Abstract
An extension of the framework of the Finite Integration Technique (FIT) including dynamic and adaptive mesh refinement is presented. After recalling the standard formulation of the FIT, the proposed mesh adaptation procedure is described. Besides the linear interpolation approach, a novel interpolation technique based on specialized spline functions for approximating the discrete electromagnetic field solution during mesh adaptation is introduced. The standard FIT on a fixed mesh and the new adaptive approach are applied to a simulation test case with a known analytical solution. The numerical accuracy of the two methods is shown to be comparable. The dynamic mesh approach is, however, much more efficient. This is demonstrated with the full scale modeling of the complete rf gun at the Photo Injector Test Facility DESY Zeuthen (PITZ) on a single computer. Results of a detailed design study addressing the effects of individual components of the gun onto the beam emittance using a fully self-consistent approach are presented.
Highlights
Memory consumption and CPU time represent the main limitations for large-scale electromagnetic field computations
In this article we address the important problem of numerical efficiency of such beam dynamics simulations using the Finite Integration Technique (FIT)
We propose an extension of this method including dynamic mesh refinement in order to locally adjust the spatial grid resolution according to the dynamics of particles and fields
Summary
Memory consumption and CPU time represent the main limitations for large-scale electromagnetic field computations This is especially the case for accelerator physics simulations involving self-consistent charged particle models based on the so-called particle-in-cell (PIC) method [1]. Simulations of this kind are an indispensable tool for the design and optimization of particle accelerators since they offer a full insight into the beam dynamics down to the particle level. We propose an extension of this method including dynamic mesh refinement in order to locally adjust the spatial grid resolution according to the dynamics of particles and fields This leads to considerable savings in the overall number of computational degrees of freedom and, reduces the computational burden in PIC simulations.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
