Pub Date : 2025-12-12DOI: 10.1016/j.compstruc.2025.108073
Dachen Gao , Hongduo Zhao , Ke Cheng , Yuxuan Xia , Haoyu Chen , Yaowen Yang
This paper presents a reaction–diffusion equation (RDE) driven level set method (LSM) for topology optimization (TO) that enforces both stress and volume constraints simultaneously. The method introduces a locally activated stress penalty that operates only where the allowable limit is exceeded, eliminating the need for global aggregation and improving fidelity in hot-spot regions. A refined in-element triangulation strategy provides accurate volume fractions without remeshing, enabling precise volume tracking on fixed structured meshes. Structural evolution is governed by RDE, enabling hole nucleation during optimization and eliminating the need for level set reinitialization. Numerical experiments in 2D and 3D demonstrate that the proposed method yields designs that satisfy the prescribed local stress limits and target volume fractions while achieving stable, efficient convergence.
{"title":"A reaction–diffusion level set method for stress-constrained topology optimization with precise volume control","authors":"Dachen Gao , Hongduo Zhao , Ke Cheng , Yuxuan Xia , Haoyu Chen , Yaowen Yang","doi":"10.1016/j.compstruc.2025.108073","DOIUrl":"10.1016/j.compstruc.2025.108073","url":null,"abstract":"<div><div>This paper presents a reaction–diffusion equation (RDE) driven level set method (LSM) for topology optimization (TO) that enforces both stress and volume constraints simultaneously. The method introduces a locally activated stress penalty that operates only where the allowable limit is exceeded, eliminating the need for global aggregation and improving fidelity in hot-spot regions. A refined in-element triangulation strategy provides accurate volume fractions without remeshing, enabling precise volume tracking on fixed structured meshes. Structural evolution is governed by RDE, enabling hole nucleation during optimization and eliminating the need for level set reinitialization. Numerical experiments in 2D and 3D demonstrate that the proposed method yields designs that satisfy the prescribed local stress limits and target volume fractions while achieving stable, efficient convergence.</div></div>","PeriodicalId":50626,"journal":{"name":"Computers & Structures","volume":"321 ","pages":"Article 108073"},"PeriodicalIF":4.8,"publicationDate":"2025-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145732463","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}
Pub Date : 2025-12-12DOI: 10.1016/j.compstruc.2025.108069
L. Faccini, F. Castellini, E. Di Gialleonardo, S. Alfi, S. Bionda, R. Corradi
Ground-borne vibrations from rail transit systems present challenges in urban areas due to their effects on structures and human comfort. This study introduces a falling mass impact setup to estimate the transmissibility between the track and a receiver in the surrounding area. This enables the calculation of the Line Source Transfer Mobility (LSTM), required by the U.S. Federal Transit Administration (FTA) to assess ground-borne vibrations from rail vehicles. The falling mass method offers higher energy input and better repeatability than traditional impact hammers, making it suitable for evaluating soil and building transmissibility even at long distances. Transfer mobilities obtained with this method are validated against hammer-based measurements. Vibration levels produced by a modern tramcar running at 10, 30, and 50 km/h are measured at various distances and normalised using the corresponding LSTM, in line with the FTA Detailed Assessment Method. Force Density Levels (FDLs) are calculated for each speed, with the highest values at 50 km/h. A strong consistency of FDLs across distances confirms the method’s robustness. This integrated experimental approach offers a reliable framework for characterising vibration sources and supports the assessment of vehicles and infrastructure planning in areas sensitive to vibration.
