RheoNet: A Physics-Inspired Neural Network for Predicting the Viscosity of Complex Fluids
Start Date:
- Constructed a benchmark dataset of 79 multicomponent polymeric fluids with measured viscosities from shear rate sweeps.
- Developed parametric symbolic regression with genetic algorithms to discover novel constitutive viscosity models.
pip install parametric-sr: SijieFu/parametric-symbolic-regression
- Designed a hybrid neural network architecture that integrates constitutive models as a physics prior to predict viscosity.