formulations yielded similar results for the wheel rolling radius, except for the regions of contact discontinuity and flange contact.

MODELING AND SIMULATION OF A LANDING GEAR SYSTEM WITH EULER-LAGRANGE VECTOR BOND-GRAPH APPROACH

Yuri Silva Souza — 2025-12-01

This work aims to develop and simulate a model of landing gear system during an aircraft touch down. During an aircraft development process, many experimental tests are made to ensure the functionality and safety of the landing gear systems. Instead of using iron bird prototypes to perform the tests, using digital twins as high-fidelity model simulations to predict their physical behavior is safer and less expensive. This work applies the vector bond-graph method to model the landing gear dynamics during touchdown. The aircraft landing gear dynamics are modeled as a multi-body system. The methodology uses modern modeling tools based on a combination of vector bond graph (VBG) representation allied to the Euler-Lagrange variational method for system dynamics modeling. The system studied consists of a planar aircraft model composed of three main modules: a nose landing gear, the airframe, and the aircraft's main landing gear. The landing gear system was first modeled using a vector bond graph representation. Once the effectiveness of this modeling methodology was proved, the resultant vector bond graph was analyzed from the Euler-Lagrange standpoint by determining the system's generalized coordinates and total Lagrangian function. The symbolic algebraic processing was applied to determine the coupled nonlinear differential equations in a symbolic form. In this case, the energy interactions of all the system elements were calculated to generate equations of motion in the form of Euler-Lagrange equations. The analytical equations were implemented in a Matlab/Simulink computational environment to simulate a typical landing condition. A vertical load was applied to the model, simulating a smooth aircraft touchdown. When the aircraft speed reached 40 m/s, and the vertical load reached the aircraft's total weight, braking force was applied to decelerate the aircraft until it stopped. The simulation results showed the effectiveness of the proposed methodology in the aeronautical field. Although the chosen study case comprises a purely mechanical system, the combined vectorized Lagrangian bond graph modeling methodology can be applied to the Multiphysics domain typical of several aeronautical systems of interest.

Experiences On Teaching Computer Graphics for Engineers and Scientists

Andre Pereira — 2025-12-01

In this paper, we describe practical experiences, acquired over several years, in the development of a graduate level course in applied computer graphics. Since the target audience of the course is engineers and scientists, we focus mainly on topics related to pre-processing and post-processing techniques, which, although fundamental for performing numerical simulations, are usually neglected in the educational process. The course structure involves and covers the following topics: representation of curves and surfaces; brief introduction to computational geometry; topological data structures; mesh generation: structured and unstructured; Delaunay triangulation and boundary propagation; modeling and representations of solids: models by space decomposition and B-Rep; overview of visualization techniques. Our main challenge is to teach very complex subjects that students will be exposed to for the first time in this course. However, students who pursue such courses are usually at a high level of maturity and understand that the topics covered are fundamental for their career and research. Therefore, to give an effective and practical character to the course, we follow an approach that is completely based on the development of practical software with user interfaces, relying on open source packages with Python APIs. We chose to use Qt and OpenGL to create interactive graphical user interfaces in a hands-on scheme. The course starts with the development of a curve collector and ends with a full-functional 2D mesh generator. Experiences acquired over the years are combined to achieve the best performance and learning process. A taste of the course can be found at https://web.tecgraf.puc-rio.br/~lfm/compgraf-251/.

Study of the influence of adaptive refinement strategy parameters to phase-field modeling of fracture

Larissa Novelli — 2025-12-01

The phase-field model has consolidated itself as a promising tool to handle complex cracks. In this model, the crack is represented in a diffuse form where the phase-field variable indicates the degradation of the material. To better approximate this variable, extremely refined meshes are required, which demands expensive computational efforts. In this work, an adaptive refinement strategy is adopted to phase-field modeling of fracture. In this strategy, the domain is initially discretized with a coarse finite elements mesh and during the analysis the crack regions are automatically replaced by a Smoothed Point Interpolation Methods (SPIMs) and refined. The SPIM is a meshfree method that possess the kronecker delta property. This property facilitates the application of boundary conditions and guarantees direct coupling with finite elements. Different criteria can be used to identify the crack regions. Furthermore, parameters such as refinement level and size of the substitution region are adopted in the strategy. For the SPIM, different smoothing domains and support nodes selection are applied. The aim of this work is to evaluate the influence of these adaptive strategy parameters on the crack propagation. Numerical simulations are performed and the results are compared with the previously refined FEM and with results available in the literature.

Implementation of the Phase Field Method for Structural Topology Optimization in a FEM Framework Developed in C++

Guilherme Teixeira Pimentel — 2025-12-01

Topological optimization, specifically with a focus on compliance minimization, seeks the best material distribution within a design domain given different loading and constraints configurations. The Phase Field Method (PFM) offers a continuous and well-posed formulation from a mathematical point of view, naturally regularizing the problem by controlling the interface thickness, without the need for artificial filters. This paper presents the implementation of the PFM for topological optimization of structures in a finite element (FEM) framework developed in C++. The main objectives are to incorporate the PFM directly into a FEM code and compare its results with other methods, such as Solid Isotropic Material with Penalization (SIMP). The theoretical foundation of the study is based on a variational formulation, combining the structural equilibrium equation with the Allen–Cahn evolution equation to govern the material distribution. The coupling of the equations is made through a staggered scheme, this implies that each equation is solved separately and the outcome of each equation is iteratively exchanged to update the other. The results are used to confirm the successful implementation of the method and to demonstrate the phase field method ability to produce smooth and physically consistent topologies.

Strategies to Enhance Column Robustness and Reduce CO2 Emissions in Progressive Collapse Scenarios: A Risk-Based Optimization Approach

Luiz Eduardo Gonçalves de Mattos — 2025-12-01

A significant challenge in structural engineering is the rational design of systems exposed to exceptional events, often referred to as low-probability, high-consequence actions, such as fires, earthquakes, explosions, and impacts. Climate change exacerbates these risks by increasing the frequency and severity of hazards, potentially exposing structural vulnerabilities. Structural robustness refers to the capacity of a structure to withstand such actions, preventing initial and localized damage from escalating into disproportionately severe consequences, such as progressive collapse. Traditional reinforcement techniques, while mitigating risks, are often costly and may be inefficient under uncertain triggering events or simultaneous failure of multiple elements. Moreover, structural reinforcements typically involve unsustainable material consumption, significantly contributing to greenhouse gas emissions, highlighting the urgent need for sustainable alternatives. Studies reveal a notable lack of multifaceted protection techniques capable of addressing different levels and stages of failure in the progressive collapse context. The main objective of this work is to perform a risk-based optimization of alternative protection techniques to mitigate progressive collapse in reinforced concrete systems, focusing on solutions that enhance structural safety through robustness, while prioritizing cost-efficiency and environmental sustainability. The proposed methodology, grounded in structural reliability, accounts for random and epistemic uncertainties in design variables and hazard probabilities, and aims to minimize global collapse probability and the associated risks in terms of costs and CO₂ emissions. The techniques explored include ultra-high-performance fiber-reinforced concretes (UHPFRC), traditional reinforcement approaches (Alternative Load Path method), and the combination of these to achieve optimized solutions. The study focuses on column reinforcement in flat slab parking garages exposed to fire and vehicular collisions, as these structures are particularly vulnerable to column loss due to the punching shear phenomenon and increased fire risk from flammable materials. The results include innovative and robust solutions capable of transforming engineering practices, advancing structural safety, sustainability, and cost-efficiency in the construction sector.

Assessment of Inherent Variability from SPT Data and Its Application to Shallow Foundations under Axial Loading

Jose Manuel Guerra Colque — 2025-12-01

This study proposes a practical framework for evaluating the inherent spatial variability of geotechnical parameters based on Standard Penetration Test (SPT) data. The approach includes the implementation and validation of numerical routines for calculating mean, standard deviation, probability distribution fitting, and the scale of fluctuation using vertical SPT profiles. The method is designed to support probabilistic and random field modeling in geotechnical applications. To illustrate the applicability of the developed routines, a case study is presented involving shallow foundations under axial loading. The geotechnical variability inferred from SPT data is incorporated into the bearing capacity assessment using random field theory. This allows for quantifying the influence of spatial variability on design parameters and failure probabilities. The study emphasizes the importance of adequate preprocessing of field data, including trend removal and normalization, and demonstrates how variability metrics—especially the scale of fluctuation—can be effectively integrated into reliability-based design approaches.

Nonlinear Analysis of Prismatic Reinforced Concrete Structures under Harmonic Loads

Adriano Louro Rocha — 2025-12-01

The dynamic analysis of structures differs from static analysis mainly due to the consideration of load variations over time. Harmonic loads, also known as cyclic loads, are among the most critical factors requiring precise evaluation to predict structural behavior accurately. These loads involve periodic variations over time, such as those induced by earthquakes, wind, or traffic, and can lead to complex responses in structures, including significant nonlinear effects. Thus, it is essential to design buildings to ensure safety and comfort, avoiding unexpected performances. In dynamic analysis, the use of computational simulations is crucial, as it provides more precise results and enables complex analyses to which a building may be subjected. This approach allows for the anticipation of potential problems and ensures that both static actions (self-weight, permanent loads) and dynamic actions (winds, vibrations, impacts) are properly considered and verified. Then, the main objective of this work was to develop a methodology for the dynamic analysis of 2D reinforced concrete truss structures, considering the physical nonlinearity of the material (cracking). To achieve this objective, Finite Element Method (FEM) was employed for the spatial discretization of the structure domain, and Newmark method was used for the temporal discretization. Physical nonlinearity was considered through the material's constitutive equation, according to ABNT NBR 6118:2023, with the update of the stiffness matrix over time. Newton-Raphson method was used to obtain the solution of the nonlinear problem in the process of verifying the reinforced concrete section. The results obtained were compared with those of well-established market software in their student, free, or temporary versions to validate the robustness and efficiency of the proposed algorithm.

Application of Artificial Intelligence in River Pollution Monitoring to Support Environmental Management

Jullia Fernandes Felizardo — 2025-12-01

Water pollution caused by visible solid waste has worsened in recent decades, compromising aquatic ecosystems, public health, and water security. Traditional monitoring methods present limitations in terms of spatial coverage and scalability. This work proposes a low-complexity model based on Convolutional Neural Networks (CNNs) for the automatic detection of visible human-eye-level waste in rivers from user-provided images. A labeled dataset was developed through the combination of different databases and manual annotation. Transfer Learning was applied using the MobileNetV2 architecture, chosen for its computational efficiency and low energy consumption, making it suitable for deployment on resource-constrained devices. Multiple independent training sessions were performed to ensure statistical robustness. The model achieved high and stable performance over 20 runs, with an average accuracy of 89.7\%, an F1-score of 90.9\%, and an area under the ROC curve (AUC) of 0.996, highlighting a strong discriminative capability. The absence of false negatives in the best-performing model reinforces its applicability in conservative environmental surveillance systems. The results indicate the feasibility, scalability, and potential of the proposed approach for automated river monitoring. Future integration with geospatial platforms and mobile devices is suggested to expand its adoption by environmental agencies and public managers.

Online Model Order Reduction for Flow Simulations problems using Dynamic Mode Decomposition

Felipe Toledo — 2026-01-01

In this study, a reduced-order model (ROM) was directly coupled with a physical solver to predict system evolution and progressively replace the full-order model (FOM) solutions. Dynamic Mode Decomposition (DMD), implemented via the PyDMD library, was employed to perform the ROM predictions. The flow around a 2D cylinder with Reynolds number Re = 100 was chosen as the case study. The FOM solver uses the FEniCSx platform. The quality of the online DMD model approximation was evaluated using the residual norm generated exclusively by the DMD prediction, ensuring that velocity and pressure predictions remained within an acceptable tolerance. Additionally, the Frobenius norm was employed to monitor the accuracy of the velocity and pressure matrix predictions, assessing DMD’s non-intrusive capability to capture and reproduce the system’s dynamics. Experiments were conducted varying the number of dynamic modes required to achieve the desired accuracy. Using DMD as an alterative to the nonlinear solver reduced the computational time for the numerical simulations up to 20% with little to no loss in the solution quality.

Modeling and Optimization of Power Transmission Capacity of Transmission Lines Using an Enhanced Evolutionary Algorithm

Ana Liz Rodrigues Ferreira — 2025-12-01

The power transmission system in Brazil faces challenges due to aging infrastructure, as it has not been updated over the decades. With over 170,000 km of transmission lines, many operate with outdated configurations, limiting their capacity. In this context, a new methodology based on Differential Evolution (DE) algorithms is proposed to optimize the geometry of conductor bundles, aiming to increase the natural power of the line (SIL), while complying with safety limits and technical regulations. This study is based on the need to modernize transmission lines and make them more efficient. The adopted methodology is based on two main steps: electromagnetic modeling of the transmission line, along with the formulation of the optimization problem, and the application of three variants of the Differential Evolution algorithm. For modeling the transmission line problem, the calculation of electric charges is based on the voltage phasors applied to the conductors using Maxwell's potential coefficient matrix. The ground-level electric field is determined using Gauss's Law and the principle of superposition. The surface electric field is calculated using successive images. The SIL is calculated from the voltage and the ratio between the positive-sequence inductance and capacitance of the transmission line. In the optimization process, three variants of the Differential Evolution algorithm are applied: the Conventional DE, which uses random combinations of individuals; the DE with Per-vector dither, which applies a perturbation to each individual; and the Per-generation dither, which applies a perturbation to the entire generation of individuals.The methodology is applied to a 345 kV transmission line with two conductors per phase. There is a significant increase in the ground-level electric field values, although they remain within regulatory limits. The optimized geometries increase the bundle radius and reduce the phase spacing, resulting in an approximate 20% increase in SIL for all strategies. All surface electric field values remain below the calculated critical value.In summary, the proposed methodology is effective in optimizing the transmission capacity of transmission lines. The optimized geometries provided significant gains in SIL while complying with all considered constraints. The methodology can be expanded to include more realistic constraints and the use of hybrid optimization methods.

A Goal-Oriented Error Estimator for Stress Intensity Factors in G/XFEM

Gabriela Fonseca — 2025-12-01

This work presents a goal-oriented error estimator specifically developed to quantify the error in the Stress Intensity Factor (SIF) for problems in Linear Elastic Fracture Mechanics. The estimator is formulated within the framework of the Generalized/eXtended Finite Element Method (G/XFEM), a widely used numerical technique for solving problems involving discontinuities and singularities. The analysis focuses on two-dimensional cases, with the SIF computed using the Interaction Integral — an energy-based method expressed as a domain integral.To estimate the error, a dual problem is constructed using the same discretization space as the primal problem. The estimated SIF error is then derived from the energy norm errors of both problems. For the energy norm estimation, the ZZ-BD recovery-based estimator is employed. This technique builds a recovered stress field using special enrichment functions and solves block-diagonal systems of equations, ensuring both computational efficiency and high accuracy.Numerical results for a mixed-mode fracture problem confirm the effectiveness of the proposed strategy, resulting in effectivity indices close to unity. Furthermore, the element-wise distribution of the estimated error closely reflects the characteristics of the Interaction Integral. All simulations were performed using INSANE, an open-source platform developed at the Department of Structural Engineering of the Federal University of Minas Gerais.

Multi-Objective Optimization of Building Demolition

Mariana Borges Oliveira — 2025-12-01

It is estimated that a significant portion of the waste generated by the construction industry originates from demolitions, a trend expected to intensify as buildings constructed in the latter half of the 20th century reach the end of their life cycle and the need for reconstruction in the same urban centres becomes increasingly necessary. In this context, the dismantling of building components, with the specific aim of enabling their reuse and recycling, has emerged as a promising strategy to mitigate the environmental impact of demolition. However, for this approach to become a viable alternative, it is essential to evaluate multiple - often conflicting - aspects, such as cost, time and sustainability. To address this problem, multi-objective optimization algorithms provide an effective framework, as they enable the identification and analysis of optimal trade-off solutions. This study aims to implement a multi-objective optimization algorithm to evaluate the best demolition strategies for a case study, based on compromise points that consider these elements simultaneously. Multiple demolition scenarios are explored, taking into account the methods currently practiced in Brazil. The findings of this study highlight the potential of optimization techniques to advance demolition practices by supporting decision-making.

CFD-Based Characterisation and Validation of Hydrogen-Sulphide Emissions in a Micro-Chamber/Thermal Extractor

Laíze Nalli de Freitas — 2025-12-01

New analytical technologies for measuring trace atmospheric pollutants are being introduced every year in the search for easier pollutant quantification. One example—already applied in some European countries—is the Micro-Chamber/Thermal Extractor™ – M-CTE250I aka “Micro-chamber”, a compact emission-analysis system whose main advantages are the small sample volume required, the low carrier-gas flow rate and rapid results. However, these emerging technologies still lack thorough validation and standardized operating protocols. The present study proposes the use of Computational Fluid Dynamics (CFD) simulations to characterise internal flow and mass-transfer mechanism, to quantify local and global mass-transfer coefficients of hydrogen sulphide (H₂S) inside the device. A three-dimensional model of a single stainless-steel slot (57 mm diameter, 14 mm height, 15 mL liquid) was built in ANSYS Fluent. The carrier gas (N₂, 50 mL min⁻¹) yields a Reynolds number of 22.8, therefore the flow is solved as laminar incompressible. Model outputs will be validated against independent decay experiments conducted at 21, 25 and 30 °C. The resulting CFD framework will support future calibration of micro-chambers, guide protocol development, and improve emission-factor estimates for quiescent wastewater sources.

Structural reliability of casing subjected to annular pressure build-up (APB) in oil and gas wells

Gilberto Lucas Leandro dos Santos — 2025-12-01

The work objective is to estimate the probability of rupture in oil and gas well casing subjected to loads induced by Annular Pressure Build-Up (APB) through First-Order Reliability Method (FORM). During the well lifecycle, temperature changes from different operations can lead to pressure variations in annular spaces. These fluctuations may compromise the structural integrity of casing and tubing strings. Due to manufacturing uncertainties, casing may exhibit variations in geometric and material properties. Along with uncertainties in geomechanical rock properties obtained from laboratory testing, these factors affect simulation accuracy. Estimating failure probability provides a more reliable safety metric than deterministic approaches. The methodology includes: i) identifying deterministic and random variables; ii) Application of FORM to estimate casing failure probabilities (burst); iii) sensitivity analysis to rank the influence of input variables; and iv) organizing and interpreting the results. This work contributes by quantifying the casing failure probability under APB conditions and identifying the most influential uncertain parameters for the studied scenario. These insights support better-informed decisions in casing design, aiming to ensure structural integrity under varying thermal and pressure conditions.

Robustness and redundancy finite element analysis of composite deck bridges

Paula Fidelis Viana — 2025-12-01

Structural robustness is a topic that still lacks consensus in both the literature and international standards such as AASHTO, ASCE, and Eurocodes. These standards do not provide clear guidelines for the application and validation of robustness and redundancy concepts in structures. Considering the need to prevent disproportionate or progressive collapse in bridges and viaducts, the impact of these concepts on structural safety must be studied and included in the design phase. This paper proposes a finite element model of the deck of a multi-girder composite bridge composed of three simply supported I-beam girders with full-web sections. The analysis, based on the conventional criteria of the AASHTO LRFD Bridge Design Specifications (2020) and assuming elastic-linear behavior, introduces a discontinuity in the girder mesh to simulate a fracture in the mid-span region. Based on this premise, the study evaluates the redistribution of forces, the performance of the structural elements under the new internal demands, and the maximum vertical displacement of the girders under critical mobile load conditions. The results are compared with theoretical definitions of robustness and redundancy, contributing to the enhancement of analytical methods in structural engineering.

MULTIVARIABLE ANALYSIS IN STRUCTURAL HEALTH MONITORING OF AN INDUSTRIAL CONVEYOR BELT

Cádmo Rodrigues — 2025-12-01

This paper presents a detailed multivariable structural health monitoring (SHM) analysis focused on an industrial conveyor belt system operating under dynamic environmental and mechanical conditions. The main objective is to explore how time- and frequency-domain features, derived from strain, acceleration, and temperature data, can provide insights into failure risks and dynamic behavior. Notably, the Fast Fourier Transform (FFT) is applied to strain gauge signals, a relatively underexplored method in this field, to identify high-frequency vibration modes correlated with operational loading. Additionally, the study quantifies the influence of thermal cycles on stress variations and integrates a probabilistic model to estimate failure risk based on alternating stress amplitude and excitation presence. The findings highlight the critical value of retaining high-frequency vibrational components, often dismissed as noise, which are shown to carry structurally relevant information. This integrative approach not only improves understanding of system dynamics but also supports predictive maintenance practices in harsh industrial environments. The study concludes with discussions on applicability, scalability, and limitations.

Geometric Modeling for Finite Elements: Advances and Contributions in Surface Intersections, Multi-Region Detection, and Topological Structures (2000–2015)

William Lira — 2025-12-01

This work presents an overview of the main contributions made between 2000 and 2015 in the field of geometric modeling applied to engineering problems, with emphasis on topological data structures, surface intersections, and automatic recognition of multi-regions. During this period, the studies conducted significantly contributed to pushing the frontier of knowledge in the area, offering conceptual and computational solutions relevant to applications based on the finite element method. The increasing demand for robust numerical simulations in engineering, especially those based on the finite element method, underscores the importance of accurate and efficient geometric representation of complex domains, which is fundamental for generating quality meshes and for the accuracy of simulations. Additionally, some of the works also addressed aspects related to mesh generation and adaptivity, expanding the scope of investigated applications and challenges. The contributions deal with topics such as the robust computation of intersections between parametric surfaces, the efficient handling of Boolean operations on non-manifold solids, and the implementation of topological data structures for the automatic detection and management of multi-regions in three-dimensional models. To present this overview, the main works published by the authors between 2000 and 2015 were selected, aiming to provide a coherent panorama of the proposed approaches, techniques, and solutions. The paper seeks to describe, in an accessible and structured manner, the main concepts, strategies, and results achieved, offering readers a comprehensive perspective of the developed contributions. The main contribution of this work is to provide a consolidated record of the solutions developed over the years, highlighting how they helped to advance the frontier of knowledge in geometric modeling for engineering. This paper establishes itself as a reference for researchers and professionals interested in the evolution of these techniques and their applications in advanced numerical simulations in engineering and related areas.

