| About this Abstract |
| Meeting |
2024 ASC Technical Conference, US-Japan Joint Symposium, D30 Meeting
|
| Symposium
|
2024 ASC Technical Conference, US-Japan Joint Symposium, D30 Meeting
|
| Presentation Title |
Efficient Design of Laminate Composites Using Ensemble Kalman Inversion Method |
| Author(s) |
Parisa Khodabakhshi, Elnaz Seylabi |
| On-Site Speaker (Planned) |
Parisa Khodabakhshi |
| Abstract Scope |
Composites are the preferred choice of material for many engineering applications, including defense, aviation, automotive industries, and civil infrastructure, owing to their potential to achieve superior performance. Proper exploration of the design space for fiber-reinforced composites is necessary for innovative and performance-based composites design. However, due to the high dimensionality of the design space, its exploration is computationally very costly. The focus of this study is the efficient design of laminate composites using an approximate Bayesian method based on the ensemble Kalman inversion (EKI) to map the homogenized laminate properties to the corresponding ply-level properties. Unlike some of the sampling-based approaches such as the Markov Chain Monte Carlo method, EKI does not suffer from the high dimensionality of the parametric space, and it provides an efficient inverse problem framework for the design of laminate composites. Prior knowledge will be incorporated in EKI as constraints in the form of equalities and inequalities to further regularize the inverse problem framework. |
| Proceedings Inclusion? |
Definite: Post-meeting proceedings |