{"title":"An experimental assessment of ground-borne vibration impact of tramcars","authors":"L. Faccini, F. Castellini, E. Di Gialleonardo, S. Alfi, S. Bionda, R. Corradi","doi":"10.1016/j.compstruc.2025.108069","DOIUrl":"10.1016/j.compstruc.2025.108069","url":null,"abstract":"<div><div>Ground-borne vibrations from rail transit systems present challenges in urban areas due to their effects on structures and human comfort. This study introduces a falling mass impact setup to estimate the transmissibility between the track and a receiver in the surrounding area. This enables the calculation of the Line Source Transfer Mobility (LSTM), required by the U.S. Federal Transit Administration (FTA) to assess ground-borne vibrations from rail vehicles. The falling mass method offers higher energy input and better repeatability than traditional impact hammers, making it suitable for evaluating soil and building transmissibility even at long distances. Transfer mobilities obtained with this method are validated against hammer-based measurements. Vibration levels produced by a modern tramcar running at 10, 30, and 50 km/h are measured at various distances and normalised using the corresponding LSTM, in line with the FTA Detailed Assessment Method. Force Density Levels (FDLs) are calculated for each speed, with the highest values at 50 km/h. A strong consistency of FDLs across distances confirms the method’s robustness. This integrated experimental approach offers a reliable framework for characterising vibration sources and supports the assessment of vehicles and infrastructure planning in areas sensitive to vibration.</div></div>","PeriodicalId":50626,"journal":{"name":"Computers & Structures","volume":"321 ","pages":"Article 108069"},"PeriodicalIF":4.8,"publicationDate":"2025-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145732464","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}
Pub Date : 2025-12-11DOI: 10.1016/j.compstruc.2025.108076
Yan Gu , Wenzhen Qu , Chuanzeng Zhang , Vladimir Babeshko , Yuri V. Petrov
This study presents a novel boundary element method (BEM) framework for the accurate and efficient numerical solution of elastodynamic problems. By reformulating the time-derivative terms as equivalent body forces, the method enables the use of static fundamental solutions for dynamic analysis, thereby eliminating the need for frequency-domain transformations or the construction of complex time-dependent Green’s functions. In the temporal domain, instead of directly approximating the time-differentiation operators, a more stable time-spectral integration technique based on orthogonal polynomial expansions is introduced. This scheme can, in principle, achieve arbitrary-order of accuracy in the temporal space and eliminate the strict time-step limitations inherent in conventional finite-difference-based schemes. Moreover, the resulting coefficient matrix is time-independent and therefore needs to be computed only once for the entire time-marching process. To evaluate domain integrals, discontinuous triangular elements are employed for spatial discretization, and a scaled coordinate transformation (SCT) technique is employed to address singularities arising from source-field point coincidences. Preliminary numerical experiments in elastodynamic analysis demonstrate that the proposed framework is robust and flexible for long-time dynamic simulations, particularly in problems involving rapid transients or complex geometries.
{"title":"High-order time-spectral BEM for efficient elastodynamic analysis","authors":"Yan Gu , Wenzhen Qu , Chuanzeng Zhang , Vladimir Babeshko , Yuri V. Petrov","doi":"10.1016/j.compstruc.2025.108076","DOIUrl":"10.1016/j.compstruc.2025.108076","url":null,"abstract":"<div><div>This study presents a novel boundary element method (BEM) framework for the accurate and efficient numerical solution of elastodynamic problems. By reformulating the time-derivative terms as equivalent body forces, the method enables the use of static fundamental solutions for dynamic analysis, thereby eliminating the need for frequency-domain transformations or the construction of complex time-dependent Green’s functions. In the temporal domain, instead of directly approximating the time-differentiation operators, a more stable time-spectral integration technique based on orthogonal polynomial expansions is introduced. This scheme can, in principle, achieve arbitrary-order of accuracy in the temporal space and eliminate the strict time-step limitations inherent in conventional finite-difference-based schemes. Moreover, the resulting coefficient matrix is time-independent and therefore needs to be computed only once for the entire time-marching process. To evaluate domain integrals, discontinuous triangular elements are employed for spatial discretization, and a scaled coordinate transformation (SCT) technique is employed to address singularities arising from source-field point coincidences. Preliminary numerical experiments in elastodynamic analysis demonstrate that the proposed framework is robust and flexible for long-time dynamic simulations, particularly in problems involving rapid transients or complex geometries.</div></div>","PeriodicalId":50626,"journal":{"name":"Computers & Structures","volume":"321 ","pages":"Article 108076"},"PeriodicalIF":4.8,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145731206","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}
Pub Date : 2025-12-11DOI: 10.1016/j.compstruc.2025.108071
Mordechay Buzaglo , Nicolò Pollini
A novel first-order equivalent static loads approach for optimization of structural dynamic response, F-ESL, is presented and compared with the basic equivalent static load formulation, ESL. F-ESL simplifies dynamic optimization problems by converting them into a series of static optimization sub-problems. The ESL algorithm in its original formulation does not have a guaranteed capability of reaching, or recognizing, final designs that satisfy the necessary first-order optimality conditions. F-ESL addresses this limitation by including first-order terms directly into the equivalent static load definition. This new mathematical information guides the optimization algorithm more effectively toward solutions that satisfy both feasibility and optimality conditions. Using reproducible numerical examples, we show that F-ESL overcomes the known limitations of the original ESL, often with a few outer function evaluations and fast convergence. At the same time, F-ESL maintains ESL simplicity, robustness, and ease of implementation, providing practitioners with an effective tool for structural dynamic optimization problems.
