This article reviews the application and progress of deep learning in efficient numerical computing methods.Deep learning,as an important branch of machine learning,provides new ideas for numerical computation by cons...This article reviews the application and progress of deep learning in efficient numerical computing methods.Deep learning,as an important branch of machine learning,provides new ideas for numerical computation by constructing multi-layer neural networks to simulate the learning process of the human brain.The article explores the application of deep learning in solving partial differential equations,optimizing problems,and data-driven modeling,and analyzes its advantages in computational efficiency,accuracy,and adaptability.At the same time,this article also points out the challenges faced by deep learning numerical computation methods in terms of computational efficiency,interpretability,and generalization ability,and proposes strategies and future development directions for integrating with traditional numerical methods.展开更多
Based on the principle of residual deformation induced by superposition of the welding residual stress and working stress, the welding heat source efficiency has been determined by measuring displacement changes of sp...Based on the principle of residual deformation induced by superposition of the welding residual stress and working stress, the welding heat source efficiency has been determined by measuring displacement changes of specimens under loading and unloading in tensile tests, and combining with calculating welding parameters. Meanwhile, the welding heat source eficiencies obtained are compared with those of the measuring-calculating method. The research results show that the welding heat source efficiencies are almost the same as those obtained by the measuring-calculating method. Therefore, the welding heat source efficiency can be determined accurately by this method, and a new determining method of the heat source efficiency for the welding heat process calculating has been provided.展开更多
Sound radiation of thin plates is a common problem in engineering. Hashimoto proposed the discrete calculation method(DCM) to deal with the problem. The calculation of the radiation impedance of the rectangular elemen...Sound radiation of thin plates is a common problem in engineering. Hashimoto proposed the discrete calculation method(DCM) to deal with the problem. The calculation of the radiation impedance of the rectangular element is more cumbersome than that of the circular one, so the discrete rectangular radiation element is approximated by the circular one. However, error is also introduced. The formula developed by Sha has been employed to get self- and mutual-radiation impedances of rectangular radiation element. Numerical study was performed to verify error introduced by the approximation Hashimoto adopted. Experimental researches on sound radiation of a 2 mm-thick and a 4 mm-thick magnesium alloy plates were also carried out to evaluate the errors introduced by the approximation. The experimental results indicate that the circular approximation Hashimoto adopted overestimates the sound radiation efficiency. The maximum error levels of the radiation efficiencies of the2 mm-thick and 4 mm-thick magnesium alloy plates are up to 0.15 and 0.12, respectively. The effect of element aspect ratio on the sound radiation efficiency is also remarkable.展开更多
To obtain the precise calculation method for the peak energy density and energy evolution properties of rocks subjected to uniaxial compression(UC)before the post-peak stage,particularly at s0.9sc(s denotes stress and...To obtain the precise calculation method for the peak energy density and energy evolution properties of rocks subjected to uniaxial compression(UC)before the post-peak stage,particularly at s0.9sc(s denotes stress and sc is the peak strength),extensive UC and uniaxial graded cyclical loading-unloading(GCLU)tests were performed on four rock types.In the GCLU tests,four unloading stress levels were designated when σ<0.9σc and six unloading stress levels were designated forσ≥0.9σc.The variations in the elastic energy density(ue),dissipative energy density(ud),and energy storage efficiency(C)for the four rock types under GCLU tests were analyzed.Based on the variation of ue whenσ≥0:9σc,a method for calculating the peak energy density was proposed.The energy evolution in rock under UC condition before the post-peak stage was examined.The relationship between C0.9(C atσ≥0:9σc)and mechanical behavior of rocks was explored,and the damage evolution of rock was analyzed in view of energy.Compared with that of the three existing methods,the accuracy of the calculation method of peak energy density proposed in this study is higher.These findings could provide a theoretical foundation for more accurately revealing the failure behavior of rock from an energy perspective.展开更多
