Binary black holes with non-orthogonal reference metrics

arXiv:2611.14647 · gr-qc · Submitted 2026-11-10 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: General Relativity and Quantum Cosmology. The day in general relativity and quantum cosmology.

Kai: I'm Kai.

Mira: And I'm Mira, here with Lev.

Kai: We're looking at the paper: "Binary black holes with non-orthogonal reference metrics".

Mira: Here's the gist of the paper. Binary black holes with non-orthogonal reference metrics demonstrate a new general non-orthogonal reference 3-metric implementation for numerical relativity simulations The work shows that despite a large deformation in the reference geometry to allow a large change in grid refinement to enable the simulation to span the relevant length scales, no pathologies are observed This approach avoids reflections caused by discontinuous refinement changes and can lead to more accurate waveforms when simulating binary black hole mergers on a single, structured grid in general non-Cartesian coordinates The flexibility of tuning structured grids through the reference metric is shown by employing spheroidal modifications of the spherical fisheye map to reduce the number of grid points in the z direction when the orbit is in the xy-plane Validation Against Exact Solutions The implementation was first validated by numerically evolving exact Kasner initial data with a fisheye-style reference 3-metric, demonstrating numerical results that converge to the exact solution with increasing grid resolution The Kasner “Apples-with-Apples” test showed that the fractional error in the recovered Kasner exponent, p2, and the Hamiltonian constraint and momentum constraint along the y axis at t ≈ 4.95 overlap closely as resolution increases The higher resolution curves rescaled by the fourth-order convergence factor show expected fourth-order convergence even at the fisheye transitions Waveform Accuracy and Convergence The authors reproduced the last orbit of a binary black hole simulation from Campanelli et al.

16: using a single-layer fisheye grid, obtaining results similar to the published work For the last orbit simulation, differences between resolutions agree closely in phase and show similar amplitudes over the plotted interval, consistent with fourth-order convergence during the late inspiral and lower-order convergence during merger and ringdown The comparison of high resolution waveforms obtained with spheroidal A and B against the spherical high resolution run suggests that retaining fine resolution over a larger region around the binary, as in spheroidal A, helps resolve the inspiral The covariant BSSN evolution equations are implemented using a non-orthogonal reference metric, where the evolved fields are tensors The code constructs the reference metric as gamma hatij = delta mn ∂xmCart / ∂x iCart ∂x jCart ∂x nCart / ∂x jCart (29) The implementation avoids the need for tensor rescaling by setting all tensor rescaling factors to unity and their derivatives to zero, since the coordinates are nonsingular The storage of precomputed reference metric fields increases memory usage but avoids their repeated evaluation during the evolution, boundary-condition applications, and diagnostic calculations Flexibility and Future Directions The general reference metric implementation in the open-source NR code NRPy will facilitate the exploration of structured grids suited to specific problems as well as their dynamical extension through modifications of the shift vector and regridding The authors plan to extend the covariant BSSN implementation to time-dependent reference metrics, to evolve BHs and GWs on time-dependent cosmological backgrounds The work demonstrates the flexibility of reference metrics in modifying structured grids through two modifications of the eight-layer spherical fisheye grid, denoted spheroidal A and spheroidal B The single-grid approach using a fixed fisheye reference metric is not a practical replacement for AMR as it involves much higher computational and memory costs However, the results are consistent with Ref.

16: and suggest approximate noise levels of order 10-9 M-1 for the fisheye grid The constraint violations do not degrade convergence at the four inner fisheye transitions and do not appear to cause large constraint violations The results are consistent with Ref.

16: Appendix Details The forward fisheye map uses transition centers Rj and width parameters sj measured in computational radius, and the normalization c is calculated directly from the boundary condition for any trial values of Rj and sj The inverse map specifies transition centers and widths directly in physical coordinates, so no parameter conversion is required For the inverse spheroidal map, the computational polar and equatorial extents generally differ due to the non-linear nature of the map The forward fisheye map requires solving coupled conditions to determine Rj and sj using a damped Newton method with a numerical Jacobian Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj The physical outer boundaries along the coordinate axes are located at ±L along each coordinate axis Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The computational extent along each axis is

−L, L: Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The physical width wj phys is the separation between the mapped endpoints of the computational interval

Rj − sj, Rj + sj: Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The computational extent along each axis is

−L, L: Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale <ref:2611.

Title and authors: Kai: So, looking at the title "Binary black holes with non-orthogonal reference metrics," I think the authors are setting up a new framework for how we handle these complex binary systems in simulations.

Mira: They're essentially proposing a different way to define that reference metric, moving beyond the traditional orthogonal setup that has been dominant in numerical relativity simulations.

Lev: From an engineering standpoint, the title implies they are trying to simplify the evolution equations by making them more flexible about the underlying geometry.

Kai: The authors validate this new approach by first evolving exact Kasner initial data using a fisheye-style reference three-metric and showing that their numerical results converge to the exact solution as we increase our grid resolution <ref:2611.14647#pg1,a fisheye-style reference 3-metric>.

Mira: That validation against Kasner data, especially showing convergence with increasing resolution, suggests their method is numerically sound for testing the concept.

Lev: And then they take it further by applying this to binary black holes in the moving-puncture formulation, tracing them from inspiral through merger and ringdown.

Kai: So they're not just doing a toy problem; they're showing this works for the full evolution of a binary black hole merger process.

The paper's summary: Mira: The core summary is that this new non-orthogonal reference metric allows grid refinement of structured grids to be changed smoothly through the reference metric, which leads to a consistent set of discrete equations that avoids spurious reflections.

Kai: It means instead of abrupt changes in grid spacing causing artifacts, you can use a deformed geometry defined by this reference metric to control refinement dynamically.

Lev: So it's about using the geometric setup itself as the tuning mechanism for the grid resolution, which sounds like a powerful way to handle the large length scales involved in these mergers.

Mira: The paper also points out that even with a large deformation in the reference geometry introduced to allow for a large change in grid refinement, no pathologies are observed during the simulation.

Kai: That addresses a major concern; they show you can make big changes to the geometry without getting those unwanted numerical instabilities or reflections.

The paper's improvements: Lev: The improvements they highlight really center around flexibility and consistency, especially when it comes to how the evolution equations are written.

Kai: They're able to avoid needing tensor rescaling factors by setting all those rescaling factors to unity and their derivatives to zero because the coordinates are nonsingular.

Mira: That's a huge practical improvement because it cuts out a lot of repeated calculations during the evolution, which saves memory and time on real hardware.

Lev: And they show how this general reference metric implementation can be adapted for future work, specifically by modifying the shift vector or through regridding to handle time-dependent and curved reference metrics later on.

Kai: They demonstrate flexibility by using two modifications of an eight-layer spherical fisheye grid, called spheroidal A and spheroidal B, to tune resolution specifically in the z direction when the orbit is in the xy plane.

Mira: So they've shown that this single fixed grid approach, when tuned correctly with these modifications, can be a way to achieve derefinement comparable to adaptive mesh refinement.

Conclusion: Kai: To wrap up on "Binary black holes with non-orthogonal reference metrics," the authors have demonstrated a general non-orthogonal reference three-metric implementation that works for binary black hole mergers <ref:2611.14647#pg1,Binary black holes with non-orthogonal reference metrics>.

Mira: The key findings are that this method avoids pathologies like reflections when changing grid refinement, and it lets you tune structured grids smoothly using these geometric deformations.

Lev: For practical runs on real hardware, the overhead is in storing the precomputed reference metric fields, but they show results consistent with previous high-resolution work for binary black hole simulations.

Kai: So, even though this single-grid approach isn't a direct replacement for adaptive mesh refinement because of its computational cost, it gives us a reliable way to get accurate waveforms.

Mira: It suggests that approximate noise levels of order ten to the minus nine per unit of mass can be achieved with this fisheye grid setup, consistent with earlier results for binary black hole simulations.

Lev: And they confirmed that constraint violations don't degrade convergence at those four inner fisheye transitions, and they don't seem to cause large problems.

Kai: So, the main implication is that we can get accurate waveforms on a single fixed grid configuration using this non-orthogonal reference metric, which is consistent with prior work like Campanelli et al. sixteen <ref:2611.14647#pg3>.

Mira: It opens a pathway for future numerical evolutions involving time-dependent and curved reference metrics, which might be useful for studying black holes on cosmological backgrounds.

Nishita Jadoo, Zachariah B. Etienne

University of Idaho · West Virginia University

gr-qc

Submitted: 2026-11-10

Updated: 2026-11-10

Comments: 15 pages, 14 figures

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 89/100

The gist: Binary black holes with non-orthogonal reference metrics demonstrate a new general non-orthogonal reference 3-metric implementation for numerical relativity simulations The work shows that despite a…

Key concepts

Non-orthogonal Reference Metric
This is a mathematical tool used in numerical relativity that defines the geometry of spacetime for the simulation. Unlike standard methods, this metric is not orthogonal, which allows the simulation to handle complex geometries and allow for grid refinement without causing numerical instabilities or reflections.
Fisheye Map
A specific type of mapping used to transform computational coordinates into physical coordinates. In this context, it helps define how a single, structured grid can be stretched or deformed to better cover the relevant length scales of the binary black hole system.
Spheroidal Modifications (A and B)
These are specific adjustments applied to the fisheye map. They allow researchers to tune the resolution of a structured grid, particularly reducing points in one direction (like z) when the black hole orbit lies in a plane (xy-plane), offering flexibility over the grid structure.

