Towards robust life cycle assessments: Adapting the Pedigree Matrix for Time Series Data

Digital twins represent virtual replicas of real-world production systems, providing a foundation for data-driven process analysis and optimisation. A key application of digital twins is supporting life cycle assessments (LCA), which evaluate the environmental impacts of products and processes. Traditionally, LCA have relied on literature-based or estimated data, which are often outdated, incomplete, and difficult to verify. These static data sources cannot reflect dynamic changes in production and therefore limit the precision and relevance of environmental assessments. In contrast, continuous, sensor-based time series data generated directly on the shop floor offer a significantly higher level of accuracy, traceability, and timeliness. They capture actual process conditions with standardised measurement methods and quantifiable measurement uncertainty, enabling LCA to reflect real-time, site-specific environmental performance. This opens the opportunity to improve the transparency, precision, and continuous updateability of environmental assessments within digital twins. However, existing methods for assessing data uncertainty, such as the pedigree matrix, were designed for literature- or estimate-based data and fail to capture the specific quality features of time series data. Classical pedigree dimensions, such as source credibility or estimation character, are inadequate for empirically measured, traceable data. This paper therefore aims to adapt selected dimensions of the pedigree matrix to account for the characteristics of time series data, and to develop an uncertainty score based on these adaptations. This score will allow the information-technical uncertainty of continuous production data to be systematically and transparently quantified. By formally linking data quality to LCA model validity, companies can make more informed decisions on which data to use for robust, up-to-date, and dynamic sustainability assessments in the context of digital twins.