Tool to Predict Transient Spatial-Temporal Boundary Conditions for Processing Autoclave-cured Composite Parts

Global Engineering and Materials, Inc. (GEM), along with its team members, Professor Jinhui Yan at University of Illinois Urbana-Champaign (UIUC), National Institute for Aviation Research (NIAR) at Wichita State University, Professor Dianyun Zhang at Composite Manufacturing and Simulation Center at Purdue University (PU), and Professor Yuri Bazilevs at Brown University, proposes to develop an integrated multiphysics and data-driven autoclave process simulation toolkit (SMARTCLAVE) for multiple composite parts with curved geometries. The novelties of the proposed autoclave process simulation and response prediction tool (SMARTCLAVE) include: 1) a coupled immersogeometric thermal CFD and aerodynamics-aided heat transfer module with a consistent LES turbulence model (without ad hoc viscosity) to capture the turbulence-induced local boundary conditions and the resulting thermal and pressurization response of multiple parts in an autoclave; 2) a hybrid shell and solid element modeling approach in CFD and its associated immersogeometric modeling for a multi-layer composite curing system; 3) an efficient coupling strategy between the CFD module and the customized Abaqus for temperature and pressure mapping; 4) micro-macro characterization of the degree of cure, a temperature-dependent constitutive model, and mechanical performance with fabrication-induced defects; 5) an application programming interface connecting to the commercial composite process simulation software COMPRO for defects prediction; 6) hybrid physics-informed neural networks (PINNs) for data fusion to enhance the accuracy of the high-fidelity model and generate surrogate models for process tailoring; and 7) optimization of autoclave processing and placement of multiple parts in an autoclave. The resulting SMARTCLAVE toolkit will be validated first by using data collected from PU and NIAR at coupon and component levels. The high-fidelity toolkit along with its multi-level PINNs will be demonstrated using data collected from representative composite structures during and after autoclaving.