OLAF: fix VTK writing at the last simulation step

[RBergua]

Feature or improvement description
In the OLAF output section, the WrVTK flag lets the user control visualization output:
0: NoVTK
1: Write VTK at each time step defined by VTK_fps
2: Write VTK at the beginning and end of the simulation

Separately, grid outputs (nGridOut) are controlled independently, with each grid defining its own TStart, TEnd, and DTOut.

Problems fixed:
1. Setting WrVTK > 0 forced every grid output to write at the last step, even though grid outputs should only follow their own schedule (TStart/TEnd/DTOut), independent of the wake's final write logic.
2. Setting WrVTK = 1 incorrectly forced an extra output at the final step, even though only WrVTK = 2 should include a start/end write.

Changes:
1. Grid outputs (nGridOut) are now written only according to their own schedule, independent from the wake WrVTK flag.
2. WrVTK = 1 no longer forces an output at the final step.

[RBergua]

Thanks to the failed r-test I realized that one condition is not properly handled at the moment. The failed test ad_Sphere_OLAF uses TMax=2, DT=1, WrVTK= 1, and VTK_fps=1.

From here, we would expect to write VTK outputs at: 0, 1, 2. However, the results at the last step are skipped. The proposed final step catch-up logic was failing when the last VTK_fps would exactly land at TMax for WrVTK= 1.

I updated the logic in commit 0873538. For reference, I attach the expected behavior below:

1. If TMax=5, DT=0.1, WrVTK= 1, and VTK_fps=1, OLAF will write at instants: 0, 1, 2, 3, 4, 5.
2. If TMax=5.4, DT=0.1, WrVTK= 1, and VTK_fps=1, OLAF will write at instants: 0, 1, 2, 3, 4, 5.
3. If TMax=5.4 and WrVTK= 2, OLAF will write at instants: 0, 5.4.