An Efficient High-fidelity Design Tool for Advanced Tailorable Composites

To harness the potential of advanced tailorable composites for lightweighting aerospace structures with enhanced performance, AnalySwift proposes to develop an efficient high-fidelity Design tool for Advanced Tailorable Composites (DATC). Building upon our accomplishments in Phase I, DATC will be developed based on the efficient high-fidelity constitutive modeling capability of mechanics of structure genome (MSG) and its companion code SwiftComp, the versatile structural analysis capabilities of two finite element analysis (FEA) packages Abaqus and MSC.Patran/Nastran, the general-purpose optimizer Dakota, and the state-of-the-art machine learning (ML) package TensorFlow. DATC will integrate all these tools into a unified and intuitive design framework, facilitating design setups and enabling innovative designs of structures made of tow-steered composites. MSG computes the location-dependent shell properties for FEA and Dakota performs optimization for varying fiber orientations, ply coverages (varying ply thickness and ply drops), and different materials with manufacturing constraints. ML provides ultra-efficient surrogate models to accelerate the optimization from days to hours. The associated computer codes are developed with an open architecture to allow users to add new functionalities for specific problems. Several realistic aerospace structures will be employed to verify the validate the developed tool and demonstrate the benefits from tailorable composites in reducing the structural weights and/or improving the load-bearing capacity. We expect to release DATC by the end of the Phase II as a user-friendly graphic user interface (GUI) plug-in for MSC.Patran/Nastran and Abaqus so that engineers familiar with these two FEA codes can easily use DATC to carry out analysis, parametric studies, and design optimizations of highly tailorable composite structures.nbsp;nbsp;nbsp;