I guess if the shroud were octagonally shaped then that would probably come close enough to an approximate circle to eliminate really counterproductive flow patterns near the corners of the square:
(Note – there are some errors in this post that I haven’t had time to fix yet, but I’m sure that if you know mechanical engineering you can easily correct the errors yourself. I think this idea might have potential once the errors are corrected. Note also that the torque tube will probably remain fixed with respect to the stationary tower rather than rotating around it. Also note that the struts each need to be connected by a vertical lattice (near the stationary tower) to keep them separated… that is, to prevent the load that tends to bend the ends of the struts towards each other from being transferred to the rest of the structure, thereby defeating the fundamental purpose of the idea.)
(Okay, here’s a pic with some errors corrected, but with no explanation:
This is a 3 bladed turbine, but I have drawn only two blades in order to make the illustration easier to understand. And I realize there are a lot of “legitimate” mechanical designs to realize this concept, likely using gears instead of tires and so forth. But I’m not a mechanical engineer, and so I just want to draw something that will give the real designers an idea they can play with.
Because the tower does not rotate, the rotor can be very tall, very slender, and it can spin at high rpm without becoming centrifugally unstable. But can’t the stationary tower can bend just as much as the rotating tower? And if the stationary tower bends, won’t this cause the rotating part of the structure to become centrifugally unstable just as if the tower were rotating? No. To see this, consider what happens when the middle of a rotating tower bows in response to the lifting forces transmitted to the tower from the airfoil by the middle strut. In this case, the middle of the rotating tower bows in the downwind direction, but its rotational axis does not change. Therefore the mass of the rotating tower has been displaced from the rotational axis, and centrifugal force now acts to cause even more bowing, and the rotor has become unstable. But when the middle of the stationary tower bows in the downwind direction, the rotational axis of the middle struts and airfoils moves along with it. And so although the rotor’s axis of rotation is no longer straight, it is at least centrifugally stable.
Another advantage of this design is that the guy wires are not connected to the tower through bearings. This should provide a big reduction in mechanical losses, since the bearings at the top of a traditionally guyed Darrieus bear a very heavy load – the rotor’s overturning moment. Of course, the overturning moment must be supported somewhere by some bearings. This design has bearings inside the rings that the struts attach to. So is there any advantage in this compared to the traditionally guyed Darrieus? I’m not a mechanical engineer, so I don’t know. Maybe there’s no advantage at all, but I’m wondering if the approach here isn’t better because it is easier to influence the bearings at design time. For one thing, you can spread the load over as many bearings as you want, while the traditional design requires two sets of bearings – one at the top of the tower and one at the bottom. For another thing, the guy wires in the traditional design are not only trying to torque the bearings about a horizontal axis, they are also doing this cyclically, from very low torque to very high torque several times a second. Surely this can’t be good. Of course, the present design also places a cyclic load on the bearings – there’s no way to avoid that. But at least it’s a “typical” load in that it doesn’t try to twist the bearings to a new axis. So maybe this is a better approach. It seems to me that mechanical losses will be decreased by eliminating the torquing thing, but again, I don’t really have the background to know if this claim is accurate.
There’s not much detail in the diagram, and much has been omitted. But I’ve drawn all the mechanisms involved so many times on this blog that I don’t think I’ll draw them again. But let me explain how it works. Heavy cables connect the blimps to the ground and carry the large loads. The symmetrical airfoils travel from the ground to the blimp and then back to the ground again. These airfoils are supported at either ends by moving cables (not shown) that turn a bunch of pulley wheels. The pulley wheels are suspsended from the heavy cables. Power is transmitted from the airfoils through the pulley system to a couple of generators at ground level. Each rail car has one of these generators on top of it.
Alternatively, replace the airfoils in the diagram with relatively small diameter Darrieus or H rotors. Power is still transmitted to the ground mechanically with a pulley system.
