AMR-Wind coupled output time step

[rthedin]

When running OpenFAST coupled with AMR-Wind, seems like the output time step is coming from the AMR-Wind input time.fixed_dt. I had the following settings:

DT = 0.01
DT_Out = Default

time.fixed_dt  = 0.05
time.initial_dt  = 0.025
time.delay_time  = 100

And the .out file had frequency of 0.05 s. Note that during the delay_time, AMR-Wind is actually running at initial_dt, and not at fixed_dt; however, it seems like the ouput is always following fixed_dt.

Bonus questions/issue: What is happening during the delay time when the ratio dt_amr/dt_OF is not an integer?

[andrew-platt]

I looked into the code a little bit. When OpenFAST is initialized, AMR-Wind passes in both dt and dt_out. The dt_out value passed takes precedence over the dt_out value from the OpenFAST input file.

My suspicion is that AMR-Wind is passing the incorrect value for dt_out (probably time.fixed_dt).

Edit: dug into the code a bit deeper. The above applies only to coupling where the FAST_Sizes routine is called. However AMR-Wind does not use this call, but instead initializes with FAST_ExtInfw_Init which does not have any way to pass or over-ride the dt_out value from OF. This leads me to the concern that AMR-Wind is not actually subcycling OF in this case.