About this Abstract |
Meeting |
TMS Specialty Congress 2024
|
Symposium
|
2nd World Congress on Artificial Intelligence in Materials & Manufacturing (AIM 2024)
|
Presentation Title |
Explainable Deep Learning Model for Defect Detection During Autoclave Composite Curing Process |
Author(s) |
Deepak Kumar, Yongxin Liu, Sirish Namilae |
On-Site Speaker (Planned) |
Sirish Namilae |
Abstract Scope |
Applications of artificial intelligence (AI) in manufacturing enable new opportunities for automated quality inspection. The practical application of AI in composite processing is currently constrained by the absence of real processing data and the deep learning approach's explainability. In this study, we have used a custom-built autoclave with viewports and an interior lighting setup to develop a novel dataset of the composite curing process using digital image correlation (DIC). Later, using a unique explainable AI technique, a zero-bias deep neural network (ZBDNN) model is developed by transforming the final dense layer of the standard DNN model into a dimensionality reduction layer and a similarity matching layer. This model is then used to identify defects during autoclave composite processing. Several visualization techniques, including Voronoi and t-SNE (t-distributed Stochastic Neighbor Embedding), are utilized to demonstrate that the ZBDNN model was able to identify composite processing defects with high accuracy. |
Proceedings Inclusion? |
Definite: Other |