Designing, Building and Testing a Darivonius VAWT

[Adriaan Kragten]

In The Netherlands in Schoondijke, we have had a testfield for small windturbines. Results of testing from 2008 up to 2012 is given in the Dutch report: "Resultaten testveld kleine windturbines Schoondijke" from 11 December 2012. You can find this report if you type the tittle in Google. One of the wind turbines tested was the Turby. This is a nice looking 3-bladed H-Darrieus rotor with twisted blades. However, test results of this wind turbine are not given in the report because the rotor was not rotating for a too long part of the testing period. This has probably the following reason.

The Turby was not equipped with a Savonious rotor for starting so it must be started by using the generator as a motor. This happens many times a day at low wind speeds and during starting the windturbine is using power from the grid. This consumed power reduces the nett power which is generated. The wind turbine is not rotating as long as the wind speed is that low that the starting procedure isn't activated. Turby went backrupt after some years as the turbine was very expensive and produced only a little amount of energy per year.

[mbouwer]

Concerning limiting the rotations per minute of a Darrieus;
Let us regard the possibility of pitching the blades through the hollow shaft.

[MagnetJuice]

Thank you mbouwer, but your design is incompatible with the design that I am contemplating.

I have seen pictures of the base that you built and posted on the other thread. That could be the beginning of an interesting VAWT.

It seems like you have a particular design in mind, you should start your own thread for that.

Ed

[oneirondreamer]

In The Netherlands in Schoondijke, we have had a testfield for small windturbines. Results of testing from 2008 up to 2012 is given in the Dutch report: "Resultaten testveld kleine windturbines Schoondijke" from 11 December 2012. You can find this report if you type the tittle in Google. One of the wind turbines tested was the Turby. This is a nice looking 3-bladed H-Darrieus rotor with twisted blades. However, test results of this wind turbine are not given in the report because the rotor was not rotating for a too long part of the testing period. This has probably the following reason.

The Turby was not equipped with a Savonious rotor for starting so it must be started by using the generator as a motor. This happens many times a day at low wind speeds and during starting the windturbine is using power from the grid. This consumed power reduces the nett power which is generated. The wind turbine is not rotating as long as the wind speed is that low that the starting procedure isn't activated. Turby went backrupt after some years as the turbine was very expensive and produced only a little amount of energy per year.

I was invited to review a round of testing done on licensed Gorlov turbines, done by Coastal Hydro Power, in BC.   CHP bought 5 rotors from Gorlov, for testing in water, with a flow augmenter.   4 of the rotors came apart in testing.   Cp was below 0.20.   Strong vibration occurred in testing when under load.   An H rotor Darius of the same swept area was tested in it's place and reached Cp 0.24.    These rotors did self start, but only at significant flow velocity.

The head researcher on the project, Dr. Brian Kirke, UQueensland, requested and received the original data from which the 35% (Cp 0.35) was claimed by Gorlov.   Dr. Kirke was not able to find support for 35%, and found that the testing was conducted in too small of a water race for the size of turbine tested, and the results are essentially meaningless.     However the test results of CHP on the Gorlov which Dr. Kirke was responsible for did find that it came close to Cp 0.2 (in Water).   

This points out the discrepancy between high Reynolds number operation (water, or very large wind) and low Reynolds numbers found in smaller wind turbines with narrow blades.    Here's a well done report showing a Turby in Air only ever reaching Cp 0.03.    I think also that dynamic stall plays a large roll in limiting the effectiveness of small bladed Darius turbines working in low Reynolds number environments.   

 https://repository.library.northeastern.edu/files/neu:1671/fulltext.pdf

Edited to correct grammar (two should have been too)   
Edited again to correct my opinion, which has changed.   The report I quote here is really garbage, no attention has been paid to correct wind tunnel blockage, and their attempt at a boundaryless wind tunnel is a farce.    The turbulence from the fan would have destroyed any chance of reasonable accuracy.    Sorry for posting it.  I'm leaving it up as an example of why academic papers deserve close examination.

[mbouwer]

MagnetJuice,

It's all about your goal to build a practical VAWT.
And if the pitching of the blades could bring advantages.

Rinus

[MagnetJuice]

Drew, are you sure that you provided a link to the test of the Turby turbine? Because I can't find the word Turby anywhere in that document, and the turbine that they tested was a small turbine (14 inches wide and 10 inches tall) that they build themselves.

