Update default TwoPunctures parameters, or at least default parameters in BBH gallery example

[Roland Haas]

@Zach Etienne reports that increasing TP resolution in the galley example helps reducing initial data noise in the simulations (in a forthcoming paper).

Thus is might be good to increase the resolution and regenerate all data for the gallery page and Zenodo (and ask Barry and Ian to update it with a new version).

Most likely this is too late for ET_2023_05 but could be done for ET_2023_11.

[Zach Etienne]

There are two major improvements I would recommend. The first is for TwoPunctures, and the second for the BBH gallery example:

1. For most binary black hole simulations, the early evolution constraint violations are dominated by the initial data, when using the default TwoPunctures grid resolution of 30x30x16. I strongly recommend increasing the default resolution to at least 48x48x20.
2. The BBH gallery example has a rather unusual choice of initial lapse, setting it to the average value of 1 - m_i / (2r_i) for both punctures i. More typically the initial lapse is set to the “pre-collapsed” value of psi^{-2} which goes like 1 - M/r at large radii for total binary mass M. This is the same asymptotic behavior that the will lapse relax to, and so this choice results in less gauge dynamics than the “averaged lapse” value. TwoPunctures supports initial_lapse = "psi^n" and chooses a perfectly fine default value of n=initial_lapse_psi_exponent=-2 .

Note that the BBH gallery example is negatively impacted by (1) as well.