By James R. Stalker

The atmosphere made analogous to a spinning bucket

Visualize a bucket filled with water half the way to facilitate mixing, covered with a lid. Now, imagine this bucket spin out of control, jump, wobble, momentarily stop, and randomly resume any of these actions. Begin drawing a drop of water from a location within this bucket and measure its properties such as its speed. Put the water drop back and wait for ten minutes and draw another drop from the same exact location as before and measure its speed. Keep repeating this measurement effort for a while. Now, use these measurements to characterize the chaotic behavior of the entire fluid (water) in the bucket.

Wind measurement limitations explained

How successful will you be in reaching your objective? What is wrong with this approach? Can we use properties of water drops at a location to characterize the entire fluid? You don’t have to be an accomplished expert on fluid dynamics to clearly understand the limitations of such measurements.
 

Let’s apply this analogy to a real world situation. The water bucket here is the atmosphere and the bucket has hundreds of kilometers in diameter and ten’s of kilometers deep. The fluid here is the air, along with some water that changes into gas, liquid, and solid forms. Similar to measurements of water drops, wind energy developers, using conventional measurement platforms, are using wind measurements at a single location or two, to characterize the entire fluid behavior. It is the author’s hope that even non-technical folks (e.g., most financiers that bear the brunt of the risk resulting from such measurement limitations) can fathom this conceptual simplicity of very complex fluid behavior.

Limited wind measurements for secondary benefits (e.g., model validation) at best

Let’s dissect this analogy a bit further. Let’s assume a measurement platform that measures wind data with 100% accuracy at that location. One hundred per cent accuracy is a generous assumption indeed. Since the underlying causes (“wind forces”–read Dr. Stalker’s article for further details) are unknown, extrapolation methods to characterize wind behavior at other locations and heights (i.e., “within the spinning bucket”) can produce much larger errors as there are no a priori wind measurements at these other locations. On the other hand, if robust fluid models that are capable of capturing the effects of the above mentioned chaotic moves on the water drop’s speed and providing quantitative information with 90+% accuracy at hundreds of locations and heights, such models would deliver a much more accurate picture of the fluid behavior. This complex comparison of error statistics, between measurements and modeled wind speed, is illustrated in the following Table:

Note: Percentages shown in the above Table are accuracy rates of wind speed measurements and modeled wind speed values. 

So, financial risk takers, if you can get only one type of measurements and one type only, they ought to come from robust fluid models. If you can afford a second type of measurements, then ask for site wind measurements–as many as you can afford from the standpoints of budget and time. From the perspective of the complex fluid behavior of the atmosphere, the current approaches of single location wind measurements can only seem analogous to “groping” in the dark since such approaches represent only a secondary step at best and clearly can not represent a primary step in wind energy assessment efforts as concenptually shown by the spinning bucket analogy.

A critical take away from this article should be to “not put the cart in front of the horse.”

In the interest of full disclosure, and for ensuring technical soundness of the topic covered in this article, Dr. James Stalker would like to state that he owns Wind Forces, a Divsion of RESPR, that provides wind energy assessment services globally.  Dr. Stalker had over a decade of atmospheric computational fluid dynamics (CFD) research experience, without any commercial interest in wind energy, prior to founding RESPR/Wind Forces.

I just returned from a two-day workshop on wind resource assessment. The global wind energy community does not seem to be ready to face the difficult challenges posed by wind. In this context, the wind energy community should go beyond the current practices of just wanting to measure wind. Afterall, wind is the net effect and the wind forces are the underlying causes. Should one just ingore all the causes and expect to do well by focusing on the effect alone? Unfortunately, that is exactly what the wind industry has been doing, i.e., focusing on the effect, so far.

This discussion elaborates on what is beyond our current horizon (wind forces) and why these forces are paramount to the global wind energy community’s successful deployment of wind projects.

Wind resource assessment is just one of many pieces of the wind project development puzzle. However, it is understandably one of the most important aspects of wind project development.

As mentioned above, the global wind energy community has focused on measuring wind at potential wind sites. It is from this type of wind information, other energy and financial analyses are made to ascertain wind project viability.

Unless one is developing a single-turbine wind project with the measurement height matching the hub height, has enough time to measure (a year or longer), and has the appetite to assume significant risk involved in long-term energy projections, wind measurements made at a location for a year are limiting.

The above limitations do not result from any inaccuracies of the measurements themselves. Such measurements are demonstrated to be incredibly accurate these days and should be utilized as much as the project time line, budget, etc. allow.

So where do these limitations come from? These limitations result when one uses the location, height, and time-specific wind measurements to estimate wind information at other locations, heights, and for other time periods. These limitations occur due to the fact that there is no information about the (atmospheric) wind forces (e.g., pressure gradient force, gravitational force, etc.) that shape the wind behavior at the site being considered in the first place. In other words, unless one has a quantitative understanding of the underlying wind forces, the above limitations will not be properly identified and overcome.