Ed

[MagnetJuice]

mbouwer, yes, I have considered pitching the blades if I decide to build an H-Darrieus. I probably use centrifugal forces to activate a mechanism to pitch the blades after a predetermined wind speed to limit the RPM in very high winds. That will come later. For now, I would like to concentrate on the type, size and configuration of the turbine to build and test.

Whatever I build, it has to conform to my design goals of being easy to build, using commonly found materials, strong and safe.

Ed

[oneirondreamer]

Drew, are you sure that you provided a link to the test of the Turby turbine? Because I can't find the word Turby anywhere in that document, and the turbine that they tested was a small turbine (14 inches wide and 10 inches tall) that they build themselves.

Ed

You are correct Ed, this is not a test of the turby.    It was my understanding that turby used the Gorlov geometry right after Gorlov stopped paying EU patent fees.   I could be incorrect in that. After looking at the paper a bit more closely, and seeing the blockage factor that they used (almost 100%!), this report is really useless.   I will add a note to this effect where I posted it. 

I think this report from NREL, doing real world testing of the Windspire turbine, is more useful, though not for helical geometry's, however I'm unable to point to any third party testing of helical darius variants. .   It shows a peak Cp for a small commercial VAWT of almost 0.20   

If anyone has copies of the SWCC testing of the helical Darius's by UGE, I'd sure love copies, or any other helical turbines of smaller size.    UGE had turbines in testing with SWCC, but pulled them and had the test reports withdrawn.   It's my recollection that they ranged from Cp 0.11 to 0.17 but I can't find the reports anywhere. . 

[SparWeb]

The NREL Report on the Mariah WS is a good read.   You can feel the eyes rolling in some spots. 

[Adriaan Kragten]

My report KD 416 about drag machines has been reviewed. Table 1 has been extended for more values of lambda. The calculated Cp-lambda and Cq-lambda curves are now given in figure 2 and figure 3. So now it isn't necesarry to follow all calculations to see how badly a drag machine works. If you only look at figures 2 and 3 you will get a good impression.

[MattM]

At this point, chasing VAWT statistics is akin to tilting at windmills.  It just seems like VAWT are never going to amount to much efficiency on the available evidence.  Any wind speeds that are useful to VAWT are far better captured with HAWT.  The articulation of an HAWT into clean flow is going to always be an advantage even if it means an insignificant higher wind to get turning.  Clearly the commercial designs steer towards horizontals for good reason.  The scale of the vertical design will never change that.

[oneirondreamer]

At this point, chasing VAWT statistics is akin to tilting at windmills.  It just seems like VAWT are never going to amount to much efficiency on the available evidence.  Any wind speeds that are useful to VAWT are far better captured with HAWT.  The articulation of an HAWT into clean flow is going to always be an advantage even if it means an insignificant higher wind to get turning.  Clearly the commercial designs steer towards horizontals for good reason.  The scale of the vertical design will never change that.

I agree that VAWT's may never amount to anything.   It's possible that I've mismeasured, and fooled a lot of professionals into beliveing results I've obtained.   This has happened before, I'd suggest with the Gorlov, and the Lenz.  I'd even suggest that this is much more likely than the idea that my results are accurate.   

If that's the case, then I suggest that tilting at these windmills is very important to keep people from wasting their time.    I'm in conversation with someone who's trying to promote the building of Lenz turbines, precisely because no one has "tilted" at it.    The tendancy is to publish good results, and sit on poor ones.    This leads people to review the literature, and find things that look promising but seem to have been abandoned (could apply to Lenz or Gorlov).      This seems like exactly the right place to do some tilting, and I'm hoping that some of you will have a go at tilting at my project as well.   If I've made some mistake in my calculations, it'd be better for everyone to know.   I'm tired of looking at the Ugrinskies of the internet and not being able to find any good info. 

I'll start a new discussion, and post some of my test data, as well as some of the reviewers opinions

Best Wishes,
Drew

[MagnetJuice]

The previous 3 comments have some assumptions that are not correct:

Adriaan assumes that I am building a drag machine.

MattM assumes that I am out to prove that VAWT's are better than HAWT's

Drew concluded “VAWT's may never amount to anything” based on the results of tests that he has done on one particular kind of VAWT.

I understand why that happens. It reminds me of the Telephone Game that we played as kids. All the kids line up, one kid whispers a phrase to the kid next to him. Then that kid whispers the phrase to the other kid next to him and that continues all the way to the end of the line. The last kid says the phrase aloud and everyone burst into laughter because the original phrase totally changed.

