Yu Wang , Yue Li , ShunPeng Zhu , Ran Gu , Xinhao Shu , Zhengwei Fan , Shufeng Zhang
{"title":"SA-PINN: A self-attention enhanced physics-informed neural network for multiaxial fatigue life prediction with small samples","authors":"Yu Wang , Yue Li , ShunPeng Zhu , Ran Gu , Xinhao Shu , Zhengwei Fan , Shufeng Zhang","doi":"10.1016/j.advengsoft.2026.104124","DOIUrl":"10.1016/j.advengsoft.2026.104124","url":null,"abstract":"<div><div>Multiaxial fatigue life prediction holds significant practical importance in the reliability analysis of equipment engineering structures. To address the high dependence on experimental data and low accuracy of long-term fatigue prediction in existing fatigue life prediction methods, this article proposes a self-attention mechanism enhanced physics-informed neural network (SA-PINN). This method introduces physical information into the loss function of artificial neural network (ANN) and jointly optimizes with the self-attention mechanism during training to enhance the accuracy of fatigue life prediction under small-sample conditions. Specifically, SA-PINN dynamically adjusts the weights of input features using the self-attention mechanism in the early stages of the network, performing global modeling of time-series stress-strain features. Under the constraint of the physical information loss functions, it further establishes accurate long-range feature dependencies, enhancing prediction accuracy and physical interpretability of the prediction results. Experiments were performed on six types of small samples material dataset. The results show that the Root Mean Square Error (<em>RMSE</em>) of the prediction results of the SA-PINN model is 0.153, and the coefficient of determination (<em>R</em><sup>2</sup>) is 0.942. Compared with ANN, RMSE is reduced by 0.09 and <em>R</em><sup>2</sup> is increased by 0.15. It significantly improves the accuracy and reliability of fatigue life prediction in small samples and multiple working conditions, and provides a new paradigm for fatigue life prediction.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"216 ","pages":"Article 104124"},"PeriodicalIF":5.7,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146193217","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A resolution-independent DeepONet-ABC framework for inverse characterization of composite materials","authors":"Enying Li , Hu Wang","doi":"10.1016/j.advengsoft.2026.104116","DOIUrl":"10.1016/j.advengsoft.2026.104116","url":null,"abstract":"<div><div>This study proposes a novel inverse method based on Approximate Bayesian Computation (ABC) integrated with a Resolution Independent Strategy (RIS) to determine the material parameters of Carbon Fiber Reinforced Polymer (CFRP). The proposed framework addresses a key challenge in materials science and engineering, particularly for CFRP composites. Unlike existing approaches, the method explicitly incorporates sample locations during modeling, which makes the sampling distribution for inversion less sensitive and relaxes the strict requirement for precise sensor placement. By integrating ABC with RIS, the framework successfully infers key parameters that govern the behavior of CFRP. The accuracy and effectiveness of the proposed method have been rigorously validated through numerical experiments, comparing posterior means and uncertainties of the inferred parameters with experimental data. An excellent agreement among experimental observations, posterior model predictions, and finite element simulations demonstrates the robustness of the methodology. These results underscore the potential of the DeepONet-augmented ABC inverse approach for the accurate characterization of CFRP material properties, offering a powerful tool to enhance the precision of inverse analysis in this field.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"215 ","pages":"Article 104116"},"PeriodicalIF":5.7,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191140","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Longlong Song , Pingchu Fang , Tong Gao , Jihong Zhu , Xiaohua Nie , Weihong Zhang
{"title":"Multistage design of high-aspect-ratio wings: Geometrically nonlinear topology optimization with aero-structural coupling","authors":"Longlong Song , Pingchu Fang , Tong Gao , Jihong Zhu , Xiaohua Nie , Weihong Zhang","doi":"10.1016/j.advengsoft.2026.104119","DOIUrl":"10.1016/j.advengsoft.2026.104119","url":null,"abstract":"<div><div>High-aspect-ratio wings for solar-powered UAVs (Unmanned Aerial Vehicles) require ultra-lightweight structural designs capable of withstanding large deformations induced by aerodynamic loads. Traditional topology optimization methods often overlook the interplay between geometric nonlinearity and aero-structural coupling, limiting their effectiveness for flexible wing systems. This study proposes a multi-material topology optimization framework that integrates geometric nonlinearity with weak aero-structural coupling. The three-field approach, incorporating directional length scale control and deformation control, is employed for the geometrically nonlinear multi-material optimization. A nodal displacement-based inverse modeling method enables efficient data transfer between aerodynamic and structural meshes, while a deformation-driven load update strategy reduces computational costs. A straight wing with four case studies demonstrates that incorporating aero-structural coupling shifts high-stiffness material toward the wing root, optimizing load transfer and reducing deformation. The reconstructed truss-based design achieves an 11.29 % mass reduction via size optimization while maintaining bending and torsional constraints. This framework bridges structural optimization with aero-structural coupling, providing a systematic approach for designing lightweight, main spar-ribs based, and deformation-resistant wings.