Using a Single InflowWind.dat File for Different Turbine Configurations

Wind & Water Inflow Turbulence
1

Hi, I am a new OpenFast user and our application requires the usage of the same InflowWind.dat file for both a 5MW and 1.5MW Wind Turbine. Our issue arises in the fact that turbine-specific parameters are required for the generation of this wind profile, such as hub height, rotor radius, and shaft length. For these reasons, we are questioning whether such an implementation is possible. Additionally, we are seeking clarification on the purpose of the Matrix Height variables in the TurbSim input file, and whether those parameters are turbine-dependent.

Thanks,

2

Dear @Mert.Timur,

Can you clarify why you’d want to use the same turbulent wind file for two turbines of quite different size? Given the presence of shear, the wind speeds across the rotor will differ anyway, so, there is really no reason to use the same wind file.

One challenge is that the domain size and spatial-temporal requirements are turbine-size dependent. If you make a wind file that follows the modeling guidance for spatial-temporal resolution for the 5-MW turbine, this would too coarse for the 1.5-MW turbine. And if you follow the modeling guidance for spatial-temporal resolution for the 1.5-MW turbine, this would be quite fine resolution and it would be computationally expensive to generate this across the domain size needed for the 5-MW turbine.

Best regards,

3

Dear @Jason.Jonkman,
The reason we aim to use the same turbulent wind file for both the 5 MW and 1.5 MW turbines is due to our machine learning approach based on the Cycle Generative Adversarial Network. Our goal is to establish a consistent training environment that uses equivalent boundary conditions for both models. This is necessary to ensure methodological consistency in our forward and backward mapping processes, where we generate synthetic operational data for a scaled 5 MW turbine using the 1.5 MW turbine results and vice versa.

I understand that the spatial-temporal requirements vary significantly depending on the turbine size, which could indeed complicate the comparability between the two models. However, the use of equivalent turbulence data is crucial for our scenario. Minor variations in wind conditions could disrupt the model’s ability to correctly identify patterns and accurately generate scaled synthetic data. The potential solution we are considering is using a larger domain size and higher resolution in the .bts file that can sufficiently cover both turbine models.

We would like to clarify how the matrix height variables in the TurbSim input file impact the generated wind profile. Do they inherently depend on the specific turbine configuration (hub height, rotor radius), or can we decouple these variables to create a more generalized wind profile that fits both turbine sizes?

Based on your experience, do you believe our approach can provide meaningful insights, or would it be better to generate separate turbulence files for each turbine? According to our investigation, two different Inflow files with separately generated .bts files would exhibit varying turbulence characteristics due to the stochastic nature of the turbulence modeling in TurbSim. Each time TurbSim generates a .bts file, the turbulence is created through random processes based on random seeds, resulting in different turbulence patterns even with identical settings. However, for our application scenario, an equivalent turbulence pattern for both wind turbines is essential.

Thanks,

4

Dear @Mert.Timur,

If you are using the standard IEC formulations for shear profile, turbulence spectra, and spatial coherence, these models will depend on the hub-height and the mean wind speed at hub height. The formulations will not depend on the size of the wind domain or rotor diameter.

It sounds like you want to generate wind data with a domain size needed for the 5-MW turbine with the resolution needed for the 1.5 MW turbine so that the wind domain is compatible for both turbines. In this case, you’ll need to use hub height of the 5-MW turbine.

Best regards,

5

Dear @Jason.Jonkman,
Thanks for your reply. Just to confirm, for the wind domain to be compatible with both turbines we need the following:

  1. Adjust the resolution to the one needed for the 1.5MW by setting the NumGridX and NumGridY to 7 (according to the TurbSim sample input file).

  2. The HubHeight, GridWidth, and GridHeight should be according to the 5 MW configuration.

At this point, we care the most about matching the wind domain for both turbines even if that requires expensive computational power given that we will run the simulation once.

6

Dear @Mert.Timur,

I agree, except that NumGridY and NumGridZ should be much larger than 7. Our modeling guidance says that NumGridY and NumGridZ should be set to ensure a spatial resolution of max chord, which for the WindPACT 1.5-MW baseline wind turbine is under 3 m.

Best regards,