Improving the computational efficiency of multi-physics simulation and constructing a real-time online simulation method is an important way to realise the virtual-real fusion of entities and data of power equipment w...Improving the computational efficiency of multi-physics simulation and constructing a real-time online simulation method is an important way to realise the virtual-real fusion of entities and data of power equipment with digital twin.In this paper,a datadriven fast calculation method for the temperature field of resin impregnated paper(RIP)bushing used in converter transformer valve-side is proposed,which combines the data dimensionality reduction technology and the surrogate model.After applying the finite element algorithm to obtain the temperature field distribution of RIP bushing under different operation conditions as the input dataset,the proper orthogonal decomposition(POD)algorithm is adopted to reduce the order and obtain the low-dimensional projection of the temperature data.On this basis,the surrogate model is used to construct the mapping relationship between the sensor monitoring data and the low-dimensional projection,so that it can achieve the fast calculation and reconstruction of temperature field distribution.The results show that this method can effectively and quickly calculate the overall temperature field distribution of the RIP bushing.The maximum relative error and the average relative error are less than 4.5%and 0.25%,respectively.The calculation speed is at the millisecond level,meeting the needs of digitalisation of power equipment.展开更多
[目的]通过增加储能回路可以将核聚变堆产生的间隙能量转化为连续能量,若进一步转变为经济社会发展最常用的电能,则需要选择合适的动力循环技术路线。[方法]文章以中国聚变工程实验堆(China Fusion Engineering Test Reactor,CFETR)为边...[目的]通过增加储能回路可以将核聚变堆产生的间隙能量转化为连续能量,若进一步转变为经济社会发展最常用的电能,则需要选择合适的动力循环技术路线。[方法]文章以中国聚变工程实验堆(China Fusion Engineering Test Reactor,CFETR)为边界,采用Aspen HYSYS软件和REFPROP物性计算方法分别建立了超临界二氧化碳(SCO_(2))分流再压缩布雷顿循环模型和氦气布雷顿循环模型,研究了循环中主要参数对循环效率的影响,并对蒸汽朗肯循环、SCO_(2)分流再压缩布雷顿循环和氦气布雷顿循环进行对比分析。[结果]推荐SCO_(2)布雷顿循环作为热核聚变电站动力循环技术路线。[结论]对热核聚变电站发电岛动力循环技术路线的分析及结论,可为后续聚变发电技术研究以及工程设计提供重要参考。展开更多
Aiming at increasing the calculation efficiency of the pseudospectral methods, a multiple- interval Radau pseudospectral method （RPM） is presented to generate a reusable launch vehicle （RLV） ＇s optimal re-entry t...Aiming at increasing the calculation efficiency of the pseudospectral methods, a multiple- interval Radau pseudospectral method （RPM） is presented to generate a reusable launch vehicle （RLV） ＇s optimal re-entry trajectory. After dividing the optimal control problem into many intervals, the state and control variables are approximated using many fixed- and low-degree Lagrange polyno- mials in each interval. Convergence of the numerical discretization is then achieved by increasing the number of intervals. With the application of the proposed method, the normal nonlinear program- ming （NLP） problem transcribed from the optimal control problem can avoid being dense because of the low-degree approximation polynomials in each interval. Thus, the NLP solver can easily compute a solution. Finally, simulation results show that the optimized re-entry trajectories satisfy the path constraints and the boundary constraints successfully. Compared with the single interval RPM, the multiple-interval RPM is significantly faster and has higher calculation efficiency. The results indicate that the multiple-interval RPM can be applied for real-time trajectory generation due to its high effi- ciency and high precision.展开更多
为能够在合理计算规模下准确模拟空爆自由场冲击波特征的网格划分方案,获得可靠的计算结果,本文运用验证与确认(verification and validation,V&V)方法,通过开展网格敏感性分析,确认了炸药空爆计算模型的数值解、外推估值、网格收...为能够在合理计算规模下准确模拟空爆自由场冲击波特征的网格划分方案,获得可靠的计算结果,本文运用验证与确认(verification and validation,V&V)方法,通过开展网格敏感性分析,确认了炸药空爆计算模型的数值解、外推估值、网格收敛指标与比例距离的定量关系,给出了满足不同网格收敛指标要求的最大比例网格尺寸随比例距离位置的变化关系。据此,在比例距离Z为0~40 m·kg^(-1/3)范围内,给出了划分渐变网格的优化方案。针对1维、2维、3维计算模型,分别比较了细网格、粗网格、渐变网格方案的计算精度与计算耗时情况,讨论了本文提出网格划分方法的适用性。同时,采用本文建议的网格优化方案,给出了动爆冲击波毁伤飞机数值模拟场景的应用算例。结果表明,本文建议的网格优化方案可在几乎不降低计算精度的前提下显著提升计算效率,可为空爆自由场或者近似计算场景的数值模型网格划分提供参考。展开更多
文摘This article reviews the application and progress of deep learning in efficient numerical computing methods.Deep learning,as an important branch of machine learning,provides new ideas for numerical computation by constructing multi-layer neural networks to simulate the learning process of the human brain.The article explores the application of deep learning in solving partial differential equations,optimizing problems,and data-driven modeling,and analyzes its advantages in computational efficiency,accuracy,and adaptability.At the same time,this article also points out the challenges faced by deep learning numerical computation methods in terms of computational efficiency,interpretability,and generalization ability,and proposes strategies and future development directions for integrating with traditional numerical methods.
文摘Based on the principle of residual deformation induced by superposition of the welding residual stress and working stress, the welding heat source efficiency has been determined by measuring displacement changes of specimens under loading and unloading in tensile tests, and combining with calculating welding parameters. Meanwhile, the welding heat source eficiencies obtained are compared with those of the measuring-calculating method. The research results show that the welding heat source efficiencies are almost the same as those obtained by the measuring-calculating method. Therefore, the welding heat source efficiency can be determined accurately by this method, and a new determining method of the heat source efficiency for the welding heat process calculating has been provided.
