This study investigates the application of Factorization Machines with Quantum Annealing (FMQA) to address the crystal structure problem (CSP) in materials science. FMQA is a black-box optimization algorithm that combines machine learning with annealing machines to find samples to a black-box function that minimize a given loss. The CSP involves determining the optimal arrangement of atoms in a material based on its chemical composition, a critical challenge in materials science. We explore FMQA's ability to efficiently sample optimal crystal configurations by setting the loss function to the energy of the crystal configuration as given by a predefined interatomic potential. Further we investigate how well the energies of the various metastable configurations, or local minima of the potential, are learned by the algorithm. Our investigation reveals FMQA's potential in quick ground state sampling and in recovering relational order between local minima.
{"title":"Machine learning supported annealing for prediction of grand canonical crystal structures","authors":"Yannick Couzinie, Yuya Seki, Yusuke Nishiya, Hirofumi Nishi, Taichi Kosugi, Shu Tanaka, Yu-ichiro Matsushita","doi":"arxiv-2408.03556","DOIUrl":"https://doi.org/arxiv-2408.03556","url":null,"abstract":"This study investigates the application of Factorization Machines with\u0000Quantum Annealing (FMQA) to address the crystal structure problem (CSP) in\u0000materials science. FMQA is a black-box optimization algorithm that combines\u0000machine learning with annealing machines to find samples to a black-box\u0000function that minimize a given loss. The CSP involves determining the optimal\u0000arrangement of atoms in a material based on its chemical composition, a\u0000critical challenge in materials science. We explore FMQA's ability to\u0000efficiently sample optimal crystal configurations by setting the loss function\u0000to the energy of the crystal configuration as given by a predefined interatomic\u0000potential. Further we investigate how well the energies of the various\u0000metastable configurations, or local minima of the potential, are learned by the\u0000algorithm. Our investigation reveals FMQA's potential in quick ground state\u0000sampling and in recovering relational order between local minima.","PeriodicalId":501369,"journal":{"name":"arXiv - PHYS - Computational Physics","volume":"10 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141937473","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this work, we investigate the implications of the differential Hebbian learning rule known as Input-Correlations (ICO) learning in the classification of synchronization in coupled nonlinear oscillator systems. We are investigating the parity-time symmetric coupled Duffing oscillator system with nonlinear dissipation/amplification. In our investigation of the temporal dynamics of this system, it is observed that the system exhibits chaotic as well as quasiperiodic dynamics. On further investigation, it is found that the chaotic dynamics is distorted anti-phase synchronized, whereas the quasiperiodic dynamics is desynchronized. So, on the application of the ICO learning in these two parametric regimes, we observe that the weight associated with the stimulus remains constant when the oscillators are anti-phase synchronized, in spite of there being distortion in the synchronization. But when the oscillators exhibit quasiperiodic dynamics, there is erratic evolution of the weight with time. So, from this, it could be ascertained that the ICO learning could be made use of in the classification of synchronization dynamics in nonlinear systems.
{"title":"Classification of synchronization in nonlinear systems using ICO learning","authors":"J. P. Deka","doi":"arxiv-2408.04024","DOIUrl":"https://doi.org/arxiv-2408.04024","url":null,"abstract":"In this work, we investigate the implications of the differential Hebbian\u0000learning rule known as Input-Correlations (ICO) learning in the classification\u0000of synchronization in coupled nonlinear oscillator systems. We are\u0000investigating the parity-time symmetric coupled Duffing oscillator system with\u0000nonlinear dissipation/amplification. In our investigation of the temporal\u0000dynamics of this system, it is observed that the system exhibits chaotic as\u0000well as quasiperiodic dynamics. On further investigation, it is found that the\u0000chaotic dynamics is distorted anti-phase synchronized, whereas the\u0000quasiperiodic dynamics is desynchronized. So, on the application of the ICO\u0000learning in these two parametric regimes, we observe that the weight associated\u0000with the stimulus remains constant when the oscillators are anti-phase\u0000synchronized, in spite of there being distortion in the synchronization. But\u0000when the oscillators exhibit quasiperiodic dynamics, there is erratic evolution\u0000of the weight with time. So, from this, it could be ascertained that the ICO\u0000learning could be made use of in the classification of synchronization dynamics\u0000in nonlinear systems.","PeriodicalId":501369,"journal":{"name":"arXiv - PHYS - Computational Physics","volume":"128 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141937400","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The demand for pseudopotentials constructed for a given exchange-correlation (XC) functional far exceeds the supply, necessitating the use of those commonly available. The number of XC functionals currently available is in the hundreds, if not thousands, and the majority of pseudopotentials have been generated for the LDA and PBE. The objective of this study is to identify the error in the determination of the mechanical and structural properties (lattice constant, cohesive energy, elastic constants, and bulk modulus) of crystals calculated by DFT with such inconsistency. Additionally, the study aims to estimate the performance of popular XC functionals (LDA, PBE, PBEsol, and SCAN) for these calculations in a consistent manner.
