Array bounds issue in Morrion.f90

[andrew-platt]

Bug description

There is an array bounds issue in the Morrison.f90 code at line 4561 (dev branch).

To Reproduce

Steps to reproduce the behavior:

1. Compile with gfortran (9) with bounds checking turned on, release mode.
   Add compile flags: -fcheck=all -pedantic -fbacktrace -fbounds-check -Wuninitialized

2. Run test case Revert Cavitation-related pull request #19 (5MW_OC3Trpd_DLL_WSt_WavesReg)

3. Error shows up as:

19: At line 4561 of file /Users/aplatt/software-development/GitHub/openfast/modules/hydrodyn/src/Morison.f90
19: Fortran runtime error: Array bound mismatch for dimension 1 of array 'm' (6/3)
19:
19: Error termination. Backtrace:
19: #0  0x11182839d
19: #1  0x111828f85
19: #2  0x111829416
19: #3  0x1102b38ef
19: #4  0x1102633a9
19: #5  0x10f4c8066
19: #6  0x10f4da69b
19: #7  0x10f4e95e9
19: #8  0x10f366482
19: #9  0x10f366e2e
19: #10  0x10f2ff70c
19: #11  0x1107e431e

This was compiled on a mac using gfortran 9 and the dev branch, hash 2d9230d (2d9230d)

My suspicion is that the array m%D_F_AM_M(:,J) should be specified as m%D_F_AM_M(1:3,J) or m%D_F_AM_M(4:6,J), but I don't know the code well enough to know what it is doing there. The RHS of that line is a matrix multiply of a 3x3 and a vector size 3.

[HaymanConsulting]

Good catch. The HydroDyn Manual, section 7.5.1.1.6 details the calculation for the Distributed Added Mass loads. In general, it uses a 6x6 added mass matrix multiplied by a vector of size 6. However in our implementation, only the upper left 3x3 of the Added Mass matrix has non-zero elements, so we can in fact do what Andy proposed which is to rewrite the code to use m%D_F_AM_M(1:3,J) = ... The moment elements (m%D_F_AM_M(4:6,J) ) are all zeroes.

[andrew-platt]

@ghaymanNREL thanks for weighing in on this! I'll add this into my next PR (hopefully today).

[rafmudaf]

fixed by PR #297