Dynamic earth pressure induced by machine foundations on a neighboring retaining wall is analyzed with emphasis on factors which control the intensity and location of the design forces. The meshless local Petrov-Galer...Dynamic earth pressure induced by machine foundations on a neighboring retaining wall is analyzed with emphasis on factors which control the intensity and location of the design forces. The meshless local Petrov-Galerkin (MLPG) method is used to analyze the problem for a variety of retaining wall and machine foundation geometries. The soil medium is assumed to be homogeneous and visco-elastic. The machine foundation is idealized as a harmonic sinusoidal dynamic force often encountered in practice. A number of analyses have been made to reveal the effect of the loading frequency, the location and size of the foundation and the soil shear wave velocity on the distribution and magnitude of the dynamic earth pressure. Results indicate that there is a critical frequency and a critical location for which the passive pressure takes the maxima in the entire duration of the dynamic load.展开更多
A method which adopts the combination of least squares support vector machine（LS-SVM） and Monte Carlo（MC） simulation is used to calculate the foundation settlement reliability.When using LS-SVM,choosing the traini...A method which adopts the combination of least squares support vector machine（LS-SVM） and Monte Carlo（MC） simulation is used to calculate the foundation settlement reliability.When using LS-SVM,choosing the training dataset and the values for LS-SVM parameters is the key.In a representative sense,the orthogonal experimental design with four factors and five levels is used to choose the inputs of the training dataset,and the outputs are calculated by using fast Lagrangian analysis continua（FLAC）.The decimal ant colony algorithm（DACA） is also used to determine the parameters.Calculation results show that the values of the two parameters,and δ2 have great effect on the performance of LS-SVM.After the training of LS-SVM,the inputs are sampled according to the probabilistic distribution,and the outputs are predicted with the trained LS-SVM,thus the reliability analysis can be performed by the MC method.A program compiled by Matlab is employed to calculate its reliability.Results show that the method of combining LS-SVM and MC simulation is applicable to the reliability analysis of soft foundation settlement.展开更多
Machine learning interatomic potentials(MLIPs)have introduced a new paradigm for atomic simulations.Recent advancements have led to universal MLIPs(uMLIPs)that are pre-trained on diverse datasets,providing opportuniti...Machine learning interatomic potentials(MLIPs)have introduced a new paradigm for atomic simulations.Recent advancements have led to universal MLIPs(uMLIPs)that are pre-trained on diverse datasets,providing opportunities for universal force fields and foundational machine learning models.However,their performance in extrapolating to out-of-distribution complex atomic environments remains unclear.In this study,we highlight a consistent potential energy surface(PES)softening effect in three uMLIPs:M3GNet,CHGNet,and MACE-MP-0,which is characterized by energy and force underprediction in atomic-modeling benchmarks including surfaces,defects,solid-solution energetics,ion migration barriers,phonon vibration modes,and general high-energy states.The PES softening behavior originates primarily from the systematically underpredicted PES curvature,which derives from the biased sampling of near-equilibrium atomic arrangements in uMLIP pre-training datasets.Our findings suggest that a considerable fraction of uMLIP errors are highly systematic,and can therefore be efficiently corrected.We argue for the importance of a comprehensive materials dataset with improved PES sampling for next-generation foundational MLIPs.展开更多
文摘Dynamic earth pressure induced by machine foundations on a neighboring retaining wall is analyzed with emphasis on factors which control the intensity and location of the design forces. The meshless local Petrov-Galerkin (MLPG) method is used to analyze the problem for a variety of retaining wall and machine foundation geometries. The soil medium is assumed to be homogeneous and visco-elastic. The machine foundation is idealized as a harmonic sinusoidal dynamic force often encountered in practice. A number of analyses have been made to reveal the effect of the loading frequency, the location and size of the foundation and the soil shear wave velocity on the distribution and magnitude of the dynamic earth pressure. Results indicate that there is a critical frequency and a critical location for which the passive pressure takes the maxima in the entire duration of the dynamic load.
文摘A method which adopts the combination of least squares support vector machine（LS-SVM） and Monte Carlo（MC） simulation is used to calculate the foundation settlement reliability.When using LS-SVM,choosing the training dataset and the values for LS-SVM parameters is the key.In a representative sense,the orthogonal experimental design with four factors and five levels is used to choose the inputs of the training dataset,and the outputs are calculated by using fast Lagrangian analysis continua（FLAC）.The decimal ant colony algorithm（DACA） is also used to determine the parameters.Calculation results show that the values of the two parameters,and δ2 have great effect on the performance of LS-SVM.After the training of LS-SVM,the inputs are sampled according to the probabilistic distribution,and the outputs are predicted with the trained LS-SVM,thus the reliability analysis can be performed by the MC method.A program compiled by Matlab is employed to calculate its reliability.Results show that the method of combining LS-SVM and MC simulation is applicable to the reliability analysis of soft foundation settlement.
基金funded by the U.S.Department of Energy,Office of Science,Office of Basic Energy Sciences,Materials Sciences and Engineering Division under Contract No.DE-AC0205CH11231(Materials Project program KC23MP)supported by the computational resources provided by the Extreme Science and Engineering Discovery Environment(XSEDE),supported by National Science Foundation grant number ACI1053575+1 种基金the National Energy Research Scientific Computing Center(NERSC),a U.S.Department of Energy Office of Science User Facility located at Lawrence Berkeley National Laboratorythe Swift Cluster resource provided by the National Renewable Energy Laboratory(NREL).
文摘Machine learning interatomic potentials(MLIPs)have introduced a new paradigm for atomic simulations.Recent advancements have led to universal MLIPs(uMLIPs)that are pre-trained on diverse datasets,providing opportunities for universal force fields and foundational machine learning models.However,their performance in extrapolating to out-of-distribution complex atomic environments remains unclear.In this study,we highlight a consistent potential energy surface(PES)softening effect in three uMLIPs:M3GNet,CHGNet,and MACE-MP-0,which is characterized by energy and force underprediction in atomic-modeling benchmarks including surfaces,defects,solid-solution energetics,ion migration barriers,phonon vibration modes,and general high-energy states.The PES softening behavior originates primarily from the systematically underpredicted PES curvature,which derives from the biased sampling of near-equilibrium atomic arrangements in uMLIP pre-training datasets.Our findings suggest that a considerable fraction of uMLIP errors are highly systematic,and can therefore be efficiently corrected.We argue for the importance of a comprehensive materials dataset with improved PES sampling for next-generation foundational MLIPs.
