Why it pays to invest in quality wind resource assessments

Developing wind energy projects require investment. To gain investment, investors want an idea of how ‘good’ the investment is – a wind resource and energy yield assessment. To achieve this assessment, measurements and analysis are required.

Go to the profile of David Pullinger
Aug 11, 2017
Upvote 0 Comment

Author(s): David Pullinger 


The decision on the wind resource assessment approach is a compromise between stakeholders – on one side the investors looking to minimise cost, and on the other side scientists who aim to provide the most accurate quantification of the site's potential. The role of the wind analyst is that of an engineer – finding the compromise between cost and benefit on a site by site basis. This work covers onshore wind projects in the >1 MW size in an international setting. The results demonstrate that current industry best practice generates high levels of return on investment for an onshore wind energy project and provides a case study on the benefits of investing in quality wind resource assessments.


High levels of financial investment are required in order to develop a wind energy project. One stage in the process of assessing the risk associated with such investment is a wind resource assessment. The assessment requires a significant investment of both cost and time; this report presents the context related to wind energy costs, the step-by-step processes and key factors involved within a wind resource assessment. A case study is then presented to give evidence demonstrating the importance of increased levels of financial investment within the wind resource assessment and the results of a cost–benefit analysis for a specific site.

Context: investment

Onshore wind energy projects require significant investment, typically in the order of £1 M per installed MW [1]. In addition, wind energy projects take several years of planning before construction and ultimately operation – the period during which the project begins to yield financial returns. A typical project lifecycle and key project milestones [2] are summarised in Fig 1.

Fig 1: Key project phases for a typical wind energy development

Throughout the project development process it is important to understand the potential wind resource of the site, as this allows for an evaluation of the future energy production and a prediction of the profitability of the project. At very early stages (site finding and feasibility on Fig 1), the wind resource is often estimated using crude wind resource maps, however in order to more accurately value the project a more confident prediction is required. This can be achieved through a wind resource assessment – which is one of the early significant investment decisions – requiring anywhere upwards of £30,000 of capital, but often in the order of £100,000 depending on the size of the project. The investment includes the cost of wind measurement equipment, installation, maintenance and also the cost of a bankable wind resource assessment report. The wind resource assessment is almost always required by potential investors in a project. This investment is required several years before a wind energy project becomes operational and therefore the return on this investment is slow to be realised.

Context: wind resource assessment

The wind resource assessment (sometimes called an energy yield assessment) is one stage in the development of a wind energy project. The output of the wind resource assessment is a probabilistic estimation of the energy that can be produced by wind turbines installed at the project site. An outline of the process is given in Fig 2.

Fig 2: Outline of a standard wind resource assessment

Each stage follows a simple structure of inputs, analysis and output – with the quality of the inputs and also the process or analysis directly impacting on the certainty of the output.


To confidently understand the wind speeds at a site, best practice within the industry is to take measurements. These measurements traditionally consist of anemometers mounted on meteorological masts, however in recent years LiDAR (light detection and ranging) and SoDAR (sonic detection and ranging) devices have also been used. Current international standards [3,4] relate to measurements from cup anemometers, however the use of remote sensing will be included in future revisions.

Best practice guidelines [3–5] recommend that site measurements are taken at multiple heights with calibrated anemometers on a mast at least two-thirds of the hub height of the proposed wind turbine (but ideally covering the whole turbine rotor). Due to seasonal variations in the wind distribution, a minimum of 12 months of site measurements is typically required in order to characterise the site-specific wind resource. In addition to wind speed, wind direction and other climatic variables should also be measured. The uncertainty associated with the site measurements is related to the quality of the recording devices (class of anemometer or accuracy of LiDAR or SoDAR), the data availability and the period of measurements recorded. The result of the site measurements is a time series of wind and other meteorological variables which are cleaned to remove any erroneous values from the sensors (due to sensor failures, sensor icing or degradation).

Long-term wind resource assessment

Site measurements are typically only available for a period of one to two years. Whilst this is a good basis for a wind resource assessment, it is still a limited period of time compared to the likely 20-year operational lifetime of a wind energy project. Wind speeds during individual years may be significantly above or below the long-term mean, so a long-term adjustment of the site-measured data is therefore desirable.

Typically long-term adjustments are made to the measured data through a comparison of the site measurements to reference data (from long-term sources such as surface weather observation networks or global reanalysis datasets). Measure-correlate-predict methods are often used to perform the adjustment [6,7]. The uncertainty associated with the long-term adjustment varies depending on the data available (the period and quality/consistency) and also the techniques applied. The result of the long-term wind resource assessment is a long-term wind distribution representative of the measurement height and location.

