Advances in Engineering Software最新文献_第6页

Influence of groundwater on seismic response of nuclear power plant soil-structure system 地下水对核电厂土-结构体系地震反应的影响

IF 5.7 2区工程技术 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Advances in Engineering Software

Pub Date : 2025-09-13 DOI: 10.1016/j.advengsoft.2025.104023

Hao Lv

The construction of coastal nuclear power plants (NPPs) on lithologically robust foundations is geographically limited, driving a shift toward inland non-rock sites. Ensuring seismic resilience of such sites has become critical for nuclear safety. Near coasts or rivers, groundwater table (GWT) fluctuations significantly influence soil-pore water distribution, thereby affecting soil seismic response and NPP performance. To analyze the influence of groundwater table on the seismic response of the nuclear power plant, this paper uses the saturated porous medium model and considers the interaction of the saturated soil and structure. The free field of the horizontally layered site of dry soil-saturated soil is obtained by the transfer matrix method, and combined with the transmission boundary, the wave input of soil-structure interaction (SSI) analysis is realized. Then, the partitioned parallel calculation method of SSI is used to analyse the saturated SSI. The soil, along with its groundwater, is characterized using the generalized saturated porous medium model. The simulation of the combined lumped-mass explicit finite element and transmission boundary is accomplished through a self-programmed FORTRAN code. On the other hand, the structural analysis is carried out using ANSYS, employing an implicit finite element approach. Taking a nuclear power plant as an example, the dynamic response of the soil-foundation-nuclear power plant system is analysed at five sites with different GWTs. In this case, the goal is an attempt to determine the effect of the depth of the GWT on the soil-foundation-nuclear power plant system under seismic action.

沿海核电站（NPPs）在岩石坚固的基础上的建设在地理上是有限的，这推动了向内陆非岩石地点的转变。确保这些场址的抗震能力已成为核安全的关键。在沿海或河流附近，地下水位（GWT）的波动会显著影响土壤孔隙水的分布，从而影响土壤的地震反应和核电厂的性能。为了分析地下水位对核电站地震反应的影响，本文采用饱和多孔介质模型，考虑饱和土与结构的相互作用。采用传递矩阵法获得干土-饱和土水平层状场地的自由场，并结合透射边界，实现土-结构相互作用（SSI）分析的波输入。然后，采用SSI的分区并行计算方法对饱和SSI进行了分析。采用广义饱和多孔介质模型对土壤及其地下水进行了表征。通过自编的FORTRAN代码实现了集总质量显式有限元和传输边界的组合仿真。另一方面，利用ANSYS进行结构分析，采用隐式有限元方法。以某核电站为例，分析了5个不同gwt场址的地基-基础-核电站系统的动力响应。在这种情况下，目标是试图确定地震作用下GWT深度对地基-基础-核电站系统的影响。

{"title":"Influence of groundwater on seismic response of nuclear power plant soil-structure system","authors":"Hao Lv","doi":"10.1016/j.advengsoft.2025.104023","DOIUrl":"10.1016/j.advengsoft.2025.104023","url":null,"abstract":"<div><div>The construction of coastal nuclear power plants (NPPs) on lithologically robust foundations is geographically limited, driving a shift toward inland non-rock sites. Ensuring seismic resilience of such sites has become critical for nuclear safety. Near coasts or rivers, groundwater table (GWT) fluctuations significantly influence soil-pore water distribution, thereby affecting soil seismic response and NPP performance. To analyze the influence of groundwater table on the seismic response of the nuclear power plant, this paper uses the saturated porous medium model and considers the interaction of the saturated soil and structure. The free field of the horizontally layered site of dry soil-saturated soil is obtained by the transfer matrix method, and combined with the transmission boundary, the wave input of soil-structure interaction (SSI) analysis is realized. Then, the partitioned parallel calculation method of SSI is used to analyse the saturated SSI. The soil, along with its groundwater, is characterized using the generalized saturated porous medium model. The simulation of the combined lumped-mass explicit finite element and transmission boundary is accomplished through a self-programmed FORTRAN code. On the other hand, the structural analysis is carried out using ANSYS, employing an implicit finite element approach. Taking a nuclear power plant as an example, the dynamic response of the soil-foundation-nuclear power plant system is analysed at five sites with different GWTs. In this case, the goal is an attempt to determine the effect of the depth of the GWT on the soil-foundation-nuclear power plant system under seismic action.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"211 ","pages":"Article 104023"},"PeriodicalIF":5.7,"publicationDate":"2025-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145049822","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}

引用次数: 0

Mooring tension estimation for multi-connected floating photovoltaic arrays via LSTM networks 基于LSTM网络的多连接浮式光伏阵列系泊张力估计

IF 5.7 2区工程技术 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Advances in Engineering Software

Pub Date : 2025-09-13 DOI: 10.1016/j.advengsoft.2025.104037

Jihun Song , Chungkuk Jin , Do Kyun Kim , Donghwi Jung , Seungjun Kim

Accurately estimating mooring‑line tension is essential for the safe operation of large, multiconnected floating‑photovoltaic (FPV) arrays, yet installing load cells on every line is impractical. This study develops and evaluates data‑driven tension estimators that use only motion responses generated from time‑domain hydrodynamic simulations. A long short‑term memory (LSTM) network trained on displacements provides the reference performance. When trained instead on raw accelerations, the model performs noticeably worse, reflecting the spectral mismatch between acceleration and tension signals. Adding directional spreading to the training data restores robustness for the displacement‑based model under oblique seas, but offers limited benefit for the acceleration‑based model. In this study, a physics‑guided LSTM is proposed to reduce reliance on displacement sensors, in which a learnable filter transforms accelerations into displacement‑like features. This hybrid model narrows the performance gap, achieving stable and robust prediction performance. The proposed model attains accuracy comparable to displacement‑based estimation, demonstrating its effectiveness with accelerometer input alone and highlighting its potential as a cost‑efficient tool for structural health monitoring of large‑scale FPV systems.

准确估计系泊线张力对于大型多连接浮动光伏（FPV）阵列的安全运行至关重要，然而在每条线上安装称重传感器是不切实际的。本研究开发和评估仅使用时域流体动力学模拟产生的运动响应的数据驱动张力估计器。基于位移训练的长短期记忆（LSTM）网络提供了参考性能。当在原始加速度上进行训练时，模型的表现明显更差，反映了加速度和张力信号之间的频谱不匹配。在训练数据中加入方向扩展可以恢复斜海下基于位移的模型的鲁棒性，但对基于加速度的模型的好处有限。在这项研究中，提出了一个物理引导的LSTM来减少对位移传感器的依赖，其中一个可学习的滤波器将加速度转换为类似位移的特征。该混合模型缩小了性能差距，实现了稳定、稳健的预测性能。所提出的模型达到了与基于位移的估计相当的精度，证明了其单独使用加速度计输入的有效性，并突出了其作为大型FPV系统结构健康监测的成本效益工具的潜力。

{"title":"Mooring tension estimation for multi-connected floating photovoltaic arrays via LSTM networks","authors":"Jihun Song , Chungkuk Jin , Do Kyun Kim , Donghwi Jung , Seungjun Kim","doi":"10.1016/j.advengsoft.2025.104037","DOIUrl":"10.1016/j.advengsoft.2025.104037","url":null,"abstract":"<div><div>Accurately estimating mooring‑line tension is essential for the safe operation of large, multiconnected floating‑photovoltaic (FPV) arrays, yet installing load cells on every line is impractical. This study develops and evaluates data‑driven tension estimators that use only motion responses generated from time‑domain hydrodynamic simulations. A long short‑term memory (LSTM) network trained on displacements provides the reference performance. When trained instead on raw accelerations, the model performs noticeably worse, reflecting the spectral mismatch between acceleration and tension signals. Adding directional spreading to the training data restores robustness for the displacement‑based model under oblique seas, but offers limited benefit for the acceleration‑based model. In this study, a physics‑guided LSTM is proposed to reduce reliance on displacement sensors, in which a learnable filter transforms accelerations into displacement‑like features. This hybrid model narrows the performance gap, achieving stable and robust prediction performance. The proposed model attains accuracy comparable to displacement‑based estimation, demonstrating its effectiveness with accelerometer input alone and highlighting its potential as a cost‑efficient tool for structural health monitoring of large‑scale FPV systems.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"211 ","pages":"Article 104037"},"PeriodicalIF":5.7,"publicationDate":"2025-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145049823","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}

引用次数: 0

Structural topology optimization considering material anisotropy induced by additive manufacturing processes 考虑增材制造工艺诱导材料各向异性的结构拓扑优化

IF 5.7 2区工程技术 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Advances in Engineering Software

Pub Date : 2025-09-12 DOI: 10.1016/j.advengsoft.2025.104021

Chao Wang , Di Lou , Zunyi Duan , Wenfeng Du , Jianhua Rong , Bin Xu

This work proposes a structural topology optimization method to consider material anisotropy induced by additive manufacturing processes. To quantify the relationship between manufacturing processes and mechanical properties of formed materials, the building direction angle is introduced into a transversely isotropic material model as a design variable. An anisotropic material model related to the building direction is thus established. A parallel optimization framework for structural topology and building direction is proposed by extending the classical compliance minimization formulation. And, to be applicable to gradient-based optimization algorithms, sensitivities related to density and angle variables are derived separately. Especially, to overcome the convergence difficulties caused by the periodic angle variables, an adaptive reduction strategy for the feasible region of angle variables is proposed. Typical numerical examples verify the rationality of the proposed method. The results show that the building direction related process-induced anisotropy significantly affects the optimized structural properties. The fluctuation of the trigonometric functions related to the angle variables would lead to obvious iteration oscillation in the optimization process, which makes the optimization difficult to converge. The proposed adaptive reduction strategy is proven effective in addressing this challenge. Besides, typical numerical properties of the co-optimization of structural topology and building direction are also revealed.

本文提出了一种考虑增材制造工艺引起的材料各向异性的结构拓扑优化方法。为了量化制造工艺与成形材料力学性能之间的关系，在横向各向同性材料模型中引入了建筑方向角作为设计变量。建立了与建筑方向相关的各向异性材料模型。通过扩展经典柔度最小化公式，提出了结构拓扑和建筑方向并行优化框架。并且，为了适用于基于梯度的优化算法，分别推导了与密度和角度变量相关的灵敏度。特别地，为了克服周期角变量带来的收敛困难，提出了一种角变量可行域的自适应约简策略。典型数值算例验证了该方法的合理性。结果表明，与建筑方向相关的过程引起的各向异性对优化后的结构性能有显著影响。在优化过程中，与角度变量相关的三角函数的波动会导致明显的迭代振荡，使得优化难以收敛。所提出的自适应减少策略被证明是有效的解决这一挑战。此外，还揭示了结构拓扑与建筑方向协同优化的典型数值特性。

{"title":"Structural topology optimization considering material anisotropy induced by additive manufacturing processes","authors":"Chao Wang , Di Lou , Zunyi Duan , Wenfeng Du , Jianhua Rong , Bin Xu","doi":"10.1016/j.advengsoft.2025.104021","DOIUrl":"10.1016/j.advengsoft.2025.104021","url":null,"abstract":"<div><div>This work proposes a structural topology optimization method to consider material anisotropy induced by additive manufacturing processes. To quantify the relationship between manufacturing processes and mechanical properties of formed materials, the building direction angle is introduced into a transversely isotropic material model as a design variable. An anisotropic material model related to the building direction is thus established. A parallel optimization framework for structural topology and building direction is proposed by extending the classical compliance minimization formulation. And, to be applicable to gradient-based optimization algorithms, sensitivities related to density and angle variables are derived separately. Especially, to overcome the convergence difficulties caused by the periodic angle variables, an adaptive reduction strategy for the feasible region of angle variables is proposed. Typical numerical examples verify the rationality of the proposed method. The results show that the building direction related process-induced anisotropy significantly affects the optimized structural properties. The fluctuation of the trigonometric functions related to the angle variables would lead to obvious iteration oscillation in the optimization process, which makes the optimization difficult to converge. The proposed adaptive reduction strategy is proven effective in addressing this challenge. Besides, typical numerical properties of the co-optimization of structural topology and building direction are also revealed.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"211 ","pages":"Article 104021"},"PeriodicalIF":5.7,"publicationDate":"2025-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145049820","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}

引用次数: 0

Research on coal wall parameter calibration and high precision model construction based on discrete element method 基于离散元法的煤壁参数标定及高精度模型构建研究

IF 5.7 2区工程技术 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Advances in Engineering Software

Pub Date : 2025-09-10 DOI: 10.1016/j.advengsoft.2025.104032

Xin Jin , Dongpo Han , Guochao Zhao , Lijuan Zhao

The accuracy of discrete element coal wall model significantly influences the characterization of coal-rock breaking mechanisms and equipment dynamic response in virtual prototype simulation. Based on coal-rock samples from Ordos Wenyu Mine of Yanzhou Coal Mining, key Tavares UFRJ parameters affecting particle compressive strength were identified through Plackett-Burman test and steepest ascent experiment. Breakage parameters were calibrated using optimal latin hypercube sampling (OLHS) and gaussian process regression (GPR). Hertz-Mindlin with Bonding parameters were then calibrated via uniaxial compression simulation. Model accuracy was verified through discrete element method-multi flexible body dynamics (DEM-MFBD) coupling simulation. Results indicate that D0, E Infinity, and Phi are the most significant parameters with influence rates of 38.5 %, 30.5 %, and 18.6 % respectively. The relative error between simulated and experimental particle compressive strength is below 4.56 %, while uniaxial compression simulation shows maximum relative error below 9.80 %. Comparing tri-axial load curves during shearer drum cutting, the maximum relative error of mean values between experimental and simulation data is 3.72 %, with maximum root mean square error (RMSE) of 4.60 %, outperforming traditional models and validating the model's accuracy and reliability for dynamic cutting process simulation.

在虚拟样机仿真中，离散单元煤壁模型的准确性对煤岩破碎机理表征和设备动态响应具有重要影响。以兖州煤矿鄂尔多斯文玉矿煤岩样品为研究对象，通过Plackett-Burman试验和最陡爬坡试验，确定了影响颗粒抗压强度的Tavares UFRJ关键参数。采用最优拉丁超立方体抽样（OLHS）和高斯过程回归（GPR）对断裂参数进行了标定。然后通过单轴压缩模拟校准带有键合参数的Hertz-Mindlin。通过离散元法-多柔体动力学（DEM-MFBD）耦合仿真验证了模型的准确性。结果表明，D0、E∞和Phi是最显著的参数，其影响率分别为38.5%、30.5%和18.6%。模拟颗粒抗压强度与实验颗粒抗压强度的相对误差小于4.56%，单轴压缩模拟颗粒抗压强度的最大相对误差小于9.80%。对比采煤机滚筒切削过程的三轴载荷曲线，实验值与仿真值的最大平均值相对误差为3.72%，均方根误差（RMSE）最大为4.60%，优于传统模型，验证了该模型用于动态切削过程仿真的准确性和可靠性。

{"title":"Research on coal wall parameter calibration and high precision model construction based on discrete element method","authors":"Xin Jin , Dongpo Han , Guochao Zhao , Lijuan Zhao","doi":"10.1016/j.advengsoft.2025.104032","DOIUrl":"10.1016/j.advengsoft.2025.104032","url":null,"abstract":"<div><div>The accuracy of discrete element coal wall model significantly influences the characterization of coal-rock breaking mechanisms and equipment dynamic response in virtual prototype simulation. Based on coal-rock samples from Ordos Wenyu Mine of Yanzhou Coal Mining, key Tavares UFRJ parameters affecting particle compressive strength were identified through Plackett-Burman test and steepest ascent experiment. Breakage parameters were calibrated using optimal latin hypercube sampling (OLHS) and gaussian process regression (GPR). Hertz-Mindlin with Bonding parameters were then calibrated via uniaxial compression simulation. Model accuracy was verified through discrete element method-multi flexible body dynamics (DEM-MFBD) coupling simulation. Results indicate that D0, E Infinity, and Phi are the most significant parameters with influence rates of 38.5 %, 30.5 %, and 18.6 % respectively. The relative error between simulated and experimental particle compressive strength is below 4.56 %, while uniaxial compression simulation shows maximum relative error below 9.80 %. Comparing tri-axial load curves during shearer drum cutting, the maximum relative error of mean values between experimental and simulation data is 3.72 %, with maximum root mean square error (RMSE) of 4.60 %, outperforming traditional models and validating the model's accuracy and reliability for dynamic cutting process simulation.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"211 ","pages":"Article 104032"},"PeriodicalIF":5.7,"publicationDate":"2025-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145027449","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}

引用次数: 0

A 3D vehicle-bridge interaction framework integrating energy-conserving Hamilton’s principle and stabilized Newmark-β method 结合节能Hamilton原理和稳定Newmark-β方法的三维车桥相互作用框架

IF 5.7 2区工程技术 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Advances in Engineering Software

Pub Date : 2025-09-06 DOI: 10.1016/j.advengsoft.2025.104022

Xinfeng Yin , Yang Quan , Linsong Wu , Tuerdi Kaiersaer , Zhou Huang

This study proposes a novel 3D (Three-dimensional) VBI (Vehicle-bridge interaction) system modeling framework based on Hamilton's principle, coupled with an improved Newmark-β method for solving dynamic responses. By considering the kinetic and potential energies of the system, Hamilton's principle accurately describes the coupled vibrations between vehicles and bridges. The dynamic equations of the VBI system are derived by constructing a Euler-Bernoulli beam theory models and vehicle a spring-damped system models, incorporating 3D road surface irregularities and random traffic loads. The coupled dynamic equations ensure energy conservation under complex traffic loads. An improved Newmark-β method is employed to solve the nonlinear dynamic responses, ensuring numerical stability and accuracy. Theoretical validation demonstrates the model's superior accuracy in describing bridge mid-span displacement and vehicle vertical displacement. Numerical simulations and case comparisons further highlight the advantages of Hamilton's principle. For example, at a vehicle speed of 40 km/h, the maximum deviation of the simulated mid-span displacement from the measured value is only 0.42 mm, with a coefficient of determination (R²) reaching 0.92 and the mean absolute error (MAE) significantly reduced to 0.24.

本研究提出了一种新的基于Hamilton原理的三维VBI （Vehicle-bridge interaction）系统建模框架，并结合改进的Newmark-β方法求解动力响应。通过考虑系统的动能和势能，汉密尔顿原理准确地描述了车辆和桥梁之间的耦合振动。通过建立欧拉-伯努利梁理论模型和车辆弹簧阻尼系统模型，推导了考虑三维路面不规则性和随机交通荷载的VBI系统动力学方程。耦合动力学方程保证了复杂交通荷载下的节能。采用改进的Newmark-β法求解非线性动力响应，保证了数值的稳定性和精度。理论验证表明，该模型在描述桥梁跨中位移和车辆竖向位移方面具有较好的准确性。数值模拟和实例比较进一步突出了汉密尔顿原理的优点。例如，在车速为40 km/h时，模拟的跨中位移与实测值的最大偏差仅为0.42 mm，决定系数（R²）达到0.92，平均绝对误差（MAE）显著降低至0.24。

{"title":"A 3D vehicle-bridge interaction framework integrating energy-conserving Hamilton’s principle and stabilized Newmark-β method","authors":"Xinfeng Yin , Yang Quan , Linsong Wu , Tuerdi Kaiersaer , Zhou Huang","doi":"10.1016/j.advengsoft.2025.104022","DOIUrl":"10.1016/j.advengsoft.2025.104022","url":null,"abstract":"<div><div>This study proposes a novel 3D (Three-dimensional) VBI (Vehicle-bridge interaction) system modeling framework based on Hamilton's principle, coupled with an improved Newmark-<em>β</em> method for solving dynamic responses. By considering the kinetic and potential energies of the system, Hamilton's principle accurately describes the coupled vibrations between vehicles and bridges. The dynamic equations of the VBI system are derived by constructing a Euler-Bernoulli beam theory models and vehicle a spring-damped system models, incorporating 3D road surface irregularities and random traffic loads. The coupled dynamic equations ensure energy conservation under complex traffic loads. An improved Newmark-<em>β</em> method is employed to solve the nonlinear dynamic responses, ensuring numerical stability and accuracy. Theoretical validation demonstrates the model's superior accuracy in describing bridge mid-span displacement and vehicle vertical displacement. Numerical simulations and case comparisons further highlight the advantages of Hamilton's principle. For example, at a vehicle speed of 40 km/h, the maximum deviation of the simulated mid-span displacement from the measured value is only 0.42 mm, with a coefficient of determination (R²) reaching 0.92 and the mean absolute error (MAE) significantly reduced to 0.24.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"211 ","pages":"Article 104022"},"PeriodicalIF":5.7,"publicationDate":"2025-09-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145005126","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}

引用次数: 0

Optimization design method of C30/C40 fly ash concrete based on machine learning and elite retention genetic algorithm 基于机器学习和精英保留遗传算法的C30/C40粉煤灰混凝土优化设计方法

IF 5.7 2区工程技术 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Advances in Engineering Software

Pub Date : 2025-09-01 DOI: 10.1016/j.advengsoft.2025.104019

Mingyue Hao , Yue Li , Xiwang Chen , Kun Ni , Wei Li

This paper establishes an intelligent optimization design method for fly ash (FA) concrete considering 28-day compressive strength, slump, and carbon emissions based on machine learning (ML) and elite retention genetic algorithm (EGA). The results demonstrate that the Extreme Gradient Boosting (XGB) model achieves high accuracy in predicting compressive strength, while Gradient Boosting (GB) shows higher accuracy and generalization ability in predicting slump. The water-to-binder ratio and cement content have a significant impact on the compressive strength of FA concrete. Reducing the water-to-binder ratio or increasing cement content helps improve compressive strength. The dosage of superplasticizer and the water content are key factors in controlling the slump. Properly increasing the dosage of superplasticizer and water content can effectively improve the slump of concrete. The FA concrete intelligent design system developed based on the XGB model, GB model, and EGA algorithm can efficiently obtain the optimal preparation parameters and accurately predict the corresponding performance. Furthermore, the carbon emissions of the optimized C30 and C40 FA concrete decrease by 12.72 % and 17.44 % respectively compared to the baseline concrete. Finally, the experimental results verify the prediction accuracy and generalization ability of the XGB and GB models, with the relative prediction error of C30 and C40 FA concrete both being less than 8 %.

本文建立了一种基于机器学习（ML）和精英保留遗传算法（EGA）的考虑28天抗压强度、坍落度和碳排放的粉煤灰混凝土智能优化设计方法。结果表明，极端梯度增强（XGB）模型在预测抗压强度方面具有较高的精度，梯度增强（GB）模型在预测坍落度方面具有较高的精度和泛化能力。水胶比和水泥掺量对FA混凝土抗压强度有显著影响。降低水胶比或增加水泥掺量有助于提高抗压强度。高效减水剂的用量和含水量是控制坍落度的关键因素。适当增加减水剂掺量和水掺量，可有效改善混凝土坍落度。基于XGB模型、GB模型和EGA算法开发的FA混凝土智能设计系统可以高效地获得最优的制备参数并准确预测相应的性能。此外，优化后的C30和C40 FA混凝土的碳排放量分别比基准混凝土减少12.72%和17.44%。最后，实验结果验证了XGB和GB模型的预测精度和泛化能力，C30和C40 FA混凝土的相对预测误差均小于8%。

{"title":"Optimization design method of C30/C40 fly ash concrete based on machine learning and elite retention genetic algorithm","authors":"Mingyue Hao , Yue Li , Xiwang Chen , Kun Ni , Wei Li","doi":"10.1016/j.advengsoft.2025.104019","DOIUrl":"10.1016/j.advengsoft.2025.104019","url":null,"abstract":"<div><div>This paper establishes an intelligent optimization design method for fly ash (FA) concrete considering 28-day compressive strength, slump, and carbon emissions based on machine learning (ML) and elite retention genetic algorithm (EGA). The results demonstrate that the Extreme Gradient Boosting (XGB) model achieves high accuracy in predicting compressive strength, while Gradient Boosting (GB) shows higher accuracy and generalization ability in predicting slump. The water-to-binder ratio and cement content have a significant impact on the compressive strength of FA concrete. Reducing the water-to-binder ratio or increasing cement content helps improve compressive strength. The dosage of superplasticizer and the water content are key factors in controlling the slump. Properly increasing the dosage of superplasticizer and water content can effectively improve the slump of concrete. The FA concrete intelligent design system developed based on the XGB model, GB model, and EGA algorithm can efficiently obtain the optimal preparation parameters and accurately predict the corresponding performance. Furthermore, the carbon emissions of the optimized C30 and C40 FA concrete decrease by 12.72 % and 17.44 % respectively compared to the baseline concrete. Finally, the experimental results verify the prediction accuracy and generalization ability of the XGB and GB models, with the relative prediction error of C30 and C40 FA concrete both being less than 8 %.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"210 ","pages":"Article 104019"},"PeriodicalIF":5.7,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144922123","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}

引用次数: 0

On the effectiveness of multigrid preconditioned iterative methods for large-scale frequency response topology optimization problems 多网格预处理迭代法求解大规模频响拓扑优化问题的有效性研究

IF 5.7 2区工程技术 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Advances in Engineering Software

Pub Date : 2025-08-30 DOI: 10.1016/j.advengsoft.2025.104017

Yongxin Qu , Niels Aage , Quhao Li

Large-scale static topology optimization of mechanical structures has been successfully realized for linear problems, including giga-voxel resolution aircraft wings and suspension bridges. Wherein the multigrid preconditioned conjugate gradient method (MG-CG) plays an important role in the repetitive solution of the state equations. However, research on large-scale dynamic topology optimization, e.g., frequency response problems, is still limited. Since the coefficient matrix of the dynamic equation is no longer a positive definite symmetric matrix, yet an indefinite, non-Hermitian and complex matrix, the conjugate gradient method (CG) is no longer applicable and the standard weapon-of-choice, the geometric multigrid preconditioner is no longer guaranteed to work. It is therefore of interest to investigate which iterative methods, if any, posses excellent generality and low computational-cost. In this paper, the effectiveness of several typical preconditioned iterative methods is studied, including conjugate gradient method, biconjugate gradient stabilized method (BICGSTAB), induced dimensionality reduction (IDR), generalized minimum residual method (GMRES). A detailed comparison and analysis of iterative methods' convergence, mesh dependence, and sensitivity to stiffness distribution in dealing with indefinite problems is given first. Then, despite its known disabilities, the geometric multigrid method is applied as a preconditioner for GMRES, BICGSTAB and IDR, i.e., MG-GMRES, MG-BICGSTAB and MG-IDR, to facilitate the efficient solution of large-scale frequency response analysis. In addition, the influence of several smoothers, including damped Jacobian iteration, successive over relaxation, symmetric SOR, and incomplete LU factorization, on the convergence of geometric multigrid iterative methods is also discussed. Numerical examples show that MG-BICGSTAB deals with low-frequency problems well, but for the whole frequency range, MG-GMRES with ILU smoother converges quickly and steadily, even if the model is extremely large. Furthermore, the effectiveness of the proposed procedure is further verified in dynamic topology optimization with up to 2.8 million degrees of freedom using a standard desktop computer.

机械结构的大规模静态拓扑优化已经成功地实现了线性问题，包括千兆体素分辨率的飞机机翼和悬索桥。其中，多网格预条件共轭梯度法（MG-CG）在状态方程的重复求解中起着重要作用。然而，对频率响应等大规模动态拓扑优化问题的研究仍然有限。由于动力方程的系数矩阵不再是正定对称矩阵，而是不定的非厄米复矩阵，因此共轭梯度法（CG）不再适用，几何多网格预调节器也不再保证起作用。因此，研究哪些迭代方法（如果有的话）具有良好的通用性和较低的计算成本是有意义的。本文研究了几种典型的预条件迭代方法的有效性，包括共轭梯度法、双共轭梯度稳定法、诱导降维法、广义最小残差法。首先对求解不确定问题的迭代方法的收敛性、网格依赖性和对刚度分布的敏感性进行了详细的比较和分析。然后，尽管存在已知的缺陷，将几何多重网格方法作为GMRES、BICGSTAB和IDR的前置条件，即MG-GMRES、MG-BICGSTAB和MG-IDR，以促进大规模频响分析的有效求解。此外，还讨论了阻尼雅可比迭代、逐次过松弛、对称SOR和不完全LU分解等平滑项对几何多网格迭代方法收敛性的影响。数值算例表明，MG-BICGSTAB可以很好地处理低频问题，但对于整个频率范围，即使模型非常大，具有ILU平滑的MG-GMRES也可以快速稳定地收敛。此外，在标准台式计算机上，进一步验证了该方法在高达280万自由度的动态拓扑优化中的有效性。

{"title":"On the effectiveness of multigrid preconditioned iterative methods for large-scale frequency response topology optimization problems","authors":"Yongxin Qu , Niels Aage , Quhao Li","doi":"10.1016/j.advengsoft.2025.104017","DOIUrl":"10.1016/j.advengsoft.2025.104017","url":null,"abstract":"<div><div>Large-scale static topology optimization of mechanical structures has been successfully realized for linear problems, including giga-voxel resolution aircraft wings and suspension bridges. Wherein the multigrid preconditioned conjugate gradient method (MG-CG) plays an important role in the repetitive solution of the state equations. However, research on large-scale dynamic topology optimization, e.g., frequency response problems, is still limited. Since the coefficient matrix of the dynamic equation is no longer a positive definite symmetric matrix, yet an indefinite, non-Hermitian and complex matrix, the conjugate gradient method (CG) is no longer applicable and the standard weapon-of-choice, the geometric multigrid preconditioner is no longer guaranteed to work. It is therefore of interest to investigate which iterative methods, if any, posses excellent generality and low computational-cost. In this paper, the effectiveness of several typical preconditioned iterative methods is studied, including conjugate gradient method, biconjugate gradient stabilized method (BICGSTAB), induced dimensionality reduction (IDR), generalized minimum residual method (GMRES). A detailed comparison and analysis of iterative methods' convergence, mesh dependence, and sensitivity to stiffness distribution in dealing with indefinite problems is given first. Then, despite its known disabilities, the geometric multigrid method is applied as a preconditioner for GMRES, BICGSTAB and IDR, i.e., MG-GMRES, MG-BICGSTAB and MG-IDR, to facilitate the efficient solution of large-scale frequency response analysis. In addition, the influence of several smoothers, including damped Jacobian iteration, successive over relaxation, symmetric SOR, and incomplete LU factorization, on the convergence of geometric multigrid iterative methods is also discussed. Numerical examples show that MG-BICGSTAB deals with low-frequency problems well, but for the whole frequency range, MG-GMRES with ILU smoother converges quickly and steadily, even if the model is extremely large. Furthermore, the effectiveness of the proposed procedure is further verified in dynamic topology optimization with up to 2.8 million degrees of freedom using a standard desktop computer.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"210 ","pages":"Article 104017"},"PeriodicalIF":5.7,"publicationDate":"2025-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144918919","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}

引用次数: 0

Prediction and analysis of temperature recovery after arch closure grouting of Baihetan Arch Dam 白鹤滩拱坝闭拱注浆后温度恢复预测与分析

IF 5.7 2区工程技术 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Advances in Engineering Software

Pub Date : 2025-08-28 DOI: 10.1016/j.advengsoft.2025.104012

Jingyu Yan , Feng Wang , Yuhang Li , Chunbo Ma , Junchi Zhou , Hu Yu

Arch closure grouting is an essential procedure for attaining structural completion during the construction of a concrete arch dam. After the closure, due to the continuous heat emission from cement hydration, the internal temperature of concrete rises rapidly, which affects stress distribution and structural stability. In order to accurately predict the temperature evolution of concrete pouring blocks after arch closure, this paper conducted a comparative study using neural networks and finite element methods. First, a hybrid model, CNN-BiLSTM, was constructed. This model integrates a Convolutional Neural Network (CNN) and a Bidirectional Long Short-Term Memory (BiLSTM). The Weighted Mean of Vectors algorithm (INFO) was then introduced to optimize the model parameters. The temperature variation trend of concrete pouring blocks after arch closure was predicted using this approach. Simultaneously, considering the factors such as external temperature, cooling water, adiabatic temperature rise and concrete age, a three-dimensional finite element model of concrete pouring blocks was established to simulate the temperature field distribution of concrete. The comparison results indicate that both methods can achieve the prediction accuracy required by the project (with an error of less than 2 °C). Among them, the finite element simulation performs better in terms of stability (with a difference of less than 1 °C from the measured value). At the same time, the INFO-CNN-BiLSTM model exhibits significant temperature fluctuations during certain periods and demonstrates insufficient generalization ability. However, it offers the advantage of high computational efficiency.

