Pub Date : 2025-12-31DOI: 10.1016/j.enganabound.2025.106618
Zhuoheng Wang , Wenxuan Xie , Junseok Kim , Yibao Li
Reliability-based topology optimization is the approach that incorporates uncertainty quantification into topology optimization to achieve robustness and reliability. In this paper, we propose a reliability-based topology optimization algorithm that integrates a phase-field model with uncertainty quantification techniques. Material uncertainty in Young’s modulus is modeled as a random field and reduced via the Karhunen–Loève expansion. A sequential optimization and reliability assessment strategy is adopted to decouple the optimization and reliability evaluation for reducing computational complexity. The topology optimization problem is solved by using the coupled finite element and finite difference approach, in which time integration is performed using a second-order Crank–Nicolson scheme. Reliability analysis is conducted by using the inverse reliability analysis method, with the limit-state function approximated by a stochastic response surface method. Monte Carlo simulations are performed to validate the accuracy of the computed failure probabilities. Numerical results confirm the efficiency and robustness of the proposed method in generating reliable structures under material uncertainty.
{"title":"Efficient phase field structural design algorithm for reliability-based topology optimization with material uncertainties","authors":"Zhuoheng Wang , Wenxuan Xie , Junseok Kim , Yibao Li","doi":"10.1016/j.enganabound.2025.106618","DOIUrl":"10.1016/j.enganabound.2025.106618","url":null,"abstract":"<div><div>Reliability-based topology optimization is the approach that incorporates uncertainty quantification into topology optimization to achieve robustness and reliability. In this paper, we propose a reliability-based topology optimization algorithm that integrates a phase-field model with uncertainty quantification techniques. Material uncertainty in Young’s modulus is modeled as a random field and reduced via the Karhunen–Loève expansion. A sequential optimization and reliability assessment strategy is adopted to decouple the optimization and reliability evaluation for reducing computational complexity. The topology optimization problem is solved by using the coupled finite element and finite difference approach, in which time integration is performed using a second-order Crank–Nicolson scheme. Reliability analysis is conducted by using the inverse reliability analysis method, with the limit-state function approximated by a stochastic response surface method. Monte Carlo simulations are performed to validate the accuracy of the computed failure probabilities. Numerical results confirm the efficiency and robustness of the proposed method in generating reliable structures under material uncertainty.</div></div>","PeriodicalId":51039,"journal":{"name":"Engineering Analysis with Boundary Elements","volume":"183 ","pages":"Article 106618"},"PeriodicalIF":4.1,"publicationDate":"2025-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884143","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 : 2025-12-31DOI: 10.1016/j.enganabound.2025.106615
Zhongting Xu , Zhengjie Sun , Shengliang Zhang
This paper proposes a novel meshless Galerkin energy-preserving method for conservative evolutionary partial differential equations (PDEs). By combining the meshless Galerkin approach for spatial discretization with the average vector field method, we establish a general framework for constructing meshless energy-preserving schemes applicable to a wide range of PDEs. Moreover, we present explicit schemes for the Korteweg–de Vries and nonlinear Schrödinger equations, analyzing their energy conservation properties and convergence results. A salient feature of our method is the use of problem-driven kernels, with periodic kernels for one-dimensional periodic problems and localized kernels for two-dimensional ones. Numerical results demonstrate the high accuracy of our method on scattered nodes and highlight the benefits of the meshless approach. It also achieves superior energy and mass preservation compared to mesh-dependent methods.
{"title":"Meshless energy-conserving schemes for conservative partial differential equations using kernel-based Galerkin methods","authors":"Zhongting Xu , Zhengjie Sun , Shengliang Zhang","doi":"10.1016/j.enganabound.2025.106615","DOIUrl":"10.1016/j.enganabound.2025.106615","url":null,"abstract":"<div><div>This paper proposes a novel meshless Galerkin energy-preserving method for conservative evolutionary partial differential equations (PDEs). By combining the meshless Galerkin approach for spatial discretization with the average vector field method, we establish a general framework for constructing meshless energy-preserving schemes applicable to a wide range of PDEs. Moreover, we present explicit schemes for the Korteweg–de Vries and nonlinear Schrödinger equations, analyzing their energy conservation properties and convergence results. A salient feature of our method is the use of problem-driven kernels, with periodic kernels for one-dimensional periodic problems and localized kernels for two-dimensional ones. Numerical results demonstrate the high accuracy of our method on scattered nodes and highlight the benefits of the meshless approach. It also achieves superior energy and mass preservation compared to mesh-dependent methods.</div></div>","PeriodicalId":51039,"journal":{"name":"Engineering Analysis with Boundary Elements","volume":"183 ","pages":"Article 106615"},"PeriodicalIF":4.1,"publicationDate":"2025-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884139","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 paper, Part 3 of a series, extends the previously developed strong-form hybrid radial basis function-finite difference (RBF-FD) method to model the thermomechanics of continuous casting (CC) of steel. Part 1 introduced the method for non-linear thermomechanics, and Part 2 applied it to the cooling of steel bars on a cooling bed. Here, a one-way coupled thermo-mechanical slice model is developed, where the temperature field provides thermal loading for the mechanical analysis. The previously introduced 2.5D formulation is adapted to include strand straightening. A visco-plastic material model, validated in our recent work, is used to describe material behaviour close to the mushy-zone temperatures. The analysis focuses on the solidified part of the strand, while the space discretisation remains constant and liquid-zone results are disregarded. Two hot-tearing criteria—temperature-based and stress-based are implemented and compared with experimental data, showing that the stress-based approach produces more realistic predictions. The influence of electromagnetic stirring, steel grade composition, and casting speed on hot-tearing susceptibility is explored. This study represents the first strong-form RBF-FD solution of CC thermomechanics, demonstrating that the hybrid RBF-FD can efficiently handle complex non-linear thermo-mechanical behaviour and allows for further process optimisation and defect mitigation in industrial CC.
{"title":"A hybrid radial basis function-finite difference method for modelling two-dimensional thermo-elasto-plasticity, Part 3: Application to thermo-mechanical modelling of continuous casting of steel billets","authors":"Gašper Vuga, Boštjan Mavrič, Tadej Dobravec, Božidar Šarler","doi":"10.1016/j.enganabound.2025.106619","DOIUrl":"10.1016/j.enganabound.2025.106619","url":null,"abstract":"<div><div>This paper, Part 3 of a series, extends the previously developed strong-form hybrid radial basis function-finite difference (RBF-FD) method to model the thermomechanics of continuous casting (CC) of steel. Part 1 introduced the method for non-linear thermomechanics, and Part 2 applied it to the cooling of steel bars on a cooling bed. Here, a one-way coupled thermo-mechanical slice model is developed, where the temperature field provides thermal loading for the mechanical analysis. The previously introduced 2.5D formulation is adapted to include strand straightening. A visco-plastic material model, validated in our recent work, is used to describe material behaviour close to the mushy-zone temperatures. The analysis focuses on the solidified part of the strand, while the space discretisation remains constant and liquid-zone results are disregarded. Two hot-tearing criteria—temperature-based and stress-based are implemented and compared with experimental data, showing that the stress-based approach produces more realistic predictions. The influence of electromagnetic stirring, steel grade composition, and casting speed on hot-tearing susceptibility is explored. This study represents the first strong-form RBF-FD solution of CC thermomechanics, demonstrating that the hybrid RBF-FD can efficiently handle complex non-linear thermo-mechanical behaviour and allows for further process optimisation and defect mitigation in industrial CC.</div></div>","PeriodicalId":51039,"journal":{"name":"Engineering Analysis with Boundary Elements","volume":"183 ","pages":"Article 106619"},"PeriodicalIF":4.1,"publicationDate":"2025-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884141","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 : 2025-12-31DOI: 10.1016/j.enganabound.2025.106624
Yu Tian , Fu-Ren Ming , Hao Chen , Xiang-Li Fang , Ping-Ping Wang
The metal jet formed by a conventional shaped charge has a high penetration depth but a narrow damage range. To expand the damage range of the metal jet, an annular shaped charge can be employed. In this paper, the underwater explosions of annular shaped charges are simulated by a graphic processing unit accelerated axisymmetric Riemann-SPH method, and the accuracy is verified by experiments. The hole-cutting effect using an annular shaped charge in an underwater explosion is analyzed, and the characteristics of the annular shaped charge are compared with a spherical segment shaped charge. Furthermore, the effects of the liner thickness on the annular jet are explored. It is revealed that the damage mode inflicted on the plate by an explosive formed projectile (EFP) is impact penetration, whereas that by an annular jet is cutting. The plate’s breach size caused by the annular jet reaches 1.79 times the charge radius and 3.15 times that of EFP. Our findings also reveal that there is an optimal dimensionless maximum liner thickness (/) to maximize the breach size , which is 0.12 for a charge mass of 60kg. This paper can provide support for the optimization design of the shaped charge.