Terminology

Summary

Binary black holes with non-orthogonal reference metrics demonstrate a new general non-orthogonal reference 3-metric implementation for numerical relativity simulations The work shows that despite a large deformation in the reference geometry to allow a large change in grid refinement to enable the simulation to span the relevant length scales, no pathologies are observed This approach avoids reflections caused by discontinuous refinement changes and can lead to more accurate waveforms when simulating binary black hole mergers on a single, structured grid in general non-Cartesian coordinates The flexibility of tuning structured grids through the reference metric is shown by employing spheroidal modifications of the spherical fisheye map to reduce the number of grid points in the z direction when the orbit is in the xy-plane

Validation Against Exact Solutions

The implementation was first validated by numerically evolving exact Kasner initial data with a fisheye-style reference 3-metric, demonstrating numerical results that converge to the exact solution with increasing grid resolution The Kasner “Apples-with-Apples” test showed that the fractional error in the recovered Kasner exponent, p2, and the Hamiltonian constraint and momentum constraint along the y axis at t ≈ 4.95 overlap closely as resolution increases The higher resolution curves rescaled by the fourth-order convergence factor show expected fourth-order convergence even at the fisheye transitions

Waveform Accuracy and Convergence

The authors reproduced the last orbit of a binary black hole simulation from Campanelli et al. [16] using a single-layer fisheye grid, obtaining results similar to the published work For the last orbit simulation, differences between resolutions agree closely in phase and show similar amplitudes over the plotted interval, consistent with fourth-order convergence during the late inspiral and lower-order convergence during merger and ringdown The comparison of high resolution waveforms obtained with spheroidal A and B against the spherical high resolution run suggests that retaining fine resolution over a larger region around the binary, as in spheroidal A, helps resolve the inspiral

Methodology and Implementation Details

The covariant BSSN evolution equations are implemented using a non-orthogonal reference metric, where the evolved fields are tensors The code constructs the reference metric as gamma hatij = delta mn ∂xmCart / ∂x iCart ∂x jCart ∂x nCart / ∂x jCart (29) The implementation avoids the need for tensor rescaling by setting all tensor rescaling factors to unity and their derivatives to zero, since the coordinates are nonsingular The storage of precomputed reference metric fields increases memory usage but avoids their repeated evaluation during the evolution, boundary-condition applications, and diagnostic calculations

Flexibility and Future Directions

The general reference metric implementation in the open-source NR code NRPy will facilitate the exploration of structured grids suited to specific problems as well as their dynamical extension through modifications of the shift vector and regridding The authors plan to extend the covariant BSSN implementation to time-dependent reference metrics, to evolve BHs and GWs on time-dependent cosmological backgrounds The work demonstrates the flexibility of reference metrics in modifying structured grids through two modifications of the eight-layer spherical fisheye grid, denoted spheroidal A and spheroidal B

Conclusion on Practicality

The single-grid approach using a fixed fisheye reference metric is not a practical replacement for AMR as it involves much higher computational and memory costs However, the results are consistent with Ref. [16] and suggest approximate noise levels of order 10-9 M-1 for the fisheye grid The constraint violations do not degrade convergence at the four inner fisheye transitions and do not appear to cause large constraint violations The results are consistent with Ref. p. 9 [16]

Appendix Details

The forward fisheye map uses transition centers Rj and width parameters sj measured in computational radius, and the normalization c is calculated directly from the boundary condition for any trial values of Rj and sj The inverse map specifies transition centers and widths directly in physical coordinates, so no parameter conversion is required For the inverse spheroidal map, the computational polar and equatorial extents generally differ due to the non-linear nature of the map The forward fisheye map requires solving coupled conditions to determine Rj and sj using a damped Newton method with a numerical Jacobian Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj The physical outer boundaries along the coordinate axes are located at ±L along each coordinate axis Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The computational extent along each axis is [−L, L] Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The physical width wj phys is the separation between the mapped endpoints of the computational interval [Rj − sj, Rj + sj] Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The computational extent along each axis is [−L, L] Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The normalization c is calculated from the boundary condition for any trial values of Rj and sj Away from well-separated transitions, dr/drˆ ≃ caj, giving a physical grid spacing ∆x iCart,j ≃ caj∆ˆx i along coordinate axis i The stretching factors determine the relative grid spacings while c sets their overall scale <ref:2611.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved numerical relativity simulations can now stably evolve binary black hole mergers using non-orthogonal reference metrics without requiring tensor rescaling, which removes an ad hoc modification needed in earlier formulations.

  2. The system can validate its implementation by numerically evolving exact Kasner initial data with a fisheye-style reference 3-metric, demonstrating convergence to the exact solution with increasing grid resolution.

  3. It can extract smooth waveforms up to l = 8 on a grid without mesh refinement boundaries, even when using a large deformation in the reference geometry that allows for large changes in grid refinement.

  4. The system can tune structured grids through the reference metric by employing spheroidal modifications of the spherical fisheye map to reduce the number of grid points in the z direction when the orbit is in the xy-plane, supporting smoothly derefined structured grids.

  5. The implementation allows for switching to different reference metrics at the outset without requiring further changes to evolution or boundary equations, avoiding approximations needed in non-covariant formulations.

  6. The system can handle time-dependent and curved reference metrics, paving the way for evolving black holes and gravitational waves on time-dependent cosmological backgrounds.

  7. The system can use a single fixed grid configuration, achieved through eight fisheye layers with specific transition centers and widths, to achieve the same level of derefinement as an adaptive mesh refinement (AMR) grid for BBH simulations.

  8. The system can demonstrate flexibility in modifying structured grids by using spheroidal A and spheroidal B modifications to tune resolution in the z direction when the orbit is in the xy-plane.

Sources

Related papers