Over the past several years, the Sheerwind Invelox ducted turbine has managed to rack up a couple of million in grants and other investments according to reports and has built two small prototypes. According to its press releases it has a New Zealand distributor and a couple of pilot projects targeted for Dubai and Minnesota.
It makes the extraordinary claim of generating six times the energy of a comparable wind turbine, but in fact is likely generating about eighteen times less for perhaps ten times the material. Briefly, they propose putting a big funnel in the sky and channeling accelerated wind down to a small wind turbine at ground level.
Sheerwind has also racked up a lot of predictable hype from tech and penny stock sites as well as support from those opposed to actually effective wind generation technologies. Recently however the Sheerwind founder and Invelox inventor, Dr. Daryoush Allaei, co-authored a paper with Dr. Yiannis Andreopoulos, Professor of Energy Research of the Department of Mechanical Engineering of the City College of New York. It’s primarily a computation fluid dynamics (CFD) study but it includes statements in a credible journal about its output from field studies. The combination means that it is worth an assessment of their extraordinary claims to see if they stand up.
Real data comparison
First up, let’s look at their field comparison to see what it says. Apparently they are using a tiny wind turbine from Sunforce rated at 600 W for testing both inside the Invelox funnel and free-standing on the mast. Since they are capturing much more swept area with their device than the swept area of the wind turbine, it’s at best a disturbingly naive comparison; more on this later.
Importantly, their tested Invelox device per the papers Table 3 is 18 meters (59 feet) at ‘hub’ height while the comparison wind turbine mast is only 10 meters (33 feet), and the wind is always stronger further off the ground. How much of a difference does that really make? Well, if the wind is an average of 4 meters per second (roughly 9 mph) at 10 meters — which is what the Sheerwind paper indicates –, it will be around 5 meters per second (11 mph) at 18 meters. As the energy in the wind goes up by the cube of the velocity, that means the wind energy available at 18 meters is going to be about two times the wind energy available at 10 meters.
The combination of much larger swept area much higher off of the ground means that they are pretending to make an apples-to-apples comparison while actually making a grapes-to-watermelon comparison. They are both fruit, but that’s about it.
The intake to this funnel in cross-section to the wind is 40 feet by 20 feet (about 12 meters by 6 meters). That’s a swept area of 800 square feet (73 square meters). An actual comparison of value would be to a wind turbine with an equivalent swept area.
But that’s not what Sheerwind did.
Instead, they make a comparison between a device with a swept area of 800 square feet and one with about 28 square feet. That’s a factor of 28 difference between the areas presented to the wind. The wind turbine they compared to fits in the tiniest part of that massive structure, the constricted nozzle on the lower right.
If a conventional wind turbine were scaled up to 800 square feet, it would, all else being equal, generate 28 times more electricity from a given wind resource. Even pretending that putting the small wind turbine at close to half the height is reasonable, Sheerwind is only managing to exceed its capacity by a factor of six if their numbers are to be believed.
But of course there is still more. They have been claiming 600% for years, but what does the paper say?
The results show INVELOX generated 80–560% more electrical energy than the traditional WTGs. P-Day 8 means partial data was collected on the eighth day. The total average energy production improvement of INVELOX over 8 days is about 314%.
So they don’t actually get to 600% at all. And some days are generating less than twice as much energy with 28 times the swept area. On average, they are running only three times the generation according to their own numbers, or nine times less energy generation than a real comparison to a wind turbine with an equivalent swept area would have elicited. And remember that fully two-thirds of that is accounted for by the greater height of their device, the two-times factor shown above for differential wind velocity. An apples-to-apples comparison would likely see eighteen times less generation, requiring an input funnel perhaps eight times larger than a football field to produce the same electricity.
Eighteen times less generation. Not six times more.
Let’s look at a scale comparison of how deceptive Sheerwind is being:
Note the vast difference in scale and height of the comparable wind generation devices in the top and bottom picture? Well, that’s just one of the lessons to learn from this.
The second is that it wouldn’t matter if the Sheerwind device generated less electricity if it were a lot cheaper to build. But it won’t be and can’t. The conventional turbine has a hollow metal tower which will use a lot less material than the funnel of the Sheerwind device. And it’s freestanding on its foundation, unlike the Sheerwind which requires a heavy supporting framework. The Sheerwind device will take probably ten times the material and a lot longer to install, as current pre-fab utility-scale wind turbines can be up in under three days.
The third is less obvious. Utility scale wind turbines survive hurricane-force winds simply by pitching their blades so that they don’t catch the wind. The Sheerwind is more of a sheet-metal tent which cannot be feathered. It will have to be engineered to be much more robust than Sheerwind acknowledges to survive high winds. Any production version will be even more expensive than the simple materials comparison that the prototype version affords.