{"title":"First-order equivalent static loads for dynamic response structural optimization","authors":"Mordechay Buzaglo , Nicolò Pollini","doi":"10.1016/j.compstruc.2025.108071","DOIUrl":"10.1016/j.compstruc.2025.108071","url":null,"abstract":"<div><div>A novel first-order equivalent static loads approach for optimization of structural dynamic response, F-ESL, is presented and compared with the basic equivalent static load formulation, ESL. F-ESL simplifies dynamic optimization problems by converting them into a series of static optimization sub-problems. The ESL algorithm in its original formulation does not have a guaranteed capability of reaching, or recognizing, final designs that satisfy the necessary first-order optimality conditions. F-ESL addresses this limitation by including first-order terms directly into the equivalent static load definition. This new mathematical information guides the optimization algorithm more effectively toward solutions that satisfy both feasibility and optimality conditions. Using reproducible numerical examples, we show that F-ESL overcomes the known limitations of the original ESL, often with a few outer function evaluations and fast convergence. At the same time, F-ESL maintains ESL simplicity, robustness, and ease of implementation, providing practitioners with an effective tool for structural dynamic optimization problems.</div></div>","PeriodicalId":50626,"journal":{"name":"Computers & Structures","volume":"321 ","pages":"Article 108071"},"PeriodicalIF":4.8,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145731205","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}
Pub Date : 2025-12-08DOI: 10.1016/j.compstruc.2025.108074
Pengfei Zhang , Qiang Zhang , Weiwei Xie , Bo Yu
A physics-informed Gaussian process regression (PI-GPR) model for predicting the probabilistic flexural strength of corroded reinforced concrete (RC) columns was developed based on the multi-level embedding strategy. According to the moment equilibrium conditions, a physical model representing the flexural mechanism of corroded RC columns was established first. Subsequently, the multi-level embedding strategy was adopted to develop the PI-GPR model by constraining the mean, kernel, and loss functions hierarchically. Meanwhile, an adaptive trade-off between physical priors and data features was achieved by adding the dynamic coefficients that are continuously updated as training data changes. Then the PI-GPR model was applied to predict the probabilistic flexural strength of corroded RC columns. Finally, the effectiveness of the PI-GPR model was validated by comparing it with traditional prediction methods. Analysis results show that the PI-GPR model, which integrates the physical constraints with data-driven learning, demonstrates excellent performance and robust uncertainty quantification. Compared with traditional GPR, the PI-GPR ensures high physical consistency of predictions, which not only achieves an average improvement of 20% in confidence interval coverage when the training set proportion was only 30%, but also reduces the mean absolute error and root mean square error for edge samples by 23% and 21%, respectively.
{"title":"Predicting probabilistic flexural strength of corroded reinforced concrete columns based on physics-informed GPR model","authors":"Pengfei Zhang , Qiang Zhang , Weiwei Xie , Bo Yu","doi":"10.1016/j.compstruc.2025.108074","DOIUrl":"10.1016/j.compstruc.2025.108074","url":null,"abstract":"<div><div>A physics-informed Gaussian process regression (PI-GPR) model for predicting the probabilistic flexural strength of corroded reinforced concrete (RC) columns was developed based on the multi-level embedding strategy. According to the moment equilibrium conditions, a physical model representing the flexural mechanism of corroded RC columns was established first. Subsequently, the multi-level embedding strategy was adopted to develop the PI-GPR model by constraining the mean, kernel, and loss functions hierarchically. Meanwhile, an adaptive trade-off between physical priors and data features was achieved by adding the dynamic coefficients that are continuously updated as training data changes. Then the PI-GPR model was applied to predict the probabilistic flexural strength of corroded RC columns. Finally, the effectiveness of the PI-GPR model was validated by comparing it with traditional prediction methods. Analysis results show that the PI-GPR model, which integrates the physical constraints with data-driven learning, demonstrates excellent performance and robust uncertainty quantification. Compared with traditional GPR, the PI-GPR ensures high physical consistency of predictions, which not only achieves an average improvement of 20% in confidence interval coverage when the training set proportion was only 30%, but also reduces the mean absolute error and root mean square error for edge samples by 23% and 21%, respectively.</div></div>","PeriodicalId":50626,"journal":{"name":"Computers & Structures","volume":"321 ","pages":"Article 108074"},"PeriodicalIF":4.8,"publicationDate":"2025-12-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145731211","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}
Pub Date : 2025-12-06DOI: 10.1016/j.compstruc.2025.108054
Sharana Kumar Shivanand , Bojana Rosić
We propose novel scale-invariant error estimators for the Monte Carlo and multilevel Monte Carlo estimation of mean and variance. For any linear transformation of the distribution of the quantity of interest, the computation cost across fidelity levels is optimized using a normalized error estimate, which is not only fully dimensionless but also remains robust to variations in the characteristics of the distribution. We demonstrate the effectiveness of the algorithms through application to a mechanical simulation of linear elastic bone tissue, where material uncertainty incorporating both heterogeneity and random anisotropy is considered in the constitutive law.
{"title":"Scale-invariant Monte Carlo and multilevel Monte Carlo estimation of mean and variance: An application to simulation of linear elastic bone tissue","authors":"Sharana Kumar Shivanand , Bojana Rosić","doi":"10.1016/j.compstruc.2025.108054","DOIUrl":"10.1016/j.compstruc.2025.108054","url":null,"abstract":"<div><div>We propose novel scale-invariant error estimators for the Monte Carlo and multilevel Monte Carlo estimation of mean and variance. For any linear transformation of the distribution of the quantity of interest, the computation cost across fidelity levels is optimized using a normalized error estimate, which is not only fully dimensionless but also remains robust to variations in the characteristics of the distribution. We demonstrate the effectiveness of the algorithms through application to a mechanical simulation of linear elastic bone tissue, where material uncertainty incorporating both heterogeneity and random anisotropy is considered in the constitutive law.</div></div>","PeriodicalId":50626,"journal":{"name":"Computers & Structures","volume":"321 ","pages":"Article 108054"},"PeriodicalIF":4.8,"publicationDate":"2025-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145689748","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}
Pub Date : 2025-12-05DOI: 10.1016/j.compstruc.2025.108068
Alberto M.B. Martins , Luís M.C. Simões , João H.J.O. Negrão
This article presents an optimization-based approach to assist in the design of bowstring tied-arch concrete bridges. This approach comprises an automated procedure to define initial designs and a gradient-based algorithm, from which local optimum solutions are obtained, and the least-cost solution is selected as the optimum design. The finite element method is used for the three-dimensional analysis considering several load cases, geometrical nonlinearities and time-dependent effects. The design is posed as a cost minimization subject to constraints on the displacements and stresses defined according to the Eurocodes provisions. A constraint aggregation approach is adopted to solve the problem by minimizing a convex scalar function. The discrete direct method for sensitivity analysis provides the algorithm with the structural response to changes in the design variables. The design variables are the arch and deck cross-sectional sizes, the hangers and tendons cross-sectional areas and prestressing forces, the geometry of the arch, the hangers’ layout and the number of hangers’ anchoring points in the deck. The optimization of a 120 m single-span bridge illustrates the features and applicability of the proposed approach. Optimum solution with Nielsen layout, deck slenderness of 1/150 and arch rise-to-span ratio of 1/5.
