Hi All,
does anybody has experience in generating turbulent wind fields starting from measured wind data using turbsim.
In BLADED, there is a function named “Matching a measured single-point time history”, which enables to get at least closer to real data.
In this case I´m going to get data from a nacelle mounted LIDAR (5 beams) and would like to reproduce it in simulation as accurately as possible.
Thank you for your imputs and suggestions.
Florian
Hi, Florian.
I recently updated TurbSIm to allow us to simulate full-field wind based on measurements at one or multiple points. The documentation isn’t complete, but most of what you need to run and understand the new “TIMESR” model is there. if you would like to test this version, here’s a link: nwtc.nrel.gov/Alphas
This model assumes the time series are at a fixed point in space (basically to replicate multiple anemometers on a tower in front of the rotor) , so if you’ve got a rotating lidar you’ll need to do some preprocessing. Actually, this model doesn’t do much (any) preprocessing of the time series input data, so be careful what you give it. I know, for instance, that files with NaN in the time series can give strange results.
HI Bonnie,
thank you very much for the new turbsim version. I have not yet used it with the LIDAR data. But I have used it to model a vertical and horizontal wind veer (direction shear).
It works well as long as you specify all 3 directions in the time serie file. The option with only the horizontal component runs through but is apparently producing errors in the wind field which lead to problems during the FAST run. This happens especially when applying very low turbulence intensity as a workaround for detwind.
Best Regards,
Florian
hi bonnie;
Is it possible to implement LADAR system in public version of FAST?. if yes would you pl refer the document on implementing. If it could not be implemented whats your suggestion to replicate the LIDAR system with the TurbSim for feedforward control . Kindly help me with this.
Hi, Srinivasa.
FAST v8.10 has a basic LIDAR module that you can use with the FAST8-Simulink interface; it was originally developed by CU-Boulder grad students for feed-forward control. It is not officially documented, yet, and won’t be until I can fix a few things in it. However, here are some guidelines (relative to what is specified in the FAST v8.10 readme document):
S-Function Parameters
Setting NumAdditionalInputs = 3 causes the code to simulate a CW lidar, whose measurement is set in an output channel called “WindMeas1”. When NumAdditionalInputs = 3, the code expects this third parameter to contain an array of initialization data for the lidar module:
[3, UseLidar, LidRadialV]
UseLidar should be 0 (single point measurement; no lidar) or 1 (CW lidar). Future version will allow more flexibility.
LidRadialV should be 0 (returns estimate of velocity in x direction) or 1 (returns radial velocity). Other values for LidRadialV will be treated as 0.
S-Function Inputs
When NumAdditionalInputs is 3, FAST_SFunc expects the following additional inputs:
9) Lidar focus position (w.r.t lidar) in X direction (required for continuous wave lidar)
10) Lidar focus position (w.r.t lidar) in Y direction (required for continuous wave lidar)
11) Lidar focus position (w.r.t lidar) in Z direction (required for continuous wave lidar)
Hi Bonnie;
I need a small help. I have followed the steps you have give in previous post regarding LIDAR measurement. But i have a doubt of where to give the initializing data which you have given in previous post, either we can add as lidar data in “servodyn” of the primary input file or any other file. I think those there lidar focus position inputs 9),10), and 11) could be added in the multiplexer before s_function. But where to initialize the lidar type and type of measurements given below:
{
S-Function Parameters
Setting NumAdditionalInputs = 3 causes the code to simulate a CW lidar, whose measurement is set in an output channel called “WindMeas1”. When NumAdditionalInputs = 3, the code expects this third parameter to contain an array of initialization data for the lidar module:
[3, UseLidar, LidRadialV]
UseLidar should be 0 (single point measurement; no lidar) or 1 (CW lidar). Future version will allow more flexibility.
LidRadialV should be 0 (returns estimate of velocity in x direction) or 1 (returns radial velocity). Other values for LidRadialV will be treated as 0.
}}
sorry for the trouble, feeling difficult since there is no document on this !!! thank you !!!
Hi,
the problem is the last post are also my questions. Could you please advise on them?
The LIDAR can only be used in the Simulink interface with FAST. Perhaps this document will help answer where you need to initialize things:
Dear Bonnie & All,
I’m currently also trying to use the Lidar module and use its measurements for eventual FF + FB control.
I think that I have followed all the possible instructions from doc’s and this forum but I’m still having troubles.
When I try to ‘fill’ the output channel WindMeas1 with simulated Lidar measurements, I get a message :
Error in port widths or dimensions. Output port 1 of 'OpenLoop/FAST Nonlinear Wind Turbine/Mux1' is a one dimensional
vector with 8 elements.
Error in port widths or dimensions. Input port 1 of 'OpenLoop/FAST Nonlinear Wind Turbine/S-Function' is a one
dimensional vector with 11 elements.
Can You please tell me where exactly do I specify the focus locations needed as input from the S_function?
And on which place could I extend the CW Lidar to a 4 beam CW Lidar?
Is there may be some updated documentation?
Thank You for Your support in advance!
BR
Filip
Hi Filip,
Because the S-Function parameters have been changed to indicate 3 additional inputs, the FAST S-Function is expecting 3 more inputs in addition to the original 8. These inputs are the x, y, and z coordinates of the lidar measurement point. Here is an example of how they could be input to the S-Function:
To extend the lidar to 4 beams, you could change the measurement point inputs to time series so that the lidar sequentially scans different measurement locations.
Also, when testing this I noticed that the range weighting along the beam does not seem to be implemented. This needs to be looked into further.
Eric
Dear Eric,
Thanks for your support,
The idea with the focus locations as time series works perfect.
Now I’m looking at the range weighting along the beam (because, as you say, outputs are point ‘measurements’ instead of averaged LOS velocities) but no solution yet. I’ll let you know if I succeed to come further.
BR
Filip
Hello everybody！
I want to use lidar module for my feed forward control signal，and I have a question about the use of Lidar module. Actually, I’d like to output wind speeds at multiple locations through lidar simulations, such as （-200,0,90）、（-50,0,90）、（-50,-30,120）、（-50,30,90）、（-50,30,60）、（-50,-30,60）.I found Wind Sensor Measurements in InflowWind sheet of OutListParameters.xlsx，as indicated in the attachment. Although I looked at the source code, I didn’t understand what these parameters meant， and wonder how to do that.
Dear @Lei.Xue,
Are you using the LIDAR module in FAST v8 for Simulink or something else? There is some documentation for the LIDAR module in FAST v8 for Simulink here: FAST/UsingLidarWithSimulink.docx at master · old-NWTC/FAST · GitHub.
Best regards,