{"title":"Uncertainty of DFT calculated mechanical and structural properties of solids due to incompatibility of pseudopotentials and exchange-correlation functionals","authors":"Marcin Maździarz","doi":"arxiv-2408.03835","DOIUrl":"https://doi.org/arxiv-2408.03835","url":null,"abstract":"The demand for pseudopotentials constructed for a given exchange-correlation\u0000(XC) functional far exceeds the supply, necessitating the use of those commonly\u0000available. The number of XC functionals currently available is in the hundreds,\u0000if not thousands, and the majority of pseudopotentials have been generated for\u0000the LDA and PBE. The objective of this study is to identify the error in the\u0000determination of the mechanical and structural properties (lattice constant,\u0000cohesive energy, elastic constants, and bulk modulus) of crystals calculated by\u0000DFT with such inconsistency. Additionally, the study aims to estimate the\u0000performance of popular XC functionals (LDA, PBE, PBEsol, and SCAN) for these\u0000calculations in a consistent manner.","PeriodicalId":501369,"journal":{"name":"arXiv - PHYS - Computational Physics","volume":"129 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141937532","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We present an extensive review of the two-dimensional finite difference Hartree--Fock (FD HF) method, and present its implementation in the newest version of X2DHF, the FD HF program for atoms and diatomic molecules. The program was originally published in Comput. Phys. Commun. in 1996, and was last revised in 2013. X2DHF can be used to obtain HF limit values of total energies and multipole moments for a wide range of diatomic molecules and their ions, using either point nuclei or a finite nuclear model. Polarizabilities ($alpha_{zz}$) and hyperpolarizabilities ($beta_{zzz}$, $gamma_{zzzz}$, ${A}_{z,zz}$, ${B}_{zz,zz}$) can also be computed by the program with the finite-field method. X2DHF has been extensively used in the literature to assess the accuracy of existing atomic basis sets and to help in developing new ones. As a new feature since the last revision, the program can now also perform Kohn-Sham density functional calculations with local and generalized gradient exchange-correlation functionals with the Libxc library of density functionals, enabling new types of studies. Furthermore, the initialization of calculations has been greatly simplified. As before, X2DHF can also perform one-particle calculations with (smooth) Coulomb, Green-Sellin-Zachor and Krammers-Henneberger potentials, while calculations with a superposition of atomic potentials have been added as a new feature. The program is easy to install from the GitHub repository and build via CMake using the x2dhfctl script that facilitates creating its single- and multiple-threaded versions, as well as building in Libxc support. Calculations can be carried out with X2DHF in double- or quadruple-precision arithmetic.