{"title":"Numerical study of hole-cutting effect using annular shaped charges in underwater explosions","authors":"Yu Tian , Fu-Ren Ming , Hao Chen , Xiang-Li Fang , Ping-Ping Wang","doi":"10.1016/j.enganabound.2025.106624","DOIUrl":"10.1016/j.enganabound.2025.106624","url":null,"abstract":"<div><div>The metal jet formed by a conventional shaped charge has a high penetration depth but a narrow damage range. To expand the damage range of the metal jet, an annular shaped charge can be employed. In this paper, the underwater explosions of annular shaped charges are simulated by a graphic processing unit accelerated axisymmetric Riemann-SPH method, and the accuracy is verified by experiments. The hole-cutting effect using an annular shaped charge in an underwater explosion is analyzed, and the characteristics of the annular shaped charge are compared with a spherical segment shaped charge. Furthermore, the effects of the liner thickness on the annular jet are explored. It is revealed that the damage mode inflicted on the plate by an explosive formed projectile (EFP) is impact penetration, whereas that by an annular jet is cutting. The plate’s breach size caused by the annular jet reaches 1.79 times the charge radius and 3.15 times that of EFP. Our findings also reveal that there is an optimal dimensionless maximum liner thickness <span><math><mi>λ</mi></math></span> (<span><math><mrow><mi>λ</mi><mo>=</mo><mi>T</mi></mrow></math></span>/<span><math><msub><mrow><mi>R</mi></mrow><mrow><mi>z</mi></mrow></msub></math></span>) to maximize the breach size <span><math><msub><mrow><mi>d</mi></mrow><mrow><mi>p</mi></mrow></msub></math></span>, which is 0.12 for a charge mass of 60kg. This paper can provide support for the optimization design of the shaped charge.</div></div>","PeriodicalId":51039,"journal":{"name":"Engineering Analysis with Boundary Elements","volume":"183 ","pages":"Article 106624"},"PeriodicalIF":4.1,"publicationDate":"2025-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884238","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 : 2025-12-30DOI: 10.1016/j.enganabound.2025.106622
Yicheng Liu , Yilin Li , Tao Liu
This work introduces a meshless computational methodology for the valuation of European-style options under the Bates stochastic volatility jump-diffusion framework. The governing dynamics are described by a two-dimensional partial integro-differential equation (PIDE) that incorporates both mixed derivative operators and a nonlocal integral component. The numerical strategy is constructed within the Radial Basis Function-Finite Difference (RBF–FD) paradigm, employing a modified multiquadric kernel that facilitates the analytical determination of localized differentiation weights. Spatial discretization is achieved through the method of lines, where the integral operator is approximated by means of quadrature rules combined with accurate interpolation techniques. Numerical experiments reveal that the proposed solver delivers superior accuracy while lowering computational cost in comparison with classical schemes.
