Designing, Building and Testing a Darivonius VAWT

[Adriaan Kragten]

Report KD 416 about drag machines has been changed again. A new chapter 3 has been added and figure 4 gives the Cp-lambda curve for only the cup which is moving in the direction of the wind. This situation is about similar to a sail boat with a spinnaker sailing in the direction of the wind. But for a real drag machine this is completely different as there are always cups moving against the wind. Then the best you can get is the Cp lambda curve as given in figure 2 of KD 416.

If you use sails in stead of half hollow spheres, the situation can be much better as now in stead of drag, lift is used as driving force during a part of the revolution. But keeping the sail in the optimum position with respect to the spoke for every postion of a spoke with respect to the wind direction, is a difficult task.

[MattM]

VAWTs seem to run optimally in clean air.  How much efficiency is typically lost when your airflow becomes unstable across the X and Z axis?

[Adriaan Kragten]

Although the maximum Cp of a sail windmill can be larger than for of a pure drag machine and may be even larger than for a Savonious rotor, it has many disadvantages which are:
1) The optimum tip speed ratio will be smaller than one. This has as disadvantage that a lot of energy is wasted in wake rotation (see KD 35 figure 4.2) and that a large accelerating gearing is needed to drive a generator at a sufficient high rotational speed.
2) A mechanism is needed to make that a sail has the optimum position with respect to the relevant spoke for every position of that spoke.
3) A lot of material is needed for a certain swept rotor area.
4) A sail will flap twice during every revolution. This will make that the sails will wear very fast.

The main advantage is that the rotor will have a large starting torque but this advantage is only important if the load has a large starting torque which is the case if the load is a positive displacement pump. But this is the use for which these kind of windmills have originally been developed. For electricity generation, a much higher optimum tip speed ratio is needed other wise it isn't possible to use a direct drive PM-generator. So if one especially wants a VAWT, one can better chose for a Darrieus rotor and puts one mental energy in solving the starting problem (and the other problems inherent to Darrieus rotors).

[clockmanFRA]

Okay, lets step away from what's known!

I am still sticking with the Chinese VAWT concept, however its possible to use ridged aerofoils instead of canvas, however it may be possible to change the aerofoil shape of each blade as it rotates, but that's another step.

I am thinking simple here.  So using 'Manet Juice' first Image that shows the sail position for each position of the VAWT rotation.

6 sail VAWT freely rotating as a modern yacht sails concept.
Make the VAWT frame central axis tube hollow, fit another shaft in side with a good WIND VANE on the top of the whole VAWT construction.

 Rigidly attached to the wind vane at the top is a CAM in horizontal plane, this cam is carefully shaped so that each sail has a Rose bearing linkage attached to a roller that is kept spring loaded on the CAM. So in essence the CAM constantly follows/ leads the sails individually for there best wind capturing positions .

Problems .... getting the cam to stay still as the VAWT revolves around it. Keeping the wind vane into the wind, how big the wind vane.  Getting the CAM to its ultimate shape could take some time.  The sail boom arm with linkage will need to cross over itself as the sail Gybes, and changes tack, so hence using Rose bearings on the linkages.

Just some thoughts.

[electrondady1]

just to review the original design criteria :

"I'm starting this topic for the purpose of encouraging discussion from Forum members and hopefully guests that will join the Forum to share their ideas. The goal would be the gathering of ideas and suggestions that can lead to the design, construction and testing of a practical VAWT.

This VAWT should be:

•   Easy to build
•   Using mostly inexpensive and commonly found materials
•   Safe to use and operate
•   Robust to withstand strong winds"

[MagnetJuice]

clockman idea was already spinning in my mind and I was already researching wind sails. Sounds very interesting. I would like to experiment with that sometime. A windmill like that could be useful for heating a shop or a barn. First, we have to design and build something more practical.

Thank you for the reminder to stay on track electrondady.

Ed

[oneirondreamer]

I made an apples to oranges comparison when I posted the link to the Taiwanese turbine.    In looking at the SWCC doc again, I realize that the version tested by SWCC is their 3000 machine, which is 4m diameter and 10m^2, which reaches Cp 0.2 (and costs about 15,000$!) and is actually rated at 1500W, their 300 series machine that is 1.24m did and claims to be 300W has not been certified, even though it's what they are selling the most of...   So it's not likely Cp 0.20, and I'd suggest it's below the critical Reynolds number that makes a Darius worthwhile.   