基金the National Technology Research and Development Program in the 12th Five Year Plan of China(No.2011BAE22B05)the Canada-ChinaUSA Collaborative Research and Development Project(No.2011DFA50900)
文摘Sound radiation of thin plates is a common problem in engineering. Hashimoto proposed the discrete calculation method(DCM) to deal with the problem. The calculation of the radiation impedance of the rectangular element is more cumbersome than that of the circular one, so the discrete rectangular radiation element is approximated by the circular one. However, error is also introduced. The formula developed by Sha has been employed to get self- and mutual-radiation impedances of rectangular radiation element. Numerical study was performed to verify error introduced by the approximation Hashimoto adopted. Experimental researches on sound radiation of a 2 mm-thick and a 4 mm-thick magnesium alloy plates were also carried out to evaluate the errors introduced by the approximation. The experimental results indicate that the circular approximation Hashimoto adopted overestimates the sound radiation efficiency. The maximum error levels of the radiation efficiencies of the2 mm-thick and 4 mm-thick magnesium alloy plates are up to 0.15 and 0.12, respectively. The effect of element aspect ratio on the sound radiation efficiency is also remarkable.
基金the National Natural Science Foundation of China(Grant Nos.52104133 and 52304227)the Postdoctoral Foundation of Henan Province(Grant No.HN2022015)are appreciated.
文摘To obtain the precise calculation method for the peak energy density and energy evolution properties of rocks subjected to uniaxial compression(UC)before the post-peak stage,particularly at s0.9sc(s denotes stress and sc is the peak strength),extensive UC and uniaxial graded cyclical loading-unloading(GCLU)tests were performed on four rock types.In the GCLU tests,four unloading stress levels were designated when σ<0.9σc and six unloading stress levels were designated forσ≥0.9σc.The variations in the elastic energy density(ue),dissipative energy density(ud),and energy storage efficiency(C)for the four rock types under GCLU tests were analyzed.Based on the variation of ue whenσ≥0:9σc,a method for calculating the peak energy density was proposed.The energy evolution in rock under UC condition before the post-peak stage was examined.The relationship between C0.9(C atσ≥0:9σc)and mechanical behavior of rocks was explored,and the damage evolution of rock was analyzed in view of energy.Compared with that of the three existing methods,the accuracy of the calculation method of peak energy density proposed in this study is higher.These findings could provide a theoretical foundation for more accurately revealing the failure behavior of rock from an energy perspective.
基金supported by China Postdoctoral Science Foundation,Grant 2024M753544Science and Technology Project of CSG,Grant GDKJXM2022106.
文摘Improving the computational efficiency of multi-physics simulation and constructing a real-time online simulation method is an important way to realise the virtual-real fusion of entities and data of power equipment with digital twin.In this paper,a datadriven fast calculation method for the temperature field of resin impregnated paper(RIP)bushing used in converter transformer valve-side is proposed,which combines the data dimensionality reduction technology and the surrogate model.After applying the finite element algorithm to obtain the temperature field distribution of RIP bushing under different operation conditions as the input dataset,the proper orthogonal decomposition(POD)algorithm is adopted to reduce the order and obtain the low-dimensional projection of the temperature data.On this basis,the surrogate model is used to construct the mapping relationship between the sensor monitoring data and the low-dimensional projection,so that it can achieve the fast calculation and reconstruction of temperature field distribution.The results show that this method can effectively and quickly calculate the overall temperature field distribution of the RIP bushing.The maximum relative error and the average relative error are less than 4.5%and 0.25%,respectively.The calculation speed is at the millisecond level,meeting the needs of digitalisation of power equipment.
文摘Aiming at increasing the calculation efficiency of the pseudospectral methods, a multiple- interval Radau pseudospectral method （RPM） is presented to generate a reusable launch vehicle （RLV） ＇s optimal re-entry trajectory. After dividing the optimal control problem into many intervals, the state and control variables are approximated using many fixed- and low-degree Lagrange polyno- mials in each interval. Convergence of the numerical discretization is then achieved by increasing the number of intervals. With the application of the proposed method, the normal nonlinear program- ming （NLP） problem transcribed from the optimal control problem can avoid being dense because of the low-degree approximation polynomials in each interval. Thus, the NLP solver can easily compute a solution. Finally, simulation results show that the optimized re-entry trajectories satisfy the path constraints and the boundary constraints successfully. Compared with the single interval RPM, the multiple-interval RPM is significantly faster and has higher calculation efficiency. The results indicate that the multiple-interval RPM can be applied for real-time trajectory generation due to its high effi- ciency and high precision.
文摘为能够在合理计算规模下准确模拟空爆自由场冲击波特征的网格划分方案,获得可靠的计算结果,本文运用验证与确认(verification and validation,V&V)方法,通过开展网格敏感性分析,确认了炸药空爆计算模型的数值解、外推估值、网格收敛指标与比例距离的定量关系,给出了满足不同网格收敛指标要求的最大比例网格尺寸随比例距离位置的变化关系。据此,在比例距离Z为0~40 m·kg^(-1/3)范围内,给出了划分渐变网格的优化方案。针对1维、2维、3维计算模型,分别比较了细网格、粗网格、渐变网格方案的计算精度与计算耗时情况,讨论了本文提出网格划分方法的适用性。同时,采用本文建议的网格优化方案,给出了动爆冲击波毁伤飞机数值模拟场景的应用算例。结果表明,本文建议的网格优化方案可在几乎不降低计算精度的前提下显著提升计算效率,可为空爆自由场或者近似计算场景的数值模型网格划分提供参考。
