DMD-Galerkin Model Order Reduction for Cardiac Propagation Modeling

摘要

Numerical simulation of cardiac propagation is a valuable tool for biomedical research. Due to the inhomogeneous, anisotropic conductive anatomy and complex nonlinear ionic current, numerical modeling of electrical activities in the heart is computationally demanding. Here, model order reduction is used to reduce the simulation time with a minimal effect on the accuracy. The semi-implicit finite difference method is used to discretize the governing equation of the monodomain (reaction-diffusion) model. The dynamic mode decomposition (DMD) is used in combination with the Galerkin projection to reduce the order of the original system. The reduced-order model is obtained by projecting the original system onto a subspace spanned by DMD basis vectors. Numerical results confirm the model order reduction decreases the simulation time by a factor of 5.96 while modifying the computed activation time, maximum time derivative and conduction velocity by 1.24%, 0.129%, and 0.639%, respectively.