Is the Savonius another arrangement of the Tesla Turbine ?

[oneirondreamer]

Hello Folks,

We've had quite a dump of snow here, so I've done an explainer video on my turbine, and how it relates to the Tesla disk turbine.

I don't ask anyone to accept my way of thinking.   I have tested this fairly extensively, and one thing is clear, maximum power production happens when the leading edge speed, and wind velocity are close to the same.   TSR 0.8 - 0.9   

If you'd like to argue about how it's working, I'm game, however if you are going to argue and ignore my measurements, we don't have much to talk about. 

[Bruce S]

I don't ask anyone to accept my way of thinking.   I have tested this fairly extensively, and one thing is clear, maximum power production happens when the leading edge speed, and wind velocity are close to the same.   TSR 0.8 - 0.9   

If you'd like to argue about how it's working, I'm game, however if you are going to argue and ignore my measurements, we don't have much to talk about.

Kinda harsh on the statement so early on the posting aren't ya?

Bruce S

[oneirondreamer]

I don't ask anyone to accept my way of thinking.   I have tested this fairly extensively, and one thing is clear, maximum power production happens when the leading edge speed, and wind velocity are close to the same.   TSR 0.8 - 0.9   

If you'd like to argue about how it's working, I'm game, however if you are going to argue and ignore my measurements, we don't have much to talk about.

Kinda harsh on the statement so early on the posting aren't ya?

Bruce S

we've all got a limited time on this earth, I've wasted plenty of mine and others arguing about things I didn't understand.   I spend my patience on other things, nor do I want to waste others time.

[MattM]

The helix blade is pretty interesting.  The point that is perpindicular to the wind/waterflow is constantly traveling as it turns.  I'd think you could use several linked together as long as they were all slightly out of phase with the other so that the drag from one will literally affect the others at a minimum.

[Bruce S]

I don't ask anyone to accept my way of thinking.   I have tested this fairly extensively, and one thing is clear, maximum power production happens when the leading edge speed, and wind velocity are close to the same.   TSR 0.8 - 0.9   

If you'd like to argue about how it's working, I'm game, however if you are going to argue and ignore my measurements, we don't have much to talk about.

Kinda harsh on the statement so early on the posting aren't ya?

Bruce S

we've all got a limited time on this earth, I've wasted plenty of mine and others arguing about things I didn't understand.   I spend my patience on other things, nor do I want to waste others time.

Maybe it's the impression given by the word argue? We generally try to quell arguments before they get out of hand. We prefer discussions.

The use of Reynold's numbers, Tesla and Fibonacci sequencing is an interesting idea in fluid dynamics. Not sure boundary layer works in this case tho, will need to work out the numbers and design after work today.

AND yes I have built air powered Tesla boundary layer turbines using the old HDDs . I have access to a whole lot of them!
Keep up the testing! I'll keep watching

Cheers
Bruce S

[electrondady1]

 something i noticed coming here over the last 15 years . people with vertical axis mill prototypes are always looking to patent or prove how great their design is . but they never have one up producing power on a day to day basis  in their own life, like horizontal axis mills or water mills do. its always about selling the idea to someone and cashing in.  they never post saying, "hey this thing has been making x amount of watts every month for the last 4 years i am really happy about it" .

     

[oneirondreamer]

something i noticed coming here over the last 15 years . people with vertical axis mill prototypes are always looking to patent or prove how great their design is . but they never have one up producing power on a day to day basis  in their own life, like horizontal axis mills or water mills do. its always about selling the idea to someone and cashing in.  they never post saying, "hey this thing has been making x amount of watts every month for the last 4 years i am really happy about it" .

     

Yes, this is what I thought too after reading the Savonius Blackwell report, if these things actually reach a Cp of 0.21, where's the working hobbiest models, where's the commercial products?  After a decade of looking, I've found 3 reports of real world data collecting on Savonius turbines, and each was very disappointed that the output was a fraction of the Cp 0.21 expected.
    The sad reality is that the Savonius, the Darius, and even the best conventional HAWT's work poorly at Reynolds numbers less than 1,000,000. 

