The accurate mechanical analysis of thick-walled pressure vessel structures composed of advanced materials,such as hyperelastic and functionally graded materials(FGMs),is critical for ensuring their safety and optimiz...The accurate mechanical analysis of thick-walled pressure vessel structures composed of advanced materials,such as hyperelastic and functionally graded materials(FGMs),is critical for ensuring their safety and optimizing their design.However,conventional numerical methods can face challenges with the non-linearities inherent in hyperelasticity and the complex spatial variations in FGMs.This paper presents a novel hybrid numerical approach combining Physics-Informed Neural Networks(PINNs)with Finite Element Method(FEM)derived data for the robust analysis of thick-walled,axisymmetric,heterogeneous,hyperelastic pressure vessels with elliptical geometries.A PINN framework incorporating neo-Hookean constitutive relations is developed in MATLAB.To enhance training efficiency and accuracy,the PINN’s loss function is augmented with displacement data obtained from high-fidelity FEM simulations performed in ANSYS.The methodology is rigorously validated by comparing PINN-predicted displacement and von Mises stress fields against ANSYS benchmarks for various scenarios of FGMconfigurations(with material properties varying according to a power law)subjected to internal and external pressurization.The results demonstrate excellent agreement between the proposed hybrid PINN-FEMapproach and conventional FEMsolutions across all test cases,accurately capturing complex deformation patterns and stress concentrations.This study highlights the potential of data-augmented PINNs as an effective and accurate computational tool for tackling complex solid mechanics problems involving non-linearmaterials and significant heterogeneity,offering a promising avenue for future research in engineering design and analysis.展开更多
In modern ZnO varistors,traditional aging mechanisms based on increased power consumption are no longer relevant due to reduced power consumption during DC aging.Prolonged exposure to both AC and DC voltages results i...In modern ZnO varistors,traditional aging mechanisms based on increased power consumption are no longer relevant due to reduced power consumption during DC aging.Prolonged exposure to both AC and DC voltages results in increased leakage current,decreased breakdown voltage,and lower nonlinearity,ultimately compromising their protective performance.To investigate the evolution in electrical properties during DC aging,this work developed a finite element model based on Voronoi networks and conducted accelerated aging tests on commercial varistors.Throughout the aging process,current-voltage characteristics and Schottky barrier parameters were measured and analyzed.The results indicate that when subjected to constant voltage,current flows through regions with larger grain sizes,forming discharge channels.As aging progresses,the current focus increases on these channels,leading to a decline in the varistor’s overall performance.Furthermore,analysis of the Schottky barrier parameters shows that the changes in electrical performance during aging are non-monotonic.These findings offer theoretical support for understanding the aging mechanisms and condition assessment of modern stable ZnO varistors.展开更多
This study develops a surrogate super-resolution(SR)framework that accelerates finite element method(FEM)-based computational fluid dynamics(CFD)using deep learning.High-resolution(HR)FEM-based CFDremains computationa...This study develops a surrogate super-resolution(SR)framework that accelerates finite element method(FEM)-based computational fluid dynamics(CFD)using deep learning.High-resolution(HR)FEM-based CFDremains computationally prohibitive for time-sensitive applications,including patient-specific aneurysm hemodynamics where rapid turnaround is valuable.The proposed pipeline learns to reconstruct HR velocity-magnitude fields fromlow-resolution(LR)FEM solutions generated under the same governing equations and boundary conditions.It consistsof three modules:(i)offline pre-training of a residual network on representative vascular geometries;(ii)lightweightfine-tuning to adapt the pretrained model to geometric variability,including patient-specific aneurysm morphologies;and(iii)an unstructured-to-structured sampling strategy with region-of-interest upsampling that concentrates resolution in flow-critical zones(e.g.,the aneurysm sac)rather than the full domain.This targeted reconstruction substantiallyreduces inference and post-processing cost while preserving key HR flow features.Experiments on cerebral aneurysmmodels show that HR velocity-magnitude fields can be recovered with accuracy comparable to direct HR simulationsat less than 1%of the direct HR simulation cost per analysis(LR simulation and SR inference),while adaptation to newgeometries requires only lightweight fine-tuning with limited target-specific HR data.While clinical endpoints andadditional variables(e.g.,pressure or wall-based metrics)are left for future work,the results indicate that the proposedsurrogate SR approach can streamline FEM-based CFD workflows toward near real-time hemodynamic analysis acrossmorphologically similar vascular models.展开更多
