Session: 04-01 Emerging Technologies: Smart AM

Paper Number: 93987

93987 - Additive Manufacturing Process Monitoring Based on Compressed Sensing and Physics-Constrained Dictionary Learning 

Compressed sensing is a new technique for data acquisition, which reconstructs the original signal from a small number of measurements. Compressed sensing takes advantage of the sparsity of data in the reciprocal space and achieves data compression. The performance of compressed sensing however depends on the design of measurement and basis matrices. To maximize the reconstruction accuracy, dictionary learning methods have been developed to customize basis matrices for specific signals instead of using predefined generic ones as in traditional compressed sensing. Nevertheless, the theoretically optimal results from dictionary learning can be difficult to obtain because of the restrictions in physical realization, such as the number of sensors, physical sizes of sensors, and sensor accessibility in the manufacturing environment. In this work, a physics-constrained dictionary learning (PCDL) approach is developed to optimize the measurement and basis matrices separately with the considerations of these restrictions. The uniqueness of the PCDL is that there is only one non-zero entry in each row in the optimized measurement matrix so that the physical locations for the sensor placement are directly determined. Additional constraints of sensor accessibility are also incorporated. The PCDL approach is also extended to solve the compression and classification problems simultaneously for process monitoring and diagnostics. A Bayesian optimization method is applied to obtain better measurement matrices. The developed framework of compressed sensing with PCDL is demonstrated with thermal imaging for additive manufacturing process monitoring. High-resolution images are reconstructed with the basis matrix and the recovered coefficient vectors to allow for efficient monitoring. The different machine states can also be identified simultaneously based on the classification matrix and the coefficient vectors to enable efficient diagnostics.

Presenting Author: Yan Wang Georgia Institute of Technology

Presenting Author Biography: Yan Wang is a Professor of Mechanical Engineering and leads the Multiscale Systems Engineering Research Group at Georgia Institute of Technology. His research areas include compute-aided design, computer-aided manufacturing, modeling and simulation, materials design, and uncertainty quantification.

Authors: