An efficient algorithm for multiscale flow simulation of dilute polymeric solutions using bead-spring chains

Abstract

In this study, a computationally efficient algorithm for multiscale flow simulation of dilute polymer solutions using a bead-spring chain description of polymer molecules is presented. The algorithm combines a computationally efficient extension of the earlier BCF-based semi-implicit method (i.e., approximately four-fold speed up) for multiscale flow simulations using a bead-spring dumbbell description [M. Somasi, B. Khomami, Linear stability and dynamics of viscoelastic flows using time-dependent stochastic simulation techniques, J. Non-Newtonian Fluid Mech. 93 (2000) 339–362] with a highly CPU efficient predictor–corrector scheme for BD simulation of bead-spring chains [M. Somasi, B. Khomami, N. Woo, J. Hur, E. Shaqfeh, Brownian dynamics simulations of bead-rod and bead-spring chains: numerical algorithms and coarse-graining issues, J. Non-Newtonian Fluid Mech. 108 (2002) 227–255]. The fidelity and computational efficiency of the parallel implementation of the algorithm are demonstrated via three benchmark flow problems, namely, plane Couette flow, Poiseuille flow and 4:1:4 axisymmetric contraction–expansion flow. The algorithm shows linear speed up with the number of processors and more importantly with the number of chain segments. In addition, the proposed algorithm is approximately 50 times faster in comparison to the only existing fully implicit method [J. Ramirez, M. Laso, Size reduction methods for the implicit time-dependent simulation of micro-macro viscoelastic flow problems, J. Non-Newtonian Fluid Mech. 127 (2005) 41–49].