A highly efficient H1-Galerkin mixed finite element method （MFEM） is presented with linear triangular element for the parabolic integro-differential equation. Firstly, some new results about the integral estimation ...A highly efficient H1-Galerkin mixed finite element method （MFEM） is presented with linear triangular element for the parabolic integro-differential equation. Firstly, some new results about the integral estimation and asymptotic expansions are studied. Then, the superconvergence of order O（h^2） for both the original variable u in H1 （Ω） norm and the flux p = u in H（div, Ω） norm is derived through the interpolation post processing technique. Furthermore, with the help of the asymptotic expansions and a suitable auxiliary problem, the extrapolation solutions with accuracy O（h^3） are obtained for the above two variables. Finally, some numerical results are provided to confirm validity of the theoretical analysis and excellent performance of the proposed method.展开更多
高位滑坡对建筑集群的冲击破坏时常导致严重的人员伤亡,基于光滑粒子流体动力学-离散元法-有限元法(smoothed particle hydrodynamics-discrete element method-finite element method,SPH-DEM-FEM)耦合的数值模型,开展了高位滑坡对框...高位滑坡对建筑集群的冲击破坏时常导致严重的人员伤亡,基于光滑粒子流体动力学-离散元法-有限元法(smoothed particle hydrodynamics-discrete element method-finite element method,SPH-DEM-FEM)耦合的数值模型,开展了高位滑坡对框架结构建筑群的冲击过程、建筑结构破坏机理、冲击力时程与框架柱关键点应力和弯矩等动力机制研究。研究结果表明:SPH-DEM-FEM耦合数值方法能够有效地模拟碎石土滑坡中土(SPH)石(DEM)混合物的抛射弹跳、爬高绕流冲击运动过程。考虑了常规建筑垂直、平行于滑坡流向的三排建筑组合布局,位于滑坡近端的纵向排列建筑表现为连续性倾倒破坏,横向排列的建筑则呈现整体倾倒破坏;因前排建筑群对滑坡冲击能量的耗散及滑坡自身摩擦耗能,位于滑坡后端建筑表现为引流面墙体和前排柱发生局部破坏,结构保持稳定,损毁程度依次为上游无建筑缓冲耗能的建筑>有横向排列的建筑>有纵向排列的建筑;纵向、横向排列的建筑冲击力衰减幅度分别31%、21%。横向框架建筑整体倾倒的损毁机制表现为框架柱的直接剪断或节点塑形铰链失效;纵向框架建筑连续性倾倒的损毁机制表现为前排框架柱的失效引起后排框架柱轴向压力和极限弯矩增加,持续冲击荷载超过其极限弯矩致使后排框架柱发生弯曲破坏,最终结构倾倒。系统能量在动能、内能和摩擦耗能间转化,其中摩擦耗能占65.5%,结构耗能占23.6%,动能快速下降与内能急剧增加是建筑破坏的关键特征。展开更多
This article presents a micro-structure tensor enhanced elasto-plastic finite element(FE)method to address strength anisotropy in three-dimensional(3D)soil slope stability analysis.The gravity increase method(GIM)is e...This article presents a micro-structure tensor enhanced elasto-plastic finite element(FE)method to address strength anisotropy in three-dimensional(3D)soil slope stability analysis.The gravity increase method(GIM)is employed to analyze the stability of 3D anisotropic soil slopes.The accuracy of the proposed method is first verified against the data in the literature.We then simulate the 3D soil slope with a straight slope surface and the convex and concave slope surfaces with a 90turning corner to study the 3D effect on slope stability and the failure mechanism under anisotropy conditions.Based on our numerical results,the end effect significantly impacts the failure mechanism and safety factor.Anisotropy degree notably affects the safety factor,with higher degrees leading to deeper landslides.For concave slopes,they can be approximated by straight slopes with suitable boundary conditions to assess their stability.Furthermore,a case study of the Saint-Alban test embankment A in Quebec,Canada,is provided to demonstrate the applicability of the proposed FE model.展开更多
Fatigue crack growth is a critical phenomenon in engineering structures,accounting for a significant percentage of structural failures across various industries.Accurate prediction of crack initiation,propagation path...Fatigue crack growth is a critical phenomenon in engineering structures,accounting for a significant percentage of structural failures across various industries.Accurate prediction of crack initiation,propagation paths,and fatigue life is essential for ensuring structural integrity and optimizing maintenance schedules.This paper presents a comprehensive finite element approach for simulating two-dimensional fatigue crack growth under linear elastic conditionswith adaptivemesh generation.The source code for the programwas developed in Fortran 95 and compiled with Visual Fortran.To achieve high-fidelity simulations,the methodology integrates several key features:it employs an automatic,adaptive meshing technique that selectively refines the element density near the crack front and areas of significant stress concentration.Specialized singular elements are used at the crack tip to ensure precise stress field representation.The direction of crack advancement is predicted using the maximum tangential stress criterion,while stress intensity factors are determined through either the displacement extrapolation technique or the J-integral method.The simulation models crack growth as a series of linear increments,with solution stability maintained by a consistent transfer algorithm and a crack relaxation method.The framework’s effectiveness is demonstrated across various geometries and loading scenarios.Through rigorous validation against both experimental data and established numerical benchmarks,the approach is proven to accurately forecast crack trajectories and fatigue life.Furthermore,the detailed description of the program’s architecture offers a foundational blueprint,serving as a valuable guide for researchers aiming to develop their specialized software for fracture mechanics analysis.展开更多
Curved beams with complex geometries are vital in numerous engineering applications,where precise vibration analysis is crucial for ensuring safe and effective designs.Traditional finite element methods(FEMs) often st...Curved beams with complex geometries are vital in numerous engineering applications,where precise vibration analysis is crucial for ensuring safe and effective designs.Traditional finite element methods(FEMs) often struggle to accurately represent the dynamic characteristics of these structures due to the limitations in their shape function approximations.To overcome this challenge,the current study introduces an innovative finite element(FE)-based technique for the undamped vibrational analysis of curved beams with arbitrary curvature,employing explicitly derived interpolation functions.Initially,the exact interpolation functions are developed for circular are elements with the force method.These functions facilitate the creation of a highly accurate stiffness matrix,which is validated against the benchmark examples.To accommodate arbitrary curvature,a systematic transformation technique is established to approximate the intricate curves with a series of circular arcs.The numerical findings indicate that increasing the number of arc segments enhances accuracy,approaching the exact solutions.The analysis of free vibrations is conducted for both circular and non-circular beams.Mass matrices are derived using two methods:lumped mass and consistent mass,where the latter is based on the interpolation functions.The effectiveness of the proposed method is confirmed through the comparisons with the existing literature,demonstrating strong agreement.Finally,several practical cases involving beams with diverse curvature profiles are analyzed.Both natural frequencies and mode shapes are determined,providing significant insights into the dynamic behavior of these structures.This research offers a dependable and efficient analytical framework for the vibrational analysis of complex curved beams,with promising implications for structural and mechanical engineering.展开更多
Physics-informed neural networks(PINNs)have prevailed as differentiable simulators to investigate flow in porous media.Despite recent progress PINNs have achieved,practical geotechnical scenarios cannot be readily sim...Physics-informed neural networks(PINNs)have prevailed as differentiable simulators to investigate flow in porous media.Despite recent progress PINNs have achieved,practical geotechnical scenarios cannot be readily simulated because conventional PINNs fail in discontinuous heterogeneous porous media or multi-layer strata when labeled data are missing.This work aims to develop a universal network structure to encode the mass continuity equation and Darcy’s law without labeled data.The finite element approximation,which can decompose a complex heterogeneous domain into simpler ones,is adopted to build the differentiable network.Without conventional DNNs,physics-encoded finite element network(PEFEN)can avoid spectral bias and learn high-frequency functions with sharp/steep gradients.PEFEN rigorously encodes Dirichlet and Neumann boundary conditions without training.Benefiting from its discretized formulation,the discontinuous heterogeneous hydraulic conductivity is readily embedded into the network.Three typical cases are reproduced to corroborate PEFEN’s superior performance over conventional PINNs and the PINN with mixed formulation.PEFEN is sparse and demonstrated to be capable of dealing with heterogeneity with much fewer training iterations(less than 1/30)than the improved PINN with mixed formulation.Thus,PEFEN saves energy and contributes to low-carbon AI for science.The last two cases focus on common geotechnical settings of impermeable sheet pile in singlelayer and multi-layer strata.PEFEN solves these cases with high accuracy,circumventing costly labeled data,extra computational burden,and additional treatment.Thus,this study warrants the further development and application of PEFEN as a novel differentiable network in porous flow of practical geotechnical engineering.展开更多
