Page 1 of 1

Inhomogen turbulent windfield

Posted: Tue Jan 16, 2018 5:18 am
by Simon.Wiedemann
Hello all,

is it possible to create inhomogen turbulent windfields somehow?

Aim is to simulate the loads of a turbine that is partly in the wake of an other turbine. This should be done ideally by reduced windspeed and or turbulence in a part of the wind field.

Best regards,

Simon Wiedemann

Re: Inhomogen turbulent windfield

Posted: Wed Jan 17, 2018 1:03 am
by Jason.Jonkman
Dear Simon,

The TurbSim preprocessor for FAST (https://nwtc.nrel.gov/TurbSim) can be used to generate full-field stochastic inflow turbulence to drive the aerodynamic loading of wind turbines. However, TurbSim does not specifically account for the wakes from upstream wind turbines.

The release of our new FAST.Farm tool for modeling fully wind farms, including all wind turbines and wake/array effects is expected later this spring.

Best regards,

Re: Inhomogen turbulent windfield

Posted: Wed Apr 17, 2019 11:35 am
by Arnold.Ramsland
Hello,
I would expect that FAST.Farm will require a significant increase in computing power. If so, do you have a recommendation for the type of computer that will be necessary?

Also, I noted in your response of Jan 17, 2018 the following statement: "the release of our new FAST.Farm tool for modeling fully wind farms, including all wind turbines and wake/array effects is expected later this spring".

Will the user be able to store the FAST.Farm calculated wind profile at each location in the wind farm as a separate wind file in .bts or .wnd format)?

Best regards,

Arnold Ramsland

Re: Inhomogen turbulent windfield

Posted: Wed Apr 17, 2019 7:45 pm
by Jason.Jonkman
Dear Arnold,

FAST.Farm can be compiled and run in serial or parallel mode. Parallelization has been implemented in FAST.Farm through Open Multi-Processing (OpenMP), which allows FAST.Farm to take advantage of multi-core computers by dividing computational tasks among the cores/threads within a node (but not between nodes) to speed up a single simulation. In parallel mode, each wind turbine (each instance of OpenFAST) can be run in parallel on separate threads at the same time the ambient wind is being read in another thread. Thus, the fastest simulations require at least one more core than the number of wind turbines in the wind farm.

I'm not sure what you mean by "wind profile", but FAST.Farm can output the ambient and disturbed wind (including wakes) at specified points throughout the domain. Further, to visualize the ambient wind and wake interactions across the wind farm, FAST.Farm includes visualization capability through the generation of output files in VTK format. The VTK files generated by FAST.Farm can be read with standard open-source visualization packages such as ParaView or VisIt.

Best regards,

Re: Inhomogen turbulent windfield

Posted: Mon Jul 01, 2019 10:18 am
by Arnold.Ramsland
Dear Jason

Thanks for the response. Sorry for the late response, however I must have forgotten to hit the submit button in April.
Based upon your response, I am planning to build my own computer with a water-cooled, 16 core Threadripper processor.

The visualization of output files in VTK format would be a great benefit. My usage of "wind profile" was merely to describe the wind data included in the Turbsim-generated file, e.g., *.bts. I was merely asking if Fast-farm could generate separate *.bts files for those turbines affected by the wake of other turbines. Perhaps this information could be used to expand the wind farm. For example, the first run might include a 4 x 4 matrix starting with row 1. If Fast-farm were to generate new *.bts files for the four turbines in row four, a second 4 x 4 matrix could possibly be generated for rows 4 - 7.

Thanks,

Arnold

Re: Inhomogen turbulent windfield

Posted: Mon Jul 01, 2019 5:26 pm
by Jason.Jonkman
Dear Arnold,

FAST.Farm cannot currently generate TurbSim binary (*.bts) files, but I'm sure this feature could be added with a bit of work. It may be easiest to generate VTK files for the YZ-slice you want and then convert these data to *.bts format.

Best regards,

Re: Inhomogen turbulent windfield

Posted: Thu Jul 11, 2019 9:59 am
by Arnold.Ramsland
Thanks. Also, I think your group is providing a very valuable resource to start-up companies like mine. I would have no way of testing out new ideas without FAST, MLife, and Turbsim.

Arnold