Classes & TalksThursday, November 19
ME 250 Seminar Series: Dr. Peng Edward Wang
Thursday, November 19 · 11 a.m.-12 p.m. · Winston Chung Hall
- When
- Thursday, November 19 · 11 a.m.-12 p.m.
- Where
- Winston Chung Hall205/206 · Riverside, CA
About this event
Toward Generalizable and Physics-Aware AI for Intelligent Welding
Abstract: Artificial intelligence (AI) has demonstrated considerable potential for advancing process monitoring, quality assurance, and automation in manufacturing. However, translating AI from laboratory demonstrations to industrial production remains challenging due to limited labeled data, variability across machines and materials, and the lack of physical interpretability in conventional data-driven models. These challenges are particularly pronounced in welding, where complex and highly dynamic thermal processes directly influence product quality. In this seminar, I will present our recent research on developing credible and generalizable AI for intelligent welding, drawing on industrial applications in automotive manufacturing. I will first discuss how advanced visual perception techniques enable robust weld segmentation and defect detection, followed by our development of multi-scale self-supervised learning frameworks that leverage large-scale unlabeled production data to improve quality prediction and defect detection in laser welding and resistance spot welding. Through industrial case studies, I will demonstrate how these approaches enhance model generalizability across varying manufacturing conditions while reducing dependence on costly labeled data. Finally, I will briefly introduce our ongoing research on physics-aware surrogate foundation models for efficient prediction of thermal dynamics in laser welding, highlighting their potential to support virtual sensing and transferable quality prediction.
Listed by UC Riverside (Localist), gathered by the San Bernardino events calendar.
Does something look wrong?
Location
Winston Chung Hall205/206 · Riverside, CA
Get directions