Refining Refinement in Binary Black Hole Simulations

To produce high-accuracy simulations of binary black hole inspirals at high mass ratios, we use the code Dendro-GR. This solves Einstein's equations on an adaptive grid, using wavelets to determine where increased refinement is needed. Simulations of black hole binaries with large mass ratios require balancing waveform accuracy with computational efficiency. Non-physical features of the solutions—such as junk radiation and an outward-propagating gauge wave—complicate the task of choosing optimal refinement strategies. In this talk, I describe how we alter refinement regions to address these challenges, vastly reducing cpu hours needed for binary simulations with large mass ratios.