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"215 ","pages":"Article 104119"},"PeriodicalIF":5.7,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191231","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Explainable AI-driven prediction and interpretation of aerodynamic interference effect in complex high-rise building clusters","authors":"H.C. Deng, Z.Y. Zhang, Z.R. Shu, X.H. He","doi":"10.1016/j.advengsoft.2026.104121","DOIUrl":"10.1016/j.advengsoft.2026.104121","url":null,"abstract":"<div><div>The rapid urban densification intensifies aerodynamic interference among high-rise buildings, which complicates the wind-resistant design and structural safety. However, the underlying flow mechanisms in complex building clusters remain under-explored, mainly due to their nonlinear and configuration-dependent behavior. This study integrates wind tunnel experiments with an explainable artificial intelligence (XAI) framework to provide high-fidelity prediction and physical interpretation of aerodynamic interference within triangular high-rise clusters. Systematic experiments varying streamwise and transverse spacing and rotational angles produced detailed surface pressure datasets. Four AI models, i.e., Support Vector Regression, Decision Tree, Random Forest, and XGBoost, were trained to predict mean and fluctuating pressure coefficients, with XGBoost yielding the best overall performance. Model interpretability, achieved through SHapley Additive exPlanations (SHAP), revealed that transverse spacing governs regime transitions between shielding and resonance amplification, while streamwise spacing primarily influences fluctuating pressures through aerodynamic damping. SHAP analysis also identified pronounced three-dimensional pressure non-uniformity and a rotation-induced converging nozzle effect that increases mean pressures while moderating fluctuations. The proposed XAI-assisted framework establishes a data-driven approach for uncovering aerodynamic interference mechanisms, thus providing insights for resilient and performance-informed wind design of high-rise building clusters.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"215 ","pages":"Article 104121"},"PeriodicalIF":5.7,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Kai Cui , Bohan Wu , Wenbin Xiao , Ran Yuan , Yi He
{"title":"A unified constitutive model for natural soils and a novel stress integration algorithm with the trust region method","authors":"Kai Cui , Bohan Wu , Wenbin Xiao , Ran Yuan , Yi He","doi":"10.1016/j.advengsoft.2026.104120","DOIUrl":"10.1016/j.advengsoft.2026.104120","url":null,"abstract":"<div><div>Constitutive models and stress integration algorithms are an essential part of computational modelling in geotechnics; selecting an appropriate constitutive model and a reliable stress integration algorithm are usually the key to successfully predicting the behaviour of geostructures. This paper develops a unified constitutive model, i.e., CASM-PSS, for natural structure clay and sand within the critical state framework, considering the influence of noncoaxial behaviour. By introducing the inclined yield surface, incorporating the influence of direction of stress increment on plastic flow rule, and introducing the super-subloading surface in the unified clay and sand model (CASM), the proposed model able to unified describe the mechanical behaviour of natural clay and sand. Novelty of the proposed model lies in its ability to consider the effects of anisotropy, destructuration, and direction of stress increment of natural soils. The CASM-PSS is then implemented into a finite difference software through the user-defined model platform (UDM), by a new semi-implicit integration algorithm integrated with the trust region method. Performance of the CASM-PSS is investigated and validated by a series of experimental and DEM evidence on soils under different loading paths, including drained and undrained conditions of triaxial compression and simple shear tests. Then, the newly proposed model is numerically applied to simulate centrifugal model testing and practical engineering problems, i.e., China Tsinghuayuan Tunnel, to identify effects of soil structure, anisotropy and noncoaxiality in boundary value problems.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"215 ","pages":"Article 104120"},"PeriodicalIF":5.7,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191228","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"An open-source computational wind load evaluation workflow for tall building design","authors":"Christopher J. Howlett, Girma T. Bitsuamlak","doi":"10.1016/j.advengsoft.2026.104117","DOIUrl":"10.1016/j.advengsoft.2026.104117","url":null,"abstract":"<div><div>Wind load evaluation (WLE) of tall buildings is a complex process as multiple factors influence the final results. The reliability of the evaluation is dependent on the analysis accuracy of each factor. Traditionally, wind tunnels have been the primary method of performing WLE studies but recent growth in computational resources has led to a wider adoption of computational-based wind engineering. However, for these computational studies to be used in research and practice, it is imperative that they maintain a high level of accuracy. The current study presents the development of an open-source computational-based WLE workflow for the design of tall buildings. The proposed workflow adapts established methods from wind tunnel studies and integrates them into a computational fluid dynamics (CFD)-based pipeline. Each step within the WLE analysis is discusses, along with the relevant background information. Separating each step of the WLE pipeline into a unique and modular component provides engineers and researchers with the foundational knowledge required to perform computational WLE studies. Finally, the effectiveness of the proposed workflow is validated using a WLE study performed in the wind tunnel.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"215 ","pages":"Article 104117"},"PeriodicalIF":5.7,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191141","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Rongdong Wang , Yang Li , Changming Li , Zhonghua Wang , Detao Wan , Gang Yang , Dean Hu
{"title":"Multi-objective optimization design of a pressure swirl nozzle for sodium spray in SFRs via adaptive surrogate model","authors":"Rongdong Wang , Yang Li , Changming Li , Zhonghua Wang , Detao Wan , Gang Yang , Dean Hu","doi":"10.1016/j.advengsoft.2026.104113","DOIUrl":"10.1016/j.advengsoft.2026.104113","url":null,"abstract":"<div><div>To address the challenge of safely and efficiently disposing of radioactive liquid sodium from sodium-cooled fast reactors (SFR), this study proposes an adaptive data-driven multi-objective optimization framework for pressure swirl nozzle designs. The spray performance is characterized by spray angle (<em>ω</em>) and outlet velocity (<em>v</em>), which are typically conflicting. The wider <em>ω</em> often leads to lower <em>v</em>. Balancing these objectives is crucial for effective dispersion. A deep neural network (DNN) surrogate model is developed to capture the nonlinear mapping between geometric variables and spray performance metrics. The model is trained on computational fluid dynamics (CFD) data and iteratively refined using an adaptive sampling strategy based on uncertainty. The well-trained surrogate model achieves a determination coefficient (<em>R<sup>2</sup></em>) above 93%, and a maximum normalized squared error (<em>NSE</em>) below 0.03 on the testing set. The NSGA-II algorithm is employed to perform multi-objective optimization, generating a Pareto front that reveals the trade-off relationship between spray angle and outlet velocity. Three representative design schemes including maximum angle, maximum velocity, and Pareto inflection scheme were selected and validated via CFD. Compared to the baseline design, the nozzle scheme improved spray angle by up to 96.25% and outlet velocity by up to 37.18%. Furthermore, Sobol sensitivity analysis indicates that the outlet diameter contributes over 40% of the total first-order sensitivity for both objectives, demonstrating its dominant role in nozzle performance. This work provides an efficient and accurate design methodology for optimizing high-temperature sodium spray systems and can be extended to other complex multiphase flow applications.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"215 ","pages":"Article 104113"},"PeriodicalIF":5.7,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191139","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Shang Lv , Wei Sun , Xuedong Sun , Yu Zhang , Hui Zhang , Dongxu Du , Hui Li
{"title":"Free vibration characteristics and veering analysis of auxetic metamaterial plates","authors":"Shang Lv , Wei Sun , Xuedong Sun , Yu Zhang , Hui Zhang , Dongxu Du , Hui Li","doi":"10.1016/j.advengsoft.2026.104118","DOIUrl":"10.1016/j.advengsoft.2026.104118","url":null,"abstract":"<div><div>Honeycomb-reinforced composite sandwich plates (CSPs) have been widely used in the aerospace field due to their excellent mechanical properties, especially the Auxetic Hexagonal Honeycomb (AHH) sandwich plates with negative Poisson’s ratio effect. However, their vibration characteristics and modal coupling mechanisms have not been studied; therefore, this paper establishes a systematic research method that integrates modelling, fabrication, verification, and analysis. First, this paper develops a semi-analytical dynamic model for the AHH-CSP structure, based on the first-order shear deformation theory (FSDT) and the layer-wise zigzag theory. Second, AHH-CSP specimens are designed and fabricated, and their natural frequencies and modal shapes are determined via impact tests. Then, this paper verifies the accuracy and applicability of the semi-analytical model by combining it with finite element simulations, existing literature, and experimental results, and systematically analyzes honeycomb parameters on structural vibration characteristics. The results show that coupled vibration occurs between different modes of the AHH-CSP structure, thereby inducing frequency veering and mode shape exchange and revealing the inherent patterns of modal coupling. The findings provide both theoretical foundations and experimental support for the engineering application of novel honeycomb sandwich plate structures with negative Poisson’s ratio.