ISE Seminar “3D Point Cloud Data Modeling, Analysis and Control for Quality Qualification and Improvement”


10:40 am-11:30 am
Add to Outlook/iCal
Add to Google Calendar


Weil 406
1949 Stadium Dr
Gainesville, FL 32611


ISE Seminar “3D Point Cloud Data Modeling, Analysis and Control for
Quality Qualification and Improvement in Additive Manufacturing”

Dr. Jianjun Shi
The Carolyn J. Stewart Chair and Professor
H. Milton Stewart School of Industrial and System Engineering
Georgia Institute of Technology

Additive Manufacturing (AM) has transformed the way we design, prototype, and produce complex parts with unprecedented geometries. However, inherent process challenges persist, such as dimensional accuracy, part quality, and process stability. The functionality and quality of the final product in AM are heavily reliant on 3D printing process conditions and material properties. Advancements in contactless 3D scanning technology have made in situ 3D profiling of printed parts readily available. This data, represented as high-dimensional, unstructured 3D point cloud data, holds immense potential for enhancing process control and quality assurance in AM. To leverage this potential, novel methodologies for data fusion and analytics must be developed to effectively model, analyze, and control 3D profile data alongside other heterogeneous process sensing data. In this presentation, a series of recent research endeavors will be explored, including:
• functional qualification of 3D-printed parts via physical and digital twins using contrastive learning and hard sampling techniques;
• 3D profile evolution modeling considering heterogeneous inputs by using the Koopman operator theory and machine learning algorithms;
• predictive control of 3D profiles propagation using nonlinear dynamic model with heterogeneous active control inputs; and
• dynamic 3D shape morphing behavior modeling, optimization and control using continuous normalizing flow methods for 4D printing.
All methodologies discussed are based on advanced data fusion and machine learning techniques, integrating physical and engineering domain knowledge. The seminar will present both theoretical investigations and experimental validations, highlighting the potential for significant advancements in additive manufacturing processes.


Hosted by