[Adriaan Kragten]

    The sad reality is that the Savonius, the Darius, and even the best conventional HAWT's work poorly at Reynolds numbers less than 1,000,000.
[/quote]

It's true that below a Reynolds value of 10^5 almost all airfoils get very high drag coefficients. But for a HAWT, you don't need a very large chord or a very high design tip speed ratio to reach the critical Reynolds number already at an acceptable low wind speed. The Reynolds number for a certain wind speed can be calculated using formula 5.5 of my public report KD 35. In this formula you can see that the Reynolds value is proportional to the product of the local tip speed ratio and the chord (if the factor 4/9 is neglected). Savonious rotors have a much larger chord than a HAWT for the same rotor diameter but also a much lower tip speed ratio. The combined effect is that the Reynolds value is about the same if the rotor diameter is about the same. However, to build a HAWT with a rotor diameter of 4 m is nothing special but to build a Savonious rotor with a rotor diameter of 4 m is almost impossible. So to realize the critical Reynolds number at an acceptably low wind speed is much easier for a HAWT than for a Savonious rotor. This may explain the low maximum Cp of small Savonious rotors.

For Savonious rotors used in water it is much easier to reach the critical Reynolds number than if used in air as the kinematic viscosity of water is about a factor 15 lower (1.004 * 10^-6 m^2/s for water and 15 * 10^-6 m^2/s for air).

[MagnetJuice]

That's interesting Adriaan.

Drew, I value your opinion because you have been studying these things for a long time.

I would love for you to give us your views on how low and high Reynolds numbers affect the operation of Savonius and Darrieus VAWT's.

Ed

[Adriaan Kragten]

That's interesting Adriaan.

Drew, I value your opinion because you have been studying these things for a long time.

I would love for you to give us your views on how low and high Reynolds numbers affect the operation of Savonius and Darrieus VAWT's.

Ed

In my public report KD 601, the functioning of a 3-bladed H-Darrieus rotor is explained. I have used the aerodynamic characteristics of the symmetrical airfoil NACA 0015 which are given in figure 1 for a range of Reynolds numbers. The rotor has a diameter of 2 m and the blade chord is 200 mm. For this chord, I found that the Cd/Cl ratio is low enough if the rotor runs at its optimum tip speed ratio of 4.2 and if the wind speed is 5 m/s. But for a smaller chord, a lower tip speeds ratio or a lower wind speed, you easily get a Reynolds value lower than 10^5.

Another point is that the measured characteristics of the NACA 0015 airfoil can be doubted because the used wind tunnel had a rather high turbulence rate. A high turbulence rate of the wind tunnel has about the same effect on the characteristics as increase of the Reynolds number. It has been shown that if asymmetrical NACA airfoils are measured in a wind tunnel with a low turbulence rate, that then the airfoil is much more sensible to stalling. But as far as I know, no measurements are available for the NACA 0015 airfoil measured in a wind tunnel with a low turbulence rate. The turbulence in a wind tunnel is created by the fan which drives the flow but it is reduced if the wind tunnel has a part with a bunch of small pipes followed by a large expansion chamber.

[oneirondreamer]

I think Adriaan and I agree on some things here,

that small bladed turbines (chord 200mm and less) suffer from poor Reynolds numbers, and that the high TSR's of HAWT's and some VAWT's give them a significant Reynolds number advantage vs low TSR VAWT's like the Savonius.   

I think Adriaan is pointing out with his report KD 601, that the minimium size Darius for effective Reynolds numbers is 2m diameter with 200mm blade chord, and that this is true for HAWT's as well, below 2m diameter Reynolds numbers are difficult and a flat sheet, or slightly curved blade (section of pipe) may work better than a fancy airfoil.

I think that if we looked in Adriaan's extensive HAWT archives we'd find that for the same amount of material, we could build a larger, more effective HAWT but that still are better to start at larger than 2m diameter.