{"title":"An RBF–FD method for pricing under the Bates model: Handling stochastic volatility and jump processes","authors":"Yicheng Liu , Yilin Li , Tao Liu","doi":"10.1016/j.enganabound.2025.106622","DOIUrl":"10.1016/j.enganabound.2025.106622","url":null,"abstract":"<div><div>This work introduces a meshless computational methodology for the valuation of European-style options under the Bates stochastic volatility jump-diffusion framework. The governing dynamics are described by a two-dimensional partial integro-differential equation (PIDE) that incorporates both mixed derivative operators and a nonlocal integral component. The numerical strategy is constructed within the Radial Basis Function-Finite Difference (RBF–FD) paradigm, employing a modified multiquadric kernel that facilitates the analytical determination of localized differentiation weights. Spatial discretization is achieved through the method of lines, where the integral operator is approximated by means of quadrature rules combined with accurate interpolation techniques. Numerical experiments reveal that the proposed solver delivers superior accuracy while lowering computational cost in comparison with classical schemes.</div></div>","PeriodicalId":51039,"journal":{"name":"Engineering Analysis with Boundary Elements","volume":"183 ","pages":"Article 106622"},"PeriodicalIF":4.1,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884138","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 : 2025-12-29DOI: 10.1016/j.enganabound.2025.106605
Minas Kouroublakis , Nikolaos L. Tsitsas , Yehuda Leviatan
This paper presents a time-domain implementation of the Method of Auxiliary Sources (MAS) combined with the Standard Impedance Boundary Condition (SIBC) for electromagnetic scattering problems involving cylindrical scatterers with finite but moderate conductivity. The proposed approach focuses on solving the two-dimensional problem using a first-order SIBC, which is valid when the conductivity is sufficiently higher than the maximum spectral frequency times the dielectric permittivity of the scatterer. This regime includes moderately conductive materials — such as carbon-based composites, conductive polymers, and doped dielectrics — that are increasingly used in real-world radio-frequency applications, including wearable electronics, electromagnetic interference shielding, and biomedical sensors. Under the above validity conditions, the interaction between the incident wave and the scatterer is dominated by surface effects, allowing for an efficient and accurate modeling strategy without the need to compute internal fields. The theoretical formulation of the time-domain MAS-SIBC method is developed, followed by extensive numerical testing on various geometries whose cross section is a closed curve. Such geometries include circular, elliptical, super-circular, rounded-triangular, and inverted-elliptical scatterers. A planar geometry is also tested. All results are validated against analytical solutions and commercial frequency-domain solvers, demonstrating the accuracy and practical potential of the proposed method. The findings suggest that time-domain MAS-SIBC offers a promising and computationally efficient approach for modeling scattering from materials even with moderate conductivity.
{"title":"A time-domain method of auxiliary sources for efficient analysis of transient electromagnetic scattering by moderately conductive cylinders","authors":"Minas Kouroublakis , Nikolaos L. Tsitsas , Yehuda Leviatan","doi":"10.1016/j.enganabound.2025.106605","DOIUrl":"10.1016/j.enganabound.2025.106605","url":null,"abstract":"<div><div>This paper presents a time-domain implementation of the Method of Auxiliary Sources (MAS) combined with the Standard Impedance Boundary Condition (SIBC) for electromagnetic scattering problems involving cylindrical scatterers with finite but moderate conductivity. The proposed approach focuses on solving the two-dimensional problem using a first-order SIBC, which is valid when the conductivity is sufficiently higher than the maximum spectral frequency times the dielectric permittivity of the scatterer. This regime includes moderately conductive materials — such as carbon-based composites, conductive polymers, and doped dielectrics — that are increasingly used in real-world radio-frequency applications, including wearable electronics, electromagnetic interference shielding, and biomedical sensors. Under the above validity conditions, the interaction between the incident wave and the scatterer is dominated by surface effects, allowing for an efficient and accurate modeling strategy without the need to compute internal fields. The theoretical formulation of the time-domain MAS-SIBC method is developed, followed by extensive numerical testing on various geometries whose cross section is a closed curve. Such geometries include circular, elliptical, super-circular, rounded-triangular, and inverted-elliptical scatterers. A planar geometry is also tested. All results are validated against analytical solutions and commercial frequency-domain solvers, demonstrating the accuracy and practical potential of the proposed method. The findings suggest that time-domain MAS-SIBC offers a promising and computationally efficient approach for modeling scattering from materials even with moderate conductivity.</div></div>","PeriodicalId":51039,"journal":{"name":"Engineering Analysis with Boundary Elements","volume":"183 ","pages":"Article 106605"},"PeriodicalIF":4.1,"publicationDate":"2025-12-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884137","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 : 2025-12-27DOI: 10.1016/j.enganabound.2025.106616
Zhanjun Shao , Zefeng Liu , Xuan Peng , Yufei Chen , Delei Yang , Linxin Peng , Ping Xiang
In previous studies of train–track–bridge coupled systems (TTBCS), box girder bridges were often simplified as one-dimensional Euler beam models, which limited the investigation of the effects of material uncertainties in the top plate, webs, and bottom plate on dynamic responses. To address this limitation, this study introduces a multi-plate box girder bridge model and proposes a stochastic dynamic computation framework combining the meshless method and new point estimate method (NPEM). Two-dimensional random field modeling is applied to the material parameters of each plate, and Karhunen–Loève (KL) expansion is employed to transform these fields into random variables. Subsequently, NPEM is utilized to compute the statistical characteristics of the dynamic responses of the TTBCS. By comparing the effects of different random fields on the dynamic responses, the sensitivity of the responses to various material parameters is quantitatively analyzed. The results indicate that the vertical displacement of the bridge is most sensitive to the thickness uncertainty of the bottom plate, followed by the Young’s modulus. The lateral displacement is most sensitive to the thickness uncertainty of the traffic-side web, followed by the Young’s modulus. The density random field exerts the greatest influence on the bridge natural frequency. In the absence of seismic effects, uncertainties in the substructure have minimal influence on vehicle accelerations.
