Multifidelity multiscale modeling of nanocomposites for microstructure and macroscale analysis

A high-fidelity multiscale modeling framework that integrates information from atomistic simulations pertaining to polymer chain sliding and bond dissociation is utilized to study damage evolution and failure in carbon nanotube (CNT)-reinforced nanocomposites. The nanocomposite constituents (microfiber, polymer, and CNTs) are explicitly modeled at the microscale using representative unit cells (RUCs). The modeled constituents are subsequently employed in a multiscale framework to describe damage initiation and propagation in these systems under transverse loading. Two CNT architectures, randomly dispersed and radially grown, are investigated. Damage initiation sites and damage evolution trends are studied, with results indicating that the presence of CNTs causes a unique stress state at the sub-microscale. This can lead to accelerated damage progression, which can be mitigated by architectural reconfiguration of the CNTs. Additionally, the Schapery potential theory is extended to develop an orthotropic nonlinear damage model that captures global behavior of the nanocomposite RUCs in a computationally efficient manner, and can be utilized as a numerical surrogate for structural scale nanocomposite analysis.