在混凝土拱坝施工中，闭拱灌浆是保证结构完工的重要工序。闭合后，由于水泥水化持续放热，混凝土内部温度迅速升高，影响应力分布和结构稳定性。为了准确预测混凝土浇筑块合拱后的温度演变，本文采用神经网络和有限元方法进行了对比研究。首先，构建了CNN-BiLSTM混合模型。该模型集成了卷积神经网络（CNN）和双向长短期记忆（BiLSTM）。然后引入向量加权平均算法（INFO）对模型参数进行优化。利用该方法预测了混凝土浇筑块合拱后的温度变化趋势。同时，考虑外部温度、冷却水、绝热温升和混凝土龄期等因素，建立混凝土浇筑块三维有限元模型，模拟混凝土温度场分布。对比结果表明，两种方法均能达到工程要求的预测精度（误差小于2°C）。其中，有限元模拟在稳定性方面表现较好（与实测值相差小于1℃）。同时，INFO-CNN-BiLSTM模型在特定时期温度波动较大，泛化能力不足。然而，它具有计算效率高的优点。

{"title":"Prediction and analysis of temperature recovery after arch closure grouting of Baihetan Arch Dam","authors":"Jingyu Yan , Feng Wang , Yuhang Li , Chunbo Ma , Junchi Zhou , Hu Yu","doi":"10.1016/j.advengsoft.2025.104012","DOIUrl":"10.1016/j.advengsoft.2025.104012","url":null,"abstract":"<div><div>Arch closure grouting is an essential procedure for attaining structural completion during the construction of a concrete arch dam. After the closure, due to the continuous heat emission from cement hydration, the internal temperature of concrete rises rapidly, which affects stress distribution and structural stability. In order to accurately predict the temperature evolution of concrete pouring blocks after arch closure, this paper conducted a comparative study using neural networks and finite element methods. First, a hybrid model, CNN-BiLSTM, was constructed. This model integrates a Convolutional Neural Network (CNN) and a Bidirectional Long Short-Term Memory (BiLSTM). The Weighted Mean of Vectors algorithm (INFO) was then introduced to optimize the model parameters. The temperature variation trend of concrete pouring blocks after arch closure was predicted using this approach. Simultaneously, considering the factors such as external temperature, cooling water, adiabatic temperature rise and concrete age, a three-dimensional finite element model of concrete pouring blocks was established to simulate the temperature field distribution of concrete. The comparison results indicate that both methods can achieve the prediction accuracy required by the project (with an error of less than 2 °C). Among them, the finite element simulation performs better in terms of stability (with a difference of less than 1 °C from the measured value). At the same time, the INFO-CNN-BiLSTM model exhibits significant temperature fluctuations during certain periods and demonstrates insufficient generalization ability. However, it offers the advantage of high computational efficiency.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"210 ","pages":"Article 104012"},"PeriodicalIF":5.7,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144908427","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}

引用次数: 0

A machine learning-based inverse analysis procedure for concrete softening law prediction using non-experimental datasets 基于机器学习的非实验数据集混凝土软化规律预测逆分析方法

IF 5.7 2区工程技术 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Advances in Engineering Software

Pub Date : 2025-08-27 DOI: 10.1016/j.advengsoft.2025.104016

H. Rodrigo Amezcua , A. Gustavo Ayala , Carlos E. González

This paper studies the mechanical behaviour of concrete as one of the most widely used quasi-brittle construction materials emphasizing on the importance of knowing its mechanical parameters and their evolution during the inelastic stage, i.e., the softening law. The softening curve, which describes the response of the material under damage or cracking, is critical for predicting the behaviour of concrete structures subjected to extreme loads. Experimental tests are commonly employed to obtain this information either directly or indirectly. Some of the indirect methods are based on inverse analysis and/or artificial intelligence techniques, both of which capable of predicting the mechanical parameters of concrete from the experimental results of one test, e.g., a notched beam subjected to vertical loads. However, an important drawback of these procedures is that they require a large dataset constructed from data gathered in multiple experiments in order to be developed. Consequently, most existing methods are tailored to specific types of experiments and even limited to certain specimen dimensions. Additionally, these procedures primarily focus on predicting mechanical parameters rather than determining the softening law. To address these limitations, this paper proposes a machine learning-based algorithm for the inverse analysis of an experimental test capable of predicting both the softening law and the mechanical parameters of concrete. By generating a non-experimental dataset through the Sequentially Linear Analysis (SLA) procedure, the proposed algorithm can be applied to other experimental setups suitable for analysis with SLA. The results of the application example demonstrate the effectiveness of the proposed approach.

混凝土作为一种应用最广泛的准脆性建筑材料，本文对其力学性能进行了研究，强调了解其非弹性阶段的力学参数及其演变，即软化规律的重要性。软化曲线描述了材料在损伤或开裂下的反应，对于预测混凝土结构在极端载荷下的行为至关重要。通常采用实验测试来直接或间接地获得这一信息。一些间接方法基于逆分析和/或人工智能技术，这两种方法都能够从一次测试的实验结果中预测混凝土的力学参数，例如，承受垂直载荷的缺口梁。然而，这些程序的一个重要缺点是，它们需要一个从多个实验中收集的数据构建的大型数据集才能进行开发。因此，大多数现有的方法都是针对特定类型的实验量身定制的，甚至仅限于某些样品尺寸。此外，这些程序主要侧重于预测力学参数，而不是确定软化规律。为了解决这些限制，本文提出了一种基于机器学习的算法，用于能够预测混凝土软化规律和力学参数的实验测试的逆分析。通过顺序线性分析（SLA）过程生成非实验数据集，该算法可以应用于其他适合使用SLA进行分析的实验设置。应用实例的结果表明了该方法的有效性。