It looks to me that the Darius design, helical, or straight, or tropenskien, has a minimum effective diameter of 4m, using a symmetric airfoil, given the real world commercial data we can see, and not the unrealistic wind tunnel performance.    That any aerodynamic analysis may predict something more effective, doesn't make it actually work, just shows that they analytic technique used may not meet the real world in some scales.   

[MagnetJuice]

I think that we have different concepts of what "makes a Darrieus worthwhile"

Our goal is to design and build a windmill that the average person with limited tools and materials can put together. We are not looking for Kilowatts, just 200 or 300 hundred watts.

At this moment, there are dozens of manufacturers selling small vertical turbines to the public and they are not 4 meters in diameter. Most of the ones being sold are frauds and belong in the trash pile. However, if you look carefully and know a little about power generation, you can find good ones. The good ones cost more, but they produce some power.

I don't want to buy a windmill. I want to design and build one. That is one of the reasons that I started this thread. I have the mechanical and electrical skills to build a working windmill, but I chose to seek the opinion and advice of the people in this Forum that are by far more knowledgeable than I am. With the help that I have gotten so far, I can build something a lot better that I could have built on my own.

But the main reason that I'm doing it here is because I know that there are many people that want to build a small windmill but don't know what to build. I want those people to benefit from our collective knowledge.

Ed

[MattM]

Sounds like common household products are key to your goals.  Material like foamboard or canvas certainly are in the realm of simple and affordable.  Maybe something like poor man's fiberglass?  Maybe stick with belt drive pulley off a central shaft to drive a commonly found PMG-based electric scooter motor.  Keep it to two or three common farm supply bearings.  Bonus if it uses a common squirrel cage assembly for the drag device as they are easy to salvage from old AC units.  Try to build using common plumbing parts as much as possible.  Common PVC endcaps screwed to the darrieus blades is simple and easy to duplicate.  Seems like building the turbine is the easy part.  Building the electronics is the not so DIY part.

[MagnetJuice]

Designing, Building and Testing a Darivonius VAWT

WOW, here we are, about a month into this project. The designing part will soon be ending and the building will begin.
 
After I have considered the great advice and suggestions, here is my tentative plan:

The turbine will be 72 inches tall by 48 inches in diameter (24 Sq. Ft – 2.23 m Sq.) For the Darrieus section, it will have three blades approximating the profile of a NACA0015 or NACA0018. I say approximate, because I will build those using common parts and easy construction.
 
I will buy a solid steel shaft 1 ¼ by 72 inches long (32mm x 1.83 meters). I chose 48 inches for the diameter because I want to use one sheet of 4 x 8 plywood for the bottom and top rotors of the turbine.

I have not yet decided on the configuration and the number of blades for the drag section in the center that will help start the turbine spinning. The ratio between the Darrieus and the Savonius-like center will be according to Adriaan's recommendations of 0.381 times the diameter of the Darrieus. It will be about 18 inches in diameter.

This is what I will be doing in the near future:

To determine the configuration and number of blades for the drag section, I will build a small specimen to do extensive testing. That small specimen will be 36 inches tall and 18 inches in diameter.

I made these patterns for possible configurations for the center section. The blade is an approximation of the CFD optimized Benesh blade. The small specimen for testing I will build in a way that I can easily change the number and the position of the blades. That way I can try many configurations to find the one that works best. These patterns are for three blades but I will be testing two blades also.

I would really appreciate your opinions about these patterns. Don't be afraid to be wrong, because as far as I know, nobody has evaluated anything like this, so no matter what you say, you can't be wrong. The one with five blades, how would that behave concerning TSR and RPM?


After I determine the best configuration for the center section, I will build the full size center section and test it using the alternator that I will build for the turbine.

I'm beginning to get excited about this.

Today I received my anemometers. I checked my fan and the highest speed I got was 18 MPH. I might have to build a shroud for the fan to concentrate the flow to see if I can get higher wind speed.

These anemometers are nice little gadgets, very sensitive. I went outside and the wind speed was 3 MPH. Then I came inside and I tested my normal walking pace. It was about 2.4 MPH. Next thing I want to do is give it to one of my kids when they are upstairs so they can measure how fast they come down when my wife calls them for ice cream. 

Ed

[MattM]

Just an observation, but it looks like 5 blades forms dead spots.  If you considered five, why not four?

[electrondady1]

i have a lot on my plate right now.
 i have eleven 60 ft. white ash trees that have been attacked by the Chinese emerald ash beetle and are  in the process of dying . they are becoming a hazard for me and my neighbors.  so i can't get involved too much in your project.

just want to repeat this one more time.
once you go squirrel cage you will never go back.
there is a whole different esthetic involved.
from any distance at all, you can't tell the machine is even rotating.
because of the the overall shape ( a cylinder ) the machine is loaded evenly upwind and down wind.
 the torque is constant .

[Adriaan Kragten]

For a nomal Savonious rotor with two blades and an overlap in between both blades, there will be a position of the rotor for which there is a flow through the rotor. Because of this flow, the rotor can work as a lift machine during a part of a revolution. Therefore, the maximum Cp and the optimum tip speed ratio of a Savonious rotor can be higher than for a pure drag machine. If you make a Savonious rotor with three blades, the path which the flow can follow is very complex as it has to split and I think that there will be almost no flow. So then you have a pure drag machine resulting a low maximum Cp, a low optimum and a low unloaded tip speed ratio. The problem for a Savonious rotor with five blades is even bigger.

But for a Savonious rotor with four blades there can be a certain flow if there is enough space in the centre of the rotor. If there is a certain flow for a certain position of the rotor, you will get the same flow if the rotor has rotated 90 degrees. The torque fluctuation for a Savonious rotor with four blades will be much less than for one with two blades so it might be enough to use only one Savonious rotor at the bottom or at the top of the Darrieus rotor. But you should test this rotor first and at least you should measure if the unloaded tip speed ratio is high enough.

[MagnetJuice]

electrondady, I'm sorry to hear about the infestation of your ash trees. I love life, animal and plant. It is always better living than dead. At least you will have some good firewood. Years ago, I used to buy seasoned ash wood for my wood burning stove for $40 a cord. I have good memories of when I stood in front of my wood burning stove cooking an omelet for my family. I cooked many meals using ash wood. The oak and the maple wood I used for heating but the ash I used for cooking because it was easier to control the heat from the ash wood in my stove.

Matt and Adriaan, both of you mention advantages of having four blades. That is interesting. I will make sure that I do some testing using four blades. Can you expand on that a little bit? My goal is to get the most power and torque from whatever configuration I chose for the center unit. I have been working with different configurations to have more space in the center of the rotor, but without the advantage of Computational fluid dynamics programs or wind tunnels, the only thing that I can do is go with my instincts.

Thanks again for your input,

Ed

[Adriaan Kragten]

The squirrel cage rotor has originally been designed as a fan to produce an air flow at a certain pressure. For this use, the air enters axial in the centre, then it bends 90 degrees outwards, then it enters radial in between the fans and it leaves the rotor almost tangential. There is a cover around the rotor shaped like a snail's shell which collects all separate flows to make one outcoming air flow. For this use, it is possible to give the fans that shape that a fan has the correct angle at the inside such that it points in the direction of the relative wind speed. The fans are curved such that a fan has also the correct angle at the outside. So it works just the same as a centrifugal pump. This fan use of a squirrel cage rotor can therefore have a rather high aerodynamic efficiency.

However, if such a fan is used as a vertical axis wind turbine, the flow enters the rotor at the front side and leaves the rotor at the back side. If you make the speed diagram at the front side of the rotor at the outside entrance, the fans can have the correct angle for the relative wind speed. But if you make the speed diagram at the back side of the rotor at the inside entrance, the fans have complete the wrong direction. So at this position, the fans will block passage of the flow and this will make that almost no flow will pass the rotor. If there is almost no flow through the rotor, the rotor can't work as a lift machine.

