Concurrent vs Serial Multiscale Modelling of Intelligent MFC Composite CFRP-PZT (General Concept)

Description

In my recent work, I’ve explored two distinct approaches for multiscale modeling of an Intelligent Composite Microfiber Composite (MFC) that used piezoelectric principles. Here, a PZT sensor is mounted onto a Carbon Fiber Reinforced Polymer (CFRP) specimen, bonded by an adhesive layer to create an integrated, responsive composite system. I’ve investigated both serial and concurrent multiscale modeling techniques to understand the unique properties across different scales.

Serial Multiscale Modeling Approach
The serial method is relatively straightforward. Each scale—from macro to micro—is analyzed independently, with data transferred between scales through homogenization or Molecular Dynamics (MD) simulations. This process allows the microstructural properties to inform the macro-level response without direct real-time linkage, simplifying data management but potentially limiting fidelity in capturing simultaneous interactions.

Concurrent Multiscale Modeling Approach
Conversely, concurrent multiscale modeling presents a complex but comprehensive alternative. In this technique, data between scales—such as stress tensors—must be transferred in real-time to reflect simultaneous interactions. Based on existing literature, some methods rely on Cauchy stress tensors and other sophisticated techniques, which add complexity. For this project, I’ve implemented concurrent modeling in ABAQUS through the Submodeling technique:

• At the macroscale, the overall MFC structure is modeled, capturing global behavior.
• Moving to the mesoscale, I analyzed CFRP and PZT components, with CFRP as a woven structure and PZT as a cubic material.
• Microscale modeling then focuses on the unit cells of both CFRP and PZT.

Each level is linked using time-dependent boundary conditions via submodeling, with stress tensors exchanged simultaneously between scales.