Fix potential error when requesting wind below the ground (also fix build issue with SD syntax)

[bjonkman]

Feature or improvement description
This fixes a couple of issues:

• the SubDyn.f90 code did not build on my Windows Intel 2019 compiler due to the way the string constant was specified on a continuation line.
• InflowWind could potentially add shear below the ground or return NaN if trying to compute the power law below the ground. This change sets the mean velocity to 0 below ground.

Related issue, if one exists
None reported on OpenFAST code.

Impacted areas of the software
Any place that requests inflow below the ground; likely not a common occurrence.

Also note the question in IfW_FlowField.f90's GetMeanVelocity function: I am not sure why it is not adding the mean horizontal or vertical shear there.

Test results, if applicable
This should not change any existing r-test cases since they likely don't request wind speeds below z=0.

[bjonkman]

Seems like there is an intermittent failure with the r-tests again; this time with the rtest-interfaces check. I reran the failing case on 7a35023 and it passed the second time (with no code changes)

[deslaughter]

@bjonkman, thanks for taking a look at this. You're right that GetMeanVelocity isn't doing everything it needs to do. At one point I was pretty sure that VLinShr and HLinShr weren't used by Grid3D as they're set to 0 in all except one place. The Grid3D_AddMeanVelocity subroutine isn't used and should be removed. The GetMeanVelocity function should be removed and replaced with CalculateMeanVelocity.

[bjonkman]

@deslaughter, I think the horizontal and vertical shear that are added in those routines is mostly for the native Bladed format 3D grid files; that's something Envision added, but I don't think OpenFAST has any tests for, yet.

[deslaughter]

@bjonkman, that does explain why it wasn't caught in testing. If you could send a small native Bladed wind file, I'd be happy to add it to the regression tests.

[andrew-platt]

Will the checks for above ground cause issues with MHK turbines? Or are those currents based on the ground located at the seafloor?

[andrew-platt]

I didn't know you could continue a string to the next line this way.

[bjonkman]

I don't typically use this format for continuing a string, but it's legal: https://www.intel.com/content/www/us/en/docs/fortran-compiler/developer-guide-reference/2023-1/free-source-form.html

[andrew-platt]

I learned something new today :)

[bjonkman]

Will the checks for above ground cause issues with MHK turbines? Or are those currents based on the ground located at the seafloor?

Basically, this just says if we are asking for mean velocities below the ground or MSL, they should return 0. The power law case returns NaN if z is negative and the power-law coefficient is less than 1 (typical use case), so that wouldn't make sense if z < 0. The log profile also doesn't make sense mathematically if Z <=0.

I'm not sure how they've set up modeling for MHK turbines, but I don't think this should hurt.

[andrew-platt]

Good point. That would result in nonsensical results.

[hkross]

We leave InflowWind alone for MHK and adjust the query points from the glue code before the are passed to InflowWind to put the origin at the seabed (as if it was the ground). So I don't think this will cause an issue.