Clarify print to screen and fix seed averaging

[dzalkind]

Purpose

Make the reported AEP trustworthy, or absent, instead of quietly wrong.

WEIS only reconstructs a meaningful AEP from OpenFAST when the AEP DLC is set up over a range of wind speeds. The previous gate (n_ws_aep > 0) also accepted a plain DLC 1.1 run, even at a single wind speed, so aeroelastic.AEP was computed from a degenerate probability weighting and then fed to both the screen output and financese_post. Users running a single wind speed saw an AEP that could be confusing.

• Add a derived DLC_driver -> use_openfast_aep flag in gc_LoadInputs: true only when the AEP DLC is present and covers more than one wind speed. The AEP DLC is the opt-in; no new user-facing option.
• glue_code connects aeroelastic.AEP to financese_post.turbine_aep and the screen output only under that flag, otherwise rotorse.rp.AEP (WISDEM), which is the more correct value in nearly all setups.
• Do not write outputs['AEP'] at a single wind speed; warn instead. The power curve channels (V, P, Cp, Ct, Omega, pitch, Thrust) are unaffected and still reported.
• Outputs_2_Screen skips the AEP / Blade Mass / LCOE / Tip Defl. lines when the input is unconnected (previously, from_openfast runs left all four at their 0.0 default and printed 0.0000000000).
• Remove a dead branch in get_AEP: idx_pwrcrv and n_ws_aep come from the same DLC-label filter, so len(idx_pwrcrv) == 0 could never be reached under n_ws_aep > 0.
• Fix seed averaging in avg_seeds, which summed only the first two seeds of each group while dividing by n_seeds_AEP. Correct at 2 seeds, biased low and dropping data for 3+. Now vec.reshape(-1, n_seeds_AEP).mean(axis=1).

Note for reviewers: LCOE will change for setups that run DLC 1.1 without an AEP DLC, since financese_post now takes WISDEM's AEP. Power curve outputs with 3+ turbulent seeds will also change, as they were previously mis-averaged.

Type of change

• Bugfix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (non-backwards-compatible fix or feature)
• Code style update (formatting, renaming)
• Refactoring (no functional changes, no API changes)
• Documentation update
• Maintenance update
• Other (please describe)

Testing

• pytest weis/dlc_driver/test/test_DLC_generator.py passes
• Built the OpenMDAO model through final_setup() and queried get_source for outputs_2_screen_weis.aep and financese_post.turbine_aep across seven DLC setups. With a
  • WISDEM-built model (examples/03_design_with_openfast): AEP DLC over 2 wind speeds routes to aeroelastic.AEP; AEP DLC at 1 wind speed, DLC 1.1 at 2 and at 1 wind speed, and a DLC 1.3-only run all route to rotorse.rp.AEP.
  • With from_openfast: True (examples/05_control_optimization): DLC 1.1 leaves the screen AEP unconnected so the line is skipped, while an AEP DLC over 2 wind speeds still routes to aeroelastic.AEP.
• Seed averaging verified separately on a 3-seed run, and screen-output suppression verified on a from_openfast run.

Checklist

• I have run existing tests which pass locally with my changes (targeted: DLC generator test plus the model-build checks above; full suite not run)
• I have added new tests or examples that prove my fix is effective or that my feature works
• I have added necessary documentation