Detailed analysis helps plan profitable wind farms

Credit: Siemens

Will a wind farm deliver high enough yields for profitable operation? In the past, over-optimistic predictions have often resulted in lower yields than expected. New requirements for wind reports have improved the prospects of success for both operators and investors, write Thomas Arnold and Thomas Zirngibl

The profitability of a wind farm essentially depends on the question of which wind conditions prevail at the chosen site.

Wind measurements and detailed analyses of the local wind and weather data form the basis for clarifying whether the wind speeds at the site are frequent and high enough for a profitable wind farm.

However, in spite of existing methodological requirements, in many cases the energy output of wind farms falls short of that promised in yield predictions, with profit margins being lower than anticipated. This, in turn, has caused decreasing confidence and increasing doubts over the informative value and reliability of the documents among banks and investors.

Figure 1A. An example of a wind profile. Wind speed (x axis) is put in relation to the height above ground level (y axis). Generally, wind speed increases with altitude. However, the degree of this increase varies from site to site and depends on a host of factors, including topography (e.g., forest areas as shown in Figure 1B)

The root cause of the problem has been the sometimes extensive room for discretion and interpretation in the preparation of wind reports. To remedy this area of concern, the FGW, the German Public Association of the Renewable Energy Sector, has revised its Technical Guideline Part 6 (TR6), introducing stricter requirements.

The reason for the move was that realistic forecasts require a minimum of suitable information at and around the selected site. Various aspects and methodologies play a role in obtaining this information, concerning both the quality of the input data used in the calculations and the type and method of project- and site-specific wind measurements and their evaluation.

It is important to use reference data from other wind turbines (known as ‘validation turbines’) and/or wind measurements conducted no more than ten kilometres from the future windfarm site. The requirements of the revised TR6 are even stricter for sites involving terrain with complex topography. If the terrain between the points of measurement for the database and the site of the planned turbine shows gradients of over 10 per cent or differences in altitude of more than 50 metres, only data which have been collected no more than two km from the project site may be considered in the yield prediction and wind report.

Figure 2. The wind profiles of a site at different seasons. The figure clearly shows several effects, including that wind blows more strongly in winter. The increase in wind speed at higher layers of the atmosphere is likewise more pronounced in winter.

The revised guideline also includes more specific requirements for the ratio between the actual hub heights used in measurement and the hub heights of the planned turbines. Given this, the new requirements of the TR 6 essentially concern the informative value, reliability and representative nature of wind measurement and validation turbines and also, in particular, the use of wind measurements collected with laser-aided LIDAR (light detection and ranging) systems.

After all, key questions that have frequently played a minor role in the past are: Is the measurement period representative? Are data and results statistically validated and enable conclusions to be drawn for a service life of 20 years? In other words: do the measurements map the real-life conditions?

Figure 3. An example of the frequency distribution of wind directions for an entire year (left) and a three-month measurement (right), given in per cent per 30° sector. Where large variations in distribution occur as here, corrective actions are permitted according to FGW TR6 provided the data are filtered primarily from the sectors of the secondary directions of wind while the data from the main directions of wind remain unchanged. Following the corrections, as a minimum requirement the data set must correspond to an effective measurement period of at least three months to ensure the results are in conformity with the FGW requirements and thus valid.

New, cost-effective methods that provide a realistic description of variable parameters even in short measurement periods are in demand. Wind speed over the course of the day, over the course of the year, by height above ground level and by wind direction are primary among these parameters; they vary strongly from site to site and, as unknown variables in the equation, have the greatest influence on the actual yields of a wind farm. The updated TR6 places greater emphasis on this aspect, with the situation being reflected in the stricter requirements.

TÜV SÜD’s wind experts have developed an improved measurement method that satisfies the requirements of the revised TR6 and enables informative and reliable yield predictions and wind reports to be delivered in a relatively short period and in compliance with the Technical Guidelines.

If the measured data or yield data from the site or the vicinity of the site do not permit an expert opinion to be drawn up in conformity with the guideline when taken alone, perhaps because they violate the above distance criterion of 10 or 2 km or the height criterion of two-thirds of the planned hub height, they will be complemented with additional data.

In this case, a LIDAR measurement as a stand-alone system at the site may be suitable for gaining additional data. Parallel to LIDAR measurement, the method uses regional data from the German Weather Service (Deutscher Wetterdienst, DWD) to remove the variables from the equation. This firstly ensures that the period mapped by the data is typical for the wind speed over the course of the year.

Wind speeds are greatly influenced by cross-regional atmospheric layering, which results from differences in temperature between near-ground and higher layers of air. The occurrence and frequency of this atmospheric layering are subject to seasonal fluctuations. Secondly, this approach permits more specific consideration to be taken of the conditions of the selected site, as the hills, forests, valleys and minor mountain ranges of its topography may have a massive influence on yields depending on the wind direction.

Figure 4. The frequency distributions of the various atmospheric layering systems for measurement periods of one entire year (green) and three months (blue) based on the data in the table. The deviation from the annual average is given in the “Difference” column. As stable and neutral layering systems result in similar wind profiles, the values in the “Difference” column in classes 1 to 4 and 5 to 6 can be added (last column, “Evaluation”). The maximum limit for this total is set at 5 per cent, ensuring that unstable layers are adequately considered.

The task at hand, therefore, is to gain a set of data within a relatively short period, which enables any existing data from the vicinity to be reviewed and validated, trends to be confirmed and incorrect assumptions to be corrected. The new method developed by TÜV SÜD does just this, and is in many cases a cost-effective and efficient alternative to long-term 12-month wind measurement with a wind mast.

Generally, three to six months are sufficient for LIDAR measurements to be conducted and a dataset to be built up in line with FGW requirements. This data set then forms the basis of yield predictions and reliable wind reports for owners, operators and investors.

Thomas Arnold is Team Leader, Measurements and Technical Testing for Wind Energy Turbines, and Thomas Zirngibl is Team Leader, Site Assessment & Technical Due Diligence, at TÜV SÜD Industrie Service GmbH