|
A flow simulation with a Bird-Carreau viscosity law is defined; the power index is low and a Picard iteration is therefore asked for the viscosity. The calculation diverges quickly, despite the selected Picard scheme? How can this be circumvented?
This may happen if the Bird-Carreau viscosity involves a low power index (e.g. 0.15), and a cut-off shear-rate far below than the typical shear rate of the flow. Indeed, under such circumstances, the viscosity drops by several orders of magnitudes between the first iteration (run at the zero-shear viscosity) and the subsequent ones. Applying an evolution scheme on the power index or on the time-constant of the Bird-Carreau law may improve the stability of the solver.
|
|