What remains is the flow going around the rotor and one side of the rotor will have more drag than the other side because of the cambered shape of the fans. So in reality a squirrel cage rotor works mainly as a drag machine which includes a low Cp and a low optimum tip speed ratio. A Savonious rotor with only four blades can be seen as a kind of squirrel cage rotor with a low number of fans and it will suffer from the same problem if the blades are shaped incorrectly.

If you look at the rotating blade turbine explained in my public report KD 417, you see that the blades have the correct position at both sides of the rotor but this is realised because the blade itself rotates with half the rotational speed of the rotor. If the blade position is fixed, I expect that a completely correct blade shape isn't possible but you should try different shapes to find the best geometry. The centre of the rotor must be free to allow passage of the flow in two directions. So the four blades can't be that long at the inside as for a 2-bladed Savonious rotor. I think that the blades can be made from a cylinder but that you use only 145 degrees of the cylinder. The diameter of the cylinder should be about a factor 0.6 of the diameter of the rotor. I have made a composte drawing with this geometry and it looks rather optimal. This drawing is added as an attachment.
 

[MagnetJuice]

Very enlightening Adriaan. I will work with different blades to try to find out the optimum configuration. Tomorrow I will post some images for consideration.

Thank you

Ed

[MattM]

The most efficient designs are the least disruptive of the airflow.  The squirrel cage is one of the best case scenarios for a VAWT.  The blades have the smallest camber in the z-plane and they work best with minimal blade thickness on the x-plane.  The width on the y-planes is limited by material stiffness.

One of the lessons of HAWT design is that diameter and blade pitch are the biggest control factors on power.  This would mean that you can get fancy and the results won't vary much from a simpler design for the same blade diameter and pitch.  On the HAWT you also rely on blade pitch and diameter, only your movement is on a different axis.  The drop in efficiency is due to relative position of the blade to flow and trying to squeeze out flow from the perpendicular positions only adds an infinite limit of complexity.  Better to keep it simple.

[Adriaan Kragten]

The squirrel cage is one of the best case scenarios for a VAWT. 

I strongly doubt this statement. A squirrel cage rotor may run smoothly but this says nothing about the flow pattern through the rotor and about the maximum Cp which can be obtained. Can you show any reliable measured maximum Cp values at a reasonable tip speed ratio in an open wind tunnel or in real wind to prove your statement that it is one the best case scenarios for a VAWT? Even if the maximum Cp would be acceptable, the amount of material which is needed to build such a rotor is very high and the rotor can't be protected against a high thrust force at very high wind speeds.

[MattM]

I don't have an engineering degree and the research is definitely your skillset over any other members here.  But before you doubt it, did you test it?  I only casually spent about two winters on trying to make a VAWT work.  At that time I had a sheet metal shop at my daily disposal and tried an awful lot of designs.  The consistent best results was keeping the air inside the diameter as clean as possible and keeping the blades along the diameter.  I never really liked building blades with round profiles and found sigma or ogee shapes easiest to create with the machinery I was using.  Darrieus designs actually were pretty interesting, but they needed help to get over the initial drag.  Squirrel cages had zero startup issues and we're very forgiving of chaotic wind patterns.

[Adriaan Kragten]

I don't have an engineering degree and the research is definitely your skillset over any other members here.  But before you doubt it, did you test it? 

No, I didn't test a squirrel cage rotor and that is because I don't believe in this kind of VAWT. But I have tested the rotating blade turbine and I have built a small H-Darrieus rotor which worked very badly because of too low Reynolds numbers. To my opinion it isn't necessary to test a certain principle if the negative aspects of it can be found by logic thinking. In my report KD 601, I have given the speed diagrams for a H-Darrieus rotor for twelve positions of the blade. If you follow the same procedure for a squirrel cage rotor, you will see that it is possible to give the blades a position and a camber at the front side of the rotor which is optimal for a certain (rather low) tip speed ratio. However, if you make the speed diagram at the inside of the back side of the rotor you will see that the position of the blade is completely wrong because the direction of the relative wind speed W is about perpendicular to the blade. So it is easy for the flow to enter the rotor but it is very difficult to leave it and this means that there will be almost no flow through the rotor. So this turbine works only on the flow around the rotor and this means that it works as a drag machine. In my report KD 416 I have proven that drag machines have a very low maximum Cp value, a very low optimum tip speed ratio and use a lot of material and that therefore it is a waste of time and money to develop them.

