Metrics or equations? Power curves or simulations?

I will try to give a politically correct answer then an answer following the circumstances.

  1. Equations and simulations are useful to help companies to elaborate then improve a project, make some predictions, avoid unnecessary manufacturing work. Power curves are the key test reports, simulations helping to see the correlation with the measurements. Balancing between metrics (power curves in first) and equations is useful.

  2. I agree with @Kitewinder about the high imbalance between metrics and equations. Indeed take all the AWE publications over the years: equations galore versus very few power curves. I looked for some, but only found: figures 15 and 18, video from 3:07, table 8, that is, less than the fingers of one hand.

On the other hand (!), to my knowledge, there is not a single scientific study on Kiwee, which is the only AWES on the market. Moreover, it is not said that it would be necessary, since the system has been built and works well, and the power curves are known even if they are not formalized.

This topic does not go against the equations and scientific studies, but it may have to be recognized that a judicious choice of themes will be necessary for what remains of the AWE wave, such like a functional design which has yet to be found. And power curves showing actual measurements should increase (even when they are not very favorable), and cover the intensity of production over time.

A power curve from real data could only come after the system is working well. I don’t thing we have such systems in AWE except the KiweeOne. A power curve for that one is interesting, but also of limited value for extrapolating to utility scale, as it is of a smaller power rating.

A synthetic power curve may be made in advance of building. Though often one fails to meet the desired power output, and there may be other issues that turn out to be hard other than producing power, that must be solved first.

Anyways, as spectators to the current evolution of AWE, we can only observe what any of the development teams decide to release, we are in no position to make demands…

Note that there is a 10-year time between the Figure 15 (2022) and the Table 8 (which describe a similar yo-yo system, each using a flexible wing (that is much larger for the Figure 15)). The results show comparable duration and energy consumption for the reel-in phase, with the average power being positive to a proportionally comparable extent.

Similarly, the results in Figure 18 overlap with those in the video (Ampyx on the left) for rigid wings in yo-yo mode.

After more than 10 years conclusions can be drawn.

Could you somehow make these more accessible? Eg screen shot image directly in the post?

I was asked to attach the citation to the final paper to the pre-print (pdf page containing figures 15 and 18), which I did, my link pointing to a post containing these two connected pieces. The first initial post was "Autonomous Airborne Wind Energy systems: accomplishments and challenges".

So I’m doing it again:


Now the video (read the document from 3:07) was often connected within the forum. Here is a direct link:

For the table 8 (representing some measured data rather than power curves), click on the link below then on the pdf link at the bottom of the page:

At the same time I will sketch out a temporary conclusion to add to my initial comment, which you can of course confirm or deny.

Figures 15 and table 8 (flexible kites in yo-yo mode) provide positive values in spite of the significant amount of energy and time during the reel-in phase. SkySails system (figure 15) generated an average of 92 kW at 12 m/s wind speed, which is one of the higher results in AWE. It is perhaps a reason why now Skysails sell off shipping to focus on AWES.

Concerning figure 18 and the video for rigid wings in yo-yo mode (and also a previous Makani wing), I already commented them (message just before the message below), in first about the (too) high energy consumption during reel-in phase. You replied

After many years of development, don’t you think that the necessary steps you describe could have been achieved, leading to power curves showing the real potential, and in particular reel-in phase with less energy consumption?

Could have: maybe
Could be expected to have: maybe but less likely

I think you must remember that these companies answer to the shareholders, and there exists information that you don’t see. Also, while it is your priority to get a power curve, it may not be the priority of these companies.

Look at wind power history and you see that these things are not really abnormal

Honestly, I don’t think I can answer that, although I have some ideas about it.

Hi @dougselsam , as you are familiar with the history of wind energy, please could you give your advice?

Recording a power curve (power vs windspeed graph) is normal in wind turbine development. However, since kite-reeling does not deliver continuous power, it could be a little more complicated. But for utility-scale wind turbines, the windspeed is averaged over ten (10) minutes, so I don’t see anything preventing the generation of a power curve for any system that operates for five hours at a time. Over the years, I’ve experienced developers of vertical-axis drag-based machines wanting to be taken seriously just on the basis that they rotate, without even having a generator attached. Some of us also may remember Santos postulating that the wind industry had become too dependent on power-meters, when responding to demands that he back up his various claims of kite-power breakthroughs with no numbers to back up his claims.
Certainly, at some point, since the entire point of any of these clean-energy projects is the production of energy, the most relevant question is how much energy is being produced. Not all companies may want to share that information, but the data should be there in some form, I would think.
In general, I would say, if a wind energy system developer can go on for years without sharing output data, there is probably a reason, since power output is the main goal. Experience would suggest that no publicized power output data could indicate a lack of impressive numbers to share, which could indicate a general problem with power generation.

Interesting information. If you can reply, do you mean that for Kitemill prototypes the shareholders are well informed about the schedule of the company, not content to demand production at the moment T, but participating in the development process of the complete project?

The shareholders would get information about progress and delays, while the public would only be notified when something is ready to be shared in public. There is nothing special here. Sharing information in public is not straightforward due to rules on insider trading, not particular to Kitemill. I am not the one dealing with this, so maybe direct further enquiries down the proper channels.

If you are looking for admitting that Kitemill is in some way «playing» the public, I can only say this is far from the case. I would testify that any news Kitemill shares reflects well what happens inside the company. But of course not everything is shared with everyone. And you guys seem a good example why that is maybe not a good idea.

So if you want the most detailed and updated status update from Kitemill, I suggest you look at the news section of the home page. It should be quite accurate.

You are referring to «mature technology» windmills, where you make a windmill with a slight or no changes from previous iterations. In that case the windmill will probably go straight knto 24/7 operation and after a year or so you will have a power curve.

I was referring to the days when windmills were not mature technology, and once build maybe broke down after just a few days worth of use. Or maybe was never used. Even though they contained useful innovations that were later commonplace.

The experimental wind farm in Balaclava (in Crimea) with capacity of 100 kW was developed under direction of inventor Yuri Kondratyuk and installed there in 1931. Before the war, it produced electricity for the Balaclava-Sevastopol tramline. During the war it was destroyed.[7]

Actually, it is easy to get a power curve in an afternoon, if you have good winds that day. It is a very simple thing, using wind speed and electrical output sensors, and feeding the data to an excel spreadsheet. The spreadsheet will feed you back a graph with a typical shaped cloud of data dots, and a superimposed average power/windspeed line (power curve). In truck testing, I can take a power curve in about 1 minute, driving with my left hand while writing down output and speed with my right hand. Well, it does help to make a few runs in each direction, so let’s say 5 or 10 minutes for a good quality power curve using a truck.

Also, I’ve never heard much about wind energy in Russia, but, the Danish Concept, which is how windfarms got their first real start here in California, resulted from farmers that needed power, building the simplest design anyone could imagine, and they worked right away. The first home powered by a windmill was by a guy named “Brush” who made a giant (but low) version of a typical farm water-pumping windmill, and I believe it found immediate use powering his home. Some wind systems just work the moment they are put up, and keep working.

Classic “Only 2 types AWES” BS references
If you want another
Ollie published a power curve for the tests on early Daisy Kite Turbines in his PhD
The early soft ones weren’t very impressive power or efficiency admittedly. Rigid ones were impressive.

1 Like

Yes, and there are a lot of documents with metrics and power curves or power data for Daisy and showing high efficiency. Similarly I mentioned in my initial post:

It looks like rotary AWES (Kiwee, SuperTurbine ™, Daisy) go well with metrics and power data, perhaps because there are closer to the regular wind turbines than crosswind AWES, in first the ones which use reeling mode.

And because you can leave them running unattended without needing a control system to keep them in the sky, this gives you time to get more than a couple of loops worth of data

One problem to finding a power curve for an AWE rig; what is the wind speed? If the rig is of some size, the one option would be a LIDAR, already out of the budget scope of many AWE companies I suppose

I thought there was some equation by a guy called Tallak which could tell how windy it was just by looking at some string in the sky

1 Like

Very nice example of stone-age instrumentation, Roddy. See? Stone-age people were smarter than we think! :wink: I’ve long imagined an analog meter for windspeed. Take a visible arrow hanging down like a pendulum, and place a graphic scale against it, then use an anemometer to calibrate what angle = what speed, and voila - instant easy-to-read windspeed sign. It would have to pivot like a weathervane, so you would also get wind direction out of it. Very much like the rock, it could also show rain if it was wet, sun if it casts a shadow, etc. But seriously, it would be an easy way to show windspeed and direction, at whatever height it was mounted. Or one could just mount an anemometer and wind vane up there are take the data. If you are flying at a higher height than the instrumentation, that can either be factored in and the higher windspeed estimated, or just specify that the power curve is based on the windspeed at height X. if you are comparing your powercurve to a turbine at height X, that could work out OK.

Having access to LIDAR data I can only say the wind gradient only holds as an average. Today wind was constant on the LIDAR 100-500 m, then much less. Other times, there are «rivers of wind» at some altitudes…

You would need a second kite or drone at the correct altitude to do it right without LIDAR. But most AWE developers probably dont care too much because we are still just learning to fly at all, and it doesnt make sense to make a power curve until control strategies have matured to a minimum level.

AWE is not like windmills if it is computer controlled. The job is not done the moment the blades come out of the mold.

Anyways. I support the demand for power curves. WHERE ARE THEY? I DEMAND ANSWERS!

Sure, I like power curves or power data. I know what Kiwee and Daisy are up to, providing continuous power at an assumed known altitude, above all for Daisy where the wind speed measure is easily made at a correct altitude by using a mast, while an anemometer on Kiwee could do this job.

For crosswind reeling (yo-yo) or fly-gen devices it is a bit more difficult, and there are very few power data. For them putting on the kite a good old anemometer measuring airspeed when the kite flies (as it is already done), and measuring wind speed the times when the kite is stopping (between the zig-zag turns during power phase when figure-eight is applied, between reel-out and reel-in phase), then cross-check the data, such like deducing wind speed by kite airspeed when the kite glide ratio is known in the concerned range.