Attend the Smart Manufacturing Seminar Series event on Wednesday, March 31, 2021, with Shenghan Guo from Rudgers University to talk about how data can improve smart manufacturing.
Data-Driven Anomaly Detection and Predictive Analytics in Smart Manufacturing
Presented by Shenghan Guo, Department of Industrial and Systems Engineering, Rutgers University, New Brunswick
Wednesday, March 31, 2021
Attend on Zoom
Manufacturing plays a vital role in almost every sector of the U.S. economy, stretching from aerospace to pharmaceuticals and beyond. Advanced manufacturing—which includes both new manufacturing methods and production of new products enabled by innovation—is an engine of America’s economic power. However, advanced manufacturing is experiencing dramatic dynamics in the manufacturing landscape. Process instability and degradation commonly lead to part defects and system downtime, limiting a wider adoption of the advanced technology in industry. Recent development in inline sensing has enabled real-time data collection from manufacturing processes. These data, in the form of time series or thermal images, convey valuable information about process health status and part quality. Shenghan Guo’s research developed statistical and machine learning approaches that exploit the valuable information in data to facilitate process monitoring, defect prediction, and preventative maintenance. The methodology scope mainly includes (1) in-situ defect detection in laser-based additive manufacturing (AM), (2) data-driven prediction of melt pool thermal dynamics in laser-based AM, and (3) nonparametric modeling and detection of process degradation in automotive manufacturing. The developed methods have been validated with real data and demonstrated for effectiveness. They can be generalized to various manufacturing processes; the scope can be extended to in-process quality improvement and design optimization in near future.
About the speaker
Shenghan Guo is a doctoral candidate in the Department of Industrial and Systems Engineering (ISE) at Rutgers University–New Brunswick. She is expected to graduate in May 2021. She received her MS in Engineering Sciences and Applied Mathematics from Northwestern University in 2016, an MS in Financial Mathematics from the Johns Hopkins University in 2014, and her BS in Financial Engineering from Jilin University, China, in 2013. She had a remote six-month internship with Oak Ridge National Lab in 2020.
Her research focuses on data-driven predictive analytics and physics-informed statistical learning to facilitate smart manufacturing. Her methods have led to downtime reduction in powertrain manufacturing, defect prediction in laser-based additive manufacturing, and nondestructive quality evaluation in resistance spot welding. Her research papers on “in-situ porosity detection for laser-based additive manufacturing” and “nonparametric detection of process deterioration in leak test for powertrain” were featured in the ISE Magazine.
She was the winner of the Tayfur Altiok scholarship at Rutgers ISE and the department nominee of Rutgers’ Louis Bevier Dissertation Completion Fellowship in 2019. She won the 2019 IISE Quality Control and Reliability Engineering (QCRE) Data Challenge. She was the second-place winner in the 2020 IISE Data Analytics and Information Sciences (DAIS) Student Data Analytics Competition and a finalist for the 2018 INFORMS Quality, Statistics, and Reliability (QSR) Best Paper Competition. She has contributed to efforts to broaden participation in research, including supervising graduate and undergraduate researchers and giving presentations to undergraduates about her work.