{"title":"Review of the finite difference Hartree-Fock method for atoms and diatomic molecules, and its implementation in the x2dhf program","authors":"Jacek Kobus, Susi Lehtola","doi":"arxiv-2408.03679","DOIUrl":"https://doi.org/arxiv-2408.03679","url":null,"abstract":"We present an extensive review of the two-dimensional finite difference\u0000Hartree--Fock (FD HF) method, and present its implementation in the newest\u0000version of X2DHF, the FD HF program for atoms and diatomic molecules. The\u0000program was originally published in Comput. Phys. Commun. in 1996, and was last\u0000revised in 2013. X2DHF can be used to obtain HF limit values of total energies\u0000and multipole moments for a wide range of diatomic molecules and their ions,\u0000using either point nuclei or a finite nuclear model. Polarizabilities\u0000($alpha_{zz}$) and hyperpolarizabilities ($beta_{zzz}$, $gamma_{zzzz}$,\u0000${A}_{z,zz}$, ${B}_{zz,zz}$) can also be computed by the program with the\u0000finite-field method. X2DHF has been extensively used in the literature to\u0000assess the accuracy of existing atomic basis sets and to help in developing new\u0000ones. As a new feature since the last revision, the program can now also\u0000perform Kohn-Sham density functional calculations with local and generalized\u0000gradient exchange-correlation functionals with the Libxc library of density\u0000functionals, enabling new types of studies. Furthermore, the initialization of\u0000calculations has been greatly simplified. As before, X2DHF can also perform\u0000one-particle calculations with (smooth) Coulomb, Green-Sellin-Zachor and\u0000Krammers-Henneberger potentials, while calculations with a superposition of\u0000atomic potentials have been added as a new feature. The program is easy to\u0000install from the GitHub repository and build via CMake using the x2dhfctl\u0000script that facilitates creating its single- and multiple-threaded versions, as\u0000well as building in Libxc support. Calculations can be carried out with X2DHF\u0000in double- or quadruple-precision arithmetic.","PeriodicalId":501369,"journal":{"name":"arXiv - PHYS - Computational Physics","volume":"29 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141937470","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Calcium carbonate plays a crucial role in the global carbon cycle, and its phase diagram has always been of significant scientific interest. In this study, we used molecular dynamics (MD) to investigate several structural phase transitions of calcium carbonate. Using the Raiteri potential model, we explored the structural transitions occurring at a constant pressure of 1 bar with temperatures ranging from 300 K to 2500 K, and at a constant temperature of 1600 K with pressures ranging from 0 to 13 GPa. At increasing temperatures, the transitions calcite, CaCO$_3$-IV, and CaCO$_3$-V are observed and characterized. Within the calcite structure, CO$_3^{2-}$ ions are ordered between layers. As temperature increases, the calcite to CaCO$_3$-IV transition occurs, determining the partial disordering of CO$_3^{2-}$ ions. At a higher temperature, CaCO$_3$-IV transforms into CaCO$_3$-V. By applying free energy analysis, we have classified the last transition as a continuous order-disorder phase transition. At a temperature of 2000 K, it appears a `disordered CaCO$_3$' structure, characterized by low order within the calcium and carbonate sublattices and the free rotation of CO$_3^{2-}$ ions. At increasing pressures, two calcium carbonate transformations were observed. At $P=$ 2 GPa, the CaCO$_3$-IV phase undergoes a phase transition into CaCO$_3$-V, demonstrating that the model can describe the transition between these two phases as pressure and temperature-driven. Another phase transition was detected at $P=$ 4.25 GPa -- CaCO$_3$-V transits into the recently discovered CaCO$_3$-Vb phase. This transition is classified as a first-order phase transition by structural analysis and free energy-based arguments.
{"title":"Structural transitions of calcium carbonate by molecular dynamics simulation","authors":"Elizaveta Sidler, Raffaela Cabriolu","doi":"arxiv-2408.04036","DOIUrl":"https://doi.org/arxiv-2408.04036","url":null,"abstract":"Calcium carbonate plays a crucial role in the global carbon cycle, and its\u0000phase diagram has always been of significant scientific interest. In this\u0000study, we used molecular dynamics (MD) to investigate several structural phase\u0000transitions of calcium carbonate. Using the Raiteri potential model, we\u0000explored the structural transitions occurring at a constant pressure of 1 bar\u0000with temperatures ranging from 300 K to 2500 K, and at a constant temperature\u0000of 1600 K with pressures ranging from 0 to 13 GPa. At increasing temperatures,\u0000the transitions calcite, CaCO$_3$-IV, and CaCO$_3$-V are observed and\u0000characterized. Within the calcite structure, CO$_3^{2-}$ ions are ordered\u0000between layers. As temperature increases, the calcite to CaCO$_3$-IV transition\u0000occurs, determining the partial disordering of CO$_3^{2-}$ ions. At a higher\u0000temperature, CaCO$_3$-IV transforms into CaCO$_3$-V. By applying free energy\u0000analysis, we have classified the last transition as a continuous order-disorder\u0000phase transition. At a temperature of 2000 K, it appears a `disordered\u0000CaCO$_3$' structure, characterized by low order within the calcium and\u0000carbonate sublattices and the free rotation of CO$_3^{2-}$ ions. At increasing\u0000pressures, two calcium carbonate transformations were observed. At $P=$ 2 GPa,\u0000the CaCO$_3$-IV phase undergoes a phase transition into CaCO$_3$-V,\u0000demonstrating that the model can describe the transition between these two\u0000phases as pressure and temperature-driven. Another phase transition was\u0000detected at $P=$ 4.25 GPa -- CaCO$_3$-V transits into the recently discovered\u0000CaCO$_3$-Vb phase. This transition is classified as a first-order phase\u0000transition by structural analysis and free energy-based arguments.","PeriodicalId":501369,"journal":{"name":"arXiv - PHYS - Computational Physics","volume":"91 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141937401","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yasin Ameslon, Olivier J. J. Ronsin, Christina Harreiss, Johannes Will, Stefanie Rechberger Mingjian Wu, Erdmann Spiecker, Jens Harting
Interest in organic solar cells (OSCs) is constantly rising in the field of photovoltaic devices. The device performance relies on the bulk heterojunction (BHJ) nanomorphology, which develops during the drying process and additional post-treatment. This work studies the effect of thermal annealing (TA) on an all-small molecule DRCN5T: PC71 BM blend with phase field simulations. The objective is to determine the physical phenomena driving the evolution of the BHJ morphology for a better understanding of the posttreatment/morphology relationship. Phase-field simulation results are used to investigate the impact on the final BHJ morphology of the DRCN5T crystallization-related mechanisms, including nucleation, growth, crystal stability, impingement, grain coarsening, and Ostwald ripening, of the amorphous-amorphous phase separation (AAPS), and of diffusion limitations. The comparison of simulation results with experimental data shows that the morphological evolution of the BHJ under TA is dominated by dissolution of the smallest, unstable DRCN5T crystals and anisotropic growth of the largest crystals.
{"title":"Phase field simulations of thermal annealing for all-small molecule organic solar cells","authors":"Yasin Ameslon, Olivier J. J. Ronsin, Christina Harreiss, Johannes Will, Stefanie Rechberger Mingjian Wu, Erdmann Spiecker, Jens Harting","doi":"arxiv-2408.03190","DOIUrl":"https://doi.org/arxiv-2408.03190","url":null,"abstract":"Interest in organic solar cells (OSCs) is constantly rising in the field of\u0000photovoltaic devices. The device performance relies on the bulk heterojunction\u0000(BHJ) nanomorphology, which develops during the drying process and additional\u0000post-treatment. This work studies the effect of thermal annealing (TA) on an\u0000all-small molecule DRCN5T: PC71 BM blend with phase field simulations. The\u0000objective is to determine the physical phenomena driving the evolution of the\u0000BHJ morphology for a better understanding of the posttreatment/morphology\u0000relationship. Phase-field simulation results are used to investigate the impact\u0000on the final BHJ morphology of the DRCN5T crystallization-related mechanisms,\u0000including nucleation, growth, crystal stability, impingement, grain coarsening,\u0000and Ostwald ripening, of the amorphous-amorphous phase separation (AAPS), and\u0000of diffusion limitations. The comparison of simulation results with\u0000experimental data shows that the morphological evolution of the BHJ under TA is\u0000dominated by dissolution of the smallest, unstable DRCN5T crystals and\u0000anisotropic growth of the largest crystals.","PeriodicalId":501369,"journal":{"name":"arXiv - PHYS - Computational Physics","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141937478","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This work develops a new method for computing a finite quantum system's continuum states and observables by applying a subspace projection (or reduced basis) method used in model order reduction studies to ``discretize'' the system's continuous spectrum. The method extracts the continuum physics from solving Schr"odinger equations with bound-state-like boundary conditions and emulates this extraction in the space of the input parameters. This parameter emulation can readily be adapted to emulate other continuum calculations as well, e.g., those based on complex energy or Lorentz integral transform methods. Here, I give an overview of the key aspects of the formalism and some informative findings from numerical experimentation with two- and three-body systems, which indicates the non-Hermitian quantum mechanics nature of the method. A potential connection with (near-)optimal rational approximation studied in Math literature is also discussed. Further details are provided in a separate paper.