{"title":"A machine learning-based inverse analysis procedure for concrete softening law prediction using non-experimental datasets","authors":"H. Rodrigo Amezcua , A. Gustavo Ayala , Carlos E. González","doi":"10.1016/j.advengsoft.2025.104016","DOIUrl":"10.1016/j.advengsoft.2025.104016","url":null,"abstract":"<div><div>This paper studies the mechanical behaviour of concrete as one of the most widely used quasi-brittle construction materials emphasizing on the importance of knowing its mechanical parameters and their evolution during the inelastic stage, <em>i.e.</em>, the softening law. The softening curve, which describes the response of the material under damage or cracking, is critical for predicting the behaviour of concrete structures subjected to extreme loads. Experimental tests are commonly employed to obtain this information either directly or indirectly. Some of the indirect methods are based on inverse analysis and/or artificial intelligence techniques, both of which capable of predicting the mechanical parameters of concrete from the experimental results of one test, <em>e.g.</em>, a notched beam subjected to vertical loads. However, an important drawback of these procedures is that they require a large dataset constructed from data gathered in multiple experiments in order to be developed. Consequently, most existing methods are tailored to specific types of experiments and even limited to certain specimen dimensions. Additionally, these procedures primarily focus on predicting mechanical parameters rather than determining the softening law. To address these limitations, this paper proposes a machine learning-based algorithm for the inverse analysis of an experimental test capable of predicting both the softening law and the mechanical parameters of concrete. By generating a non-experimental dataset through the Sequentially Linear Analysis (SLA) procedure, the proposed algorithm can be applied to other experimental setups suitable for analysis with SLA. The results of the application example demonstrate the effectiveness of the proposed approach.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"210 ","pages":"Article 104016"},"PeriodicalIF":5.7,"publicationDate":"2025-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144902936","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}

引用次数: 0

Deep reinforcement learning-based control algorithm for flight kinematics of insect-scale flyers 基于深度强化学习的昆虫尺度飞行器飞行运动学控制算法

IF 5.7 2区工程技术 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Advances in Engineering Software

Pub Date : 2025-08-27 DOI: 10.1016/j.advengsoft.2025.104014

Seungpyo Hong , Sejin Kim , Innyoung Kim , Donghyun You

An autonomous flight control algorithm based on deep reinforcement learning (DRL) is developed for insect-scale flyers with flexible wings in complex flow environments, addressing the challenges posed by highly unsteady and nonlinear aeroelastic dynamics. Unlike conventional model-based approaches, this study employs high-fidelity computational fluid–structural dynamics (CFD-CSD) simulations that fully resolve the governing equations of both the fluid and the flyer, providing physically consistent data for training the DRL agent. To mitigate the computational cost, a novel physics-guided data augmentation strategy is introduced, which synthetically expands the training dataset by replicating CFD-CSD data across diverse virtual flight scenarios while preserving the underlying physics. This approach enables the DRL agent to learn a robust control policy that generalizes across a broad range of aerodynamic conditions, demonstrating strong control performance even in complex and untrained flow environments. This work establishes a scalable framework for the autonomous control of flexible, bio-inspired flyers under realistic aerodynamic conditions, representing a significant step toward fully autonomous insect-scale flight.

针对具有柔性翼的昆虫级飞行器在复杂流动环境中的飞行控制问题，提出了一种基于深度强化学习（DRL）的自主飞行控制算法。与传统的基于模型的方法不同，该研究采用了高保真的计算流体-结构动力学（CFD-CSD）模拟，完全解决了流体和飞片的控制方程，为训练DRL代理提供了物理上一致的数据。为了降低计算成本，引入了一种新的物理引导数据增强策略，该策略通过复制不同虚拟飞行场景的CFD-CSD数据来综合扩展训练数据集，同时保留底层物理。这种方法使DRL智能体能够学习一种鲁棒的控制策略，该策略可以在广泛的空气动力学条件下进行推广，即使在复杂和未经训练的流动环境中也能表现出强大的控制性能。这项工作为在现实空气动力学条件下自主控制灵活的仿生飞行器建立了一个可扩展的框架，代表了迈向完全自主昆虫级飞行的重要一步。

{"title":"Deep reinforcement learning-based control algorithm for flight kinematics of insect-scale flyers","authors":"Seungpyo Hong , Sejin Kim , Innyoung Kim , Donghyun You","doi":"10.1016/j.advengsoft.2025.104014","DOIUrl":"10.1016/j.advengsoft.2025.104014","url":null,"abstract":"<div><div>An autonomous flight control algorithm based on deep reinforcement learning (DRL) is developed for insect-scale flyers with flexible wings in complex flow environments, addressing the challenges posed by highly unsteady and nonlinear aeroelastic dynamics. Unlike conventional model-based approaches, this study employs high-fidelity computational fluid–structural dynamics (CFD-CSD) simulations that fully resolve the governing equations of both the fluid and the flyer, providing physically consistent data for training the DRL agent. To mitigate the computational cost, a novel physics-guided data augmentation strategy is introduced, which synthetically expands the training dataset by replicating CFD-CSD data across diverse virtual flight scenarios while preserving the underlying physics. This approach enables the DRL agent to learn a robust control policy that generalizes across a broad range of aerodynamic conditions, demonstrating strong control performance even in complex and untrained flow environments. This work establishes a scalable framework for the autonomous control of flexible, bio-inspired flyers under realistic aerodynamic conditions, representing a significant step toward fully autonomous insect-scale flight.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"210 ","pages":"Article 104014"},"PeriodicalIF":5.7,"publicationDate":"2025-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144902937","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}

引用次数: 0