BugFix: PDE shape errors

[yb6599]

Hey @Jacob-Stevens-Haas, while running the PDE experiments, I got the error

ValueError: zero-size array to reduction operation maximum which has no identity

which suggests that the PDE array shapes are not conforming with the experiment setup. This PR changes the PDE data shapes, removes array lists since there's only one trajectory, and adds a signal parameter for noise_abs calculation, as x_test needs some reshaping to compute x_dot_test.

[Jacob-Stevens-Haas]

I think we talked about this before your exam, but my understanding of this PR is that
(a) _gen_pde_data still needs to return a list of arrays for compatibility, even if its just attempting to create a single trajectory, and
(b) Using magic numbers for axes feels weird. It's hard to understand why these moveaxis commands are needed and if they're written correctly.
(c) creating a signal variable that stores the whole array in order to keep track of the axes seems excessive.

If you're still interested in doing this, can you explain what's wrong in more detail?