{"title":"Automated design optimization of bowstring tied-arch concrete bridges","authors":"Alberto M.B. Martins , Luís M.C. Simões , João H.J.O. Negrão","doi":"10.1016/j.compstruc.2025.108068","DOIUrl":"10.1016/j.compstruc.2025.108068","url":null,"abstract":"<div><div>This article presents an optimization-based approach to assist in the design of bowstring tied-arch concrete bridges. This approach comprises an automated procedure to define initial designs and a gradient-based algorithm, from which local optimum solutions are obtained, and the least-cost solution is selected as the optimum design. The finite element method is used for the three-dimensional analysis considering several load cases, geometrical nonlinearities and time-dependent effects. The design is posed as a cost minimization subject to constraints on the displacements and stresses defined according to the Eurocodes provisions. A constraint aggregation approach is adopted to solve the problem by minimizing a convex scalar function. The discrete direct method for sensitivity analysis provides the algorithm with the structural response to changes in the design variables. The design variables are the arch and deck cross-sectional sizes, the hangers and tendons cross-sectional areas and prestressing forces, the geometry of the arch, the hangers’ layout and the number of hangers’ anchoring points in the deck. The optimization of a 120 m single-span bridge illustrates the features and applicability of the proposed approach. Optimum solution with Nielsen layout, deck slenderness of 1/150 and arch rise-to-span ratio of 1/5.</div></div>","PeriodicalId":50626,"journal":{"name":"Computers & Structures","volume":"321 ","pages":"Article 108068"},"PeriodicalIF":4.8,"publicationDate":"2025-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145685360","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}
Pub Date : 2025-12-04DOI: 10.1016/j.compstruc.2025.108067
Saeed Saeedmonir , Alireza Sadeghirad
Isogeometric Analysis is known to be a powerful numerical method in bridging between Computer-Aided Design and Computational Mechanics by utilizing special spline basis functions for both geometric and field discretization, resulting in reduction of errors corresponding to geometric approximations in traditional Finite Elements Method. Moreover, it can satisfy higher-order continuity to ensure stability and accuracy. However, there are several challenges in dealing with multi-patch domains, especially in the case of non-conforming patches in establishing continuity across the patch interfaces. Existing methods such as the penalty method, Lagrange multipliers and Nitsche’s method often encounter several numerical issues such as ill-conditioned systems, saddle point problem and computational overhead. This study presents a novel variational formulation for coupling of non-conforming patches while preserving higher-order continuity without introducing new degrees of freedom. In addition, the proposed method ensures a symmetric system, guaranteeing stable and efficient implementation. The constraints are applied through the variational formulation and therefore, no other approach is utilized within the procedure. Some benchmark numerical examples are provided to demonstrate the performance, accuracy and effectiveness of the proposed method.
{"title":"A new variational approach for coupling of non-conforming patches in Isogeometric Analysis","authors":"Saeed Saeedmonir , Alireza Sadeghirad","doi":"10.1016/j.compstruc.2025.108067","DOIUrl":"10.1016/j.compstruc.2025.108067","url":null,"abstract":"<div><div>Isogeometric Analysis is known to be a powerful numerical method in bridging between Computer-Aided Design and Computational Mechanics by utilizing special spline basis functions for both geometric and field discretization, resulting in reduction of errors corresponding to geometric approximations in traditional Finite Elements Method. Moreover, it can satisfy higher-order continuity to ensure stability and accuracy. However, there are several challenges in dealing with multi-patch domains, especially in the case of non-conforming patches in establishing continuity across the patch interfaces. Existing methods such as the penalty method, Lagrange multipliers and Nitsche’s method often encounter several numerical issues such as ill-conditioned systems, saddle point problem and computational overhead. This study presents a novel variational formulation for coupling of non-conforming patches while preserving higher-order continuity without introducing new degrees of freedom. In addition, the proposed method ensures a symmetric system, guaranteeing stable and efficient implementation. The constraints are applied through the variational formulation and therefore, no other approach is utilized within the procedure. Some benchmark numerical examples are provided to demonstrate the performance, accuracy and effectiveness of the proposed method.</div></div>","PeriodicalId":50626,"journal":{"name":"Computers & Structures","volume":"321 ","pages":"Article 108067"},"PeriodicalIF":4.8,"publicationDate":"2025-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145685361","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}
Pub Date : 2025-12-04DOI: 10.1016/j.compstruc.2025.108070
Boxu Tian , Wenliang Qian , Yefei Yang
Topology optimization is a powerful computational methodology for generating high-performance structures. However, ignoring aesthetics limits the practical application. User interaction helps address this limitation, but current methods lack generality, needing separate design and sensitivity analyses for each interaction form. To overcome this challenge, we propose a Designer Preference-Driven Neural Topology Optimization (DPDNTO) method, which utilizes a unified interaction parameter to formulate a unified, mask-based loss function to control geometric features and material distribution. In addition, feature sizes are flexibly controlled by adjusting the number of neurons in the DPDNTO method. By utilizing the backpropagation mechanism of neural networks, the proposed method efficiently updates design variables and automatically balances these multi-objective tasks through a dynamic parameter strategy. To provide further intuitive visual feedback, a multimodal large model is employed to render optimized structures into conceptual visualization images. Numerical experiments demonstrate that the proposed method not only enhances the aesthetic quality of the final designs but also improves structural stress performance and linear buckling resistance. These findings establish DPDNTO as a versatile and computationally efficient paradigm, bridging the critical gap between algorithmic optimization and preference-driven aesthetics and paving the way for advancements in fields such as architecture, industrial design, and advanced manufacturing.
