Numerical relativity simulations of binary black hole inspiraling and mergers are computationally costly and storage requirements can quickly become unmanageable. By implementing a multi-domain spectral method we are able to more efficiently store metric component data when increased time resolution is desired over increased spatial metric resolution. Within the framework of a binary black hole system, multi-domain spectral methods work well using two different domain sets, one centered on each black hole, so they are able to absorb the singular behavior at each black hole's center. There is no difficulty in transferring quantities from one domain to another, or splitting the source function across two different domains, but there is no a priori choice for the relative weighting function to split a metric component. Here, we investigate what breakdown yields the highest accuracy.
Applied and Computational Mathematics (MS)
Department, Program, or Center
School of Mathematical Sciences (COS)
Matthew J. Hoffman
Joshua A. Faber
Caimano, Brian A., "Multi-Domain Spectral Methods for Data Reduction in Numerical Relativity Simulations" (2017). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus