About this Abstract |
| Meeting |
MS&T25: Materials Science & Technology
|
| Symposium
|
Advances in Dielectric Materials and Electronic Devices
|
| Presentation Title |
Machine Learning Models as a useful tool for predicting the Dielectric Breakdown Strength |
| Author(s) |
Matthew Mileski, Adib Samin |
| On-Site Speaker (Planned) |
Matthew Mileski |
| Abstract Scope |
Breakdown strength refers to the maximum voltage a material can withstand before it conducts electricity. High breakdown voltage ensures that electronic components don't fail prematurely. This is crucial for safety, reliability, and performance in applications involving high-voltage systems. In this work, a physics-informed machine learning model is developed to accurately predict breakdown strength by utilizing atomic-level descriptors supplied from ab initio simulations. Not only is breakdown strength affected by the type of material but also by geometry. To this end, another machine learning model utilizing input from a continuum mechanistic model is developed to predict this effect. This approach could accelerate the discovery of superior materials and geometries for use in future electronic components. |