Of all possible VAWT's, the Darrieus rotor has the highest maximum Cp, the highest optimum tip speed ratio and the least material required but it has starting problems. Using a Savonious rotor to make it start is an option but this must be done such that the Savonious rotor isn't producing a strong braking torque if the Darrieus rotor is working at its optimum tip speed ratio. This limits the diameter of the Savonious rotor. It isn't a real problem that the Savonious rotor has a low Cp as the main power is generated by the Darrieus rotor. But a 2-bladed Savonious rotor has a very fluctuating torque and this resulted in the idea of a 4-bladed Savonious rotor presented in my previous post. But this 4-bladed Savonious rotor isn't a squirrel cage rotor with only four blades. I don't know if this 4-bladed Savonious rotor will be strong enough to start the Darrieus rotor. I also doubt if it is useful to use only a big 4-bladed Savonious rotor.

[MagnetJuice]

I created a new blade for these images. I realized that the CFD optimized Benesh that I was using was wrong for anything other than a two blade turbine. This new blade seems to work good with three blades, but its shape can be modified to make it work even better.

I feel good about some of these patterns.

Take a look and let me know which pattern could work, and if the TAIL of the blade should be bent differently.

[Adriaan Kragten]

I will now investigate what kind of flow can be expected for the 4-bladed Savonious rotor for which I have given a figure in my previous post. The rotor has four blades and in between these four blades there are four channels which are called: A, B, C and D. In the centre of the rotor there is a part with no blades and the flow in the centre must have a certain direction. So it is assumed that only a flow is possible to opposite channels. There are four directions of the flow possible being: from A to C, from B to D, from C to A and from D to B. This means that if there is a flow from A to C or from C to A, there will be no flow from B to D or from D to B. So there is only a flow in a certain channel for half the time and that there is only a flow in a certain direction in a channel for a quart of the time, so for 90 degrees of a revolution.

This means that the speed of the flow in a certain channel has to accelerate from zero to its maximum value and then decellerate up to zero within a quart of a revolution. This can't result in a large flow, so the lift effect of a 4-bladed Savonious rotor will be limited. Fast fluctuation of the flow is more similar to a sound wave so it might be that a 4-bladed Savonious rotor produces a lot of noise if it runs fast enough. If this really happens, must be tested in practice. For the calculation of the maximum rotor diameter of the Savonious rotor it was assumed that the unloaded tip speed ratio is 1.6. So it has also to be tested if this unloaded tip speed ratio is reached.

The next question is at what position of the rotor, the flow will have in a certain direction. The rotor in my previous post is drawn such that the centre of the blade radius r is lying at the x-axis or at the y-axis. I think that for this position, there will be no flow from the front side of the rotor to the back side. I think that for this position, there will be a flow from the left side to the right side and that this flow will have about its maximum speed for this position. It this would be true, it means that the flow starts 45 degrees earlier and ends 45 degrees later. But the angles may be different for a braked rotor than for rotor which is rotating at its optimum tip speed ratio and also different for a rotor which is turning unloaded. A lot of research is needed to get a good insight in the real flow pattern through the rotor. But it seems possible to find out if this rotor can be used to start a H-Darrieus rotor by try and error.

MagnetJuice, I think that my blade geometry is better than any of the eleven options which you present in your last post because for my geometry, the flow in the centre isn't hindered and no sudden change of the flow direction is required. For 3-bladed rotors the flow pattern in the centre is even more complicated so the centre will form a large resistance for the flow.

The pressure difference over the rotor is very small and there will only a flow through the rotor if inside the rotor there is almost no resistance. If the resistance is too high, it is easier for the flow to go around the rotor. Resistance is caused by two effects, sudden change of the flow direction and sudden change of the cross sectional area of the channel. So you should draw flow lines in the channel and see if the flow is smooth. For my rotor you can see that all flow lines have a smooth S-shape and that there are no sudden changes in the cross section area. For your options 4A, 4B and 4C there is a very large change in cross sectional area and that the flow also has to bend strongly if the flow enters and leaves the central part. This will give a lot of turbulence and so a lot of friction and so the flow through the rotor will be small.