Suppose an airborne turbine can choose its altitude, as in the various blimp supported turbines proposed on this blog. In this case, the blimp may choose to hover in the clouds if it’s a cloudy day. Since the air in the clouds is wet, won’t its density be greater? And since the energy in the wind is proportional to density, then won’t this mean that the wind in the clouds will have more energy? Of course, the blimp may not be able to reach that altitude, but if it can, then here’s a possible option for increasing the amount of energy harvested.
About an hour after I posted this, I read in Wikipedia that the density of air actually decreases as a function of humidity. I was quite surprised to read this, but if you don’t believe me, check out the Wikipedia article for yourself. The reason it decreases is that “the molecular mass of water (18) is less than the molecular mass of air (around 29)”. One of my earlier posts suggested increasing the energy density of wind approaching a turbine by injecting a fine mist into the wind in some upstream location. (I was thinking of those misting devices they have to cool people down on the patios of some coffee shops and restaurants.) The Wikipedia article raises an interesting question. Is air that carries a fine mist “humid air”, or is it air that is carrying droplets of water? Because for the Wikipedia analysis to apply, the water must be in a gaseous state. And what about clouds? The stuff that falls on our heads on a rainy day certainly isn’t a gas. Well… it’s late, and I hope you guys will be able to sleep without knowing the answer to this compelling mystery… because I’m going to sleep! More on this later…
This is the 100th post on the Salient White Elephant! To celebrate the Salient Centenary, I thought I’d try to summarize the lessons learned and tricks developed up to this point.
Almost all of the AeroArchitecture (non-turbine) part of this machine is fixed. The only part of the flow manipulating structure that is not fixed is the curved blue panels. The rotational angle of these panels is regulated by the turbine controller, and is a function of wind direction. I have drawn these panels at what I believe to be approximately the correct rotational angles given wind direction, but of course this is only an intuitive guess. I think some detailed aerodynamic modeling will be required to determine the optimum angles for the panels. However, let me explain the reasoning behind my guess.
Let’s begin by pretending that the adjustable panels and the turbines are not present, and that the circular wall does not have segments cut out of it. In other words, the wind is flowing around a very tall round wall. In this case the high pressure region will be around the most upwind part of the wall. The flow accelerates to get around the wall, and I think the lowest pressure will be approximately at the 3 o’clock and 9 o’clock positions. The flow about the rest of the wall is oscillatory and unstable. Vortices are shed in alternating fashion, first from one side of the wall and then from the other. For example, the flow may detach at about 4:30, and a clockwise vortex (with a vertical axis) will spin off of the wall at that point. Then the flow will detach at about 7:30, and a counter-clockwise vortex will spin off the wall at that point. Then the pattern will repeat. So the first thing to notice is that it may be desirable to add adjustable panels at 2:30, 4:30, 7:30, and 10:30 in order to prevent the vortex shedding. I have not drawn these panels because I don’t know if they will actually be necessary. Remember that the wind is being de-energized by the 4 turbines, and maybe this will be sufficient for stabilizing the flow. In any case, let’s forget about the vortex shedding for a moment, and consider the system as depicted above (with the segments removed from the wall, the curved adjustable blue panels, and the turbines present just as depicted in the diagram).
As I said, the high pressure occurs at the 12 o’clock position. As the diagram shows, the wind is flowing more or less straight at the 12 o’clock turbine. In this case, both the adjustable curved panels and the curved parts of the wall that extend downstream from the turbine act to concentrate and accelerate the flow through the 12 o’clock turbine.
