All measurements have a degree of uncertainty regardless of the precision and accuracy methodologies and techniques. This is caused by three factors, the limitation of the measuring instrument (systematic error), the limitation of the modelling software (modelling error) and the skill of the experimenter making the measurements (random error). Random error and unexpected events can cause a single data point to be somewhat non-representative and thus, the overall forecasting reliability is violated. Unless the resource analyst can offer a degree of confidence that the wind resource falls within a specified range, it is not possible to construct a sound financial model for a wind energy investment since the AEP estimation is incomplete without an accompanying estimate of its cumulative uncertainty. Consequently, multiple measurements are usually taken, and the “best value” is represented using a variety of approaches-scenarios each with its own peculiar idiosyncrasy. However, there are other factors we must consider when deciding what is the best way to mitigate uncertainties related with the wind resource assessment such us how wind resource assessment uncertainty is quantified, how the meteorological campaign strategy impacts uncertainty itself, or how an assumption influence the overall prediction preciseness such us the windiness of the measurement period relative to the long-term mean wind speed at the site, and the accuracy of the anemometer.
The uncertainty present in all wind resource estimates is primarily related to the following factors:
- wind speed measurements
- the historical climate adjustment
- potential future climate deviations, wind shear, and
- the spatial wind resource distribution
Through experience with operational data and historical or reference data of a large portfolio of projects with various turbine technologies and measurement techniques, world leaders in wind energy engineering investments could refine and validate loss assessment for components and multi-source phenomena such as wake losses, icing losses, turbine curtailment due to sector management, extreme weather conditions, instruments quality violation, turbulence events, sound curtailment, site topography characteristics, availability, electrical losses and environmental factors.
However, to reduce the cost of uncertainty in wind energy prediction is a complicated task with many amphidromic synergies and interactions. Also, because of the financial crisis and economic slowdown, the requirement for an accurate and precise energy yield forecasting is more than a necessity to establish the economic viability and financial feasibility of wind energy investments. Besides that, due to the lack of meteorological observations in the atmospheric boundary layer, measurements and Numerical Weather Prediction (NWP) models should be jointly used to link the latent data gap that needs to be efficiently filled to estimate the wind resource potential.
Numerical modelling of wind conditions is responsible for a significant share of the uncertainty on wind plant energy yield estimation. For the last 20 years, numerical modelling of surface wind for wind energy installations has been dominated by linear flow models such us MsMicro and the Wind Atlas Analysis and Application Program (WAsP and WAsP Engineering). In the last 5 to 10 years, the so called Computational Fluid Dynamics (CFD) models have been gaining relevance within the wind energy industry. Firstly, because CFD can model the full wind flow field, unlike WAsP, and secondly, because they can capture, accurately or not, the flow distortion due to the variations in topography and terrain obstacles. Nowadays, remote sensing systems including sonic detection and ranging (sodar) and laser imaging detection and ranging (lidar) are gaining currency as a way of measuring wind at the higher heights demanded by today’s utility-scale wind turbines. Used either as a complement to met tower data or as a standalone tool, remote sensing systems are finding broader application in wind resource assessment campaigns and are also being used throughout the entire lifecycle of a wind energy project.
The main focus of the wind energy industry in order to manage and mitigate the uncertainty phenomena can be summarized as follows:
- Decrease measurement uncertainties with advanced measurement techniques based on LIDAR and sonic anemometry (with respect to conventional anemometry based on cup anemometers and wind vanes).
- Improve numerical models for wind resource assessment on complex terrain and offshore based on CFD (with respect to linearized models, only suitable in flat homogeneous terrain).
- Improve regional available wind potential assessment (wind mapping and digitization) with the aid of meteorological models of different spatial/temporal scales.
- Simulate stochastic winds and improve turbulence models for atmospheric flows and wind turbine wakes under various atmospheric and topographic-geospatial conditions and data.
- Carry out a comprehensive characterization of extreme winds and wind speed variations at different atmospheric scales.
- Determine the potential of using wind tunnel modelling as a complementary tool to field testing, for site calibration and validation of numerical models.
- Design, development and implementation of field and wind tunnel experimental databases for validation of numerical models.
- Characterization of the particular wind conditions of complex terrain and offshore environments.
- Benchmarking of all the different wind resource assessment techniques. Determine the range of application of each one through a cost versus uncertainty analysis.
While wind is never going to run out one can said that the future of wind resource assessment is data availability, data quality and their representativeness. Despite the normal variability in the wind climate, the risk of long-term climate change and the development of advanced simulation models and techniques there is a critical question to answer. Is there only one fundamental resource or many resources to be considered for an accurate and precise energy yield campaign? If there are many then one cannot convert (part of) one to another as that would violate its identity as a resource. If this is true, are there any subtle synergies that we cannot observe?
