Index (in reverse order so the latest stuff is at the top)Change of DirectionMeanwhile, back in the garage... Darrieus Test Conclusions Early Darrieus Test Results Testing the 1.2 meter Darrieus 1.2 meter Darrieus Setup Combination Rotor Prototype Problems & Solutions References Back to the Overview Change of Direction, 15 February, 2006It's been a while since the last update. Careful examination of videos of the 1.2M rotor have shown me the nature of the problem: Aerodynamic imbalance between the upwind and downwind blades. The pulse every 1/2 revolution only promises to get stronger as the rotor's speed increases. Successive modifications to the rotor over the past 6 months have served to increase the vulnerability to this vibration because they also increased its speed and loads. Raising the torque would have given better performance, but was hampered by this problem. I considered this for over a month, before deciding I should take the rotor down and work on something else for a while. Meanwhile, back in the garage..., 10 December, 2005In the meantime, I've been adjusting the big Darrieus rotor. I added weight to the red blade to try to improve the balance of the rotor, and machined a set of fittings to more solidly mount the blades to the arms. The fittings worked well; the blades are set at the optimum pitch. The balancing, on the other hand, seems to have had the opposite effect! I think I put too much weight into the red blade. Taping a bit of rod to the yellow blade corrects the vibration (mostly). There may also be an aerodynamic imbalance that will be harder to pin down. For now I've put a trim tab on the red blade, but it didn't seem to help much either way. I assembled a tachometer using a DC motor. Reading the voltage from the DC motor will tell me the RPM more directly than from the video. Test Results of my Darrieus-Type Wind Turbine, October 25, 2005A list of observations, in point form:
A truly productive rotor will require a means of eliminating or minimizing regions where "back-torque" exists. The possible solutions below must be studied.Early Darrieus Test Results, October 18, 2005 and
Video (375 kb)In the video, you can see the two rotors turning at the same RPM. The wind is increasing and the Darrieus has yet to catch up. Typically the Dar turns at 5 or 6 times the speed of the anemometer (that is, the RPM is about 25% higher, and the blade's tangential velocity is 5 times higher). Given the anemometer turns at a TSR of 0.16, I still haven't quite achieved TSR>1. The rotor response lags the wind speed by a second or two, depending on the strength of the wind gusts and lulls. The figures posted below are examples of that effect. A second-by-second graph of the motion shows the rotor has considerable inertia. This is making it hard to start, so I can see another round of lighten/stiffen re-designing in the future if I'm going to build another model. Provisions exist to mount two more blades on the hub of this model, so I have several options. Testing the 1.2 meter Darrieus Model, October 15-16, 2005Over the past week, I've collected data from the rotor, leaving it to run without any loads. I have already been able to prove that bearing losses are very important - they can quickly eat away at the speed. I currently have something like a trailer axle setup - two opposed conical bearings. As I tighten them together, it's like applying a brake to the turbine. For now, I have them loosened off, but there's no chatter, so I think I'm all right.I'm crunching numbers from my first videos. By recording both the 1.2M rotor, and the little savonius in the background, I can watch immediate wind speed, and the rotor's acceleration/deceleration speed simultaneously, frame by frame. The pitch of the blades seems to have an effect on the Tip Speed Ratio. I expected this to happen and when the data is ready to display, I can show curves for finding the best pitch angle. With more work, adjusting the length of the arms, I can give supplementary data on whether the pitch of the blades is affected. Promises, promises... 1.2 meter Darrieus Model, October 10, 2005I opted to stick with straight vertical blades because they're simply easier to make. Using styrofoam and fiberglass bonded to the surface, I have two blades 9.5" chord and 46" long. They weigh less than 2 pounds each. The inertia of the rotor should be very small, compared to its volume. This should allow starting with light winds. The airfoil shape of the blades is not symmetrical. I deliberately cambered them to discourage the "back-torque" angle that I discovered on the other model. In this area, the rotor would park itself and simply weathercock. The idea doesn't seem to have worked, but performance of this rotor has greatly improved over the model, so I don't think I'm being penalized for the cambered blades. The rotor has the following geometry:
[Oct 10] Early data (free running) are as follows:
In the scanty data above, I've noticed that higher TSR's prevail in the light winds. I don't know of a reason for this to happen - in fact, I can think of a few ways the opposite should happen. With more data collected, I will watch to see if this pattern continues. It is possible that the wind was dying when I took the low wind speed data, making the anemometer slow down, but the greater inertia of the Darrieus rotor kept it going, giving a false reading. There may also be hysteresis in the anemometer that would cause that. I have built the rotor to allow adjustment to the blade radius and pitch (no, not dynamically). Over the coming month, I will test the rotor with varying geometry to develop shape optimization curves. Combination Rotor, September, 2005Here the plywood Darrieus and the mini Savonius rotors have been mated on the same shaft. The point was to have a Darrieus rotor that is started by a Savonius, and then spins up independently. In the end, what I got was a Darrieus that started itself (good), and then bumbled along at the Savonius' characteristic speed (bad). Video of the self-start process proved to me that the acceleration of the rotor is linear. This was a big relief, as it implies that the torque acting on the rotor (when it is slow) is nearly constant, and so is the resisting torque from the bearings. This is important for future analysis. Upon disassembling this rotor, I have converted the Savonius rotor into an anemometer. By recording video of the Darrieus test rotor with the Savonius in the background, I can measure wind speed from the RPM of the Sav. The Plywood Prototype, July 2005This model was assembled over a weekend, with bits and pieces gathering dust in the garage. The blades are no more that a piece of 1/2" plywood sawed edgewise with a bevel. The point was solely to learn the basics of making such a turbine run. Obviously, the aerodynamics and performance are poor, but I was pleasantly surprised by how easy it was to make it run. A small push was usually all it needed. Lengths of the arms were varied: Diameters of 15", 21", and 30" were tried. In a 20kph wind, the following results were obtained:These results were poor from a performance point of view (TSR should be >1), but that was to be expected from the manner in which I fabricated the blades. However, this did teach me a lesson about the effect of rotor solidity, and while considering this fact, I also delved into bearing losses, moment of inertia, and Reynold's number. In no configuration would the rotor start by itself, but the least impulse is required when the diameter is small. Even in 60 kph winds it would resolutely remain weathercocked at about 45 degrees to the oncoming wind. Pitching the blades at different angles had a noticeable effect on the RPM. Also, the axis about which the blades pitch seems to matter, meaning that pitching a blade around an axis through itself has a small effect, while pitching it around an axis somewhere in on the radius has a bigger effect. I was not equipped to explore this behaviour in great detail on the first rotor.
Darrieus Turbines - Problems and SolutionsGetting a Darrieus to start up on its own as the wind rises is always an issue. Usually they stop at a point where turning further puts the blades into a zone of "back torque", where drag overcomes lift and it won't turn further. When running, it can scoot through this zone. The right geometry of turbine can start on its own, but its solidity is far too high to work efficiently. On the other hand, if its solidity is a bit lower, it won't start without a push. The generator control circuit would need to start it up when there's enough wind, and know when not to do so. This is complex, and could lead to batteries drained by a generator constantly trying to start a turbine that dies between gusts. After searching extensively on the internet, I discovered a wide variety of Darrieus
machines, shapes and sizes and combinations with other types, too. Through various means, I have found several
ways to control or augment the performance of a Darrieus rotor. I look forward to exploring each of these in turn.
References:
Updated 30 March, 2006 Created 15 September, 2005, Steven Fahey |