[MattM]

Adriaan Kragten-

A squirrel cage actually has less material involved than a Savonius, not more.  They were extremely easy to build, using simple wire-roll edges from the Pexto rotary tool on sheet metal rings at the top and bottom, with a single arm across the diameter at the top.  Wire rolling gives a lot of stiffness to the ring.  If you use stainless sheet you get a very rigid and lightweight cage.  I just used pop rivets to mount blades around the rings.  Commercial air handlers can survive decades using this simple construction method, although most of them are using galvanized that is spot welded together.  I feel more comfortable with riveting for durability, but spot welding is the faster way to manufacture in quantity.

With a squirrel cage, your resistance drops to airflow because of the large open volume for which the air to pass.  The single biggesr advantage of a squirrel cage is how it increases the frequency of exposure to airflow available for driving the rotor.  Darrieus lacks frequency and focuses on maximizing the effects of when it is in optimal position.  The squirrel cage always has some surface in that airflow.  And they are to a greater degree more omni-directional, whereas a wind direction shift is going to change your frequency for lift on Darrieus and drag on a Savonius that has fewer blades.

[electrondady1]

"No, I didn't test a squirrel cage rotor and that is because I don't believe in this kind of VAWT".
 
you have to believe adriaan https://youtu.be/HKh6XxYbbIc

[Bruce S]

Actually:
There have been tests in the past and posted here on the forum
CmeBrew had a pretty good post about a fairly large unit and it predecessor
https://www.fieldlines.com/index.php/topic,129155.25.html.

Not sure of the outcome, but there is a bit of history for the "Squirrel Cage" types.
Only thing people need to understand going into these builds is that there's no boilerplate models with which to glean information from.

Cheers
Bruce S 

[CraigM]

MagnetJuice,

I know you're working on the Savonius portion of your build but wanted to go back to reply #95 from Adriaan concerning the Darrieus airfoil shape question you asked in reply #81.

Adriaan replied about a self starting Darrieus that used an eccentric and push-rod arrangement to change the blade angle of attack when rotating. It dawned on me that Ed Lenz of Windstuffnow.com did some testing with this same type of arrangement and reported the Darrieus he built was indeed self starting.

This is from Windstuffnow.com







Adriaan – I'm not completely sure of the wind direction in the 12 position diagram. I believe it's coming from the 6 o-clock position... this is where you show a + 9 degree angle of attack at station one.

Adriaan Kragten reply #95;

In figure 2 of my public report KD 601, I give the speed diagramn for 12 positions of the blade and for a 3-bladed rotor with a NACA 0015 airfoil. In table 4, I give the angles of attack alpha for these positions if it is assumed that the tip speed ratio is 4.2 and that the wind speed in the rotor plane is 2/3 V. You can see that you have the maximum angle of attack alpha of + 9 degrees at position 1, the minimum alpha of - 9 degrees at postion 7 and that alpha is 0 degrees at positions 4 and 10. This means that you have to use a symmetrical airfoil and that the blade angle must be 0 degrees!

[ Specified attachment is not available ]

Going back to the eccentric and push-rod... the tail fin positions the eccentric to allow equal distance of the push-rod to blade at 3 o-clock and 9 o-clock position where angle of attack is zero. The eccentric offset, whether it's in front of or behind the axis of rotation for the blades determines the angle of attack change at 12 o-clock and 6 o-clock positions. The eccentric offset and push-rod remind me of radio control airplane controls used for elevator and rudder movement. Changing the push-rod attachment point on the control horn (arm) can fine tune the angle of attack change imparted on the blade.

This also looks nearly what a sailboat would be doing at a beam reach where the wind direction is perpendicular to the boat direction. The upwind position is zero because there is only drag to overcome and the downwind position is zero because tip speed is greater then wind speed. A sailboat heading directly downwind is the same as a drag based machine.

Adriaan – Was a working turbine made with this arrangement? If so how did the mechanism hold up? The mechanics of it are rather simple but it adds a bit of something that can fail over time.

I know Ed Lenz used to post in this forum. Maybe someone remembers if he elaborated on his Darrieus design.

Thanks,
CM

[MagnetJuice]

That's interesting Craig, I have been considering something similar to that, but placed under the turbine. But instead of using the push-rod to pitch the blades, I wanted to use it to activate flaps to slow down the turbine when the wind is blowing really hard. I was thinking of using weights and springs activated by centrifugal force.

I think that to help start the turbine, the simplest solution would be some Savonius-like device in the center.

Thanks for the tip; it is always good to have many options.

Ed

[Adriaan Kragten]

MagnetJuice,




Adriaan – Was a working turbine made with this arrangement? If so how did the mechanism hold up? The mechanics of it are rather simple but it adds a bit of something that can fail over time.



CM

In chapter 1 of my report KD 417 about the rotating blade turbine, I refer to the H-Darrieus rotor with oscillating blades. A rather big 3-bladed turbine of this type called the Gyromill has been tested on the testfield of Energie Anders in Hoek van Holland, The Netherlands and I have seen it running. I have also built a small H-Darrieus rotor with oscilating blades myself and it worked but I found the mechanism to complex to continue. It makes use of an eccentric and a strip connected to the blade which creates a positive blade angle when the blade is at the front side and a negative blade angle when the blade is at the back side. The main disadvantage of this system is that now a vane is needed which makes that the line through the eccentric axis and the rotor axis is in line with the wind direction. In report KD 601 I have described a system with oscillating blades which needs no eccentric but for which the oscillation at low tip speed ratios is steered by the aerodynamic force acting on the blade. This system works for any wind direction and needs no vane.

In KD 601 I have shown that the optimum tip speed ratio is about 4.2 for a fixed blade and that this results in an angle of attack + 9 degrees at the front side and of - 9 degrees at the back side if the blade angle is 0 degrees. Now assume that the blade is oscillating such that the blade angle at the front side is 5 degrees. This means that the angle of attack is now only 4 degrees and the generated lift at this angle is a factor 4/9 of the lift generated at 9 degrees. The get enough lift to slow down the wind speed in the rotor plane to 2/3 V, the chord c must be increased by a factor 9/4 and it becomes 9/4 * 200 = 450 mm. This is a very large chord for a rotor with a diameter of 2 m. So it is better that the oscillating system works only at low tip speed ratios and that the blade angle becomes zero when the optimum tip speed ratio of 4.2 is reached.

For a blade angle of 9 degrees, a symmetrical airfoil NACA 0015 in not yet stalling if the Reynolds value is high enough but at low Reynolds values it may stall. I have calculated the Reynolds value for the optimum tip speed ratio of 4.2 and for a wind speed of 5 m/s and then Reynolds is high enough. But if the rotor is running at lower tip speed ratios, Reynolds will become too low and then it is good that the angle of attack is reduced by the oscillating movement.

[Adriaan Kragten]

I think that to help start the turbine, the simplest solution would be some Savonius-like device in the center.

Ed

If you want to use a Savonious rotor which suplies enough torque and for which the torque fluctuation is limited during one revolution, I think that there are only two realistic options. One is to use a 4-bladed Savonius rotor and one is to use two 2-bladed Savonious rotors which are positioned 90 degrees with respect to each other. Assume that there is place for a certain rotor diameter D and a certain height H. For two 2-bladed Savonious rotors every rotor will then get a height 1/2 H. I think that the torque level of the two 2-bladed rotors will be higher than for the one 4-bladed rotor because for a 2 bladed rotor the flow through the rotor can be active in one direction for 180 degrees of one revolution as it is only active for 90 degrees for a 4-bladed Savonious rotor. The unloaded tip speed ratio of a squirrel cage rotor will be too low and the allowed maximum rotor diameter will be too small if you don't want that the squirrel cage rotor is working as an aerodynamic brake if the Darrieus rotor turns at its design tip speed ratio.

[Adriaan Kragten]

Adriaan Kragten-

A squirrel cage actually has less material involved than a Savonius, not more.  T

Something doesn't become good if you compare it with something which is also bad. For a nomal 2-bladed Savonious rotor, the blade area is about a factor 1.5 times the swept area of the whole rotor. But you also need a top and a bottom sheet and if these areas are also taken into account, the needed material area is about a factor 3 larger than the swept area. For a squirrel cage rotor you won't be much lower. But if I design a simple HAWT with blades made out of cambered steel sheet, I have a total blade area which is only about 15 % of the swept rotor area and the maximum Cp is also at least a factor two better. So for the same power, the amount of needed material is very much lower than for a Savonious rotor.

If a Savonious rotor or a squirral cage rotor was really a good design, the world would be full of them but in reallity you only see HAWT's, especially for very big wind turbines. If you consider building of these kind of wind turbines as a hobby and if you don't compare the investment costs with the value of the generated energy, then it is a good way to spend your time. But for commercial use, they will always loose the battle with HAWT's. Most of these commercial small VAWT's can't generate the energy value in the whole lifetime of the turbine which was needed to buy and maintain them.

[MattM]

The whole idea behind using an ogee or sigma shape of blade was because I wanted the stale air that is in front of the blades to move under the blade, which keeps the air moving centripetally.  But because the air moves considerably faster than the tangential velocity of the rotor, it's a minimal disruption of the airflow and smoke passed through it.  The interesting thing I did notice with the squirrel cage was air did not get pushed around the cylinder by the proceeding side of the drum, nor did much smoke seem to ever go through the top or bottoms of the drum.  I would have expected significant centripetal force to move air from the center out the ends.  I even tried making the arms that went from the center to the outside edge into propellers to lift air out of the middle and it was an insignificant amount of lifting due to the low rotational speeds involved.  So my theory from watching smoke go through the squirrel cage was that it's going to go through it regardless of the number of blades.  So quite frankly the number of blades do not add considerable drag, but the sheer number sure does have an impact on keeping the drum moving in lockstep with the airflow.  The part I did not see as helpful on the Savonius was blocking air moving through the center by overbuilding the width of the blades.  They worked better keeping the center clean.  And the results of messing with the Savonius is what led me to the conclusion that the squirrel cage was better than any Savonius will be.  But to say it is bad is just nonsense.  To say it is not the best is a matter of application.

[CraigM]

@ Adriaan,

Thanks for the reply, KD 417 was an interesting read. Quite the machine you built 40 plus years ago. I was a 17 year old kid when you were working on it.

I googled “cyclorotor” and found some very interesting info on using these as a thrust vectoring device. I fly fixed wing R/C and saw a cyclorotor drone, pretty cool stuff.

I have a better understanding now of how a cycloturbine (gyromill) operates and may pursue a test design of my own.

@ MagnetJuice, don't want to hijack your thread any longer. Thanks for giving me a second.

Thanks
CM

[MagnetJuice]

No problem Craig, it is not hijacking, it is all related and interesting.

I did a test with a ‘squirrel cage' roof vent. I don't know how useful this test is but it revealed something anyways.

I had bought this vent years ago at a garage sale. My plan was to build a windmill with it to power a few LED's at night for the yard, but never did anything with it.

This vent is 17 inches in diameter, which is larger than average. I placed it next to a plant so I could see how windy it was. It started spinning very easy with a slight breeze. Then I placed a five Lbs. weight on top, well balanced, and it took a stronger wind to get it to spin. After that, I placed a ten Lbs. weight on top and it turned easy when I spun it by hand, but it didn't spin by itself unless there was a strong gust of wind.


I think that if a squirrel cage device was in the center of a heavy H-Darrieus windmill, it had to be very large to get the windmill to start spinning. I think that they spin easy but don't have much torque.

I did a Google search for this type of mills and saw a lot of them spinning on Youtube but nobody said anything about generating any power.

I'm leaning towards simplifying things and just use two cups using the optimized Benesh blades for the center of the windmill. That way I can keep the ratio of Savonius to H-Darrieus at 0.381 to 1. Also I can start building without spending a lot of time testing different blade configurations. Instead of building a small Savonius windmill for the testing, I can build the full size one and do testing on that. That will speed things up so I can get the show on the road.

I'm thinking of just placing one Savonius on the bottom half. I don't know if that would be enough to get the turbine spinning. If not I can place another one on the top half also.

Forgive my crude drawing; graphics is not one of my strengths.