Knowledge discovery assistants for crash simulations with graph algorithms and energy absorption features

Abstract We propose the representation of data from finite element car crash simulations in a graph database to empower analysis approaches. The industrial perspective this work is narrow gap between uptake modern machine learning methods and current computer-aided engineering-based vehicle development workflow. main goals for are achieve searchability enable pattern trend investigations product history. In context, we introduce features enrich provide summary overview stages. These based on energy output solver and, example, filtering input by identifying essential components vehicle. Additionally, these features, fingerprints simulation studies that assist summarizing exploration design space facilitate cross-platform as well load-case comparisons. Furthermore, combine with use weighted heterogeneous visualization identify outliers cluster according their similarities. present results real-life stages an automotive company.