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"215 ","pages":"Article 104118"},"PeriodicalIF":5.7,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191229","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Free vibration analysis of in-plane bidirectional functionally graded plates with piezoelectric layers using RPT-IGA method","authors":"Zhenyang Gao , Yaqiang Xue , Yuan Gao , Chunyu Zhang","doi":"10.1016/j.advengsoft.2026.104122","DOIUrl":"10.1016/j.advengsoft.2026.104122","url":null,"abstract":"<div><div>The free vibration of in-plane bidirectional functionally graded plates with piezoelectric layers (FGPPL) is investigated in this study, where isogeometric analysis (IGA) in conjunction with refined plate theory (RPT) is utilized. This paper proposes an original transverse shear stress function through thickness distribution, formulated by sine and power functions. The RPT offers a computational advantage over the first-order shear deformation theory (FSDT) by describing the field displacement vector field with only four variables and avoiding shear correction factors, thus markedly diminishing the quantity of degrees of freedom. The non-uniform rational B-spline (NURBS) functions have strong modeling capability and high-order continuity, easily meeting the <em>C</em><sup>1</sup> continuity requirement of the RPT model. Vibration characteristics of square, skew, and elliptical plates are explored by RPT-IGA method. A novel bidirectional functionally graded material distribution of skew plate is introduced. The convergence was benchmarked against existing research data. This study finds that boundary conditions, geometric parameters, and the graded index significantly affect the dynamic response of in-plane non-uniform FGPPL.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"215 ","pages":"Article 104122"},"PeriodicalIF":5.7,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146191232","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Eric Simo , Thomas Helfer , Tymofiy Gerasimov , Christoph Lehmann , Dmitri Naumov , Tomas Krejci , Tomáš Koudelka , Jaroslav Kruis , Thomas Nagel , David Mašín
{"title":"Wrapper-based integration of a thermo-hydro-mechanical hypoplastic bentonite model in OpenGeoSys via MFront","authors":"Eric Simo , Thomas Helfer , Tymofiy Gerasimov , Christoph Lehmann , Dmitri Naumov , Tomas Krejci , Tomáš Koudelka , Jaroslav Kruis , Thomas Nagel , David Mašín","doi":"10.1016/j.advengsoft.2026.104106","DOIUrl":"10.1016/j.advengsoft.2026.104106","url":null,"abstract":"<div><div>This paper presents the implementation of a hypoplastic thermo-hydro-mechanical constitutive model for bentonite into the <span><span>OpenGeoSys</span><svg><path></path></svg></span> (OGS) simulation platform using a wrapper-based interface with the <span><span>MFront</span><svg><path></path></svg></span> code generation tool. The model, originally developed as a stand-alone model in the <span><span>Triax</span><svg><path></path></svg></span> platform, features a double-structure formulation that captures micro- and macroscale interactions, as well as suction and temperature-dependent mechanical responses of bentonite relevant to engineered barrier systems in deep geological repositories (DGR). The integration strategy preserves the validated legacy implementation of the model while enabling its use within modern multiphysics simulations through a modular <span><span>MFront</span><svg><path></path></svg></span> wrapper. Verification is performed using single-element benchmarks in <span><span>Triax</span><svg><path></path></svg></span> and <span><span>MTest</span><svg><path></path></svg></span>, demonstrating numerical consistency. The implementation is validated by simulating two laboratory experiments on Czech bentonite (BCV), showing good agreement with experimental measurements and with reference results from the <span><span>Sifel</span><svg><path></path></svg></span> code. The developed interface provides a robust and flexible framework for coupling advanced bentonite models with large-scale geotechnical simulations for application in the integrity and performance assessments of bentonite barriers in the scope of safety assessment of DGR.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"215 ","pages":"Article 104106"},"PeriodicalIF":5.7,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146057542","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}