Automated Documentation of Research Processes Using RDM

Abstract

Published research results usually represent only a fraction of the data generated at a research institute. The unpublished data created in the process of producing the final result, however, often contain valuable information that can be reused. Through research data management, all these data should be stored centrally according to the FAIR principles (Findable, Accessible, Interoperable, Reusable). However, a significant part of knowledge is often not found in the data, but in the processes that led to their generation. It is therefore important to map these processes to archive and document this knowledge in a structured way. Procedures for documenting scientific processes already exist and are actively used at research institutes. However, these are often analogue or paper-based and hence do not meet the requirements for FAIR data management. At the Institute for Microstructure Technology of the KIT, such a paper-based procedure is used to document the production of microstructure components. During their manufacturing, it is essential to adhere to the correct process parameters in order to enable error-free production. Therefore, a so-called job ticket always accompanies the production of components. On this job ticket, the correct process sequence is listed and a detailed description of the respective process step is given. Depending on the component to be produced, a distinction is made between different types of job tickets according to internal conventions. On the one hand, there are so-called green job tickets, which describe a standardised process sequence, and on the other hand, blue job tickets, which are intended to document experimental manufacturing processes. The process sequence on the blue job tickets is initially empty and is filled in during the manufacturing process. Common to both types of job tickets is that they are stored in the institute's archive after completion of the component production. However, since the job tickets are paper-based, the corresponding archive of job tickets cannot be searched quickly and, given the sheer volume of archived job tickets, represents an unmanageable collection of data. The existing system for process documentation is therefore to be implemented with the help of the research data infrastructure Kadi4Mat [1] in accordance with FAIR principles, thereby making the available process knowledge more accessible.