{"title":"Designer preference-driven topology optimization using a human-in-the-loop neural network","authors":"Boxu Tian , Wenliang Qian , Yefei Yang","doi":"10.1016/j.compstruc.2025.108070","DOIUrl":"10.1016/j.compstruc.2025.108070","url":null,"abstract":"<div><div>Topology optimization is a powerful computational methodology for generating high-performance structures. However, ignoring aesthetics limits the practical application. User interaction helps address this limitation, but current methods lack generality, needing separate design and sensitivity analyses for each interaction form. To overcome this challenge, we propose a Designer Preference-Driven Neural Topology Optimization (DPDNTO) method, which utilizes a unified interaction parameter to formulate a unified, mask-based loss function to control geometric features and material distribution. In addition, feature sizes are flexibly controlled by adjusting the number of neurons in the DPDNTO method. By utilizing the backpropagation mechanism of neural networks, the proposed method efficiently updates design variables and automatically balances these multi-objective tasks through a dynamic parameter strategy. To provide further intuitive visual feedback, a multimodal large model is employed to render optimized structures into conceptual visualization images. Numerical experiments demonstrate that the proposed method not only enhances the aesthetic quality of the final designs but also improves structural stress performance and linear buckling resistance. These findings establish DPDNTO as a versatile and computationally efficient paradigm, bridging the critical gap between algorithmic optimization and preference-driven aesthetics and paving the way for advancements in fields such as architecture, industrial design, and advanced manufacturing.</div></div>","PeriodicalId":50626,"journal":{"name":"Computers & Structures","volume":"321 ","pages":"Article 108070"},"PeriodicalIF":4.8,"publicationDate":"2025-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145685359","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}
This study presents a computational framework that integrates micromechanics, a phase-field damage model, and genetic programming-based symbolic regression to predict the elastic modulus and compressive strength of both normal and high-strength concrete. The micromechanical formulations are developed using an extended generalized self-consistent scheme, incorporating a quasi-elasto-plastic brittle behavior governed by a von Mises yield criterion. A simplified phase-field damage model is proposed, introducing a compressive phase-field variable along with an analytical approximation that links fracture energy to the macroscopic behavior of concrete. The proposed framework is validated through strong agreement among numerical simulations, experimental observations, and theoretical predictions, supporting the development of a robust theoretical database for elastic modulus, compressive strength, and associated material properties. Based on this dataset, an efficient computational strategy is developed and examined to generate simple and practical symbolic regression expressions within the Genetic Programming-based framework to derive predictive equations for elastic modulus as a function of compressive strength and aggregate characteristics. These equations are validated against established standards and experimental data, confirming their accuracy and practical relevance for structural design applications.
{"title":"A micromechanics-based computational framework for predicting the elastic modulus and compressive strength of normal and high-performance concrete","authors":"Hoang-Quan Nguyen , Gia-Khuyen Le , Tinh Quoc Bui , Thi-Loan Bui , Bao-Viet Tran","doi":"10.1016/j.compstruc.2025.108055","DOIUrl":"10.1016/j.compstruc.2025.108055","url":null,"abstract":"<div><div>This study presents a computational framework that integrates micromechanics, a phase-field damage model, and genetic programming-based symbolic regression to predict the elastic modulus and compressive strength of both normal and high-strength concrete. The micromechanical formulations are developed using an extended generalized self-consistent scheme, incorporating a quasi-elasto-plastic brittle behavior governed by a von Mises yield criterion. A simplified phase-field damage model is proposed, introducing a compressive phase-field variable along with an analytical approximation that links fracture energy to the macroscopic behavior of concrete. The proposed framework is validated through strong agreement among numerical simulations, experimental observations, and theoretical predictions, supporting the development of a robust theoretical database for elastic modulus, compressive strength, and associated material properties. Based on this dataset, an efficient computational strategy is developed and examined to generate simple and practical symbolic regression expressions within the Genetic Programming-based framework to derive predictive equations for elastic modulus as a function of compressive strength and aggregate characteristics. These equations are validated against established standards and experimental data, confirming their accuracy and practical relevance for structural design applications.</div></div>","PeriodicalId":50626,"journal":{"name":"Computers & Structures","volume":"321 ","pages":"Article 108055"},"PeriodicalIF":4.8,"publicationDate":"2025-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145657739","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}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号