At the beginning of this thread, I stated:

“A HAWT mounted on a high tower would be ideal to produce power from the wind. Not everybody has the land or the money to pay for and erect a tall tower. There are millions of people around the world that have strong surface winds and would like to be able to make a little power, and for them, a VAWT would be ideal"

I also said, “The ideal design should be a combination of the H-Darrieus and the Savonius”

I do not intend to put the HAWT manufacturers out of business.

I will continue with my research to find the best way to build a VAWT that can generate approximately 200 watts. I has to be easy to build, using commonly found materials, strong and safe.

I will post pictures of the construction here and include all the test results. That way if someone else wants to build a small VAWT, they don't have to start from scratch.

The main reason that I want the windmill to be easy and inexpensive to build is because if someone in rural India, Nicaragua or any other part of the world wants to build one, they can do so. It would be nice if a young father can build one of these windmills so his kids can use LED lights to read instead of using candles. And they can charge their phone batteries so they can stay connected to family and friends.

I appreciate all the help that I have receive so far. I have learned a lot from Adriaan and from the documents that have been suggested by others and I thank you for that.

I understand that the negative comments are a normal part of all Forum discussions. I can learn from them too. I don't get discouraged too easy, I will plough along and if I make mistakes I will consider them to be great lessons, not only for me, but also for others.

Ed

[oneirondreamer]

The previous 3 comments have some assumptions that are not correct:

Adriaan assumes that I am building a drag machine.

MattM assumes that I am out to prove that VAWT's are better than HAWT's

Drew concluded “VAWT's may never amount to anything” based on the results of tests that he has done on one particular kind of VAWT.

I understand why that happens. It reminds me of the Telephone Game that we played as kids. All the kids line up, one kid whispers a phrase to the kid next to him. Then that kid whispers the phrase to the other kid next to him and that continues all the way to the end of the line. The last kid says the phrase aloud and everyone burst into laughter because the original phrase totally changed.


Ed

Keep in mind that the telephone game as you describes looses information each time.   You seem to have lost quite a bit on the first round.  You've taken 5 words from my paragraph and ignored the 2nd word and then you've added this idea that my thinking is only informed by my own VAWT design and testing, when I've described a long history of research into many different designs.   

I think I am on the right track, and I believe my previous design is the best small VAWT ever tested.   I've posted my data, and letters from qualified people who've reviewed my data and reached their own conclusions.  If the lunatics who owned my original patent weren't crazy, it looks like it would be a booming business.  Since they are kooks, I haven't given up, and I'm now more 10 years in. 

Best Wishes
Drew

[JW]

  If the lunatics who owned my original patent weren't crazy, it looks like it would be a booming business.  Since they are kooks, I haven't given up, and I'm now more 10 years in. 

Im like 22years in. I invented a 4cycle steam engine I know this kind of thing, lots of criticism. If you believe in your self. Keep going

https://flashsteam.com/steam-engine-project-part-3 I have 3 patents but that is just allow the whole concept to work.

If you made any money on your patent you are ahead of the game

JW

[Adriaan Kragten]

The previous 3 comments have some assumptions that are not correct:

Adriaan assumes that I am building a drag machine.



Ed

In your own picture you show two cups which are mounted at the bottom of the Darrieus rotor and you have asked me if these cups can be used to start the Darrieus rotor. From what I see on this picture, these cups are half hollow cylinders so they are a part of a drag machine. What else can it be?. So I don't say that you are designing a drag machine but you use a drag machine to start the Darrieus rotor. This is only possible if the possitive Q-n curve of the drag machine is larger than the negative Q-n curve of the Darrieus rotor for low tip speed ratios. As a drag machine has an unloaded tip speed ratio of only about 0.3 (see fig. 2 KD 416) it will work as an aerodynamic brake if the Darrieus rotor forces it to drive at a higher tip speed ratio than 0.3 and this reduces the maximum Cp of the combination.

In my report KD 215 I give all the disadvantages of the Darrieus rotor. The fact that it has a negative torque at low tip speed ratios is only one of the disadvantages. Even if you can solve this disadvantage by using a Savonius rotor for starting, the other disadvantages remain. So I see it as a hopeless task to design a good working Darrieus rotor especially if it must be that simple that everyone can build it. But if you believe in it, go for it.

[MagnetJuice]

Drew, I carefully read your post again and now understand what you were trying to say.

Adriaan, I read your complete report KD 416 and now I understand that your calculations were based on anemometer cups, a purely drag device, and that you were referring only to the half cylinders in my picture.

The incorrect assumptions were on my side. I need to be more careful in the future.

My apologies to both of you.

Ed

[oneirondreamer]

Thanks Ed,

It looks like your thinking around a Darivonius VAWT is validated in the Taiwan combined Sav/Darius, in that it produces at about 20%, which is very close to what a good small HAWT would do, and the YouTube video of the American rep for them has a good YouTube video of his house mounted version generating decently.   I am surprised because of the building mounted turbine studies I'd seen showed very poor output, but this unit did make it through SWCC testing (not while on a roof).  If you built one you might improve some things?, and if you sized it right it you could buy their fancy MPPT controller for about 300$ It looks like.   Interesting, just thinking about the effects of the combined rotor.   

The https://prod-ng.sandia.gov/techlib-noauth/access-control.cgi/2012/120304.pdf paper reviewing the results of the Sandia Lab's extensive Darius work is very interesting in it's extensive computational model of resonance in the blades.   My reading of the paper gleans the idea that what they found is that to be most aerodynamically effective they needed very high aspect ratio blades.   They were most aerodynamically effective as 2 bladed rotors, but then required mechanical starting, and the resonance issues were amplified.   The did apparently reach above 40% efficiency, if I recall, with a large 3 bladed version.

All this makes me wonder if the combination of rotors somehow damps out some of the dangerous resonances.

Best Wishes
Drew

[Adriaan Kragten]

Report KD 601 about a self starting Darrieus rotor has been reviewed. The estimated Cp-lambda and Cq-lambda curves given at chapter 5 have been changed. For low lambda values, the rotor works as a drag machine and the Cq-lambda curve now has a small positive value for values of lambda smaller than 0.3. The maximum Cp has been reduced from 0.35 up to 0.3. It is explained why an unloaded Darrieus rotor may start at fluctuating high wind speeds.

[Adriaan Kragten]

Drew, I carefully read your post again and now understand what you were trying to say.

Adriaan, I read your complete report KD 416 and now I understand that your calculations were based on anemometer cups, a purely drag device, and that you were referring only to the half cylinders in my picture.

The incorrect assumptions were on my side. I need to be more careful in the future.

My apologies to both of you.

Ed

I have investigated if a drag machine can be used to start an H-Darrieus rotor. This is the case if the positive part of the Q-n curve of the drag machine is lying higher than the negative part of the Q-n curve of the H-Darrieus rotor, mirrord around the x-axis. The Q-n curves can be derived from the Cq-lambda curves according to the method as given in chapter 8 of my report KD 35. The calculated Cq-lambda curve of a drag machine is given in figure 3 of KD 416. It can be seen that this is about a straight line and that the unloaded tip speed ratio is about 0.32. The estimated Cq-lambda curve of a H-Darrieus rotor is given in figure 4 of KD 601 and it can be seen that Cq is negative for lambda in between 0.3 and 1.4. Both rotors are mounted on the same shaft, so both rotate with the same rotational speed.

The first preriquisite is that the rotational speed ratio of the drag machine for lambda is 0.32 isn't lower than the rotational speed of the Darrieus rotor for lambda is 1.4. So this means that the ratio in between the outside diameter of the drag machine and the diameter of the H-Darrieus rotor can be maximal 0.32 / 1.4 = 0.229. This means that the maximum diameter of the drag machine is 0.229 * 2 = 0.458 m if the diameter of the H-Darrieus rotor is chosen 2 m. So this first preriquisite results in a rather small maximum diameter of the drag machine.

The second preriquisite is that the positive value of Q-n curve of the drag machine for a certain rotational speed is larger than the mirrord negative value of the Q-n curve of the H-Darrieus rotor for the same rotational speed and for values of lambda of the Darrieus rotor smaller than 1.4. I expect that you need a very large cylinder height of the drag machine to fulfil this second preriquisite. Therefore I think that using a drag machine to start an H-Darrieus rotor isn't possible. But even if it would be possible, a lot of power will be dissipated in the drag machine if the H-Darrieus rotor forces the drag machine to run at a much higher tip speed ratio than 0.32. So you can better use a Savonious rotor as such a rotor has a much higher unloaded tip speed ratio.

[MattM]

Nobody is trying to be negative.  It's just that drag machines have been explored quite a bit here.  If you attach a drag machine to a lift machine the second becomes handicapped by the first.  Nothing says your smaller unit has to be directly tied in, as you can use other means to start up.

[oneirondreamer]

This forum has been very helpful, especially in it's unexpected benefits.   I'm not worried about anyone being negative.   I was shocked to find the Tawain based HI Power turbine on the SWCC certified list.   The reality that a combined Savonius Darius, could make useful power is counter intuitive, though as Adriann points out, if the diameters are careful selected and matched, it does fit within theoretical understanding, and even if it didn't, the SWCC testing reaching at least 20% wind power to charging Watts is real.    It's also interesting that the USA rep has installed one on his house, in what would normally be considered a terrible location, and it seems to be functioning fine (for now anyway).   

The things I'm curious about now, is what is the geometry of the Savonius rotor in the High-power turbine.   It's definitely not the Blackwell geometry, nor does it appear to be the Benesh.   In a way it's flatness is similar to my geometry, but it's large gap is different.   I've got some idea's around why the gap can matter so much, and how it can be eliminated (as I have).   I hope to test some of them out on this new turbine and test platform.   So I better get at it.
Best Wishes,
Drew

[Bruce S]

Drew;
I remember reading about the spacing in the Metal drum based S-type systems in the Mother Earth News mag some years ago.
It had to do with allowing the air to pass through to aide in the flow. It was a fairly detailed (at the time) explanation.

It may still be up on their website in their archives.
The Taiwan units do seem to have an air flow through them as well, I just can't see enough to tell.

Cheers
Bruce S

[oneirondreamer]

Drew;
I remember reading about the spacing in the Metal drum based S-type systems in the Mother Earth News mag some years ago.
It had to do with allowing the air to pass through to aide in the flow. It was a fairly detailed (at the time) explanation.

It may still be up on their website in their archives.
The Taiwan units do seem to have an air flow through them as well, I just can't see enough to tell.

Cheers
Bruce S

The space has been subject to a lot of discussion, however I've not found any work that I have faith in that “explains” the gap. 

It's clear that units with a gap performed better in the Blackwell report (and the Mother Earth work was mostly informed by the Blackwell report).   It's also pretty clear that the testing done in the Blackwell report was done in high blockage condition and so none of it can be considered conclusive in my opinion.   They didn't have the instrumentation or high speed camera's in that study to properly visualize the flows, and even if they had, the high blockage condition makes it all questionable.

My theory is essentially that they gap is functioning as a flow separation device, providing a point for the flow which has transited along the blade, slowing down and viscously transferring energy as it gets closer to the center of rotation (as in a Tesla disk turbine).  The sharp edge of the gap provides a point for it to separate from the foil, and if the turbine is loaded correctly, that should happen around the moment where the turbine has revolved far enough for the separated, and almost exhausted flow to be swept away by new flow from the front.   I've never seen any evidence of significant flow through the slot.   

As I read somewhere, the real problem with a wind turbine is how does it get rid of exhaust, if it's extracted all it's velocity.   What I've found is that every time I've thinned my airfoil, so that it presents the minimum frontal area, it works better.   This aligns perfectly with very low Reynolds number airfoil, they are always very fine.    This thinning, combined with giving it some edge to encourage flow separation at the appropriate time, I think is the key.   I think also that giving it a toothed leading edge may help to energize the boundary layer, allowing more energy to be transferred. 

One thing at a time.   Been distracted too much today, not good for my productivity 

If you are interested in seeing an interesting analog to my turbine work, in a pump rotor, have a look at the Harmon Pax rotor, and note the two small sharp features, about where a slot would be in a Savonius.

[MagnetJuice]

Matt, thank you for contributing.

I have looked at other possibilities for helping the Darrieus start turning in lower winds. Some use a separate motor to start the turbine, and some use the turbine's alternator and electronics to start the rotation. Those solutions work at a cost. They are using electricity to generate electricity, therefore reducing the efficiency of the system. I also found a patent by Windspire that talks about using magnetic levitation and special low friction bearings to reduce frictional drag.

All those solutions add extra components, complexity and additional cost to the system.

I am aiming at simplicity and low cost, that is why I continue to look at the possibility of using some type of drag device to help start the Darrieus.

