Workshop scientific report
Title: Multiscale Simulation: from Materials through to Industrial Usage
Location: O´Brien Centre for Science University College Dublin Room H2.38 Sept 5-6 (Mon, Tues) of and E1.17 Sept 7 (Wed) 2016 [https://goo.gl/maps/o68YpMhfX5q]
Dates: September 5, 2016 to September 7, 2016Organizers: Eoin O'Reilly, Joan Adler, Heinz A. Preisig, Shahriar Amini, A. W. J (Sander) Gielen,
State of the art
modern physics, the evolution of computational capacity allows
materials to be investigated over a wide range of length and time
scales, often referred to as multi-scale materials modelling (MMM). This
allows for the modelling of complex materials under realistic
constraints in a wide range of situations. Almost all MMM discussions
take a bottom-up approach, starting from ab-initio density functional
theory and linking upwards via higher level models to predict key
material properties at the micro- or nano-scale. From the other side,
Integrated Computational Material Engineering (ICME) starts from the
application requirements, and aims to understand how processes form
material structures, how these structures give rise to material
properties and how to select proper materials for applications based on
this information. ICME uses computational engineering as a major
component aiming to remove the need for a lot of experimentation. The
aim is to design products alongside the materials that comprise them,
and the end result is sought from investigations and simulations at
multiple length scales during the product and process development.
Outcomes of the meeting:
• Progress in addressing challenges in multiscale simulation
• Coupling and linking between “bottom-up” and “top-down” communities
• Building critical mass/consensus towards development of best practice and standardisation
• European leadership in this important area of multiscale simulations from first principles through to industrial applications
There is indeed much European activity thru the EUMMC, and this workshop is an important step in spreading information. However one aspect that succeeded even beyond expectations was the interaction with our US invited speakers. Several of these are leaders of projects with experience in standards development (e.g. OpenKIM and LAMMPS) which they generously shared with us.
Specific topics we addressed included:
1. Supporting best practices: we shared what works among those involved in this type of activity.
2. Reducing redundancy: what tools and frameworks are available and how reuse can be done most effectively.
3. Developing standards: discussion of where standardization is needed will greatly increase efficiency of the platform tools being developed, and their interoperability.
4. Integration databases: avoid redoing calculations, increase standardization and data compatibility, providing opportunities for data mining.
5. Roadmap key needs: seeking community agreement about what tools are most critical to develop, routes to interoperability and to platform sharing, and how to move forward most efficiently - may require a cooperative effort across different communities and continents to be realized.
6. Building expertise: we held a plugfest and poster session devoted to hands-on overview, training and comparison of different platforms and linkage approaches.
The 3 days were arranged so that:
• Day 1 set the scene regarding the current status and vision for multiscale simulation, including the perspective of the multiscale materials modelling and the Integrated Computational Material Engineering communities, with keynote talks from recognised research leaders, and a panel discussion on interoperability. Discussion kicked off with comments on the classification of models, why this is needed and how this could be achieved. Overall the panel and the audience appreciated very much the multiscale approach, however, it was also highlighted that providing and agreeing the classifications is an extremely difficult. Discussion was followed by a discussion on keywords and metadata. The discussion focused on the requirements of industry partners and how they will use such a framework for their simulation and design purposes.
• Day 2 shared current practice, with a morning of presentations and discussions from a cluster of EU-funded multiscale simulation projects regarding current developments and challenges for interoperability and the afternoon poster and plugfest.
• Day 3 addressed next steps and challenges, with a series of overview presentations on Roadmaps, future plans and directions for integration being used to set the scene for chaired open discussions to identify and agree critical next steps for the widespread uptake of Multiscale Simulation going from Materials Modelling through to Industrial Usage. The discussion was mainly centred around the MODA and how a MODA is defined and used. It was argued that it is not a flowchart but it gives indication of the parts involved in the modelling process of a given project. A focus of the discussion centred around interoperability and integration, model development and validation. Given the participation from organizations in Europe and USA, there was a healthy discussion on how European and US initiatives can work together. A strong commitment from both sides of the Atlantic was expressed with the idea to start with a relatively small and focused collaborations, which will help define how to best work together.
The main challenge, both in MMM and ICME, is the coupling and linking
of models describing phenomena at different scales, from the nano-scale
up to the device or macroscopic scale. These two communities need to be
brought together to discuss in detail the challenges to be met in the
structuring and realisation of integrated frameworks that facilitate
linkage of various multi-physics models to achieve the predictive design
of novel materials optimised for specified applications.
Current computational tools typically have their own non-standard schemas for input and output files, including also varying definitions of the model. At the same time, different communities often rely on distinct nomenclature to describe the same model components or physical parameters, which makes it even harder to link and couple disparate tools.
Many multiscale problems require the linking of several levels of code, that span several research and application communities, and including open-source and commercial packages. Some existing tools use their own metadata schemas. Though these schemas are themselves distinct, translating from one to the other is often time consuming and inefficient. In addition, large quantities of data are being generated across different platforms, much of which is potentially useful and re-usable by the wider community, but often in ways that are not immediately obvious in the original context.
Further CECAM workshops in this area would continue the momentum we acheived in Dublin.
A number of funding channels already support work presented at the workshop. In addition to the Cluster of 5+1 EU projects that organised the workshop, several attendees at the Workshop have European funding relevant to the workshop, including the Nomad Centre of Excellence (www. nomad-coe.eu), and a Coordination and Support Action suporting the EMMC (http://emmc.info) which seeks to build a network to capitalise on the strong European position in materials modelling and to allow industry to reap the benefits of modelling. NMBP 07-2017: Systems of materials characterisation for model, product and process optimisation and NMBP 25-2017: Next generation system integrating tangible and intangible materials model components to support innovation in industry are relevant current calls, which involve further open innovation and interoperability developments. Joint research proposals were discussed on a one-to-one basis at the meeting, but were not a topic formally on the meeting agenda.
Will these developments bring societal benefits?
are important for every aspect of modern life. Improved materials are
essential for economic development of heavy industry and
communications, for sustainability, efficient energy production and
progress in drug design. There are many funding opportunities for
research into better materials for energy production. We described some
of these in the funding section above.
There were excellent exchanges between the Europeans and our US counterparts concerning the US experience with standards for
potentials and formats relating to interoperability. The carefully selected US participants made major contributions to all discussions
and helpful suggestions for working with larger collaborations from their experience.
One benefit that has already been achieved following discussions with Aidan Thompson, a co-developer of LAMMPS , is the preparation and release of a LAMMPS patch for the REBO potential. This may seem small, but it brings the efficiency of the standardized LAMMPS code to simulations involving carbon allotropes such as nanotubes which are central to many clean energy applications, such as hydrogen storage in nanotubes.