Fix for repeated calls to calculate_wake returning incorrect values.

[bayc]

Not ready to be merged

Feature or improvement description
This bugfix adds _sorted versions of the respective variables that get sorted based on wind direction during the initialization step for use within the solver. In the finalize step after the solve completes, the sorted variables are unsorted and stored in a new variable, therefore allowing repeated calls to calculate_wake and maintaining the sorting of the required variables.

Related issue, if one exists
Closes #378

Impacted areas of the software
Most of the simulation software, though I am not 100% that I have properly covered all the initialization of the variables (@rafmudaf could use your eyes here).

Additional supporting information
The code in issue #378 now produces the expected result of:

[paulf81]

thanks @bayc ! I made the following checks:

-example 1: Results are unchanged compared to when run on main branch

• example 7: AEP result is unchanged when I add additional unnecessary calls to calculate wake
  -example 2: Visualizations all look reasonable
  -example 14: Visualization looks reasonable
• other examples: Work as expected

Looks good!

[bayc]

@rafmudaf I updated the tests here to use the _sorted versions of u, v, w and x, y, z. Let me know what you think!