The present initiative to digitize materials is of particular strategic importance for Germany as a business location, which can only be adequately addressed by pooling the efforts of all stakeholders. Due to the diversity of materials and the processes associated with their production and use, the complexity of the lifecycles of materials and the multitude of academic and industrial participants impose a huge challenge. There are many scattered activities, but there is a lack of bundling of the efforts of all stakeholders, which can eliminate redundancies and a lack of acceptance and advance a common basis in the digitalization of materials. Our goal in meeting this challenge is to contextualize material data in a way that is consistent with all stakeholders: All necessary information on the condition of the material, including production and application-related changes, must be made available via a uniform, machine-readable description. Another central task is to lay the foundation for the development and transfer of an efficient ontology for materials and the associated processes into the applications.
The innovation platform MaterialDigital was launched on 1 July 2019 to initiate appropriate measures. The platform serves primarily to structure the topic area but serves as well as an interaction and communication platform. In addition, it helps to identify concrete work tasks that can be developed in agile working groups and projects of different composition linked to the platform. The dimensions of a data room can be spanned by coupled scientific (specific materials: e.g. Cu, steel, fiber composite, functional ceramics,...) and industry-led pilot projects (along value chains: e.g. raw material across supply chains and OEMs up to recycling).
The aim of our project is to develop a sustainable platform that brings together and supports interested parties from industry and academia in the sustainable implementation of digitalization tasks for materials. At the same time, tools are being developed with which the digitalization of materials can also be put into practice and which, as added value, underpin the economic connectivity of the MaterialDigital platform as a prerequisite for a sustainable business model.
Data generation must be transparent, traceable and repeatable. Automated, upstream and downstream processing must also be able to be openly viewed using freely available software tools. It must be ensured that the flexibility of the analysis and processing of data is not restricted, but that interoperability and transparency are ensured by a standardized description based on metadata*. Data from experiments and simulations as well as the associated metadata must become seamless and fully compatible ("interoperable"). This requires a physically sound and standardized description of the representation of an experimental, simulated or theoretical material and its behavior in all relevant physical facets.
As simulations and theoretical methods play an increasingly important role, available theoretical methods need to be transparently documented and validated. In addition to validation, further development and validation of scale bridging approaches is also required.
The development of "digital twins", i.e. virtual images of a material, component or product, on which certain design ideas or impacts of properties of different materials can be checked and varied in each processing step on the computer before real production takes place, is still in its early stages, especially for complex materials, and is particularly challenging for academic stakeholders.
Concepts must be developed for the quality control of the data stored in data platforms, which quantify uncertainties or errors related to the characterization method or the material model, and which are recorded and documented together with the data.
Ownership and access rights to the data must be clarified and made transparent. The interests of the industry for the protection of data and procedures must be taken into account.
Open platforms must be created for the permanent archiving, use or exchange of data. These make it possible, for example, to use existing and freely available data to validate one's own research results.
Development of a roadmap for the digitalization of the materials and their annual extension.
In order to achieve significant progress in the digitalization of materials, close coordination of all partners and stakeholders and a high degree of coherence of the individual measures are important. By implementing this concept, a standardized exchange of data, knowledge and tools for the digitization of materials that functions across domain boundaries is to be achieved. The dynamics of digitization, the multitude of different needs and relevant stakeholders make it necessary to react in short intervals and to face changes positively and openly. An efficient management must therefore make use of new agile approaches and choose an appropriate responsive organizational form.
The platform supports identification of essential key tasks that are relevant for a successful implementation of the digitalization of materials. An important component here is public relations work directed at all relevant stakeholders, which provides information on the platform's work on specific topics and addressees and, on the other hand, collects, bundles and passes on information supporting further development of the field. With regard to the sustainability of the overall project, the platform distinguishes between tasks that can be implemented in projects of stakeholders of different constellations and structural key and service tasks that, as enablers, should promote and accelerate the implementation of the projects in a standardized environment. On the basis of a holistic concept that includes both structuring elements and tools, the platform develops a sustainable business model for the economic connectivity of the platform beyond the funding period.
In addition, the platform is organizing a technology hub that will provide partners with tools for digitizing materials in the overall process. The technical infrastructure for the realization of this hub must be developed by the platform's partners or on their behalf in the first year of funding and be maintained and supported during the lifetime of the project. The technical realization of the hub must be carried out in such a way that the needs of industrial users in particular are taken into account.
In accordance with the BMBF, the platform organizes a management advisory board which supports the focusing of the activities in the project on the maximum benefit and ensures an optimal transfer of results into industrial application. For this purpose, members of the platform develop proposals. The management advisory board meets once a year together with the members of the platform and is involved in the workshops of the project.
