Session: 12-23-01: Symposium on Multiphysics Simulations and Experiments for Solids

Paper Number: 146128

146128 - Machine Learning-Based Prediction of Multiphysical Responses in Vacuum Assisted Resin Infusion Molding Process 

The Vacuum-assisted Resin Infusion Mold (VARIM) process is widely employed in industries for manufacturing composite products like wind blades. Under specific processing conditions, defects such as micro-distortions can arise, resulting in scrapped or defective parts and significantly escalating the costs associated with wind blade production.  The VARIM process and their influence on stresses, strains, displacements, and temperature distributions provide a complex problem statement of multiphysical coupling. To enable process control and optimization, a high-fidelity digital twin for manufacturing process is needed. With this digital twin, operators can potentially preemptively address defect formation during the VARIM process. Once a validated high-fidelity multiphysics simulation methodology is established, the development of predictive tools, or "digital twins," for the VARIM process becomes feasible. In a digital twin of the VARIM process, all pertinent inputs—such as composite layup (ply orientations, etc.), geometry, process conditions (vacuum pressure, mold temperature), and material properties of fibers and matrix (viscosity, modulus, permeability, shrinkage as a function of degree of cure)—are considered. Based on these inputs, the digital twin can provide predictive capabilities regarding various manufacturing outcomes. However, multi-physics-based simulations currently necessitate extensive computation times (ranging from hours to days, or even longer), rendering them unsuitable for real-time process control. As an initial stride toward developing process control capabilities, there is a demand for rapid modeling methods and predictive tools for temperature distribution over time, facilitating their implementation in controlling infusion and curing cycles for composite blades to prevent defect formation.

In this presentation, we outline a machine learning (ML) methodology utilizing a deep convolutional neural network model to forecast the spatio-temporal temperature distribution throughout the vacuum-assisted resin infusion molding (VARIM) process. The ML model is trained using data derived from physics-based high-fidelity simulations, effectively serving as a digital twin of the blade manufacturing process once fully trained. Validation is conducted by comparing simulation outcomes with experimental data obtained from a unidirectional glass fiber composite laminate plate. Furthermore, critical processing parameters of the VARIM process are identified and linked to the final temperature mapping after the resin curing process using a CNN-based model of a similar design. This enables manufacturers to promptly anticipate the effects and final outcomes of key processing parameters, facilitating real-time adjustments to operating conditions to achieve the desired and optimized temperature profile in the composite part at various curing phases. With a predictive accuracy of 94% and computational speeds exceeding 100 times faster than physics-based simulations, the ML approach outlined in this study establishes a comprehensive framework for a digital twin pertaining to temperature distribution in the composite manufacturing process. The presented ML model allows for extensive study and optimization of the process effects. As such, this work demonstrated the great potential of the proposed ML model as a digital twin of the VARIM process.

 

Presenting Author: Dong Qian The University of Texas at Dallas

Presenting Author Biography: Dr. Dong Qian is professor and associate department head of mechanical engineering at the University of Texas at Dallas. He received his B.S. degree in Bridge Engineering from Tongji University in China in 1994, his M.S. degree in Civil Engineering from the University of Missouri in 1998 and Ph.D. degree in Mechanical Engineering from Northwestern University in 2002. Shortly after his graduation, He was hired as an Assistant Professor of mechanical engineering at the University of Cincinnati and promoted to Associate Professor with tenure in 2008. In the Fall of 2012, he joined the newly established Mechanical Engineering Department as a tenured Associate Professor at the University of Texas at Dallas and was promoted to full Professor in 2015. Dr. Qian has conducted research and published extensively in the general areas of computational mechanics and materials, including nonlinear meshfree and particle methods, hierarchical and concurrent multiscale methods, computational nanomechanics, fatigue and simulation-based life prediction, surface engineering, additive manufacturing, data-driven modeling and simulation and applications of machine learning. His work has been widely cited and appeared in high-impact journals such as Science, Nature Communication, Advanced Materials, Applied Mechanics Reviews, Journal of Applied Mechanics, Computer Methods in Applied Mechanics and Engineering. Dr. Qian is a fellow of the American Society of Mechanical Engineering. He currently serves as a member of the editorial board for the Journal of Computational Mechanics and is an associate editor for the Journal of Computer Modeling in Engineering and Sciences.

Authors:

Sahil Kamath The University of Texas at Dallas
Guiherme Caselato Gandia The University of Texas at Dallas
Niloufar Adab The University of Texas at Dallas
Hongbing Lu The University of Texas at Dallas
Dong Qian The University of Texas at Dallas