In the incremental sheet forming （ISF） process, springback is a very important factor that affects the quality of parts. Predicting and controlling springback accurately is essential for the design of the toolpath f...In the incremental sheet forming （ISF） process, springback is a very important factor that affects the quality of parts. Predicting and controlling springback accurately is essential for the design of the toolpath for ISF. A three-dimensional elasto-plastic finite element model （FEM） was developed to simulate the process and the simulated results were compared with those from the experiment. The springback angle was found to be in accordance with the experimental result, proving the FEM to be effective. A coupled artificial neural networks （ANN） and finite element method technique was developed to simulate and predict springback responses to changes in the processing parameters. A particle swarm optimization （PSO） algorithm was used to optimize the weights and thresholds of the neural network model. The neural network was trained using available FEM simulation data. The results showed that a more accurate prediction of s!oringback can be acquired using the FEM-PSONN model.展开更多
文摘The accurate mechanical analysis of thick-walled pressure vessel structures composed of advanced materials,such as hyperelastic and functionally graded materials(FGMs),is critical for ensuring their safety and optimizing their design.However,conventional numerical methods can face challenges with the non-linearities inherent in hyperelasticity and the complex spatial variations in FGMs.This paper presents a novel hybrid numerical approach combining Physics-Informed Neural Networks(PINNs)with Finite Element Method(FEM)derived data for the robust analysis of thick-walled,axisymmetric,heterogeneous,hyperelastic pressure vessels with elliptical geometries.A PINN framework incorporating neo-Hookean constitutive relations is developed in MATLAB.To enhance training efficiency and accuracy,the PINN’s loss function is augmented with displacement data obtained from high-fidelity FEM simulations performed in ANSYS.The methodology is rigorously validated by comparing PINN-predicted displacement and von Mises stress fields against ANSYS benchmarks for various scenarios of FGMconfigurations(with material properties varying according to a power law)subjected to internal and external pressurization.The results demonstrate excellent agreement between the proposed hybrid PINN-FEMapproach and conventional FEMsolutions across all test cases,accurately capturing complex deformation patterns and stress concentrations.This study highlights the potential of data-augmented PINNs as an effective and accurate computational tool for tackling complex solid mechanics problems involving non-linearmaterials and significant heterogeneity,offering a promising avenue for future research in engineering design and analysis.
文摘In modern ZnO varistors,traditional aging mechanisms based on increased power consumption are no longer relevant due to reduced power consumption during DC aging.Prolonged exposure to both AC and DC voltages results in increased leakage current,decreased breakdown voltage,and lower nonlinearity,ultimately compromising their protective performance.To investigate the evolution in electrical properties during DC aging,this work developed a finite element model based on Voronoi networks and conducted accelerated aging tests on commercial varistors.Throughout the aging process,current-voltage characteristics and Schottky barrier parameters were measured and analyzed.The results indicate that when subjected to constant voltage,current flows through regions with larger grain sizes,forming discharge channels.As aging progresses,the current focus increases on these channels,leading to a decline in the varistor’s overall performance.Furthermore,analysis of the Schottky barrier parameters shows that the changes in electrical performance during aging are non-monotonic.These findings offer theoretical support for understanding the aging mechanisms and condition assessment of modern stable ZnO varistors.