So the real world data paints a disturbing picture of vastly inappropriate comparisons and hyperbolic inflation of results, assuming that the results are even reported accurately. What about the quality of the computational fluid dynamics? Personally, I only have a rule of thumb for CFD results: they never match the real world. That’s obviously not enough, so I asked an expert for his opinion on the quality of the modelling showing in the Sheerwind paper. He wasn’t particularly complimentary.
He indicated that the basics of tool and model use were okay, but had a bunch of concerns related to the application of the tools and model.
- Higher mesh points on the models would have been preferred.
- Constant input velocity of wind as opposed to more realistic turbulent flow means the results are inaccurate.
- Steady state flow condition without a turbine inside which will lead to “uber-rosy results for any duct or venturi tube without blockage”.
- All the CFD was done without a turbine inside which leads to substantially overstated results.
All of these points are possible to do with the tools used by Drs. Allaei and Andreopoulos. Point one requires more computational horsepower than they probably had available. An additional CFD model is required for point 2, but the tools support it.
For those interested in the details of the analysis, it’s available here. Drs. Allaei and Andreopoulos vacillated on commenting on it, originally saying that they would provide a response, then deciding not to and cutting off all communication. The analysis is not published as a critique or rapid response in a peer-reviewed journal, so it is reasonable for at least Dr. Andreopoulos to not respond in detail.
Sadly, this is another case of a paper with obvious and deep flaws being published in a good peer-reviewed journal. Energy is a respected journal, multiply indexed and with an impact factor of 4.686. So how did such a poor paper make it past the fairly high standards of the journal and its peer-reviewers? Well, it’s likely that part of the reason is that this paper didn’t show up in the main journal, but in a special issue, Energy & Environment: Bringing together Economics and Engineering. Special issues have a bulk of studies coming in at roughly the same time, requiring overlapping peer reviewers and as a result quality can tend to suffer. In this case the co-author is from a relatively highly ranked school and has a number of publications in CFD as well (although none pertaining to wind generation), which tends to increase the odds of publication. And so grossly incorrect papers such as this one get published and then are used downstream to provide credibility where none exists. That’s certainly what Sheerwind is doing.
Of course, the poor performance of the Sheerwind device is no surprise. The first attempt to accelerate airflow by putting a wind turbine inside of a funnel was done 90 years ago according to Robert W. Righter’s book Wind Energy in America: A History. Yet people continue to invest in this without doing their due diligence. One of those investors has commented at length on prior analyses of the Sheerwind Invelox including my blog post on the CFD report, finally sparking Mike Bergey to comment in return (reproduced with permission).
Concentrators and ducted fans have been proposed and promoted by the dozens over the last 35 years that I have been in the industry. In your due diligence you must have missed Next-Gen Wind, Vortec, TurboDynamX, Enflo, Enco, Ring Turbine, Smart Wind, Wind Cube (Wind Sphere), WindTammer, Sky Wolf, Elena, Catching Wind Power, and OptiWind, to name a few. The fatal flaw in all these unsuccessful attempts to build a better wind turbine is the promoters failure to account for the wind’s ability and preference to go around a blockage like the entrance to a funnel. The operating environment of a wind turbine is nothing like the constraining ducting of a hose or a wind tunnel and that dooms the concept to poorer performance. And the dishonest use of the rotor area instead of the total intercepted area to inflate the calculated efficiency doesn’t change the physics.
For those not familiar with Mr. Bergey, he’s twice past-President of AWEA, served on the AWEA board for 26 years, has manufactured the top-selling small wind turbine for roughly the past 30 years and chaired governmental committees around wind energy. He’s been multiply awarded for his efforts and his direct, ongoing and committed involvement in all aspects of wind generation are second-to-none.
Compare Mr. Bergey’s experience to Dr. Allaei. According to his research publication history and consultancy, he’s an expert on vibration. Like many other ‘innovators’ in wind energy, he has no prior history of any involvement in wind energy. This is so common, that it’s a separate red flag question in my material which aims to inoculate investors against bad wind energy bets. Sheerwind managed to rack up eight red flags based on this sieve, including this statement about conventional wind turbines which is extraordinary to see in a peer-reviewed paper:
They are also expensive, unwieldy, inefficient, and hazardous to people and wildlife.
The inability of devices like Sheerwind to do anything except suck money out of investors’ wallets is well-documented and an ongoing proof of PT Barnum’s axiom.
Sheerwind makes radically inappropriate comparisons between their long-disproven approach to wind generation and actually useful wind generation. The numbers show that they are likely about eighteen times worse at generating electricity from moving air than a truly equivalent wind turbine would be, and will require an order of magnitude more material to achieve that.
The claims that they make aren’t supported by their own data, and their data is distorted beyond credible defence. Their device will produce much less electricity at much greater cost than conventional wind generators.
Potential investors: stay away. Current investors: don’t expect to see your money again.
(Author’s note: an earlier version of this article had miscalculated the energy difference available in the wind by squaring instead of cubing velocity. That is corrected throughout now.)