Session: 04-04-01: (04-04: Advances in Aerospace Structures and Materials & 04-11: Advances in Mechanics, Multiscale Models and Experimental Techniques for Composites)

Paper Number: 95714

95714 - Msg-Base Design and Analysis of Tailorable Composites 

Tailorable composites are innovative lightweighting space structures. Thanks to their tailorable fiber orientations, they have proven to have the potential for further mass reduction and performance improvement compared with unidirectional fiber-reinforced composites (UDFRCs). However, for tailorable composites, the capabilities of existing design tools are lagging behind the manufacturing techniques. The theory underpinning most existing design tools is the classical lamination theory (CLT), originally established for UDFRCs. A design process is usually guided by semi-empirical rules derived from UDFRCs, e.g., consecutive plies with the same fiber angle should not exceed a certain number, and the fiber angle difference between two adjacent plies should not exceed a certain value. Some tools idealize composite materials as black aluminum and are only applicable to quasi-isotropic stacking sequences. To harness the full potential of tailorable composites, we must develop theories and design methodologies suitable for tailorable composites and further integrate them into commercially available design tools.

The theoretical and modeling foundation is built upon the mechanics of structure genome (MSG), which is an innovative high-fidelity method for multiscale structural analysis. MSG provides a unified method to connect lower-scale structures (e.g., meta-material, layered structure, and cross-sections) and upper-scale structural model (e.g., solid, shell, and beam). This method can use the minimum information at the lower scale (structure gene (SG)) and calculate accurately the fully populated structural properties at the upper scale and stress/strain fields at the lower scale.

A multilevel parameterization and analysis modeling approach will be developed for the MSG-based optimization framework to provide flexible designs for advanced material systems including different types of materials such as metal, lamina, fiber, and matrix. The global structure is the level 0 object, where all design variables are defined. Specific values of a point design are passed top-down to each level of SG and analyses are carried out bottom-up providing structural properties. A library of SG design templates will be available along with the framework, including commonly used types of materials such as stacking of laminae and fiber reinforced matrix.

The proposed design framework will comprise the following several components: 1) the SG builder and solver (SwiftComp) at the core, 2) an SG library containing commonly used baseline or template designs of SGs, 3) the parameterization model along with the users input files for the overall structural design and optimization setup, and 4) an interface written in Python connecting aforementioned components with external global structural analysis tool (e.g., Abaqus, Nastran) and the optimizer (e.g., Dakota). Particularly, the Python interface is one of the key components and focuses in the framework, aiming to: 1) Define design concepts for composite structures and structure gene. This is a key step for parameterization of the hybrid material structures, such as the spatially varied fiber angle or layer coverage. 2) Build SG models and write input files for SwiftComp using parameters defined in the previous item. 3) Parse SwiftComp output and store sectional properties. 4) Read global structural model input files for necessary data and write sectional properties as supplement files using correct format that can be recognized by the structural analysis solver. 5) Read structural analysis output and retrieve data needed. 6) Calculate objectives and constraints that are functions of those data retrieved. 7) Control the overall analysis workflow starting from design parameters to desired structural responses.

This work is significant because it will result in not only advanced models but also a commercial-grade computational design framework for tailorable composites. Specifically: 1) As for modeling tailorable composites, the proposed models will be as efficient as but more accurate than existing theories/models. 2) With the proposed methodology, engineers will design the structure together with the material while consider varying fiber orientations and ply coverage simultaneously, with greatly enlarged design spaces. 3) With the proposed integrated computational design framework, engineers will leverage the power of commercially available tools for real structures made of tailorable composites.

Presenting Author: Su Tian Purdue University

Presenting Author Biography: PhD student, Aeronautical and Astronautical Engineering, Purdue University, <br/>Purdue University, Aeronautical and Astronautical Engineering, MS, 2015<br/>Tongji University, China, Aerospace Engineering, BS, 2013

Authors:

Su Tian Purdue University
Yufei Long Purdue University
Xin Liu University of Texas at Arlington
Liang Zhang AnalySwift
Wenbin Yu Purdue University