Inquiry About Heterogeneous Wind Direction Simulation

[mohib988]

Hello Misha,

First of all, I would like to appreciate your exceptional work and effort in this field. I am truly amazed by this library.

Lately, I have been trying to simulate FLORIS with real-world turbine data to determine the optimal yaw angles. I found that using a single wind speed and direction did not yield good results, so I attempted to simulate heterogeneous wind speeds and directions. I succeeded in simulating varying wind speeds at different locations using a heterogeneous map, but I couldn’t find any method to simulate different wind directions at different locations.

Then I came across your pull request related to this feature, where you introduced bulk_wd_change.

I wanted to ask how I can incorporate multiple wind directions using that approach, and what is the current status of that work. My turbine field is quite large, and using a single wind speed and direction is not producing accurate results.

I would deeply appreciate your guidance and help.

Regards,
Mohib

[misi9170]

Hi Mohib,

First of all, thanks for the compliment on FLORIS! Let me say, though, that developing FLORIS has been a group effort by many people, of which I'm just one (perhaps I do get the prize for being the most vocal these days, though :) ). So, it's also thanks to the other code developers as well as the community of people that are using FLORIS and making our developments worthwhile.

Now, to your question: as you've found, while wind speed heterogeneity is a well-established feature in FLORIS, wind direction heterogeneity is not so well-established. In the pull request you mention #954 , I've built a prototype for the inclusion of wind direction heterogeneity, with the basic principle that the user could supply the wind direction at various points throughout the domain; we would then solve an ODE that determines a streamline passing through those points; and the layout could be "shifted" to account for this streamline (essentially, warping the layout of the farm to account for the wind direction heterogeneity). However, this is still a prototype and I'll admit that a) I haven't had time recently to work on this, with other things being prioritized and b) there were still some wrinkles to iron out when I left off.

So, the basic answer is that when (not if...) I get back to working on #954 and it is merged, I believe it will do what you need (with one caveat that I think this first pass will allow wind direction heterogeneity in the $xy$ plane, but likely not in the $z$ direction). However, I don't have a clear timeline on that. I am finally coming into some clearer space in the coming months, and I hope to get back to working on this, but I've had that hope before... Let's say, I hope to have made progress on it in the next couple of months, but I can't promise it will be ready for merger in that timeframe.

Looking back at #954, I'm realizing that I forgot to respond to a comment from @pjireland there. My answer here is unlikely to be satisfying, but it's the best I can offer at the moment.

Thanks,
Misha

[mohib988]

Thank you @misi9170 for your quick and humble response. I deeply respect you and your community's effort in helping a naive developer like me :).
Thanks again for your commitment (sort of :) ) to returning to that work. Please take your time!

I just wanted to ask: my goal is to get the optimal yaw angles. I’m getting wind speed data and I know wind speed points which I’ve used in the heterogeneous map.

One final issue I’m facing is that I have some turbines at different elevations. One workaround I’ve tried is adding the elevation height to the hub height. However, I still have only one wind reference height, but each turbine gets wind speed from its own anemometer and height.

Can I handle this by setting z_loc in heterogeneous map to the turbine height? If not, what solution would you suggest? and can I have different Turbulence Intensities at different points?
Thanks,
Mohib

[misi9170]

Hi Mohib,

Sounds good in terms of using wind speed measurements taken at the turbines to represent the heterogeneity. There is one catch here to watch out for---the speed_multipliers can only be interpolated between points, and this can mean that the turbines on the "corners" of the farm end up with half of their rotors outside of the domain of the heterogeneity, which can lead to odd behavior of those turbines. To counteract this, I suggest that you specify a few extra speed_multiplier locations at, say, 1D away from the "corner" turbines and give those the same speed multiplier you're giving to the corner turbines, so that the wind speed at the outside half of the rotor matches that measured at the anemometer. You'll need to be a bit careful using waked turbine's wind speed measurements to set the wind speed, since you may end up double-counting the effect of wakes.

Regarding turbine elevations---we generally caution against altering the hub heights of the turbines to account for elevation differences. The reason for this is that changing the hub heights can cause wakes to pass "above" or "below" turbines, where in fact, the wake likely follows the ground surface and would still impact the downstream turbine. So, we instead take the approach of "flattening" the terrain (essentially ignoring the effect of different base elevations) and letting the wind speed heterogeneity take care of the differences in "background" wind speed caused by elevation differences.

With that in mind, I would also keep the z location of the heterogeneous map at the turbine hub height. However, if you'd like to try it out, you can indeed specify a "3D" heterogeneous wind speed, which will allow a different speed multiplier at different z locations. If you do this, any wind_shear will be ignored as it is not clear how to reconcile the shear with a user-specified vertical velocity profile.

At this stage, no, we don't have any handling for heterogeneous TI throughout the domain (but wake models generally take account of "wake-induced" turbulence).

Cheers,
Misha

[mohib988]

Thanks a ton! Really appreciate the help. Adding extra multipliers and points for the corner turbines sounds perfect, I'll as make sure to consider all your suggestions.
Once again thank you so much !!!
Cheers,
Mohib