Adriaan has been very graceful doing calculations and with his input. I think that it is a matter of finding the right combination of size and placement of the drag device.

After seeing that small Darrieus-Savonius Hybrid from Hi-VAWT spinning and producing power, I know that we are on the right track.

That turbine's rotor is 1.24 meters in diameter and 1.06 meters high and weights 25 Kgs. It turns at 835 RPM at full power to produce 300 watts. It uses two sets of modified 2-cup Savonius rotors in the center, turned at 90 degrees from each other.



Over the years, there have been improvements made to the original Savonius design that makes it more efficient.

The Benesh was an improvement over the original Savonius, and then the Ugrinsky, and more recently, the computer enhanced Benesh by California State University, Long Beach. That design using CFD was built and tested at the Boeing/CSULB low speed wind tunnel and to quote from the report, “Wind tunnel tests confirmed a 40% increase in peak power coefficients over prior art”

I don't know about you guys but I think that is very exciting.

Here are the Ugrinsky and the computer enhanced Benesh.



Ed

[Adriaan Kragten]

If you use a rotor diameter of only 1.24 m, you should calculate the Reynolds value for the chosen chord and for a moderate wind speed of maximal 5 m/s. The Reynolds value will be rather low and I am afraid that there are no symmetrical airfoils which perform well at that low Reynolds value.

A Savonious rotor is a realistic option to start the Darrieus rotor but if it is mounted at the centre of the Darrieus rotor, it has a large turbulent wake which distroys the air flow of a blade of the Darrieus rotor when it is at the backside. So this makes you end up with a maximum Cp of the combination of about 0.2 as 0.3 is possible. So I think that you should use one Savonious rotor on top of the Darrieus rotor and one with the buckets 90° rotated at the bottom other wise the torque will fluctuate too much.

The Savonious rotor should not only bridge the gap in the negative Q-n curve of the Darrieus rotor but also supply the torque of the bearing friction, the friction of the seal on the generator shaft and the torque due to iron losses if you use a generator with iron in the coils. You also need a certain torque to accelerate the rotor from stand still position. To get enough torque, the Savonious rotor must be rather large. However, the diameter should be taken not too large other wise the Savonious rotor works as a brake for the Darrieus rotor.

For determination of the maximum diameter of the Savonious rotor, you need to know the Cq-lambda curve of the Darrieus rotor and of the Savonious rotor and as no reliable measurements are available, you must make a realistic estimation. I think that what I have estimated for an H-Darrieus rotor with fixed blades at figure 3 of my report KD 601 is realistic. So this rotor has an optimum tip speed ratio of 4.2 and a maximum Cp of 0.3. Assume that for the Savonious rotor we estimate that the optimum tip speed ratio is 0.9, that the unloaded tip speed ratio is 1.6 and that the maximum Cp is 0.2.

Assume that the generator is matched such that the Darrieus rotor runs with the optimum tip speed ratio of 4.2. The point is that you don't want that the Savonious rotor works as a brake for this tip speed ratio of the Darrieus rotor but it is no problem that it works as a brake for higher tip speed ratios as the generator load prevents that the Darrieus rotor reaches this region. So the maximum rotor diameter of the Savonious rotor is a factor 1.8 / 4.2 = 0.429 times the diameter of the Darrieus rotor. If the Darrieus rotor has a diameter of 2 m (what I think is the minimum for a chord of 0.2 m), the maximum diameter of the Savonious rotor is about 0.86 m. The torque of a Savonius rotor increases about proportional which the height H of the buckets. The height can be rather large for the required torque level. To find the height H, you have to make Q-n curves of the Savonious rotor for a certain wind speed for different heigts and compare them with the Q-n curve of the Darrieus rotor for the same wind speed.

There might be new type Savonious rotors which have a higher maximum Cp than a traditional Savonious rotor with two Buckest made out of half cylinders but if this higher Cp results in a lower unloaded tip speed ratio, this reduces the maximum diameter of the Savonious rotor. So if you chose such a Savonious rotor, at least you should measure the unloaded tip speed ratio.


 

[MattM]

In all honesty, I hadn't considered the savonius being beneficial in limiting the top end rotation rate.  That's an interesting concept.

If you know your impulse necessary to start up the Darius then it shouldn't be difficult to match a small HAWT turbine that charges a capacitor to overcome that threshold.  I cannot imagine a 16-24 inch diameter HAWT wouldn't spin enough to charge up that capacitor.  If you really are set on savonius then knowing that threshold should decide that size.  Having them separate does add complexity.  On the other hand your extra weight is going to add additional stress when they are combined.  It will be interesting to see how this all turns out!

Whenever I think of savonius it makes me think about passive roof vents.  Keeping the air undisturbed inside the squirrel cage looks more promising than bucket-scoops that extend past the centerline.  You'd lose the braking effect with a squirrel cage, but they spin up easier than bucket-scoops.

[MattM]

https://www.intechopen.com/online-first/straight-bladed-vertical-axis-wind-turbines-history-performance-and-applications

[Adriaan Kragten]

If you chose for a 3-bladed H-Darrieus rotor you get a high bending moment in the blades due to the centrifugal force. The maximum bending moment depends very much on how the blades are connected to the hub. If you use only one spoke in the middle of a blade, you get the highest bending moment and I think that the maximum bending stress for this construction will be too high at high rotational speeds. You get the lowest bending moment if you use two spokes per blade and if the upper spoke is connected at about 1/5 * H from the blade top and if the lower spoke is connected at about 4/5 * H from the blade top. The spokes must have an aerodynamic airfoil to limit drag losses. Lets assume that you chose for this configuration of the spokes and that you take the geometry for a 3-bladed rotor as given in KD 601. So the rotor diameter D is 2 m, the blade length H is 1.5 m and you use a symmetrical NACA 0015 airfoil with a chord c of 0.2 m and a blade angle beta = 0 degrees. The blades are made of massive hard wood.

If the two Savonious rotors would be mounted out of the swept area of the Darrieus rotor you get a rather silly construction. So assume that the bottom of the upper Savonious rotor coincides with the upper spokes and that the height of the upper Savonious rotor is chosen 0.3 m. Assume that the top of the lower Savonious rotor coincides with the lower spokes and that the height of the lower Savonious rotor is also chosen 0.3 m. This means that the total height of the combination is still 1.5 m. I expect that if the Savonious rotors get a diameter of about 0.75 m, the generated torque is big enough.

Both Savonious rotors will now give a wake which destroys the flow around a blade of the H-Darrieus rotor when it is at the backside. However, at the blade tip you have tip losses and most of the power is therefore generated by the central part of the blade in between the spokes. This part feels almost no negative influence of the wake of the Savonious rotors for the chosen position. So I think that this is a rather optimal combination of an H-Darrieus rotor with two Savonious rotors.

[oneirondreamer]

Matt, thank you for contributing.

I have looked at other possibilities for helping the Darrieus start turning in lower winds. Some use a separate motor to start the turbine, and some use the turbine's alternator and electronics to start the rotation. Those solutions work at a cost. They are using electricity to generate electricity, therefore reducing the efficiency of the system. I also found a patent by Windspire that talks about using magnetic levitation and special low friction bearings to reduce frictional drag.

All those solutions add extra components, complexity and additional cost to the system.

I am aiming at simplicity and low cost, that is why I continue to look at the possibility of using some type of drag device to help start the Darrieus.

Adriaan has been very graceful doing calculations and with his input. I think that it is a matter of finding the right combination of size and placement of the drag device.

After seeing that small Darrieus-Savonius Hybrid from Hi-VAWT spinning and producing power, I know that we are on the right track.

That turbine's rotor is 1.24 meters in diameter and 1.06 meters high and weights 25 Kgs. It turns at 835 RPM at full power to produce 300 watts. It uses two sets of modified 2-cup Savonius rotors in the center, turned at 90 degrees from each other.

(Attachment Link)

Over the years, there have been improvements made to the original Savonius design that makes it more efficient.

The Benesh was an improvement over the original Savonius, and then the Ugrinsky, and more recently, the computer enhanced Benesh by California State University, Long Beach. That design using CFD was built and tested at the Boeing/CSULB low speed wind tunnel and to quote from the report, “Wind tunnel tests confirmed a 40% increase in peak power coefficients over prior art”

I don't know about you guys but I think that is very exciting.

Here are the Ugrinsky and the computer enhanced Benesh.

(Attachment Link)

Ed

I have the Rahai report somewhere, and the cross section of my previous turbine used something similar, however Rahai was cagey in his patent filing and published reports.  I think he hoped to commercialize the technology, and wanted to keep some cards in his pocket, but had to publish because of his grant.   