APPLICATION OF COMPUTATIONAL FLUID DYNAMICS MODELS FOR ASSESSING FIRE SAFETY IN A HISTORIC BUILDING

Aline Braga de Oliveira Menaget — 2025-12-01

Some buildings, especially those with complex architecture or historical value, face challenges in fully complying with the fire safety requirements established by traditional prescriptive codes, particularly with regard to passive protection measures such as compartmentation, the number and width of emergency exits, and the maximum egress travel distance. In such cases, implementing a performance-based fire engineering design proves to be a feasible alternative, as it allows for the development of customized solutions that meet fire safety objectives while considering the unique characteristics of each building. This article discusses the application of this approach to a heritage-listed mixed-use university building that has only one staircase serving all typical floors, with a egress travel distance exceeding the limit established by the current prescriptive regulations. The analysis was carried out using computational simulations with the Fire Dynamics Simulator (FDS) software, considering different ventilation scenarios and fire ignition locations. Key parameters such as visibility, temperature, and fractional effective dose were assessed along the evacuation route to estimate the available safe egress time. The results indicated that maintaining open windows in rooms, the presence of natural smoke extraction in the corridor, and the compartmentation of spaces significantly contribute to improving the escape conditions. It is concluded that the adoption of performance-based fire engineering design enabled the development of effective fire safety strategies without significant alterations to the building’s original architecture.

Fatigue Assessment of Steel-Concrete Composite Highway Bridges Utilising a Vehicle-Bridge Dynamic Interaction Methodology based on the Eurocode FLM 4

Marilene Lobato Cardoso — 2025-12-01

Nowadays, the increased weight and volume of traffic on highway bridges have accelerated deterioration processes, with fatigue induced by dynamic loads standing out as a critical challenge. In this scenario, the emergence of fatigue cracks, resulting from vehicle dynamic impacts, represents a major concern for engineers, necessitating reliable methods to assess the service life and support intervention decisions. This way, this research work proposes an analysis methodology to assess the fatigue performance of steel-concrete composite highway bridges, integrating traffic dynamic effects and the progressive pavement degradation in the dynamic structural response. The methodology develops a computational tool called VBI (Vehicle-Bridge Interaction) implemented in MATLAB and comprising an interface with the finite element program ANSYS, enabling the incorporation of complex finite element models for both the bridge and the vehicles. In this context, different scenarios based on the standard vehicle traffic prescribed by EN 1991-2: 2023 (FLM 4: Fatigue Load Model 4) are generated and adapted to statistical models of Brazilian vehicle classes. This approach allows for diverse truck configurations and weight distributions, reflecting real traffic patterns. The proposed analysis methodology was applied to a 40m span steel-concrete composite highway bridge, where a critical structural detail, mainly subjected to distortion-induced fatigue, is investigated through sub modelling techniques. The results highlight the influence of the dynamic effects and pavement degradation on the service life of the bridge structural components.

Predicting Remaining Useful Life of Well Tubulars

Luís Philipe Ribeiro Almeida — 2025-12-01

This work proposes a reliability-based framework to estimate the Remaining Useful Life (RUL) of well tubulars throughout the lifecycle of oil and gas wells. The RUL estimation is supported by a probabilistic analysis that accounts for uncertainties in geometrical and material properties, based on failure modes defined in the API/TR 5C3 standard. The framework is applicable to both design and monitoring phases, incorporating degradation mechanisms such as corrosion. The methodology integrates a mechanical model to calculate safety factors and failure probabilities associated with internal and external pressures and axial forces, as defined by API/TR 5C3. Structural reliability methods combined with Monte Carlo simulation are used to propagate uncertainties and to update failure probabilities over time. The framework enables the estimation of key integrity indicators, such as the failure rate and Mean Time to Failure (MTTF), essential for RUL assessment. A case study is presented, utilizing statistical data on mechanical properties, and degradation effects. The proposed methodology allows for integrity assessment under both serviceability and survival load conditions, incorporating advanced features such as the effect of corrosion in several loading scenarios. By capturing the variability of mechanical and geometrical parameters over time, the approach provides a robust basis for evaluating the mechanical condition and failure probability of casing and tubing strings. These results support more informed decision-making in RUL analysis and contribute to the improvement of well design and monitoring strategies

DEVELOPMENT OF A PYTHON PROGRAM FOR ANALYSIS OF GRILLING STRUCTURES BY FINITE ELEMENT METHOD

Diogo Marinho Cardoso — 2025-12-01

Abstract. Grid structures are commonly used in civil engineering systems such as floor slabs, bridge decks, and industrial platforms, where elements are subjected to bending in two orthogonal directions and torsion. The proper analysis of these systems requires numerical techniques capable of accurately capturing their structural behavior. This work presents the development of a computational tool in Python for the analysis of planar grid structures based on the Finite Element Method (FEM). The formulation considers structural elements with three degrees of freedom per node: two rotational and one translational, enabling the evaluation of torsional and flexural effects. The implementation involves the calculation of local stiffness matrices for each element, coordinate transformation to the global reference system, and the assembly of the global stiffness matrix. Once the structure’s geometry, boundary conditions, and loads are defined, the system of equations is solved to obtain nodal displacements. From these, internal forces such as bending moments, torsional moments, shear forces, and support reactions are derived. To verify the accuracy and reliability of the program, benchmark problems are used, and the results are compared with analytical solutions and well-established software. The main objective of this project is to offer a lightweight, open-source, and educationally oriented tool that can support engineers, students, and researchers in understanding and applying grid structure analysis, especially in academic environments where commercial solutions are not readily available.

