(YIP) Machine learning enabled structural mechanics
Abstract
The goal of the proposed research is to develop the next generation of computational methods for structural mechanics, enabling unprecedented computational capabilities for the solution of problems for lightweight and deployable structures that are extremely challenging using traditional numerical techniques. This will be achieved by adaptively combining physical laws with data through novel Machine Learning techniques leading to the development of accelerated and robust modeling frameworks, targeted for aerospace applications. In order to be able to realize this vision and leverage it to guide the design of the next generation of computational techniques for virtual twins (modeling replicas) of structures with capabilities in multiscale modeling, parameter estimation, optimization and control problems in structural mechanics, we will work towards 3 interrelated Tasks- 1) Deep Learning enabled model order reduction, 2) design optimization under imprecise probabilistic input data and 3) surrogate models for multiscale structural modeling. In most current and future problems relevant to the U.S. Air Force involving deployable and lightweight structures, there are significant computational bottlenecks that we need to overcome. Enabling a new generation of computational tools will advance areas from design and optimization of lightweight architected 3D-printed structures and composites for aerospace applications, to control problems for morphing and adaptive structures as well as uncertainty quantification in structural health monitoring.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 07, 2023
- Source ID
- FA95502210075
Entities
People
- Nikolas Bouklas
Organizations
- Air Force Office of Scientific Research
- Cornell University
- United States Air Force