Something I'm not sure about but will speculate on is that a HAWT produces most of it's power from it's tips, where it's got good reynolds numbers, so larger, non rotating nose cones, may help increase flow on the outside of the rotor.   A rotating cone might cause trouble by creating a non uniform flow field over the disk area of the rotor via the Flettner effect.

This is why Adriaan is not enthusiastic about VAWT's, he is showing us that he's literally done the research into all the most promising units he could find, and none have suggested promise.    It's a bit surprising that he's willing to continue to comment on this forum as it must be frustrating to watch people tilt at windmills he knocked over years ago. 

I agree with Adriaan, continuing on trying to develop VAWT's based on the Darius, and to be smaller than 2-3 m diameter is not likely to be fruitful for well understood reasons.   It's useful to look at the Darius's which Dr. John Dabir built over the past decade, and you will see that none are smaller than 2m diameter, and all have as deep or deeper blades than Adriaan is suggesting as minimum.     

However, I am reminded of Arthur C. Clarks quote

“ When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.”

What Adriaan is saying/showing, I think,  is that he's shown what it's possible, (in respect to the conventional understanding of how these work, his research and all the many others he's investigated), to get from a Darius, and found that specifically at around 2m dia 200m chord, where they begin to show any potential promise, that the time and materials would be better spent on a HAWT.   

However, I propose that Savonius Rotors have a more complex mechanism of action, and I would say that I've proven it, though certainly not to a scientific standard that I can expect a professional like Adriaan to accept at this time.    I am working to reduce certainty around this, and I'm thankful that Adriaan has been willing to critique and make suggestions on a project that from his perspective may seem a bit futile.   

[Adriaan Kragten]

I think Adriaan and I agree on some things here,



Something I'm not sure about but will speculate on is that a HAWT produces most of it's power from it's tips, where it's got good reynolds numbers, so larger, non rotating nose cones, may help increase flow on the outside of the rotor.   A rotating cone might cause trouble by creating a non uniform flow field over the disk area of the rotor via the Flettner effect.

It's true that the outside part of a blade of a HAWT produces most of the power but that isn't because of higher Reynolds numbers. Especially for tapered blades, the Reynolds number isn't reduced strongly if the local radius is smaller. This is because the reduction of the relative speed is almost compensated by the increase of the chord. The aerodynamic theory is developed such that the wind speed is reduced to 2/3 V over the whole rotor plane. Therefore a pressure difference is realized in between the front and the back side of the rotor plane which is the same for every small area dA (if the tip losses are neglected). So every small area dA generates the same power. This means that the outer half of the blade generates 3/4 of the power as the swept area of the outer half of the blade covers 3/4 of the total swept area of the rotor. The inner half of the blade covers only 1/4 of the total swept area of the rotor.

The real power generated by the outer half of the blade is reduced by tip losses as the pressure difference can't be maintained close to the blade tip. The real power generated by the inner half of the blade is reduced because the airfoil can't be maintained up to the center of the hub so the effective blade length is always shorter than R. Both effects give about the same reduction so even with these effects, the outer half of the blade generates about 3/4 of the power.

Nose cones have some positive effect to reduce the losses at blade root. The German Enercon wind turbine has a nose cone an a nicely curved head and blades with a very large chord and blade angle at the blade root. So there is almost no gap in between the blade root and the cone or in between the blade root and the head and this makes that the complete flow is used. These wind turbines have a maximum Cp of about 0.5 but it must be a lot of work to make these extra wide blades. I think that for small wind turbines it isn't worth while to make blades very wide and with a large blade angle at the blade root.

[MattM]

There appears to be some vertical components to these blades, so looking at it from strictly horizontally may not explain everything.  The advantage I would hope with a helical design is smooth drag rather than a pulsating drag while extracting wind energy.

[oneirondreamer]

There appears to be some vertical components to these blades, so looking at it from strictly horizontally may not explain everything.  The advantage I would hope with a helical design is smooth drag rather than a pulsating drag while extracting wind energy.

I think the vertical component is important, though I’ve looked at data from testing a Gorlov (licenced copy from Gorlov) vs straight 3 bladed Darius, and the Gorlov produced strong pulsations, and didn’t reach 0.2 Cp (testing in water off Vancouver Island), the Darius produced stronger pulsation, but higher Cp 0.24

I think a potential advantage that some helical VAWT’s may have, relates to the Reynolds number issue I’ve been digging at.    I’ll see if I can make myself understood.   This was first explained to me by a senior physicist, so no credit for me.   

In a HAWT, a blade functions in a continous fashion sweeping in a circle.    These blades will only stall if overloaded, tubulance, change of wind.   Let’s look at the phenomenon of stall, in stall fluid is no longer attached to the blade.   This is not an instantaneous effect, but a time sensitive one.   It takes a moment for the air to stall after a rapid change in angle, and as a pilot will tell you, stall recovery in a dive also takes a moment for lift to reestablish.    As your airfoil gets larger (higher reynolds number) it gets more resistant to stall and tolerates a larger angle of attack. 

Now when we look at a Darius VAWT we can see that it is likely going through stall 2x per revolution.   While you could plot a circle and say that on this portion of the circle the blade should be parting the flow in such a way as to cause lift, and that might be geometrically true on certain higher Reynolds number VAWT’s, as they get smaller the longer and longer stall recovery means that the energy availble is reduced.   Thus a 70m tall Darius Flo VAWT from NREL might have reached Cp 0.45, and yet a Mariah windspire with what is better theoretical geometry struggles to reach Cp 0.11 in NREL testing.   It is certainly interesting to note that Adriaan suggests a 2m minimum 20cm blade as a minimum, and the only third party certified VAWT starts at about 2m diameter. 

So it’s this dynamic stall behavior that I think can be influenced by a helical VAWT.    In this instance the flow attachment to the blade can maintain a continuous “lift bubble” as it travels up the blade.   This means that rather than have the momentary delay as flow establishes and attaches, a more continous flow can occur.    Something that may relate is that if my turbine is allowed to overspeed substantively, it can start a vibration of 2 pulses per rev, which I’m sure is an aerodynamic phenom.    When this happens if conditions are correct, you can see it shedding vortex’s 2x per rev, at the top of the blade (TSR1+) if the turbine is optimally loaded, nothing seems to shed. 

I had some great testing yesterday in my creek.   Conditions a bit difficult, snowing fairly hard, temps -3, but thankfully my boots continue to be a little taller than the stream (14” vs 12”).   This creek has about 10”-20 drop over a few hundred feet, and wouldn’t normally be considered suitable for any power generation.    I’ll try to get some video and analysis out today. 

[Adriaan Kragten]

In my report KD 601 I have shown that the airfoil may stall at the front and at the back side where the angle of attack is maximal. Stalling can be prevented if the tip speed ratio, the wind speed and the chord are large enough to realize a sufficient high Reynolds number. But there is always a wind speed for which the Reynolds number becomes too low.

Once stalling occurs, there is always hysteresis which means that the angle of attack must become much smaller to make that the flow follows the backside of the airfoil again. If the blades are in line with the turbine axis (straight blades), stalling will start for the whole blade at the same point of rotation and it will also stop at the same point of rotation. If the blades have a helical shape, stalling will start at the part of the blade where the angle of attack is maximal. This point will move upwards or downwards depending if the screw line is right hand or left hand. But I don't see a good reason why a helical shaped blade will stall for a smaller percentage of the time if the wind speed is that low that stalling occurs.

The tendency of stalling is larger than the tendency of cancelling it. So I think that just the opposite will happen. For a helical shaped blade, there is always a point on the airfoil where stalling starts. This stalling area will spread upwards and downwards and may ruin the flow at a large distance from the point where the angle of attack is maximal. So stalling for helical shaped blades may happen for a larger percentage of the time than for straight blades and so it may result in a lower maximum Cp at low wind speeds.

The torque supplied by a blade element varies strongly during a revolution. A helical shaped blade flattens this torque just in the same way as inclined magnet grooves flatten the fluctuation on the sticking torque of a PM-generator made from an asynchronous motor. It might be that the helical shape has also a flattening effect on the thrust variation due to beginning and ending of stalling. But making helical shaped blades for a H-Darrieus rotor is much more difficult than making straight blades and so I doubt if it is worth while.