The members of the platform, the management advisory board and representatives of the projects jointly develop a continuously evolving roadmap for the digitalization of the materials, which takes into account relevant research topics as well as the transfer into applications and economic use. This roadmap takes into account both the national strategy and international developments, especially at the European level. The platform supports the digitalization of materials through regular publication of the roadmap. The roadmap will be discussed and reflected with community actors and associations. The objective is to develop a uniform roadmap for the national strategy with international reach, which can then be extended with the stakeholders.
Data analysis methods play an important role in both the experimental and simulation-based digital description of materials, but have so far been poorly structured. The platform should contribute to a standardized description of data processing methods in materials research. The goal is the quality assurance of the processes and the output data, the acquisition and definition of their accuracy as well as the interoperability between applications.
Data-driven procedures in materials research benefit enormously from the use of data from different sources. Their combination requires a standardization of data formats, machine-readable metadata and the extracted information. In view of the abundance of different data sources and their broadness, a specific structuring of the data and associated metadata and information must be developed by the respective domain experts. In order to be able to take international advances in materials research and the internationally fragmented value chains in industry into account, a worldwide exchangeability and interconnectivity - interoperability - is necessary. A separate national approach for ontology and standardization of the description of material data therefore makes no sense, which is why the initiative would like to work closely with European initiatives such as the European Materials Modelling Council (EMMC) right from the very beginning. However, Germany can play a leading role here and thus give German industry a valuable lead as a pioneer.
The standardization of models and the analysis and simulation methods based on them is an important element in digital material development. For simulation models, this project can draw on the model classification of the European Materials Modeling Council (EMMC), which is based on the underlying physical equations. These groundbreaking efforts were supported by Fraunhofer IWM staff and translated into software tools, making them available and applicable to the entire community. An analogous classification criterion for data-based models must still be developed with the involvement of all stakeholders, to which this initiative can contribute. In any case, the (secure) area of application of the models must be specified and they must be classified themselves.
Data analysis and material models are both based on elementary algorithmic building blocks. Many of these elementary algorithms must be executable directly on the data and on the computers that provide them. The MaterialDigital platform will make algorithms available to all users as trusted hosts.
With the establishment of the operative core of the platform, potential interested parties on the supplier and customer side are invited to actively participate. These contacts should take place at the management and operational level of the respective organizations / institutions in order to generate acceptance at all levels. Workshops and meetings should be held as topic-specific as possible such that concrete aspects of an industry or a material class can be addressed. Of great importance will be the close cooperation with the technical-scientific societies, which have wide-ranging contacts in their fields as well as many years of experience in networking. The platform serves as a 'single point of contact' to channel the reception and processing of requests from potential users; the stakeholders can make their requests problem-oriented and can be brought into contact with the corresponding experts through the platform or guided to already existing solutions.
In the context of conferences, exhibitions and fairs the platform is actively presented as innovation engine in particular for small and medium-size enterprises. These events are also intended to bring together job-seeking experts and companies. The digitalization of the materials is not limited to national borders; a high degree of networking with comparable initiatives on a European and global level is sought, whereby we strive for a pioneering function for the initiative planned here.
In addition to the development of common standards, an important task of the project is the preparation of tools that enable potential users to participate efficiently in the initiative on a uniform basis. In addition, the provision of tools is highly relevant for the overall economic viability of the project. The platform will develop an essential core of software tools and provide the project partners with a nucleation point for a common software platform. Some of this work will be carried out by the platform's partners or will be commissioned. Central contact point for the coordination and exchange of information will be a website of the platform MaterialDigital on which the results will be made available to all interested parties.
The MaterialDigital platform will coordinate and advance the development of standardized software tools for the generation, storage, management, analysis and distribution of material data. The objective of these activities is to provide both data and service providers and industrial users with a standardized, flexible and easy-to-use toolbox for implementing digitization projects. The development of these tools takes into account the wide range of material science questions, the multitude of available data sources and simulation and analysis methods, but also the dynamic development of the field. In order to enable their inclusion in industrial processes, the interests of industrial users with regard to the confidentiality and security of their data and queries are already taken into account in the design of the tools.
These requirements for flexibility and confidentiality cannot be met by a central data server that stores and processes all data in one place. The infrastructure of the MaterialDigital platform is therefore realized through a network of decentralized data servers, each operated by a local institutional or industrial research group. Local hosting is a basic prerequisite for integrating industrial partners at an early stage and meeting their demands for confidentiality of their data. The decentralized servers can be addressed by a central server, which is hosted and administered by the neutral platform MaterialDigital. The central server does not only provide the links to the relevant data servers but also enables the user to use active software agents. After approval by the owner, these agents run on the local servers, aggregate or process the data and make a new data record available to the external user. This approach gives users a completely new quality of accessing dislocated material data to reveal previously inaccessible correlations. The user does not have to access the individual servers, download the data and analyze them on his local computer, but can do this in a single step with an agent. Another important advantage of this agent-based approach is that it enables the direct use of modern methods of machine learning or data mining for material science issues on much larger data sets.