Linear Numerical Schemes for a $\textbf{Q}$-Tensor System for Nematic Liquid Crystals

Justin Swain,Giordano Tierra
2024-03-26
Abstract:In this work, we present three linear numerical schemes to model nematic liquid crystals using the Landau-de Gennes $\textbf{Q}$-tensor theory. The first scheme is based on using a truncation procedure of the energy, which allows for an unconditionally energy stable first order accurate decoupled scheme. The second scheme uses a modified second order accurate optimal dissipation algorithm, which gives a second order accurate coupled scheme. Finally, the third scheme uses a new idea to decouple the unknowns from the second scheme which allows us to obtain accurate dynamics while improving computational efficiency. We present several numerical experiments to offer a comparative study of the accuracy, efficiency and the ability of the numerical schemes to represent realistic dynamics.
Numerical Analysis
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to develop efficient linear numerical methods to simulate the dynamic behavior of nematic liquid crystals, especially using the Landau - de Gennes Q - tensor theory. Specifically, the author proposes three new linear numerical schemes to meet the following challenges: 1. **Complexity of Non - linear Problems**: The Q - tensor model involves complex non - linear partial differential equation systems, and the unknowns among these equations are coupled with each other, resulting in great difficulty in solving. 2. **Balance between Computational Efficiency and Precision**: It is necessary to improve computational efficiency while maintaining high computational precision so that large - scale simulations can be carried out. ### Overview of Three Numerical Schemes 1. **The First Scheme (UES1D)**: - **Features**: Based on the energy truncation process, an unconditionally energy - stable, first - order accurate decoupling scheme is proposed. - **Advantages**: It ensures energy stability and reduces computational complexity through decoupling. - **Formula Representation**: \[ \psi^{\Delta t}(Q^{n + 1}, Q^n)=\tilde{\psi}_1(Q^{n + 1}, Q^n)+\psi_2\left(\frac{Q^{n + 1}+Q^n}{2}\right)+\tilde{\psi}_3(Q^{n + 1}, Q^n) \] where, \[ \tilde{\psi}_1(Q^{n + 1}, Q^n)=\bar{\psi}_1(Q^n)+\frac{1}{2}S_1(Q^{n + 1}-Q^n) \] \[ \tilde{\psi}_3(Q^{n + 1}, Q^n)=\bar{\psi}_3(Q^n)+\frac{1}{2}S_3(Q^{n + 1}-Q^n) \] 2. **The Second Scheme (OD2C)**: - **Features**: Using an improved second - order optimal dissipation algorithm, a second - order accurate coupling scheme is proposed. - **Advantages**: It improves the precision in time, but the computational cost is relatively high because the unknowns are still coupled. - **Formula Representation**: \[ f^{\Delta t}(Q^{n + 1}, Q^n)=f(Q^n)+\frac{1}{2}(Q^{n + 1}-Q^n):\nabla f(Q^n) \] 3. **The Third Scheme (OD1D)**: - **Features**: A new idea is introduced to decouple the unknowns in the second scheme, thereby improving computational efficiency while maintaining high precision. - **Advantages**: It retains high time precision and reduces computational complexity through decoupling. - **Formula Representation**: \[ f^{\text{LT}}(Q^{n + 1}, Q^n)=f(Q^n)+\frac{1}{2}(Q^{n + 1}-Q^n):\nabla f^{\text{LT}}(Q^n) \] ### Summary The main contribution of this paper lies in proposing three different linear numerical schemes, aiming to simulate the dynamic behavior of nematic liquid crystals efficiently and accurately. These schemes provide multiple choices to adapt to different application scenarios and requirements under the premise of ensuring different levels of energy stability and computational efficiency.