Two new research papers propose advanced methods for monitoring the health of Silicon Carbide (SiC) power modules, crucial components in electric vehicle inverters. The first paper introduces a physics-informed framework that uses cumulative damage indicators and a monotonicity constraint to predict degradation, achieving a 70% reduction in error compared to data-driven methods on Infineon Technologies data. The second paper demonstrates that these cumulative damage features, particularly when used with Neural Ordinary Differential Equations (NODEs), significantly improve transferability across different failure mechanisms like solder fatigue and wire-bond lift-off, outperforming traditional prognostics methods. AI
IMPACT These advancements could lead to more reliable electric vehicle components and improved predictive maintenance strategies in power electronics.
RANK_REASON Two academic papers published on arXiv detailing new research methods.
- arXiv
- Infineon Technologies
- Miner rule
- Neural Ordinary Differential Equations
- SiC MOSFET modules
- SiC power modules
- Silicon Carbide
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
