Deprecate AD14 (rebasing of #87)

[andrew-platt]

Removing AD14 and replacing with AD15. This PR supercedes #18 and #87 (both PRs sat for too long and got a bit stale).

This PR also contains updates based on comments in PR #87. This is currently a draft PR; the description will be updated as cases are updated.

The following cases were updated from AD14 to AD15 (see notes below for comparisons):

• [x] 5MW_ITIBarge_DLL_WTurb_WavesIrr
• [x] AOC_WSt
• [x] AWT_YFix_WSt
• [x] AWT_YFree_WTurb
• [x] SWRT_YFree_VS_EDG01
• [x] SWRT_YFree_VS_EDC01
• [x] SWRT_YFree_VS_WTurb
• [x] UAE_Dnwind_YRamp_WSt
• [x] WP_VSP_WTurb_PitchFail

5MW_ITIBarge_DLL_WTurb_WavesIrr

Note: at some point in the past, maybe around version 3.0, this case was disabled as it did not compare well when run on different platforms. This case was then never updated after WaveTMax was reduced from 3630 seconds to 600 seconds. This was done for all offshore cases to memory usage on GH actions (the shorter WaveTMax results in different waves). To verify the changes introduced by changing to AD15 were minimal, two comparisons were made. For the first comparison, the original AD14 results with WaveTMax=3630 were compared with AD15 results also using WaveTMax=3630. These results were reasonably close. For the second comparison, a new baseline using AD14 and WaveTMax=600 was compared against AD15 with WaveTMax=600 also. These comparisons are shown above.

AOC_WSt

AWT_YFix_WSt

AWT_YFree_WTurb

NOTE: the tower shadow was turned off since the Bladed wind input file did not contain any tower info (regenerating this file results in a different time series).

SWRT_YFree_VS_EDG01 Added tail fin into ED and AD15 for this case (already used AD15).

When FAST8 was introduced, the tail fin option in AeroDyn 15 was not implemented. So this test case has been unstable with AD15 and was turned off. Adding the tail fin from the updated AD15 to the model does not give very good agreement with the FAST7 version of this case (CertTest #15). This is most likely due to missing features in the AeroDyn tail fin, most notably that the TFinIndMod=1 option is not available yet (somewhat related to the FAST7 option of SubAxInd in the furl file).

Despite not matching the FAST7 results very well, we think it best to add this case as is and note that it should be updated when TFinIndMod=1 is available. For later comparison, the Test15 results from FAST7 have been added along with a note about this.

OpenFAST dev refers to the OpenFAST dev results prior to this PR, and OpenFAST AD15 tailfin are the new results.

SWRT_YFree_VS_EDC01 As with the above case, the match with the FAST7 results is not great due to the missing induction option for the tail fin. Note also that the new results are using the same AeroDyn configuration as the above case (no DBEMT even though the original case had this enabled -- this was done for convenience while debugging the tail fin and can be updated at a later time). Note: the FAST7 Test15 is really test 16 (incorrect label). OpenFAST AD14 dev refers to the OpenFAST dev results with AD14 prior to this PR, and OpenFAST AD15 tailfin are the new results.

SWRT_YFree_VS_WTurb This regression test roughly follows the expected general trends for the selected channels (not very many channels were common between the original FAST7 test and the current test version). However note that the yaw is different, which may be related to the above notes on the induction for the tail fin. Note that the shaft torque (LSShftMxs) doesn't follow as well after including the tail fin.

UAE_Dnwind_YRamp_WSt This test had previously been disabled due to instabilities. With the new AD15, the results are much smoother and follow the same trends as the AD14 results.

WP_VSP_WTurb_PitchFail: Using the same AD15 input file that the WP_VSP_WTurb case uses. This causes some minor differences in the aerodynamic power, thus rotor speed, blade pitch, and other output results changed slightly.

[andrew-platt]

Unfortunately rerunning the SWRT cases in FAST7 with SubAxInd turned off does not improve the SWRT comparisons. So there is something very different between the old FAST7 and current results.

However, for the sake of keeping this moving, we will go ahead and merge, then check model differences later.

[andrew-platt]

@jjonkman will formally review later, and we'll follow up with any corrections.