在以往的列车-轨道-桥梁耦合系统（TTBCS）研究中，箱梁桥通常被简化为一维欧拉梁模型，这限制了对顶板、腹板和底板材料不确定性对动力响应影响的研究。针对这一局限性，本文引入多板箱梁桥模型，提出了一种结合无网格法和新点估计法（NPEM）的随机动力计算框架。对每块板的材料参数进行二维随机场建模，利用karhunen - lo (KL)展开将这些场转化为随机变量。随后，利用NPEM计算TTBCS动态响应的统计特征。通过比较不同随机场对动态响应的影响，定量分析了响应对不同材料参数的敏感性。结果表明：桥梁的竖向位移对底板厚度的不确定性最敏感，其次是杨氏模量；横向位移对车辆侧腹板厚度的不确定性最敏感，其次是杨氏模量。密度随机场对桥梁固有频率的影响最大。在没有地震影响的情况下，子结构的不确定性对车辆加速度的影响最小。
{"title":"A stochastic dynamic analysis method for plate structures based on meshless method and new point estimate method: A case study of high-speed railway box girder bridges","authors":"Zhanjun Shao , Zefeng Liu , Xuan Peng , Yufei Chen , Delei Yang , Linxin Peng , Ping Xiang","doi":"10.1016/j.enganabound.2025.106616","DOIUrl":"10.1016/j.enganabound.2025.106616","url":null,"abstract":"<div><div>In previous studies of train–track–bridge coupled systems (TTBCS), box girder bridges were often simplified as one-dimensional Euler beam models, which limited the investigation of the effects of material uncertainties in the top plate, webs, and bottom plate on dynamic responses. To address this limitation, this study introduces a multi-plate box girder bridge model and proposes a stochastic dynamic computation framework combining the meshless method and new point estimate method (NPEM). Two-dimensional random field modeling is applied to the material parameters of each plate, and Karhunen–Loève (KL) expansion is employed to transform these fields into random variables. Subsequently, NPEM is utilized to compute the statistical characteristics of the dynamic responses of the TTBCS. By comparing the effects of different random fields on the dynamic responses, the sensitivity of the responses to various material parameters is quantitatively analyzed. The results indicate that the vertical displacement of the bridge is most sensitive to the thickness uncertainty of the bottom plate, followed by the Young’s modulus. The lateral displacement is most sensitive to the thickness uncertainty of the traffic-side web, followed by the Young’s modulus. The density random field exerts the greatest influence on the bridge natural frequency. In the absence of seismic effects, uncertainties in the substructure have minimal influence on vehicle accelerations.</div></div>","PeriodicalId":51039,"journal":{"name":"Engineering Analysis with Boundary Elements","volume":"183 ","pages":"Article 106616"},"PeriodicalIF":4.1,"publicationDate":"2025-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145840846","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 : 2025-12-27DOI: 10.1016/j.enganabound.2025.106620
Wenhua Zheng , Jianqiang Du , Yanchen Zhu , Yang Yuan , Zhikai Zhang , Shaokun Lan
The rapid advancement of artificial intelligence (AI) has greatly accelerated the development of AI-driven solutions in the medical field. However, research on the application of AI in Chinese Medicine has lagged behind this trend. A critical area in Chinese Medicine is the verification of Chinese herbal formulas, which plays an essential role in pharmacy and in the development of automated formula dispensing systems. At present, there is a significant lack of studies exploring the use of AI for Chinese Herbal Formula Verification (CHFV). To address this gap, we conducted a comprehensive analysis of the challenges associated with CHFV and proposed an innovative CHFV model designed to improve the verification process of Chinese herbal formulas. To mitigate the problems of class imbalance and poor label quality in Chinese herbal medicine datasets, we introduced a personalized focusing mechanism (FM) grounded in both spatial simulations and real experiments. Furthermore, to overcome the challenges posed by limited inter-class variance and substantial intra-class variance in Chinese herbal medicine features, we developed a prior knowledge matrix and confidence reconstruction algorithms, which significantly enhanced the performance of our CHFV model. The source code and data are publicly available at https://github.com/Wenhua-Zheng/CHFV.
{"title":"CHFV model: Chinese herbal formula verification model","authors":"Wenhua Zheng , Jianqiang Du , Yanchen Zhu , Yang Yuan , Zhikai Zhang , Shaokun Lan","doi":"10.1016/j.enganabound.2025.106620","DOIUrl":"10.1016/j.enganabound.2025.106620","url":null,"abstract":"<div><div>The rapid advancement of artificial intelligence (AI) has greatly accelerated the development of AI-driven solutions in the medical field. However, research on the application of AI in Chinese Medicine has lagged behind this trend. A critical area in Chinese Medicine is the verification of Chinese herbal formulas, which plays an essential role in pharmacy and in the development of automated formula dispensing systems. At present, there is a significant lack of studies exploring the use of AI for Chinese Herbal Formula Verification (CHFV). To address this gap, we conducted a comprehensive analysis of the challenges associated with CHFV and proposed an innovative CHFV model designed to improve the verification process of Chinese herbal formulas. To mitigate the problems of class imbalance and poor label quality in Chinese herbal medicine datasets, we introduced a personalized focusing mechanism (FM) grounded in both spatial simulations and real experiments. Furthermore, to overcome the challenges posed by limited inter-class variance and substantial intra-class variance in Chinese herbal medicine features, we developed a prior knowledge matrix and confidence reconstruction algorithms, which significantly enhanced the performance of our CHFV model. The source code and data are publicly available at <span><span>https://github.com/Wenhua-Zheng/CHFV</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":51039,"journal":{"name":"Engineering Analysis with Boundary Elements","volume":"183 ","pages":"Article 106620"},"PeriodicalIF":4.1,"publicationDate":"2025-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145845084","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 : 2025-12-26DOI: 10.1016/j.enganabound.2025.106614
Peng Wang , Guiyong Zhang , Da Hui , Jinxin Wu
The Immersed Smoothed Point Interpolation Method (IS-PIM) is an effective technique for simulating fluid-structure interaction (FSI) problems involving structural motions and deformations. As a boundary non-conforming method, IS-PIM introduces a fictitious fluid domain to form a continuous fluid field, thereby circumventing computationally expensive mesh reconstruction. However, its reliance on non-coincident fluid and solid nodes at the interface leads to an inherent spatially non-matching node problem, which degrades the accuracy of FSI velocity boundary conditions and induces spurious numerical oscillations.
To address this fundamental issue, this study proposes a novel velocity reconstruction algorithm (VRA). The proposed VRA strategically reconstructs the near-boundary fluid velocity field by extrapolating from the kinematically consistent fictitious fluid velocity inside the solid domain and a set of intersection points of the fluid mesh with the solid boundary. These reconstructed velocities are then imposed as boundary conditions, leading to a better satisfaction of the no-slip and no-penetration conditions at the fluid-structure interface. The proposed VRA is implemented within the IS-PIM framework and validated through several numerical benchmarks. Comparative analyses demonstrate that the proposed VRA significantly enhances the accuracy of FSI velocity boundary conditions during structural motions and deformations. This improvement consequently suppresses spurious fluid flux across the solid boundary and refines the pressure solution, thereby enabling a more accurate evaluation of the FSI forces. Furthermore, the algorithm maintains high computational efficiency, introducing minimal overhead compared to the essential interpolation of the original IS-PIM.
