Utilizing an AI material database for sustainable material selection

Sammanfattning

The rising importance of sustainability in the automotive industry has increasedthe demand for more environmentally conscious decision-making during productdevelopment. The materials chosen for a vehicle directly shape its environmentalimpact, resource efficiency, and sustainability performance across its entire lifecycle.Yet in practice, engineers regularly find themselves working with scattered materialinformation, insufficient sustainable data, tight time constraints, and increasingcomplexity in evaluating sustainable material alternatives during early-stage productdevelopment. While a number of tools and methods have been created over theyears to help bring sustainability into engineering workflows, research consistentlyshows that such tools are often not adopted or used to the extent intended in industrialpractice due to challenges related to usability, workflow integration, trustand alignment with engineering needs. This thesis investigates how an AI-enabledSustainable Material Data Ecosystem (SMDE), developed within an automotivemanufacturing context, is perceived and used by engineers in practice, as an industrialcase study in collaboration with Volvo Trucks Technology & Industrial Division(TTI).Empirical data was collected through interviews and user feedback sessions withengineers from different product development domains, and analysed qualitativelyto understand what helps and what hinders the practical use of the system withinexisting workflows. The findings show that the system has genuine potential interms of enabling more convenient access to material information, helping increasethe rate at which alternatives are explored, and being useful for early engineeringinvestigation. However, some issues also emerged during the study, including inconsistentresponse quality, excessive and irrelevant output, concerns around trustingAI-generated information, and the problem of the system’s integration into everydaywork along with other existing tools. Overall, results suggest that it is veryimportant to develop AI-based decision support systems specifically according tothe industrial needs and workflows of the engineers using it to support successfuladoption within product development environments.