I built a helical tapering Rahai, and it started up at very low speeds, but never took off and at the time my instrumentation collection hadn't started.    One of them was made of fabric and telescoped up and down.   I then got pneumonia and had to take some time off, and reread most of the literature that I had on Savonius types as I couldn't understand why I wasn't getting the wind tunnel performance, Ian Ross hadn't yet published his paper.   In rereading the Rahai report, what I found was that he tested a number of different overlaps ratio's between his two foils, in both the x and y plane.   His data showed that the best performer was with no gap, both trailing edges touching, but he only published CFD images of the two foils in a high overlap condition.   If I recall correctly, in at least one of his wind tunnel images he shows it in the zero overlap, joined trailing edges condition.    I've had discussions about this CFD work with Dr. Curran Crawford, at uVic, and with people at the Ottawa NRC wind tunnel.   This was 2 years ago, but their opinion was that this CFD work was not likely to be particularly representative of reality, because it was to complex for any CFD of the time.   The way people have been trying to get around this, Rahai is not the first to do it, but is to run a large number of simulations with the turbine fixed at different angles.   This ignores the dynamic component of the flow interacting with the rotor, which if you've looked at CFD or real world testing of Flettner rotors, is massive.   

I sure would like to find more information on the Ugrinsky

[oneirondreamer]

Here's a link to the Rahai report

https://www.researchgate.net/profile/Hamid_Rahai/publication/228818185_Development_of_Optimum_Design_Configuration_and_Performance_for_Vertical_Axis_Wind_Turbine/links/02e7e5255d9801bd9b000000/Development-of-Optimum-Design-Configuration-and-Performance-for-Vertical-Axis-Wind-Turbine.pdf?origin=publication_detail

The relevant pages are 34-36.   On page 34 you see the CFD image of the dual high overlap foils, this is the image that I believe is the basis for a lot of peoples experiments, myself included.    However on page 35, notice the A-B graph and that the B graph of zero overlap is a substantially better performer. 

The simple explaination on why it's a better performer might be that it would block far more of the wind tunnel.   It could of course actually work better, but the higher blockage might make that difficult to discern.   

What Rahai did was print 2 blades, then attach them to the bottom and top plates in different positions.   This might have been accurate if he weren't already at 80% blockage.   He didn't build a Benesh, just used other peoples high blockage condition data.   

I'm not sure that there's much worth taking away from his work, other than the idea that perhaps a finer, thinner foil, with less camber might work better.   

On another note,  I used to have a copy of a very old popular mechanics, or similar publication from the 50's, post WW2 where someone had fitted a sort of “VAWT” to a boat.   It was an attempt at a self powering Flettner rotor, but it was a flattened eclipse with sharp edges.   He claimed it worked great, but I don't think it went anywhere.   I'd love to find a copy of it again.

[Ungrounded Lightning Rod]

On another note,  I used to have a copy of a very old popular mechanics, or similar publication from the 50's, post WW2 where someone had fitted a sort of “VAWT” to a boat.   It was an attempt at a self powering Flettner rotor, but it was a flattened eclipse with sharp edges.   He claimed it worked great, but I don't think it went anywhere.   I'd love to find a copy of it again.

Here's an article on how much power you get from a Flettner rotor (in the form of pushing a boat) compared to the amount of power you put into spinning the cylinder, for various speeds and true wind angles:

https://www.boatdesign.net/attachments/net-power-generated-by-flettner-rotor-pdf.140946

That should get you the info you need to calculate what results you'd get from putting various wind turbines on the top of it to spin it for you.  (It would be interesting to see if you could have a small sail or other wind diverter to get a flettner to self-power.)

Unfortunately, you don't get the wind to spin it if it's just a cylinder.  But though you do have to actually put some power into the rotor, it can be giving you power output factors of greater than 20 times what you put in.  So you could power the spin by a turbine powered by the boat's motion and only suck out about 5% or so of the push - at least on some points of sail.

[Adriaan Kragten]

I have made a mistake in the calculation of the maximum rotor diameter of the Savonius rotor in my previous post if it is assumed that the Savonious rotor has an unloaded tip speed ratio of 1.6. The maximum diameter is a factor 1.6 / 4.2 = 0.381 times the diameter of the H-Darrieus rotor. So if the diameter of the H-Darrieus rotor is chosen 2 m, the maximum diameter of the Savonious rotor is about 0.76 m in stead of 0.86 m.