ALTERNATIVE NUMERICAL MODEL FOR VEHICLE–BRIDGE INTERACTION BASED ON THE POSITIONAL FINITE ELEMENT METHOD

GEOVANA VIVIANI DE ARAUJO — 2025-12-01

Vehicle–Bridge Interaction (VBI) refers to the dynamic responses generated during a vehicle's passage over a bridge, where both systems behave in a coupled manner, resulting in a unified dynamic response. As vehicles traverse the bridge, the mutual influence between the vehicle and the structure becomes critical for understanding the overall behavior. VBI analysis investigates how these interactions contribute to structural anomalies, fatigue damage, and long-term deterioration of the bridge system, making the comprehension of these effects essential for maintaining structural integrity.The literature commonly represents VBI using numerical models or approaches that combine analytical solutions for the vehicle dynamics with a numerical representation of the bridge. Bridge structures are typically discretized using finite elements of varying complexity, such as 2D or 3D beam elements, plate or shell elements, depending on the required modeling detail. Due to the higher computational cost of the surface models, beam representations are more prevalent, with Euler-Bernoulli kinematics the most widely used, although the Timoshenko beam model is also employed, as adopted in this study for greater generality. Vehicle models in existing VBI studies are simplified as mass-spring-damper systems that represent their mechanical components: truck, trailer and various axles. In this work, the vehicle is modeled in an alternative manner using strategically positioned plane frame finite elements that are mechanically equivalent to conventional mass-spring-damper systems.To address the nonlinear geometric effects, due to the large displacement of the vehicle regarding the bridge, the positional formulation of the finite element method is employed. Numerical solution of the nonlinear dynamical system is performed by the Newton-Raphson method with the generalized-alpha time integration scheme. Lagrange multipliers are used to account for the coupling between vehicle and bridge frame elements. The influence of pavement roughness is also considered along the vehicle path.Comparison of the simulated VBI dynamic responses with existing literature is conducted to evaluate the suitability of the proposed model in representing the problem.

Direct Strength Method Design of CFS Lipped Channel Beams under Fire-Induced Distortional and Global Buckling Modes

Alexandre Landesmann — 2025-12-01

Normal0falsefalsefalsePT-BRX-NONEX-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Tabela normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-parent:""; mso-padding-alt:0cm 5.4pt 0cm 5.4pt; mso-para-margin:0cm; mso-pagination:widow-orphan; font-size:10.0pt; font-family:"Calibri",sans-serif; mso-bidi-font-family:"Times New Roman"; mso-ansi-language:PT-BR; mso-fareast-language:PT-BR;}This work reports the findings of a comprehensive numerical investigation on the post-buckling behaviour, failure and Direct Strength Method (DSM) design of cold-formed steel (CFS) single-span simply supported lipped channel beams buckling in pure distortional and global modes at elevated temperatures (up to 800 ºC) due to fire conditions. It extends the scope of a previous investigation carried out by Landesmann and Camotim [1], by analysing a substantially larger lipped channel beam set, exhibiting various cross-section dimensions and yield stresses, selected to cover wider distortional slenderness ranges. The beams analysed (i) display two end support conditions (SCA and SCB), (ii) the Eurocode 3 (part 1.2) model to describe the temperature-dependence of the CFS material properties is adopted and (iii) the results are obtained by means of Abaqus shell finite element GMNIA. After presenting and discussing the main features of the beam distortional and global post-buckling behaviour, extensive beam failure moment sets are gathered and used to develop and validate DSM-based design approaches. The methodology followed consists of modifying the most performant available DSM-based design curves, which naturally involves the temperature-dependent reduction factors of the CFS model. A merit assessment procedure shows that the modified DSM-based strength curves predict the lipped channel beam distortional and global failure moments with remarkable accuracy and reliability, thus constituting an excellent starting point to search for a DSM-based design approach capable of handling arbitrary CFS beams failing in pure distortional and global modes at elevated temperatures.

e-Stress3D: A Web-Based Educational Tool for Interactive 3D Stress Tensor Visualization in Solid Mechanics

Matheus Amancio Miranda — 2025-12-01

We present e-Stress3D, a web-based educational tool designed to enhance the learning of Solid Mechanics in undergraduate and graduate engineering curricula. The application enables interactive visualization of the 3D stress tensor, integrating mathematical formulations with graphical representations. Key features include stress decomposition into isotropic and deviatoric components with their invariants, a dedicated module for Cauchy traction vector analysis on arbitrary planes, visualization of principal stresses, and construction of the 3D Mohr’s circle. The platform supports real-time manipulation of input parameters and provides immediate visual feedback. This interactivity promotes a deeper and more intuitive understanding of fundamental concepts. Overall, the tool contributes to the growing integration of digital tools in engineering education, responding to the increasing demand for more engaging and effective pedagogical approaches.

Mechanical Analysis of Closed-Cell Foam Microbeams Using a High-Order Beam Model Based on Modified Strain Gradient Theory

Fabio Carlos da Rocha — 2025-12-01

Closed-cell foams are lightweight cellular materials widely used in applications requiring high mechanical strength, low density, and excellent energy absorption capacity. These materials are used in impact absorption systems in the automotive industry, thermal and acoustic insulation in civil construction, structural components in the aerospace sector, biomedical prosthetics and implants, and functional elements in sensors, microactuators, and precision engineering devices. At the micro- and nanoscale, they are particularly relevant in the design of microbeams for emerging technologies. Due to scale effects that significantly influence the mechanical behavior of such structures, classical models become inadequate, necessitating the use of higher-order continuum theories for a more accurate representation. This study proposes a high-order laminated beam model based on the Modified Strain Gradient Theory, incorporating thickness-direction deformation (quasi-3D effect) and three distinct porosity distributions. The Halpin-Tsai model is employed to estimate the effective material properties. The governing differential equations and boundary conditions are derived through a variational formulation, and the analytical solution is obtained using the Navier method. The results demonstrated excellent agreement with data from the literature, confirming the accuracy and robustness of the proposed model.

Comparative analysis of boundary treatments and regularization orders in GPU-based lattice Boltzmann simulations

Bruno Yan dos Anjos — 2025-12-01

We present a two-dimensional GPU-accelerated implementation of the lattice Boltzmann method (LBM) using CUDA, tailored for computational fluid dynamics applications. The numerical framework employs the D2Q9 velocity stencil combined with a regularized Bhatnagar-Gross-Krook collision operator. Our primary objective is the systematic evaluation and optimization of distinct LBM formulations through detailed comparisons between second-order and high-order regularizations, extending the analysis to include implementations up to fourth-order moments. Such high-order regularizations are instrumental in enhancing numerical stability and accuracy. Additionally, the impact of different boundary condition approaches is rigorously assessed through comparative studies between regularized boundary conditions and incompressible regularized boundary conditions. The computational performance and numerical accuracy of each approach are investigated using the classical lid-driven cavity flow configuration as a standard benchmark problem. Exploiting the computational advantages inherent in two-dimensional simulations, we utilize high-resolution grids consisting of 1024, 2048, and 4096 lattice points. These simulations are conducted across a comprehensive range of Reynolds numbers (3,200, 10,000, 50,000, and 100,000), thus demonstrating the robustness of the solver and capability in accurately resolving complex flow features typical of highly inertial regimes.

Comparison between 3D Finite element and analytical models in railway track analysis

Guilherme Nunes Bassegio — 2025-12-01

The satisfactory performance of a railway track throughout its service life depends on an adequate design. In this process, analytical models are often employed which, despite their practicality, are based on assumptions that may not accurately represent real site conditions. A more enhanced representation of these conditions can be provided by higher-hierarchy models, such as ones employing the finite element method. A reliable finite element analysis requires a suitable constitutive model and properly defined parameters, which can be obtained from on-site data, laboratory tests, or from the literature referring to similar conditions. In this paper, the analytical model of the Winkler foundation is compared to a finite element analysis for the problem of a statically loaded railway track section, characterized by on-site data, to assess the main discrepancies that may arise between the two methods. The comparison between both models is made by evaluating the track modulus, which is an important parameter for analytical models. In the high hierarchy model, stress fields are visualized along layers of under-rail materials, helping in understanding the physical behavior of the system.

Impacts of Academic Mobility on Scientific Production: Evidence from the Lattes Platform

HIGOR MASCARENHAS — 2025-12-01

Academic mobility has been a relevant phenomenon in Brazil, influencing the trajectory and scientific productivity of researchers. This study analyzes the relationship between academic mobility and scientific production of Brazilian PhDs, using data from the Lattes Platform processed by the LattesDataXplorer framework. The methodology involved the extraction and analysis of 381,462 CVs of PhDs, considering three groups: those who did not migrate, those who migrated nationally, and those who had international mobility. Publications in conference proceedings and journals classified by Qualis CAPES (2017-2020) were examined. The results indicate that academic migration positively impacts scientific productivity. The median number of publications in conference proceedings increased after migration, with a higher volume among those who migrated within Brazil. In journals, most publications are concentrated in higher strata of Qualis (A1 and A2), being more expressive among PhDs with international experience. It is concluded that academic mobility favors collaboration and research impact, but it is not the only determining factor of productivity. Mobility incentives, especially at the national level, can strengthen scientific networks and contribute to research excellence in Brazil.

Lumped Damage Mechanics applied to steel fiber-reinforced concrete plates

David Amorim — 2025-12-01

A nonlinear description of the structural behavior of structures is a crucial subject that must be addressed. In such problems, collapse failure may occur due to strain localization phenomena. This issue can be evaluated using nonlinear theories, such as Lumped Damage Mechanics (LDM). LDM models have demonstrated high accuracy in solving various nonlinear problems, including reinforced concrete and fiber-reinforced concrete frames, two-dimensional problems (under in-plane loads), and reinforced concrete slabs. Since steel fiber-reinforced concrete (SFRC) is a widely used composite material applied in various structures, it is paramount to develop nonlinear models to assess its structural behavior. Therefore, this paper presents a study of an LDM model applied to SFRC plates under bending, where the finite element known as the constant moment triangle is used. All inelastic effects are lumped at the element’s edges by hinge lines. For the analyzed examples, the numerical results agree with experimental observations.

VALIDATION OF NUMERICAL SIMULATION OF FORCED CONVECTION IN A VENTILATED CAVITY USING OPENFOAM

Hugo Kalev Noel — 2025-12-01

This study presents a numerical investigation of laminar and turbulent forced convection in a square ventilated cavity. The main objective of the present work was to replicate the velocity and temperature fields from the reference studies using the OpenFOAM® software, that is free, open-source and without cost to analyze the fluid flow and heat transfer mechanisms under varying Reynolds and Prandtl numbers. The computational domain consists of a square cavity with two openings: an inlet located at the upper left side of vertical wall and an outlet at the lower right vertical wall. The flow is considered steady, incompressible, and Newtonian, with constant thermophysical properties. Boundary conditions include a uniform inlet velocity and a temperature of 299 K, while all cavity walls are assumed adiabatic. The initial temperature in the cavity is set to 300 K. Simulations were carried out using the ‘buoyantSimpleFoam’ solver. Numerical modifications were made to the discretization schemes and relaxation factors, differing from the original study, with the aim of assessing their impact on the stability and accuracy of the solution. The analysis includes velocity vector fields, temperature distributions, internal vortex structures, and the evaluation of local and average Nusselt numbers along the cavity walls. The results obtained are strong agreement with those reported by references, both qualitatively and quantitatively, confirming the reliability of the implemented numerical approach. The study also reinforces the suitability of OpenFOAM® as a robust and flexible platform for simulating fluid flow with heat transfer in engineering systems, even under different numerical settings.

