I´m an italian mechanical engineering student and currently I´m carrying out a 4 months internship for a German research institute dealing with wind energy.
I´m trying to engender a code similar to Turbsim by means of Modelica language and I´d like to know how to give the right coordinates to every single grid point.
At the moment I have already implemented a circulant polar grid,in which the hub point is not part of the grid,but now I would rather model a square grid in order to let the switch between the two different grids possible.
I hope having explained my matter as better as possible and I apologize for my not fluent english.
I guess I’m unsure what you are trying to do. Are you trying to create wind files similar to those generated by TurbSim so they can be used by AeroDyn?
thanks for answering me.
Actually I just would like to generate a stochastic wind field.
I’ve seen that Turbsim usually makes use of square grid with a symmetric points indexing.
Within the model I’ve implemented so far, I hav a certain number of circles around the hub and a certain number of points on every circle lie. The kinda point indexing I use is a circular one;if for instance I have just 2 circles and 4 points on each of those,on first circle you find points from 1 to 4,and on the second circle points from 5 to 8.
By the way,now I would like to have a square grid instead of a polar one in order to roughly compare the wind time series I get with my code and those ones achieved by means of Turbsim (the process is random,but the final results should not be so much scattered).
So my question deals with how to generate a square grid of points and in particular:
- calculate the coordinates of each point on the grid;
- index system of points.
I thank you in advance.
Are you asking how to interpolate from the circular grid to a rectangular grid?
I definitely recommend you use rectangular grids–they are much faster to interpolate within the turbine simulator. Way back in the day when turbines were on stiff towers and rotated at a constant rate, a rotating circular grid made sense–no interpolation was required. Ah, the good old days! But with variable-speed turbines that yaw and flexible blades and towers, that method no longer works. You may as well use a rectangular grid for the fast interpolation.
You are better off generating data at the rectangular grid points instead of interpolating from your circular grid to the rectangular grid. Every time you do an interpolation, you smooth the data and reduce the turbulence intensity.
The TurbSim source code has some comments at the top that indicate how we’ve set up the indexing of the rectangular grid. You may choose another indexing method, just make sure you keep track of the y and z coordinates associated with each point. You may also want to read Paul Veers’ paper, Three-Dimensional Wind Simulation, to get an idea of how the choice of index may affect the results.
so far I have fully followed the Veers´paper to get my own wind field simulator and it seems working,even though times series have a too small range (i.e wind speed has a too little range of values) onto some grid points.
To check my model I would like getting a square grid centered at the hub height.
Usually you can make use of two different indexing methods:
- the radial one
- the circulant one.
Roughly speaking, by having a look at the Turbsim users manual I noticed the for loop to generate points coordinates,but when I try to use that loop it does not work in my model,since Turbsim gets the grid started from the bottom, whereas I would let it started from the hub point.
Moreover,I don´t want to get a hub point on my grid,since I suppose it´s not a such interesting point for further analyses.
I hope to have explaind what I mean as good as possible.
I thank you in advance.