{"title":"A non-Hermitian quantum mechanics approach for extracting and emulating continuum physics based on bound-state-like calculations","authors":"Xilin Zhang","doi":"arxiv-2408.03309","DOIUrl":"https://doi.org/arxiv-2408.03309","url":null,"abstract":"This work develops a new method for computing a finite quantum system's\u0000continuum states and observables by applying a subspace projection (or reduced\u0000basis) method used in model order reduction studies to ``discretize'' the\u0000system's continuous spectrum. The method extracts the continuum physics from\u0000solving Schr\"odinger equations with bound-state-like boundary conditions and\u0000emulates this extraction in the space of the input parameters. This parameter\u0000emulation can readily be adapted to emulate other continuum calculations as\u0000well, e.g., those based on complex energy or Lorentz integral transform\u0000methods. Here, I give an overview of the key aspects of the formalism and some\u0000informative findings from numerical experimentation with two- and three-body\u0000systems, which indicates the non-Hermitian quantum mechanics nature of the\u0000method. A potential connection with (near-)optimal rational approximation\u0000studied in Math literature is also discussed. Further details are provided in a\u0000separate paper.","PeriodicalId":501369,"journal":{"name":"arXiv - PHYS - Computational Physics","volume":"77 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141937475","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Carlo.jl is a Monte Carlo simulation framework written in Julia. It provides MPI-parallel scheduling, organized storage of input, checkpoint, and output files, as well as statistical postprocessing. With a minimalist design, it aims to aid the development of high-quality Monte Carlo codes, especially for demanding applications in condensed matter and statistical physics. This hands-on user guide shows how to implement a simple code with Carlo.jl and provides benchmarks to show its efficacy.
Carlo.jl 是一个用 Julia 编写的蒙特卡罗仿真框架。它提供 MPI 并行调度，有组织地存储输入、检查点和输出文件，以及统计后处理。它采用简约设计，旨在帮助开发高质量的蒙特卡罗代码，尤其是针对凝聚态物质和统计物理中的高要求应用。这本实用的用户指南展示了如何使用 Carlo.jl 实现一个简单的代码，并提供了基准来显示其功效。
{"title":"Carlo.jl: A general framework for Monte Carlo simulations in Julia","authors":"Lukas Weber","doi":"arxiv-2408.03386","DOIUrl":"https://doi.org/arxiv-2408.03386","url":null,"abstract":"Carlo.jl is a Monte Carlo simulation framework written in Julia. It provides\u0000MPI-parallel scheduling, organized storage of input, checkpoint, and output\u0000files, as well as statistical postprocessing. With a minimalist design, it aims\u0000to aid the development of high-quality Monte Carlo codes, especially for\u0000demanding applications in condensed matter and statistical physics. This\u0000hands-on user guide shows how to implement a simple code with Carlo.jl and\u0000provides benchmarks to show its efficacy.","PeriodicalId":501369,"journal":{"name":"arXiv - PHYS - Computational Physics","volume":"48 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141937395","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
From a data perspective, the field of materials mechanics is characterized by a sparsity of available data, mainly due to the strong microstructure-sensitivity of properties such as strength, fracture toughness, and fatigue limit. Consequently, individual tests are needed for specimens with various thermo-mechanical process histories, even if their chemical composition remains the same. Experimental data on the mechanical behavior of materials is usually rare, as mechanical testing is typically a destructive method requiring large amounts of material and effort for specimen preparation and testing. Furthermore, mechanical behavior is typically characterized in simplified tests under uniaxial loading conditions, whereas a complete characterization of mechanical material behavior requires multiaxial testing conditions. To address this data sparsity, different simulation methods, such as micromechanical modeling or even atomistic simulations, can contribute to microstructure-sensitive data collections. These methods cover a wide range of materials with different microstructures characterized under multiaxial loading conditions. In the present work, we describe a novel data schema that integrates metadata and mechanical data itself, following the workflows of the material modeling processes by which the data has been generated. Each run of this workflow results in unique data objects due to the incorporation of various elements such as user, system, and job-specific information in correlation with the resulting mechanical properties. Hence, this integrated data format provides a sustainable way of generating data objects that are Findable, Accessible, Interoperable, and Reusable (FAIR). The choice of metadata elements has centered on necessary features required to characterize microstructure-specific data objects, simplifying how purpose-specific datasets are collected by search algorithms.
