I am trying to get a basic understanding of how TurbSim generates a wind file (any, for e.g., full-field time-series data in TurbSim/Aerodyn form).
According to my understanding, the RandSeed1 parameter initializes the pseudo random number generator and if RandSeed2 is set as RANLUX, it will generate the pseudo random numbers based on the Luxury Pseudorandom Number algorithm. These random numbers are then used to create random phases (one per frequency per grid point per wind component) - mapped between 0 to 2pi.
After this, I am a little confused regarding how the hub-height, hub-height mean wind speed, turbulence intensity, turbulence model (for e.g., IEC Kaimal) is used to create the velocity spectra. Once the the velocity spectra is created, IFFT can be used to get the time series.
Mainly I am trying to understand the red box in the below figure. Any help in understanding this would be greatly appreciated. Thank you for your time.
I agree with your comment regarding the random seeds.
The wind profile (time-series mean wind speed and direction as a function of height) is specified by the user, e.g., a power law.
The velocity spectra are analytical functions (except for user-defined, which is a table look-up) based on user-specified settings and implemented following the equations documented in the TurbSim User’s Guide and user-specified settings: https://openfast.readthedocs.io/en/main/_downloads/cb14d3e2d3533d76e405d730fea19846/TurbSim_v2.00.pdf.
Dear Dr. Jonkman,
I have checked the TurbSim user guide but I couldn’t find what I am looking for. I understand that the velocity spectra are analytical functions.
What I wanted to understand was, if I specify IECKAI (which has its own spectrum), how the random phases enter those analytical functions?
Thank you for your time!
The velocity spectra determines the magnitude of the velocity at each frequency, not the phase. The (random) phase comes in separately. If there was no spatial coherence, the phase would be randomized, as you said, at every grid point, velocity component, and frequency. When spatial coherence is enabled, the spatial coherence functions are analytical, and you can consider the phases to be random at one grid point, but the randomness at other points will be limited to some extent by the coherence function (the smaller the spacing between grid points, the less random the phasing will be).
I hope that helps.
Thank you so much for your explanation.