Identification of stagnant regions of viscoplastic fluid flows in production lines and equipment is of paramount importance owing to potential material degradation and process contamination. The present work introduces an assessment strategy to identify, classify and quantify unyielded regions with the objective of optimizing the flow conditions with the purpose of minimizing stagnant regions. Flow of Carbopol ® 980 in a T-bifurcation channel is adopted to illustrate the procedure. The rheological behavior of Carbopol ® 980 was simulated using the Herschel-Bulkley viscoplastic model regularized by Papanastasiou's exponential approach. The analysis shows that three distinct types of stagnant unyielded regions take place in the bifurcation channel depending upon the Reynolds condition. Furthermore, the rheological characteristics of the fluid indicate the existence of an ideal Reynolds condition which allows the smallest flow stagnant area at the bifurcation zone.

PurposeNumerical simulation of polymer injection processes has become increasingly common in mould design. In industry, such a task is accomplished mainly by using commercial packages. Owing to the complexities inherent of this class of problems, most commercial codes attempt to combine realistic rheological descriptions with simplified numerical models. In spite of the apparent success, such approaches are not able to capture important aspects of the flow topology. The present work aims to describe a more elaborate mathematical model based on finite volumes which is able to provide both accurate solutions and further insights on the physics of the polymer flow.Design/methodology/approachThe mathematical model comprises the momentum and energy equations and a Poisson equation for pressure to impose the incompressibility constraint. The governing equations are discretized using the finite volume method based on central, second‐order accurate formulas for both convection and diffusion terms. Artificial dissipation terms are added externally in order to control the odd‐even decoupling problem.FindingsThe numerical model was conceived within the framework of a generalized Newtonian formulation. The capability of the numerical scheme is illustrated by simulations using three distinct constitutive relations to approach the non‐Newtonian behaviour of the polymer melt: isothermal power‐law, modified Arrhenius power‐law and cross models.Originality/valueThis paper extends the computational strategies previously developed to Newtonian fluids to account for more complex constitutive relations. The velocity and temperature coupled solution for polymer melts using only second‐order accurate formulas constitute also a relevant contribution.