Andreas Gebhard

• In tribology laboratories, the management of material samples and test specimens, the planning and execution of experiments, the evaluation of test data and the longterm storage of results are critical processes. However, despite their criticality, they are carried out manually and typically at a low level of computerization and standardization. Therefore, formats for primary data and aggregated results are wildly different between laboratories, and the interoperability of research data is low. Even within laboratories, low levels of standardization, in combination with ambiguous or non-unique identifiers for data files, test specimens and analysis results greatly reduce data integrity and quality. As a consequence, productivity is low, error rates are high, and the lack or low quality of metadata causes the value of produced data to deteriorate very quickly, which makes the re-use of data, e.g. for data mining and meta studies, practically impossible. In other fields of science, these are mitigated by the use of Laboratory Information Management Systems (LIMS). However, at the moment, such systems do not exist in tribological research. The main challenge for the implementation of such a system is that it requires extensive interdisciplinary knowledge from otherwise very disparate fields: tribology, data and process modelling, quality management, databases and programming. So far, existing solutions are either proprietary, very limited in their scope or focused on merely storing aggregated results without any support for laboratory operations. Therefore, this thesis describes fundamentals of information technology, data modelling and programming that are required to build a LIMS for tribology laboratories. Based on an analysis of a typical workflow of a tribology laboratory, a data model for all relevant entities and processes is designed using object-relational data modelling and object-oriented programming and a relational database is used to provide a reference implementation of such a LIMS. It provides critical functionalities like a materials database, test specimen management, the planning, execution and evaluation of friction and wear tests, automated procedures for tribometer parameterization and data transmission, storage and evaluation and for aggregating individual tests into test sets and projects. It improves the quality and long-term usability of data by replacing error-prone human processes by automated variants, e.g. automated collection of metadata and data file transmission, homogenization and storage. The usefulness of the developed LIMS is demonstrated by applying it to Transfer Film Luminance Analysis (TLA), which is a newly developed advanced method for the analysis of the formation and stability of transfer films and their impact on friction and wear, but which produces so much data and requires such a large amount of metadata during evaluation that it can only be performed safely, quickly and reliably by integration into the presented LIMS.