[MattM]

If the blade is somewhat flexible it would explain two pulses per revolution.  It probably is the weakest point of the blade flexing at the maximum load.

Do you think maybe a double-helix would fair any better spreading the stalling point between multiple locations?

It would also be fascinating to see a triple-blade arrangement orbiting around a fixed gear like in a planetary gear system.   The tip speeds would more closely match windflow and each orbital blade would always move out of its own stagnant air.  I know, my imagination goes overboard.

[Adriaan Kragten]

It would also be fascinating to see a triple-blade arrangement orbiting around a fixed gear like in a planetary gear system.   The tip speeds would more closely match windflow and each orbital blade would always move out of its own stagnant air.  I know, my imagination goes overboard.

I think that what you mean is the rotating blade turbine which is described in my public report KD 417. The advantage of this VAWT is that it is mainly a lift machine and that therefore it will have a rather high maximum Cp. It has a high starting torque coefficient and can therefore be used to drive a load with a high starting torque like a grinding mill or a piston pump. Disadvantages are that you need a transmission with a gear ratio 1 : 2 to drive the blades such that a blade makes half a revolution for a whole revolution of the rotor and that you need a vane which keeps the blades correctly orientated with respect to the wind direction. I have built such a wind turbine very long ago (with a horizontal shaft) and there is a photo of this wind turbine at page 4 in KD 417.

[MattM]

I was thinking something more chaotic and simple.  As long as all three helical blades are not aligned and 1/3 out of phase they should never share stale flow with one section of blade let alone between blades.  Only a fixed outer ring would be necessary as centrifugal force will keep them bound as it rotates.  Could probably be wheels rather than true gears and nature's whole path of least resistance might naturally keep helical blades out of phase.  Keep it simple, reproducible.

Black represents fixed ring.  Green is fixed-location generator at sun position.  Blue is the carrier bar supporting the helicals.  Yellow marks demonstrate how blades would be out of phase and do not meant to suggest somehow they are not oneirondreamer's shape.  It's a crude picture, I know.

[oneirondreamer]

I’m going over my data from testing, and it’s very promising, however it looks like enough water got in to the alternator, that one phase was not working, and was shorted, making my V and A output reading likely missing a lot of rotor drag.    I’ll work out a better internal wiring, it should be able to be waterproof (though I don’t want the rotor drag of it having any substantial amount of water in it.   I’ve got an idea for a lard filled shaft gland, but need to get shaft supported better first.

One thing that I haven’t talked about in terms of why I’m so secure in my belief that a helical but thinned Savonius cross section is a different beast than what people are expecting is that I did some hydro testing on different versions of a rotor about 10 years ago, on the front of a fairly big boat (dual 120hp landing craft) with some good instrumentation.  The rotors were very much like what I”m working on now, but much shorter aspect ratio.   I did 3 versions of it, tightly twisted about 40deg, medium twist about 15-20, no twist.   The 15-20 twist worked very well, the 40deg twist didn’t work well, and the no twist, didn’t turn, even if started it wouldn’t remain turning.

[MattM]

When water passes through the front to the back wouldn't a helical shape at best only move half a rotation of the main shaft pretty much like any other drag VAWT? 

Or is the slope enough where it never gets tangled up moving with the water and acts more like a lift device?  My feeling is that if it acts more like a lift device than something with more turns that has enough less angle of attack to the water might open it up even more to performance.  But at some point of adding twists I think it becomes an Archimedes screw of some type.

[oneirondreamer]

Hello Folks,

I hope these next rounds of testing this system, to show  that it is working, and working best at higher speeds than would be expected in conventional thinking for a drag based system, but slower than expected in an efficient “lift” based system.   

I’ve tried to keep it lighthearted and short, so I hope you don’t find it difficult to take seriously. 

Next week I hope to be done a video on how I think it works.   I may be wrong, but I think I’ve got a good explanation and I have yet to do it justice in presentation.