If we consider a measured wind speed value of 5 m/s at a location, as an example, that wind speed was essentially forced by a specific combination of wind forces. This specific combination of forces that caused the wind speed to be 5 m/s, could have been one of numerous other possible combinations of wind forces. In other words, any other combination of forces could have caused the wind speed to be 5 m/s. However, using this specific measured wind speed (5 m/s) to estimate wind speed at other locations or heights or time periods, without the knowledge of the exact combination of the wind forces, will result in an inherent and unknown error. This is the reason why one needs to worry about the wind forces.

So, in summary, this specific measured wind speed does not contain a clue to point us to that exact combination of wind forces that produced that wind speed. One might argue that he/she has measured the wind speed they need and they really don’t need to know the combination of wind forces. One can live without knowing the exact combination of the wind forces that caused the 5 m/s wind speed, as long as the use of that wind information is just for that location, height, and time. The moment one moves away from the measurement location, however, these wind forces must be taken into account.

If this discussion is of interest to you, try to join the LinkedIn group (Wind Forces at Work!).

Successful wind energy assessment efforts require adequately established long-term (> 20 years) wind behavior at potential wind project sites. Measurements, from such traditional instruments as anemometers or from more advanced measurement platforms such as sodars or lidars will not help overcome the issue related to the time requirement for long-term wind energy analysis. Additionally, measurement platforms, both traditional and advanced, will only provide data at limited locations and heights. This spatial coverage limitation may prove to be a significant issue for wind farm sites within complex terrain locations.

When clearly known wind farm energy performance variability is of utmost importance for wind project developers and financiers alike, limited wind measurements in space and time are proven to be inadequate, expensive, and time consuming.

Fortunately, there is a cost-effective and timely alternative.  An example of such an alternative is based on an Atmospheric Simulation Technology, developed by Dr. James Stalker. This simulation technology offers location specific, height specific, and time specific wind information, not only at a much lower cost but also within a fraction of the time. For example, the 100-m height wind information can be obtained for only a few 10′s of thousands of dollars compared to much more expensive measurement efforts.  Also, yearlong assessments can be completed within weeks.

As important as the above two advantages are for simulation based wind energy assessments, the greatest benefit is in their ability to provide comprehensive spatial information within weeks. In other words, project developers do not have to rely on measurements made at a single location or use just one year measured data to determine long-term wind energy variability. In other words, wind project developers and financiers need to ask themselves this question: “Would we not measure wind for twenty years at every possible location within a prospective site if we had all the money and time in the world?” The answer would be a resounding ‘yes.’ And yet, these folks are currently relying on short-term (often one year) measurements at a handful of locations.

Well, simulation based methods offer wind information for multiple years within a small fraction of the time and for a fraction of the cost. Why not use these alternatives then? Several explanations are often offered in the wind industry why limited measurements are relied upon instead of using some of these established simulation techniques. For example, one may hear that the industry is used to seeing and being content with only yearlong measurements followed by statistical long-term projections. Another explanation might be that the simulation techniques they have employed thus far have proven to be inadequate, etc.

It will understandably take some time for the wind industry and the simulation based wind information providers to overcome the above limitations. But, in the meantime, it will make a good financial sense to combine one year measurements with multiple simulation years to quantify wind energy resource variability so one can invest in wind projects with more confidence.

Wind simulation techniques offer physically realistic explanations on why the wind speed or direction is of a specific magnitude at a specific location. Should a wind project developer be just happy to see a measured annual average wind speed value of 8 m/s, without knowing the causes behind such wind speed regime? Do wind measurements offer this type of critical information? No. This is because wind measurements just measure wind speed (the end state) and not the causes (the atmospheric forces) that lead to the observed wind. There is a simple reason why measurement efforts do not include measurements of forces because such efforts will be prohibitively expensive. Instead, measurements involve just the end states of wind speed, leaving a huge void in our understanding of the wind regime.

Fortunately, again, wind simulation techniques offer valuable insights into these forces responsible for observed wind. From a cost standpoint, a time standpoint, and from the standpoint of the critical causal information, wind simulation techniques are imperative in future wind energy resource assessment efforts.

However, wind simulation techniques must be carefully chosen to meet one’s wind project development objectives. Attention must be paid to the following factors to reduce the errors that can potentially be introduced in such techniques.

• Use highest spatial resolution possible and quality input data sets to drive these models or expect unknown and potentially large errors

• Use thoroughly tested atmospheric models or expect unknown and potentially large errors

• Use well established computer platforms or expect unknown and potentially large errors

• Employ people with the expertise to run these models or expect unknown and potentially large errors

James Stalker is an atmospheric scientist with over nineteen (19) years of meso-mircorscale modeling experience and expertise as a researcher and as an entrepreneur.