A timestepper approach for the systematic bifurcation and stability analysis of polymer extrusion dynamics

M.E. Kavousanakis,L. Russo,C. I. Siettos,A. G. Boudouvis,G.C. Georgiou
DOI: https://doi.org/10.48550/arXiv.0707.3764
2007-07-25
Abstract:We discuss how matrix-free/timestepper algorithms can efficiently be used with dynamic non-Newtonian fluid mechanics simulators in performing systematic stability/bifurcation analysis. The timestepper approach to bifurcation analysis of large scale systems is applied to the plane Poiseuille flow of an Oldroyd-B fluid with non-monotonic slip at the wall, in order to further investigate a mechanism of extrusion instability based on the combination of viscoelasticity and nonmonotonic slip. Due to the nonmonotonicity of the slip equation the resulting steady-state flow curve is nonmonotonic and unstable steady-states appear in the negative-slope regime. It has been known that self-sustained oscillations of the pressure gradient are obtained when an unstable steady-state is perturbed [Fyrillas et al., Polymer Eng. Sci. 39 (1999) 2498-2504]. Treating the simulator of a distributed parameter model describing the dynamics of the above flow as an input-output black-box timestepper of the state variables, stable and unstable branches of both equilibrium and periodic oscillating solutions are computed and their stability is examined. It is shown for the first time how equilibrium solutions lose stability to oscillating ones through a subcritical Hopf bifurcation point which generates a branch of unstable limit cycles and how the stable periodic solutions lose their stability through a critical point which marks the onset of the unstable limit cycles. This implicates the coexistence of stable equilibria with stable and unstable periodic solutions in a narrow range of volumetric flow rates.
Dynamical Systems
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the instability in the polymer extrusion process caused by the viscoelasticity of the material and the non - monotonicity of wall - slip. Specifically, the author focuses on how to use the matrix - free/time - stepper algorithms to conduct a systematic bifurcation and stability analysis of non - Newtonian fluid dynamics simulators. These algorithms can effectively handle large - scale systems and are particularly suitable for studying the stability problems of Oldroyd - B fluids in plane Poiseuille flow, especially when there is non - monotonic wall - slip. The main objectives of the paper are twofold: 1. **Introduce the matrix - free/time - stepper method**: This method enables non - Newtonian fluid dynamics simulators to efficiently perform numerical stability/bifurcation analysis, including the continuous tracking of stable states and limit cycles and their stability calculations. 2. **Apply this method**: Through specific case studies, demonstrate the applicability of this method in analyzing the systematic numerical bifurcation and stability of periodic solutions. The specific case selected is the one - dimensional plane Poiseuille flow of Oldroyd - B fluid with non - monotonic wall - slip. ### Main Findings - **Stability Analysis**: Through linear stability analysis, two Hopf bifurcation points were found. The first Hopf point appears at a volume flow rate \(Q\approx0.4135\), from which a stable limit - cycle solution branches out; the second Hopf point appears at \(Q\approx0.5215\), from which an unstable limit - cycle solution branches out. - **Existence of Periodic Solutions**: In the parameter region between the two Hopf points, there are periodic solutions. This was verified by transient simulations, that is, after a small perturbation was applied near the unstable steady - state solution, the system entered a periodic oscillation state. - **The Role of Elasticity**: Elasticity, as an energy - storage mechanism, leads to self - sustained oscillations in the unstable region. ### Methodology - **Time - Stepper Method**: The dynamic simulator is regarded as a time - stepper with black - box input - output. The large - scale nonlinear equation systems are solved by iterative linear algebra methods (such as Newton - GMRES), avoiding the need for explicit calculation of the Jacobian matrix. - **Pseudo - Arc - Length Continuation Algorithm**: It is used to track the stable and unstable branches of periodic solutions and minimize the phase shift between the solutions to be solved and the reference solution through phase - constraint conditions. ### Conclusions - This study is the first to show how the steady - state solution loses stability through sub - critical Hopf bifurcation, generating an unstable limit - cycle branch, and how the stable periodic solution loses stability at the critical point, marking the beginning of the unstable limit - cycle. - In a specific volume - flow - rate range, the stable steady - state coexists with stable and unstable periodic solutions, which may lead to sudden instability of the system. Through these analyses, this paper provides new tools and methods for understanding the complex rheological behaviors in the polymer processing process.