Abstract
In this paper, we study linearized Crank-Nicolson Galerkin finite element methods for time-dependent Ginzburg-Landau equations under the Lorentz gauge. We present an optimal error estimate for the linearized schemes (almost) unconditionally (i.e., when the spatial mesh size h and the temporal step τ are smaller than a given constant), while previous analyses were given only for some schemes with strong restrictions on the time step-size. The key to our analysis is the boundedness of the numerical solution in some strong norm. We prove the boundedness for the cases τ ≥ h and τ ≤ h, respectively. The former is obtained by a simple inequality, with which the error functions at a given time level are bounded in terms of their average at two consecutive time levels, and the latter follows a traditional way with the induction/inverse inequality. Two numerical examples are investigated to confirm our theoretical analysis and to show clearly that no time step condition is needed.
Original language | English |
---|---|
Pages (from-to) | 1183-1202 |
Number of pages | 20 |
Journal | SIAM Journal on Numerical Analysis |
Volume | 52 |
Issue number | 3 |
DOIs | |
Publication status | Published - 1 Jan 2014 |
Externally published | Yes |
Keywords
- Crank-Nicolson scheme
- Finite element methods
- Ginzburg-Landau equations
- Optimal error estimates
- Superconductivity
- Unconditional stability
ASJC Scopus subject areas
- Numerical Analysis