Data-driven Multiscale Modeling Simulation Platform

Research

Multiscale materials modeling simulation software is a CAE software connecting multiple simulation methods for the integral description of materials behavior from the nanoscale to the micro-, meso- and finally to the macro-scale. It models the complex behaviors of advanced materials, and processes the data quickly and effectively. The current global market is approximately 700 million US dollars. 


The data-driven multiscale modeling and simulation platform developed in this project utilizes multiscale analysis and statistical analysis to improve computational efficiency. The platform drastically reduces computational cost compared with traditional Finite Element analysis (x10000~40000 faster), and provides modeling capability for complex heterogenous materials and traditional materials under complex boundary conditions. With the pioneering data-driven order reduction method, SCA, the platform is able to achieve high accuracy in performance prediction (error<1%) at high speed. Its compatibility with current major commercial FEA software makes it easy to use. The powerful features of this software would drive research, product design and development in the industries of transportation, civil, energy, etc. 

Publications

Cheng, Yi, et al. "A sequential homogenization of multi-coated micromechanical model for functionally graded interphase composites." Computational Mechanics 64.5 (2019): 1321-1337.