Now let’s consider the 3 and 9 o’clock turbines. I am reasoning that the wind is moving slow and is at a relatively high pressure in the regions just inside the wall near these two turbines. This is true because the wind has been forced to flow through a very large cross-section inside the wall, and also because the wind has been de-energized somewhat as it passed through the 12 o’clock turbine. On the other hand, the velocity is at a maximum and the pressure at a minimum in the areas that are outside of the wall and outside of the 3 o’clock and 9 o’clock turbines. This is true because the wind has had to speed up to get around the wall, and also because it has yet to pass through any turbines, and so it hasn’t had energy extracted from it yet. Based on all this reasoning, I have elected to orient the adjustable panels so as to further encourage the flow that already wants to happen anyway – that is, the flow from slow high pressure inside the dam to fast low pressure outside the dam. When you consider what normally happens when wind flows through a turbine, you see that it is not exactly as I have described:
So maybe my reasoning is flawed. Of course, the rules will be changed when flow is manipulated, and in this case we are not only manipulating the flow, but we are manipulating it a great deal with a very very large structure. So I guess the only way to figure out whether my reasoning is correct is to build a sophisticated aerodynamic model and see what it says. Anyway… if you take a look at the hypothetical path of flow I’ve drawn near the left side of the diagram, you can see that the flow will be exiting the trailing edge of the adjustable panel at high velocity, just like the way air exits the trailing edge of an airplane wing. So I am reasoning that this flow will draw the dead air out of the inside of the dam and through the turbine.
Now for the 6 o’clock turbine. This one seems pretty straightforward. On the one hand, you might expect that the adjustable panels and the fixed curved parts of the wall that extend upwind of the 6 o’clock turbine should together look exactly like the flow accelerating structure around the 12 o’clock turbine. The reason I haven’t drawn it that way is because assuming we are able to successfully prevent the flow from detaching from the wall and spinning into a vortex, then the flow will be wanting to curve in the downstream direction as it continues its journey away from the wind dam in the downwind direction. In this case, the panels need to be rotated toward this path somewhat so that the flow doesn’t collide with the most downwind end of the adjustable panel and spin off of that sharp edge in a vortex.
This post describes a means for greatly increasing the capacity factor of an AeroArchitecture Wind Turbine, such as the Circular Wind Dam. The concept is very simple. The fixed structure of the wind turbine is extremely large, yet very flimsy. For example, the walls of the Circular Wind Dam are made to withstand winds that blow no faster than 20 miles per hour! Because the walls are so flimsy, they are very inexpensive. What happens when the wind blows faster than 20 miles per hour? Vents in the wall are opened to let some of the wind pass through. The turbine hits rated power when the wind speed reaches 20 miles per hour. When the wind speed exceeds 20 miles per hour, power is regulated using exactly the same mechanism – vents are opened just enough to prevent the generators from overpowering.
If you have a million dollars to spend on the fixed structure of an AeroArchitecture Wind Turbine, you spread that money over the largest possible geographical area. Every square foot added to the size of the fixed structure provides an increase in the capacity factor of the machine.
One very interesting twist to this design technique is whether to spread the cost of the fixed part of the turbine over a single very large wind turbine, or over multiple flimsy structures that are located relatively far apart. I suppose the wind is even more likely to reach cut in speed if only the average wind speed at all the different wind turbine locations needs to be considered. Of course, this also spreads the transmission lines over a larger area. So you can see that some sophisticated computer modeling would be required to develop the optimum design.
Another intriquing question is whether to use the capital cost of the fixed structure to expand in the horizontal or vertical direction. It would seem that you could increase the size of the structure more if the increased dimension is in the horizontal direction (the structure is wider) rather than in the vertical direction (the structure is taller). But of course the extra wind harvested has more energy if you expand in the vertical direction. So again, the best design choice is not obvious, and a sophisticated means for developing this aspect of the design is desirable.
And finally, I wish I could see come kind of mathematical analysis that validates or refutes the fundamental principle behind this post – that the use of flow concentrators can increase capacity factor. My reasoning is only intuitive; it is not scientific or quantitative. It is known that a flow accelerating (upwind) shroud or a flow deccelerating (downwind) shroud increases the speed of wind that passes through the rotor disk. This must be so, of course, to maintain a constant volume rate of flow. Say for example that the cut-in speed for the rotor is 10 miles per hour (mph). If a shroud multiplies velocity such that a 5 mph wind becomes a 10 mph wind at the rotor, then isn’t it true that the cut-in speed has been reduced from 10 mph to 5 mph? And if so, doesn’t this mean that a flow manipulating shroud is capable of increasing capacity factor? This may seem like a minor point when you consider that there isn’t much power in a 5 mph wind or a 10 mph wind, but doesn’t the same mechanism apply for any windspeed? For example, doesn’t this reasoning mean that all of the time that the wind blows 10 mph means that the rotor spends the same amount of time harvesting a 15 mph wind? (I realize that the flow accelerating effect probably isn’t linear, but that’s just a detail.) In fact, it seems to me that one of the challenging aspects of designing this type of machine is how you would bleed energy away from the rotor at higher wind speeds, while still allowing the rotor to safely and reliably harvest an amount of energy equal to its rated output. (This would be necessary because, if it is our desire to increase capacity factor, then we wouldn’t want to lose the productive hours at higher wind speed simply because those winds don’t occur very often. Aiming for high capacity factor, we want the machine to be producing power in the highest number of different possible conditions. As a matter of fact, seems to me that it may also be possible to add a little bit to the capacity factor on the high end of the wind speed spectrum. This is true because the rotor is small and therefore, from a mechanical perspective, it is very robust and very strong, and also because wind velocity may be regulated somewhat by opening vents in the shroud.)
One final point. If the flow manipulating shroud is very large and very expensive, then it may be desireable to add a few rotors and generators so that some of the energy that would have been allowed to pass through the shroud during high winds could be harvested by the extra rotors or generators. This would do much for the capacity factor of the machine, but since the energy is there, and since you only need a small rotor and supporting equipment to harvest it, then it may be cost effective to do so. I’m glad I included this paragraph because, if nothing else, it illustrates how much of the complexity of the design of a very high capacity wind turbine has been shifted from areas like blade resonance and tower resonance, to a much more hazy problem of optimization, such as might be solved with linear programming.
This post shows how the piston of the old water pumping windmill can be moved up to ground level:
When the piston goes down, it pushes water down through the small tube into an extremely taught rubber bulb (sphere at bottom of small tube). This causes the bulb to expand, which expands the water in the region above the lowest check valve and below the upper check valve, thus raising the water column in the larger cylinder and ejecting some water into the storage tank. Now when the piston goes up, the combination of the suction created at the top of the small tube and the pressure provided by the taught rubber bulb at the bottom of the small tube are sufficient for lifting the column of water in the small tube. This returns the system to its beginning state, and the piston is ready to begin another cycle with its downstroke.
This post suggests increasing the energy density of the wind by injecting a fine mist into the oncoming streamtube. This might be impractical for traditional wind turbines, but it may be feasible for many of the turbines proposed on this blog. For example, the Sustainable Skyscraper and the Circular Wind Dam both manipulate flow with fixed structures. In this case, it is easy to position the misting devices in areas where the flow is already concentrated. I am hypothesizing that if the mist is injected at a sufficiently distant location upwind of the turbine, then by the time the fog reaches the turbine it will have absorbed kinetic energy from the surrounding wind to the extent that the velocity of the foggy wind is the same as the wind’s original velocity (that is, the velocity it had before the fog was injected).
This post proposes to increase the efficiency of a VAWT with yawing shields. In the case of the Darrieus, for example, the drag on the blade as it travels in the upwind direction is reduced by diverting the flow with a shield. The flow that is not diverted is slowed because the channel becomes wider in the middle. The flow on the opposite side of the rotor (left side in the diagram below) is not manipulated because it is already in the direction of blade velocity, and it therefore already reduces drag. As for the Savonius, I’m not sure if the idea will work, but I’ve drawn my best guess at how to augment the efficiency of the turbine using either one or two shields.
Many of the AeroArchitectural turbines presented on this blog are quite insensitive to lightning damage. Consider, for example, the Circular Wind Dam. The rotor and electrical components of this machine are housed inside the dam. They are almost 100% impervious to lightning strikes!