文摘This study develops a surrogate super-resolution(SR)framework that accelerates finite element method(FEM)-based computational fluid dynamics(CFD)using deep learning.High-resolution(HR)FEM-based CFDremains computationally prohibitive for time-sensitive applications,including patient-specific aneurysm hemodynamics where rapid turnaround is valuable.The proposed pipeline learns to reconstruct HR velocity-magnitude fields fromlow-resolution(LR)FEM solutions generated under the same governing equations and boundary conditions.It consistsof three modules:(i)offline pre-training of a residual network on representative vascular geometries;(ii)lightweightfine-tuning to adapt the pretrained model to geometric variability,including patient-specific aneurysm morphologies;and(iii)an unstructured-to-structured sampling strategy with region-of-interest upsampling that concentrates resolution in flow-critical zones(e.g.,the aneurysm sac)rather than the full domain.This targeted reconstruction substantiallyreduces inference and post-processing cost while preserving key HR flow features.Experiments on cerebral aneurysmmodels show that HR velocity-magnitude fields can be recovered with accuracy comparable to direct HR simulationsat less than 1%of the direct HR simulation cost per analysis(LR simulation and SR inference),while adaptation to newgeometries requires only lightweight fine-tuning with limited target-specific HR data.While clinical endpoints andadditional variables(e.g.,pressure or wall-based metrics)are left for future work,the results indicate that the proposedsurrogate SR approach can streamline FEM-based CFD workflows toward near real-time hemodynamic analysis acrossmorphologically similar vascular models.
文摘泥石流作为高破坏性混合流体,其携带的块石对框架结构的冲击机制尚未得到充分研究。为揭示含块石泥石流冲击框架结构的动力响应与损伤机制,基于光滑粒子流体动力学-离散元法-有限元法(smoothed particle hydrodynamics-discrete element method-finite element method,SPH-DEM-FEM)耦合数值方法,构建流体-块石-结构的多耦合数值模型,模拟不同冲击速度和角度下框架结构的损伤过程,并结合两相溃坝试验来验证模型有效性。研究表明:泥石流冲击导致结构损伤经历“接触-扩散-反弹-堆积/冲击”四个阶段,当冲击速度超过6.00 m/s时,结构产生不可恢复损伤,且块石阻隔作用使流体上部冲击力大于下部,引发结构中部最先发生集中损伤;此外,冲击力峰值随速度和角度增大呈非线性增长,10.00 m/s与90°工况下框架柱底冲击力达497.17 kN,超过结构抗冲击承载力;最后,数值模拟与经验公式所得冲击力结果误差在13.95%~29.00%,数量级一致,验证了模型可靠性。研究结果可为泥石流高发区框架结构的抗冲击设计提供参考。
基金Project supported by the National Natural Science Foundation of China(Nos.10971203,11271340,and 11101381)the Specialized Research Fund for the Doctoral Program of Higher Education(No.20094101110006)
文摘A highly efficient H1-Galerkin mixed finite element method （MFEM） is presented with linear triangular element for the parabolic integro-differential equation. Firstly, some new results about the integral estimation and asymptotic expansions are studied. Then, the superconvergence of order O（h^2） for both the original variable u in H1 （Ω） norm and the flux p = u in H（div, Ω） norm is derived through the interpolation post processing technique. Furthermore, with the help of the asymptotic expansions and a suitable auxiliary problem, the extrapolation solutions with accuracy O（h^3） are obtained for the above two variables. Finally, some numerical results are provided to confirm validity of the theoretical analysis and excellent performance of the proposed method.