{"title":"A Workflow-Centric Approach to Generating FAIR Data Objects for Mechanical Properties of Materials","authors":"Ronak Shoghi, Alexander Hartmaier","doi":"arxiv-2408.03965","DOIUrl":"https://doi.org/arxiv-2408.03965","url":null,"abstract":"From a data perspective, the field of materials mechanics is characterized by\u0000a sparsity of available data, mainly due to the strong\u0000microstructure-sensitivity of properties such as strength, fracture toughness,\u0000and fatigue limit. Consequently, individual tests are needed for specimens with\u0000various thermo-mechanical process histories, even if their chemical composition\u0000remains the same. Experimental data on the mechanical behavior of materials is\u0000usually rare, as mechanical testing is typically a destructive method requiring\u0000large amounts of material and effort for specimen preparation and testing.\u0000Furthermore, mechanical behavior is typically characterized in simplified tests\u0000under uniaxial loading conditions, whereas a complete characterization of\u0000mechanical material behavior requires multiaxial testing conditions. To address\u0000this data sparsity, different simulation methods, such as micromechanical\u0000modeling or even atomistic simulations, can contribute to\u0000microstructure-sensitive data collections. These methods cover a wide range of\u0000materials with different microstructures characterized under multiaxial loading\u0000conditions. In the present work, we describe a novel data schema that\u0000integrates metadata and mechanical data itself, following the workflows of the\u0000material modeling processes by which the data has been generated. Each run of\u0000this workflow results in unique data objects due to the incorporation of\u0000various elements such as user, system, and job-specific information in\u0000correlation with the resulting mechanical properties. Hence, this integrated\u0000data format provides a sustainable way of generating data objects that are\u0000Findable, Accessible, Interoperable, and Reusable (FAIR). The choice of\u0000metadata elements has centered on necessary features required to characterize\u0000microstructure-specific data objects, simplifying how purpose-specific datasets\u0000are collected by search algorithms.","PeriodicalId":501369,"journal":{"name":"arXiv - PHYS - Computational Physics","volume":"2 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141937397","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The properties of constrained fluids have increasingly gained relevance for applications ranging from materials to biology. In this work, we propose a multiscale model using twin neural networks to investigate the properties of a fluid constrained between solid surfaces with complex shapes. The atomic scale model and the mesoscale model are connected by the coarse-grained potential which is represented by the first neural network. Then we train the second neural network model as a surrogate to predict the velocity profile of the constrained fluid with complex boundary conditions at the mesoscale. The effect of complex boundary conditions on the fluid dynamics properties and the accuracy of the neural network model prediction are systematically investigated. We demonstrate that the neural network-enhanced multiscale framework can connect simulations at atomic scale and mesoscale and reproduce the properties of a constrained fluid at mesoscale. This work provides insight into multiscale model development with the aid of machine learning techniques and the developed model can be used for modern nanotechnology applications such as enhanced oil recovery and porous materials design.
{"title":"Multiscale modeling framework of a constrained fluid with complex boundaries using twin neural networks","authors":"Peiyuan Gao, George Em Karniadakis, Panos Stinis","doi":"arxiv-2408.03263","DOIUrl":"https://doi.org/arxiv-2408.03263","url":null,"abstract":"The properties of constrained fluids have increasingly gained relevance for\u0000applications ranging from materials to biology. In this work, we propose a\u0000multiscale model using twin neural networks to investigate the properties of a\u0000fluid constrained between solid surfaces with complex shapes. The atomic scale\u0000model and the mesoscale model are connected by the coarse-grained potential\u0000which is represented by the first neural network. Then we train the second\u0000neural network model as a surrogate to predict the velocity profile of the\u0000constrained fluid with complex boundary conditions at the mesoscale. The effect\u0000of complex boundary conditions on the fluid dynamics properties and the\u0000accuracy of the neural network model prediction are systematically\u0000investigated. We demonstrate that the neural network-enhanced multiscale\u0000framework can connect simulations at atomic scale and mesoscale and reproduce\u0000the properties of a constrained fluid at mesoscale. This work provides insight\u0000into multiscale model development with the aid of machine learning techniques\u0000and the developed model can be used for modern nanotechnology applications such\u0000as enhanced oil recovery and porous materials design.","PeriodicalId":501369,"journal":{"name":"arXiv - PHYS - Computational Physics","volume":"91 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141937476","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：