Code crash after 2000 - 3000 iterations (FD)

[rpvin]

Hi,

So, there has been an issue with the code crashing, due to time steps becoming too small, after about 2000 - 3000 iterations with the most recent update of Rayleigh. The exact same model ran fine with commit 28de237, which passed the benchmark tests without issues last year. This is happening with some of my other models too, which I don't expect this to happen.

One of the major differences between this and the most recent update, is the way the radial grid data is read in. However, when I try using the exact radial grids for runs with the past and most recent commits, the equation_coefficient files show very similar radial grids and reference state values.