文摘高位滑坡对建筑集群的冲击破坏时常导致严重的人员伤亡,基于光滑粒子流体动力学-离散元法-有限元法(smoothed particle hydrodynamics-discrete element method-finite element method,SPH-DEM-FEM)耦合的数值模型,开展了高位滑坡对框架结构建筑群的冲击过程、建筑结构破坏机理、冲击力时程与框架柱关键点应力和弯矩等动力机制研究。研究结果表明:SPH-DEM-FEM耦合数值方法能够有效地模拟碎石土滑坡中土(SPH)石(DEM)混合物的抛射弹跳、爬高绕流冲击运动过程。考虑了常规建筑垂直、平行于滑坡流向的三排建筑组合布局,位于滑坡近端的纵向排列建筑表现为连续性倾倒破坏,横向排列的建筑则呈现整体倾倒破坏;因前排建筑群对滑坡冲击能量的耗散及滑坡自身摩擦耗能,位于滑坡后端建筑表现为引流面墙体和前排柱发生局部破坏,结构保持稳定,损毁程度依次为上游无建筑缓冲耗能的建筑>有横向排列的建筑>有纵向排列的建筑;纵向、横向排列的建筑冲击力衰减幅度分别31%、21%。横向框架建筑整体倾倒的损毁机制表现为框架柱的直接剪断或节点塑形铰链失效;纵向框架建筑连续性倾倒的损毁机制表现为前排框架柱的失效引起后排框架柱轴向压力和极限弯矩增加,持续冲击荷载超过其极限弯矩致使后排框架柱发生弯曲破坏,最终结构倾倒。系统能量在动能、内能和摩擦耗能间转化,其中摩擦耗能占65.5%,结构耗能占23.6%,动能快速下降与内能急剧增加是建筑破坏的关键特征。
基金supported by the National Natural Science Foundation of China(Grant Nos.51890912,51979025 and 52011530189).
文摘This article presents a micro-structure tensor enhanced elasto-plastic finite element(FE)method to address strength anisotropy in three-dimensional(3D)soil slope stability analysis.The gravity increase method(GIM)is employed to analyze the stability of 3D anisotropic soil slopes.The accuracy of the proposed method is first verified against the data in the literature.We then simulate the 3D soil slope with a straight slope surface and the convex and concave slope surfaces with a 90turning corner to study the 3D effect on slope stability and the failure mechanism under anisotropy conditions.Based on our numerical results,the end effect significantly impacts the failure mechanism and safety factor.Anisotropy degree notably affects the safety factor,with higher degrees leading to deeper landslides.For concave slopes,they can be approximated by straight slopes with suitable boundary conditions to assess their stability.Furthermore,a case study of the Saint-Alban test embankment A in Quebec,Canada,is provided to demonstrate the applicability of the proposed FE model.
基金funding of the Deanship of Graduate Studies and Scientific Research,Jazan University,Saudi Arabia,through Project number:JU-20250230-DGSSR-RP-2025.
文摘Fatigue crack growth is a critical phenomenon in engineering structures,accounting for a significant percentage of structural failures across various industries.Accurate prediction of crack initiation,propagation paths,and fatigue life is essential for ensuring structural integrity and optimizing maintenance schedules.This paper presents a comprehensive finite element approach for simulating two-dimensional fatigue crack growth under linear elastic conditionswith adaptivemesh generation.The source code for the programwas developed in Fortran 95 and compiled with Visual Fortran.To achieve high-fidelity simulations,the methodology integrates several key features:it employs an automatic,adaptive meshing technique that selectively refines the element density near the crack front and areas of significant stress concentration.Specialized singular elements are used at the crack tip to ensure precise stress field representation.The direction of crack advancement is predicted using the maximum tangential stress criterion,while stress intensity factors are determined through either the displacement extrapolation technique or the J-integral method.The simulation models crack growth as a series of linear increments,with solution stability maintained by a consistent transfer algorithm and a crack relaxation method.The framework’s effectiveness is demonstrated across various geometries and loading scenarios.Through rigorous validation against both experimental data and established numerical benchmarks,the approach is proven to accurately forecast crack trajectories and fatigue life.Furthermore,the detailed description of the program’s architecture offers a foundational blueprint,serving as a valuable guide for researchers aiming to develop their specialized software for fracture mechanics analysis.
