Prof. Dr. Dirk Hartmann (Simcenter Technology Siemens & TU Darmstadt)

Abstract
Thermal management is a critical challenge across numerous engineering domains. In practice, achieving optimal thermal management often is not realized due to significant costs and complexity involved, leaving system operations constrained by thermal limitations. Digital Twins - virtual replicas of physical systems that seamlessly integrate physics-based modeling with real-time data - offer a new paradigm. By enabling model-based control strategies, they hold the promise of unlocking more optimal and efficient system operations. However, most existing approaches remain limited in their scalability, hindering broader adoption.

In this talk, we introduce Autoregressive Operator Inference (AOI), a robust and highly scalable Reduced Order Modeling (ROM) technology designed to overcome these limitations. We demonstrate its practical impact through several compelling use cases drawn from the electrification stack, including electric motors and battery systems. Finally, we explore how these technologies play a pivotal role in the rapidly emerging fields of robotics and physical AI, where real-time, accurate thermal modeling will be indispensable.

10.06.2026, Time: 16:00,
Room: Seminarraum Prigogine (MPI)