2014 Aurecon recruitment banner

Thinking

Wind farm developments: Getting energy prediction right

Collgar Wind farm, Western Australia

Blair Walter, Renewable Energy Leader for Aurecon spoke with Australia’s Energy Generation magazine about wind farm developments and getting energy prediction right.

The wind is a natural system and therefore non-linear in its short term behaviour, which makes it difficult to model and predict.


Good commercial decisions require robust predictions

Making good commercial decisions at each stage of wind farm development requires robust predictions of the site’s future energy yield. But at early stages when project feasibility is unconfirmed, developers are often reluctant to spend large sums of money assessing the wind resource. Even at financial close on some projects there still does not appear to be sufficient understanding of the site wind resource to provide a reliable energy yield prediction.

Aurecon has been involved in a number of reassessment exercises in Australasia and Europe to determine actual long-term output of wind farms once they have been operating for a few years. We have observed a trend towards over-prediction in preconstruction estimates of operating wind farms, possibly caused by a tendency to invest in more over-predicted projects as they appear to offer better economic performance. Developers may drop under-predicted projects as investment options as they may not appear to be viable investments.

Over-prediction worsens when accompanied by underestimation of the uncertainties contributing to exceedance values relied on by banks and equity investors. We have observed a number of projects where the P50 from reassessment using operating data is below the preconstruction P90 energy output level. Understandably, investors continue to be concerned about wind risk and the accuracy of output predictions for new projects.

Three main challenges in wind resource assessment


The wind is a natural system and therefore non-linear in its short term behaviour, which makes it difficult to model and predict. Three of the main challenges in wind resource assessment are:

  • Calculating a long-term wind climate for the site from a short period of measured data
  • Modelling wind flow across the site in order to predict wind speed at each turbine location
  • Extrapolating data from short masts up to turbine hub-height

A good wind monitoring campaign should focus on reducing the uncertainties associated with these activities to target levels, through collecting multiple annual cycles of data, monitoring at enough locations across the site to reduce wind flow modelling errors, and monitoring at or near turbine hub-height.

In the early stages of development, the full extent of the project may be unknown. Furthermore, assessment of engineering and environmental effects and stakeholder consultation may, over time, change the project extent. This can make it difficult to design a wind monitoring campaign that results in sufficient wind data at the time of the final investment decision.

Our recent project proving ground


Aurecon has overcome these challenges using a combination of traditional and advanced technologies. For a mega wind farm (>500MW) currently under development in Australasia, Aurecon identified the site using in-house mesoscale meteorological modelling of wind speeds combined with GIS analysis of terrain. We then used this information to develop a preliminary assessment of the potential wind farm including layout of turbines, wind resource assessment and energy yield prediction – all before a single mast or land rights agreement was established. Our client was then able to understand the development potential of the whole area and begin discussions with the landowners whose properties the project covered.

Our client erected two 80 metre masts in 2009 (with Aurecon’s assistance) in central locations selected to confirm the predicted wind resource from mesoscale modelling. After initial confirmation of the wind resource, another four 80 metre masts were erected at locations selected to form a backbone of wind measurements from the site. Over the next two years, using remote sensing (LiDAR) we measured wind at several additional locations carefully selected to provide representation of all turbine positions in the layout. Due to having a backbone of quality 80 metre masts for correlation, only short monitoring periods of a few months were required at each LiDAR location.

Three years on, and with engineering assessment completed and a permitted project, we have confirmation of the wind speeds at the levels predicted in the original mesoscale modelling investigation and the wind farm layout is similar to the preliminary assessment. Due to the approach our client took on this project, with uncertainty reduction at the core of the development process, abortive development-spend was eliminated.

Using these advanced tools reduces uncertainty from the outset of the project development process, and thereby reduces the likelihood of over-prediction in the final investment analysis.

Aurecon holds a leadership position in wind energy providing consultancy services to project developers and financiers. Our wealth of global expertise and our extensive range of technical advisory services will give you a competitive advantage and help you realise maximum project value.

To top