文摘Curved beams with complex geometries are vital in numerous engineering applications,where precise vibration analysis is crucial for ensuring safe and effective designs.Traditional finite element methods(FEMs) often struggle to accurately represent the dynamic characteristics of these structures due to the limitations in their shape function approximations.To overcome this challenge,the current study introduces an innovative finite element(FE)-based technique for the undamped vibrational analysis of curved beams with arbitrary curvature,employing explicitly derived interpolation functions.Initially,the exact interpolation functions are developed for circular are elements with the force method.These functions facilitate the creation of a highly accurate stiffness matrix,which is validated against the benchmark examples.To accommodate arbitrary curvature,a systematic transformation technique is established to approximate the intricate curves with a series of circular arcs.The numerical findings indicate that increasing the number of arc segments enhances accuracy,approaching the exact solutions.The analysis of free vibrations is conducted for both circular and non-circular beams.Mass matrices are derived using two methods:lumped mass and consistent mass,where the latter is based on the interpolation functions.The effectiveness of the proposed method is confirmed through the comparisons with the existing literature,demonstrating strong agreement.Finally,several practical cases involving beams with diverse curvature profiles are analyzed.Both natural frequencies and mode shapes are determined,providing significant insights into the dynamic behavior of these structures.This research offers a dependable and efficient analytical framework for the vibrational analysis of complex curved beams,with promising implications for structural and mechanical engineering.
基金supported by the National Natural Science Foundation of China(Grant Nos.42272338 and 41827807)Department of Transportation of Zhejiang Province,China(Grant No.202213).
文摘Physics-informed neural networks(PINNs)have prevailed as differentiable simulators to investigate flow in porous media.Despite recent progress PINNs have achieved,practical geotechnical scenarios cannot be readily simulated because conventional PINNs fail in discontinuous heterogeneous porous media or multi-layer strata when labeled data are missing.This work aims to develop a universal network structure to encode the mass continuity equation and Darcy’s law without labeled data.The finite element approximation,which can decompose a complex heterogeneous domain into simpler ones,is adopted to build the differentiable network.Without conventional DNNs,physics-encoded finite element network(PEFEN)can avoid spectral bias and learn high-frequency functions with sharp/steep gradients.PEFEN rigorously encodes Dirichlet and Neumann boundary conditions without training.Benefiting from its discretized formulation,the discontinuous heterogeneous hydraulic conductivity is readily embedded into the network.Three typical cases are reproduced to corroborate PEFEN’s superior performance over conventional PINNs and the PINN with mixed formulation.PEFEN is sparse and demonstrated to be capable of dealing with heterogeneity with much fewer training iterations(less than 1/30)than the improved PINN with mixed formulation.Thus,PEFEN saves energy and contributes to low-carbon AI for science.The last two cases focus on common geotechnical settings of impermeable sheet pile in singlelayer and multi-layer strata.PEFEN solves these cases with high accuracy,circumventing costly labeled data,extra computational burden,and additional treatment.Thus,this study warrants the further development and application of PEFEN as a novel differentiable network in porous flow of practical geotechnical engineering.
基金Project(50175034) supported by the National Natural Science Foundation of China
文摘In the incremental sheet forming （ISF） process, springback is a very important factor that affects the quality of parts. Predicting and controlling springback accurately is essential for the design of the toolpath for ISF. A three-dimensional elasto-plastic finite element model （FEM） was developed to simulate the process and the simulated results were compared with those from the experiment. The springback angle was found to be in accordance with the experimental result, proving the FEM to be effective. A coupled artificial neural networks （ANN） and finite element method technique was developed to simulate and predict springback responses to changes in the processing parameters. A particle swarm optimization （PSO） algorithm was used to optimize the weights and thresholds of the neural network model. The neural network was trained using available FEM simulation data. The results showed that a more accurate prediction of s!oringback can be acquired using the FEM-PSONN model.