{"title":"A velocity reconstruction algorithm for immersed smoothed point interpolation method in fluid-structure interaction problems","authors":"Peng Wang , Guiyong Zhang , Da Hui , Jinxin Wu","doi":"10.1016/j.enganabound.2025.106614","DOIUrl":"10.1016/j.enganabound.2025.106614","url":null,"abstract":"<div><div>The Immersed Smoothed Point Interpolation Method (IS-PIM) is an effective technique for simulating fluid-structure interaction (FSI) problems involving structural motions and deformations. As a boundary non-conforming method, IS-PIM introduces a fictitious fluid domain to form a continuous fluid field, thereby circumventing computationally expensive mesh reconstruction. However, its reliance on non-coincident fluid and solid nodes at the interface leads to an inherent spatially non-matching node problem, which degrades the accuracy of FSI velocity boundary conditions and induces spurious numerical oscillations.</div><div>To address this fundamental issue, this study proposes a novel velocity reconstruction algorithm (VRA). The proposed VRA strategically reconstructs the near-boundary fluid velocity field by extrapolating from the kinematically consistent fictitious fluid velocity inside the solid domain and a set of intersection points of the fluid mesh with the solid boundary. These reconstructed velocities are then imposed as boundary conditions, leading to a better satisfaction of the no-slip and no-penetration conditions at the fluid-structure interface. The proposed VRA is implemented within the IS-PIM framework and validated through several numerical benchmarks. Comparative analyses demonstrate that the proposed VRA significantly enhances the accuracy of FSI velocity boundary conditions during structural motions and deformations. This improvement consequently suppresses spurious fluid flux across the solid boundary and refines the pressure solution, thereby enabling a more accurate evaluation of the FSI forces. Furthermore, the algorithm maintains high computational efficiency, introducing minimal overhead compared to the essential interpolation of the original IS-PIM.</div></div>","PeriodicalId":51039,"journal":{"name":"Engineering Analysis with Boundary Elements","volume":"183 ","pages":"Article 106614"},"PeriodicalIF":4.1,"publicationDate":"2025-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145840845","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 : 2025-12-24DOI: 10.1016/j.enganabound.2025.106607
Chong Zhang , Jianming Zhang , Rongxiong Xiao
In computer-aided engineering (CAE), coupled non-matching meshes are indispensable for disposing scenarios such as mesh deformations arising from nonlinear problems, domain decomposition to mitigate meshing complexity and computational scale, and gluing of non-matching interfaces across multiphysics simulations. However, the generation of high-precision conformal mesh remains challenging due to geometric and topological constraints. This paper proposes a systematic novelty automated trimming method (ATM) for non-matching surface meshes, which is independent of mesh shape (arbitrary polygons) and continuity (discontinuous or continuous). The proposed methodology extends the boundary representation (B-Rep) framework to interface trimming, constructing a topological structure of multiple “faces” to enable precise interface partitioning. The triangulation of the boundary mesh in the overlapping region of the interface is implemented according to the circumscribed circle criterion. A straightforward optimization scheme is subsequently applied to triangular cells to generate high-quality interface meshes. Compared to existing algorithms, ATM fully utilizes interface topology information, enforces geometric constraints during boundary recovery via Delaunay triangulation, and employs cell vertex relocation strategy to ensure trimmed mesh quality. Numerical examples demonstrate that ATM achieves automated, efficient, and robust trimming of arbitrary non-matching interfaces.
{"title":"An automatic trimming method for arbitrary non-matching surface meshes","authors":"Chong Zhang , Jianming Zhang , Rongxiong Xiao","doi":"10.1016/j.enganabound.2025.106607","DOIUrl":"10.1016/j.enganabound.2025.106607","url":null,"abstract":"<div><div>In computer-aided engineering (CAE), coupled non-matching meshes are indispensable for disposing scenarios such as mesh deformations arising from nonlinear problems, domain decomposition to mitigate meshing complexity and computational scale, and gluing of non-matching interfaces across multiphysics simulations. However, the generation of high-precision conformal mesh remains challenging due to geometric and topological constraints. This paper proposes a systematic novelty automated trimming method (ATM) for non-matching surface meshes, which is independent of mesh shape (arbitrary polygons) and continuity (discontinuous or continuous). The proposed methodology extends the boundary representation (B-Rep) framework to interface trimming, constructing a topological structure of multiple “faces” to enable precise interface partitioning. The triangulation of the boundary mesh in the overlapping region of the interface is implemented according to the circumscribed circle criterion. A straightforward optimization scheme is subsequently applied to triangular cells to generate high-quality interface meshes. Compared to existing algorithms, ATM fully utilizes interface topology information, enforces geometric constraints during boundary recovery via Delaunay triangulation, and employs cell vertex relocation strategy to ensure trimmed mesh quality. Numerical examples demonstrate that ATM achieves automated, efficient, and robust trimming of arbitrary non-matching interfaces.</div></div>","PeriodicalId":51039,"journal":{"name":"Engineering Analysis with Boundary Elements","volume":"183 ","pages":"Article 106607"},"PeriodicalIF":4.1,"publicationDate":"2025-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145822957","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