incorrect WRG reading?

Hi Nick,

I've been getting some unusual answers in the energy reports of some of my projects.  In particular, the mean free wind speed of a particular turbine is sometimes quite different (by ~1 m/s) to the wind speed that shows up in the display (ie, when you select the 'i' pointer and read off the wind speed), and is also quite different to what WindFarmer predicts (WindFarmer is in agreement with the OpenWind display, but not the OpenWind energy report, though only for free wind speed, the energy predictions are consistent).

Does the energy report use the individual sector Weibull A & k parameters, while the display only use the overall values?

Doing some more research, I'm wondering if the problem is due to OpenWind incorrectly reading our wrg file (which we've produced 'in house', ie. not using WAsP).

I''ve noticed that when loading 2 identical wind resource grids, I get different answers for the wind speed and energy yield of a turbine depending on which wrg file I use.  Note that I'm being a bit missleading when I say 'identical', the 2 wrgs have slight formatting differences, however both are compliant with the WAsP wrg specs.  For example, the 1st Weibull A parameter is allocated columns 81-85 in the WRG file. WAsP always seems to use columns 81-83 (if it only needs 3 columns), whereas our wrg file used columns 83-85.

Obviously we could fix the problem at our end by making our wrg file identical to the WAsP wrg file, though since OpenWind is giving different answers to WindFarmer, I thought you may be interested in the problem.

Anyway, I've copied the openWind file and the 2 'identical' wrg files to the pub folder on your ftp site, if have some time to investigate.  They're in a file called filesForNick.zip

thanks,

Dave

ps.  the Google Earth Export option still seems to be working incorrectly in the Southern Hemisphere.  If you export the turbine in this file to a kmz file, it gives it an incorrect northing (ie, it gives it a northing of 3835699 instead of 6164301, if you subtract your value from 10,000,000, it will give the correct value).

Thanks Dave,

OpenWind and WindFarmer have slightly different approaches to this. In Windfarmer, I believe, the wind speed map is recalculated every time one changes the associated TAB file as that can effectively change the average wind speeds (if there are any Garrad Hassan people reading this please feel very welcome to jump in and correct me). OpenWind does not attempt to track the data you enter and calculate and so a WRG is treated as exactly the values contained in it. For openWind, a raster is just a raster. It is not linked to its source.

So far as I am aware, openWind conforms very strictly to the WAsP file format definition. If you disagree then let me know.

I looked at your WRGs. The WAsP one is corrupted in some way. It doesn't conform to the WAsP WRG format as you can see from looking at the second to last column. This in turn means that it gives strange values in openWind. However, in the latest version of openWind, when I recreate the display layers I still only get a difference between the wind map and the energy capture of 0.07m/s. In the WRG which is correct (your home-made version) then I get a difference of 0.03m/s. Now you must remember that the calculations leading to these values are very different. One uses a single Weibull curve per direction and multiplies by the probability of that direction. The other calculates the probability of each wind speed and sums them altogether. They should be close but they will never match perfectly and I believe this is the case with WindFarmer too, at least in our tests.

However, this is all at least a little academic as you should be using TAB files for your energy capture and not Weibull curves. As openWind evolves we will be moving away from Weibull curves altogether. They are an approximation for another age in which computing resources and data storage was extremely limited.

Cheers,

Nick

Hey Nick,

I'm also having some trouble doing an energy capture using an wrg I created in my newly acquired WAsP 10. I've emailed you the wrg I'm using (as I don't know your ftp address). I created this wrg through the steps outlined in wasp using a met mast tab file, properly formatted. The message "Turbine 1 of Site layer does not have access to an appropriate WRG" comes up when I try to do an energy capture, and I believe my hierarchy is set properly (WRG> tab file, wind spd, power density, elevation, site turbine layer).

Does this have to do with not having a point WRG? I've never needed to create one of these since I've been using the openWind flow model so far and haven't needed to compare apples-to-apples weibull values, which I vaguely understand to be the need for this point wrg thing. How do I create one in WAsP and do I need it or is there some other issue with my wrg or the program?

Thanks,

Jackson

Previously nick wrote:

Thanks Dave,

OpenWind and WindFarmer have slightly different approaches to this. In Windfarmer, I believe, the wind speed map is recalculated every time one changes the associated TAB file as that can effectively change the average wind speeds (if there are any Garrad Hassan people reading this please feel very welcome to jump in and correct me). OpenWind does not attempt to track the data you enter and calculate and so a WRG is treated as exactly the values contained in it. For openWind, a raster is just a raster. It is not linked to its source.

So far as I am aware, openWind conforms very strictly to the WAsP file format definition. If you disagree then let me know.

I looked at your WRGs. The WAsP one is corrupted in some way. It doesn't conform to the WAsP WRG format as you can see from looking at the second to last column. This in turn means that it gives strange values in openWind. However, in the latest version of openWind, when I recreate the display layers I still only get a difference between the wind map and the energy capture of 0.07m/s. In the WRG which is correct (your home-made version) then I get a difference of 0.03m/s. Now you must remember that the calculations leading to these values are very different. One uses a single Weibull curve per direction and multiplies by the probability of that direction. The other calculates the probability of each wind speed and sums them altogether. They should be close but they will never match perfectly and I believe this is the case with WindFarmer too, at least in our tests.

However, this is all at least a little academic as you should be using TAB files for your energy capture and not Weibull curves. As openWind evolves we will be moving away from Weibull curves altogether. They are an approximation for another age in which computing resources and data storage was extremely limited.

Cheers,

Nick

Hey Jackson,

I'm optimising with your WRG just now and it seems fine. The only thing I can think of is did you match the hub height of your turbine type to your WRG?

Nick

Thanks Nick. You were right about the turbine hub height mismatch. I guess I never thought of that and have always used 80m, but of course you would need the turbine to be captured at the same height as the wrg model.

I guess there's no need for a point wrg in this case, and I am able to do energy captures without it...