A MACHINE-LEARNING ENABLED MULTISCALE-MULTIPHYSICS MODELING PLATFORM FOR DAMAGE DETECTION AND MITIGATION IN MULTIFUNCTIONAL COMPOSITE STRUCTURES
Abstract
This research will create a machine learning-enabled, robust multiscale-multiphysics computational platform for multifunctional piezo-composites to enable: (i) detection of the initiation and propagation of damage in the 3D microstructure using damage sensors located on the surface, and (ii) subsequent activation of electro-magnetic field induced deformation to contain the growth of local material cracking. The multifunctional composites to be modeled will consist of microstructural piezoelectric fibers embedded in a magneto-rheological elastomer (MRE) matrix. The platform will incorporate the following modules: (a) “3D coupled electro-magneto-mechanical FE model with interface failure and crack growth in the piezocomposite microstructure”; (b) "parametrically upscaled coupled constitutive damage model (PUCCDM) from microscale electro-magneto-mechanical simulations” where machine learning (ML) tools will establish PUCCDM coefficients as functions of representative aggregated microstructural parameters (RAMPs); (c) “machine learning-based correlation maps between subsurface damage indicators and surface electromechanical response”; (d) “damage mitigation through activation of magnetic field-induced deformation.” In summary, the computational platform will provide an autonomous damage detection and mitigation capability that will be of significant advantage with respect to non-destructive evaluation and self-healing in aerospace applications. The PI will engage with the AFRL and aerospace companies for transferring this technology
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 20, 2023
- Source ID
- FA95502210351
Entities
People
- Somnath Ghosh
Organizations
- Air Force Office of Scientific Research
- Johns Hopkins University
- United States Air Force