Reliability Analysis of Cold-Formed Steel Columns Subjected to Load Combinations Prescribed by Brazilian and North American Structural Codes

Giovanne de Lanna Santana — 2025-12-01

The structural safety of columns made of cold-formed steel sections, when the wind load is considered in the analysis, requires further studies for a more in-depth understanding, as most reliability studies focus on combinations of actions involving only self-weight and overload. This paper proposes the evaluation of the reliability of columns designed according to the Direct Strength Method (DSM). Using a comprehensive database, a statistical analysis of the variable "model error" was performed, defined as the ratio between the experimentally measured strength and the strength calculated via DSM. Based on this database, a statistical analysis of the test results was carried out to determine the optimal probabilistic distribution that best fits the model error data and its relevant distribution parameters. Variables related to material strength and geometric properties were incorporated into the reliability evaluation. The structural reliability levels were assessed using the FORM (First Order Reliability Method) for bars designed according to the design requirements of the standards NBR 14762 (ABNT, 2010) and AISI S100 (AISI, 2024). The results indicate that both standards provide adequate levels of reliability, with emphasis on the uniformity of the ASCE 7-22 load combinations. The Brazilian standard showed greater sensitivity to variations in loading, especially when wind is the predominant action. Most of the evaluated combinations achieved the target reliability index of β = 2.5, although in some cases, values below this threshold were observed for the Brazilian standard.

Advancing Graph Neural Networks for CO2 Plume Migration in Complex Geological Formations

Adriano Cortes — 2026-01-01

Carbon Capture and Storage (CCS) is essential for mitigating climate change by permanently storing industrial CO2 emissions in subsurface formations [1]. As net-zero targets gain urgency, accurately predicting injected CO2 behavior over long timeframes becomes critical for ensuring storage integrity and regulatory compliance. Simulating CO2 plume migration presents significant computational challenges. Conventional numerical simulators, while accurate, require prohibitive computational resources when modeling the centuries-long timescales relevant to CO2 storage, especially for uncertainty quantification or optimization studies requiring multiple simulations. Deep learning surrogate models offer orders-of-magnitude speedup while maintaining reasonable accuracy [2].Current literature features two main surrogate modeling approaches: Fourier Neural Operators (FNO) [2] and Graph Neural Networks (GNN) [3]. FNO models perform well on structured domains but are limited to Cartesian grids, inadequately representing complex geological features. In contrast, GNN approaches like MeshGraphNet (MGN) [4] can naturally handle unstructured meshes, better representing heterogeneous subsurface environments and geological boundaries critical to CO2 migration.We present our MGN-LSTM implementation for temporal prediction of CO2 plume migration in faulted reservoirs [3], incorporating algorithmic improvements for enhanced prediction stability and efficiency. Our key contribution is integrating physics-aware loss terms from FNO literature [5] into the GNN framework. By incorporating mass conservation, phase behavior, and Darcy flow principles directly into the loss function, we achieve physically consistent predictions even with limited training data. These physics-aware constraints improve model generalization to unseen geological configurations.References[1] IPCC. Special Report on Carbon Dioxide Capture and Storage. CUP, 2005.[2] Wen, G. et al. Towards a predictor for CO2 plume migration using deep neural networks. Int. J. Greenhouse Gas Control, 2021.[3] Ju, X et al Learning CO2 plume migration in faulted reservoirs with Graph Neural Networks. Computers & Geosciences, 193, 2024.[4] Pfaff, T. et al. Learning mesh-based simulation with graph networks. ICML, 2021.[5] Badawi, D., Gildin, E. Neural operator-based proxy for reservoir simulations considering varying well settings, locations, and permeability fields. J. Petroleum Science and Engineering, 2023.

SUPPORT SYSTEM FOR THE DIAGNOSIS OF THE RISK OF ANXIETY DISORDER IN CHILDREN

RENATA COSTA ROCHA — 2025-10-01

Anxiety disorders in children represent an important public health challenge, considering the subjective nature of the symptoms, the individual variability of clinical manifestations, and the lack of standardized diagnostic support tools. This gap compromises early identification and appropriate preventive interventions. This study investigated computational approaches for the multilevel classification of the risk of anxiety disorders in children, using behavioral and physiological data. The methodology involves the application and comparison of machine learning models, including Random Forest, Support Vector Machine, and Multilayer Perceptron Neural Network, in binary (presence or absence) and multilevel (mild, moderate, severe risk) paradigms. The research used a dataset of 193 children, publicly available on Harvard Dataverse by Carpenter [1], licensed under CC0 1.0 Universal. The evaluation of the models used standardized metrics aligned with the diagnostic criteria of the DSM-5, Diagnostic and Statistical Manual of Mental Disorders - 5th edition, and the ICD-11, International Classification of Diseases - 11th edition. After refinements in the methodology, the results showed significant improvement: Random Forest – accuracy 89.4%, sensitivity 80%, specificity 88.7%; Support Vector Machine – accuracy 88.5%, sensitivity 81.3%, specificity 91.9%; Multilayer Perceptron – accuracy 87.8%, sensitivity 77.7%, specificity 91.1%. Accuracy above 87% indicates excellent overall performance of the models, correctly classifying most cases. Given the topic, high sensitivity is crucial to avoid the omission of relevant cases. Sensitivity values (77% to 81%) demonstrate effective identification of positives, which is clinically important. High specificity (above 88%) shows accurate recognition of negatives, reducing false positives. These results indicate a strong predictive performance, especially for the SVM, which showed balance and robustness, being the most suitable to minimize both false negative and false positives. This highlights the feasibility of machine learning in early detection of anxiety risk in children. Continuous improvements aim to improve accuracy and clinical applicability. This study contributes to the advancement of diagnostic strategies in child mental health by offering a computational approach to support personalized, evidence-based clinical interventions.