J-OCTA Technical Week 2023
Online Video Streaming

In this talk, the current state of Modeling and Simulation as long as Materials Informatics will be discussed. Some case studies will be shared with the audience to demostrate how modeling and simulation and materials informatics are being leveraged at 3M.

We consider examples of J-OCTA software application to physics-based analysis of in-space manufacturing technologies. Firstly, we consider an atomistic model of thermal welding at the polymer-polymer interface for applications to 3D manufacturing by following diffusion of polymer chains. We show that each component of the polymer blend has its own characteristic time of diffusion and that both strain-stress and shear viscosity curves agree with experimental data. Next, nonlinear shrinkage of the metal part during manufacturing by bound metal deposition under microgravity, is considered using a multi-scale physics-based approach that spans from atomistic dynamics to full-part shrinkage. At the smallest scale we use atomistic dynamics to simulated sintering of a few Ti6Al4V nanoparticles, study sintering mechanisms and estimate important parameters including grain boundary width and diffusion rates of the atoms at the surface that are used at the other scales of modeling. Finally, we analyze selective ionic conduction through an artificial nanopore in a graphene sheet motivated by applications to desalination and energy harvesting. We build distributions of the number and orientation of water molecules in the ions hydration shells providing novel insight into the ion-water-carbon interaction at the nanopore.

Coarse grained models are quite useful for studying polymer rheology and dynamics, since many rheological properties are "universal", i.e., independent of the chemical details of the polymer chains. In this talk, first I will give a short introduction to the linear polymer rheology, especially the prediction of Rouse and Tube models, and their confirmation by molecular dynamics (MD) simulation using a corse-grained model. Next I will discuss primitive path analysis (PPA) for finding entanglement spacing of a given polymer and its application to miscible polymer blends, and discuss the tube diameter in the blends. Then I will show results of a slip-link model, a model that is further coarse grained and takes account of only entanglements. Finally I will show a few unsolved problems of non-linear rheology and present our (un-finished) work for investigating these problems.

In this talk, we introduce some examples on the analyses of adhesion phenomena for soft viscoelastic materials. We explain meso-scopic and dynamical features of deformation, debonding and fracture of such materials, and then discuss how experiment and modeling should be communicated and how a multi-scale model should be constructed.

The NEDO Project "Ultra High-Throughput Design and Prototyping Technology for Ultra Advanced Materials Development Project (U2M Project)" which was conducted from FY2016 to FY2022, mainly targets organic and polymeric materials and aimed to accelerate the speed of research and development utilizing AI technology in addition to calculation, process, characterization technologies. This presentation will introduce overview and outcomes of U2M project. Especially, an extension of functions of OCTA was conducted in U2M project, and new functions of extended OCTA such as image analysis and linkage with machine learning tools are implemented. Such functions of extended OCTA are being implemented to free version of OCTA, and the details of new functions will be introduced. In addition, COGNAC update and recent research project related a collaboration of coarse-grained simulation and machine learning will be introduced.

"GENESIS: GENeralized-Ensemble Simulation System" is a molecular dynamics simulation software developed by RIKEN Sugita team and co-workers. It is highly versatile software that not only exhibits high parallel performance on supercomputers but also supports GPUs. It allows the all-atom, coarse-grained, and quantum-classical hybrid calculations with various functions for efficient conformational sampling and free energy analysis. A wide range of applications are expected in the fields of drug discovery and material design. We will introduce various functions of GENESIS to be used for researchers not only in academia but also in industry.

ASAP is a powerful platform for materials design and atomistic modeling. The platform features several automated workflows designated to model and analyze structural, chemical, electronic, dynamic, optical properties of a wide range of systems of industrial interest (OLED, dyes, oxide electronics, polymers additives, polymer industry, large scale bio-systems). This presentation demonstrates the capabilities of ASAP such as: powerful structure builder, flexible workflows, local and remote calculation control. We will give an overview of advanced ASAP functionalities such as the TranSIESTA workflow for evaluation of transport properties, charge analysis, DFT+U and access to a dedicated terminal window enabling flexible scripting. As ASAP integrates the SIESTA package, we will discuss its latest developments such as the recently improved Linear-Scaling method that enables computations of large scale systems (several thousands of atoms).

The latest features and development trends of J-OCTA will be introduced.

The latest case studies will be introduced.

J-OCTA provides many functions for efficient modeling and calculation. In this presentation, we will introduce some examples of the use of such functions mainly COGNAC modeler and calculation execution functions with some videos.

VSOP-PS is a particle-based fluid simulation engine capable of shear calculations for particulate dispersion suspensions.We will introduce latest features added at a version 8.0 and 8.1 with some case studies.

In order to manipulate biomolecular systems in J-OCTA, we start to implement GUI pre-posts, GENESIS modeler, to execute the modules of GENESIS which is the molecular dynamics software packages developed mainly in RIKEN.

In this seminar, we will introduce the latest version of SIESTA modeler and the procedures to use it to evaluate physical properties, not only to check/evaluate analysis results, but also to use it for multi-scale analysis in J-OCTA.

Mol-infer is a function that estimates the molecular structure that satisfies a given physical property from a given physical property value. The development of mol-infer's original logic and core functionality is performed by Nagamochi Laboratory, while we are developing the front end to use the core functionality.

For CAE engineers who are interested in learning molecular simulation techniques, we will provide tips on how to use them. An overview will be given of approaches to multi-scale simulation, using different methods at the same scale, and using machine learning for property estimation.