Wind flow modelling (analysis)

To estimate the wind resource and energy yield for a wind energy project, the wind resource available at each of the wind turbine locations needs to be determined. This is generally done by extrapolating the long-term wind distribution from the measurement locations and height levels to the wind turbine locations at hub height. The extrapolation typically has a horizontal component and a vertical component, and numerical wind flow models are most commonly used for this step of the analysis. In some cases, measured wind shear is used in order to extrapolate the wind distribution to hub height, and therefore only the horizontal component is modelled.

The flow model considers the variability of the terrain across and surrounding the site, and models its impact on the wind flow accordingly. Terrain effects considered generally include elevation and surface roughness (a parameter describing the amount of resistance to wind flow closest to the ground, which may include obstacles such as buildings or dense areas of forestry). Flow models have a performance envelope (limits of capability due to assumptions and simplifications) and may not work well in all types of terrain. It is preferred to validate the flow modelling results against measurements at multiple measurement locations, however due to limitations of measurement campaigns this is not always possible. The uncertainty associated with the wind flow modelling is related to the quality and suitability of the inputs into the model (number and quality of wind inputs, accuracy of terrain inputs and extrapolation distance – both horizontal and vertical), and also the flow modelling approach chosen. The results of the wind flow modelling step are hub height wind distributions at each of the wind turbine locations.

Energy yield and loss assessment

To calculate the energy production of a wind energy project, the hub height wind distributions are combined with a power curve for the wind turbine model under consideration. A wind turbine power curve is a transfer function relating the input wind speed to the electrical power produced by the turbine. This is calculated on a per turbine basis using the results of the previous step of the process. The gross annual energy yield after terrain and wake effects is subject to system inefficiencies and losses. These losses and inefficiencies are generally project specific and fall into seven main categories (availability, wake effects, turbine performance, electrical, environmental, curtailment and other) [8]. The uncertainty of the predicted energy yield is dependent on the quality and accuracy of the inputs within the assessment.

The result of the energy yield and loss assessment is a long-term average ‘central estimate’ of the electrical energy that would be generated annually and could be sold via the electrical network. This is typically referred to as the P50 energy yield.

Uncertainty assessment

To understand the potential of a wind energy project, it is important to quantify not only the P50, but also the confidence of the prediction of this figure. The uncertainty, and therefore risk, associated with the energy prediction directly impacts the attractiveness of the project for investors.

The results arising from the previous steps of the wind resource assessment are derived from measurements, statistical data analysis and modelling processes. Some of the steps involved are subject to random variations and systematic errors; a degree of uncertainty is therefore inherent in the prediction results. A wind resource and energy yield analysis assesses and combines the individual uncertainties using statistical approaches in order to produce the resulting probability distribution of production of the wind energy project. A typical wind energy distribution is shown in Fig 3.

Fig 3: The results of a typical wind resource assessment displayed as a probability distribution

The figure shows the P50, and also the P90 (the figure of which there is 90% confidence of exceeding) which is a commonly used figure within project financing. The P90 figures are typically stated with a return period (in years) which reflects increasing certainty over longer averaging periods (reflecting the uncertainty associated with inter-annual variability of wind speeds).

The P90 figure is often used by investors in projects to assess the attractiveness of any investment, this figure is a direct input to the financial model. By reducing the uncertainty of a project, the P90 is increased and therefore the project becomes more attractive for investment.


Wind energy projects require significant investment of capital. One significant early stage contributor to the total investment is the wind resource assessment of the site. This report along with the following case study demonstrate that the impact of the level of investment in the wind resource assessment, both inputs and analysis approach, directly affects the confidence in the future energy yield of a wind energy project. Some approaches to lowering wind resource assessment uncertainty have been identified for a specific site within this work, and the benefits of additional investment in the measurement campaign are clearly assessed.

The increased confidence is represented as an increase in the project's P90 value. As this figure is used within financing of projects, any increase yields a higher predicted rate of return and a high return on the initial investment. By increasing the predicted rate of return the asset value is increased making the project more viable and hence more attractive for equity investors. Each wind energy project is unique, however the benefit of improved wind resource assessment approaches should be assessed at an early stage to ensure that optimal investments can be made.


  1. EWEA: ‘The Economics of Wind Energy – a report by the European Wind Energy Association’ (EWEA, 2009), pp. 30–32.
  2. Deloitte: ‘Establishing the investment case – wind power’ (Deloitte, 2014), p. 5.
  3. IEC 61400-12-1: ‘Wind turbines – Part 12-1: Power performance measurements of electricity producing wind turbines’ , 2006.
  4. IEA 11: ‘Recommended Practices for wind turbine testing and evaluation, wind speed measurement and use of cup anemometry’, 2003.
  5. MEASNET: ‘Evaluation of site-specific wind conditions’, 2009.
  6. ‘Measure-Correlate-Predict Methods: Case Studies and Software Implementation’, available at: http://www.emd.dk/files/windpro/Thoegersen_MCP_EWEC_2007.pdf, accessed 10 July 2015.
  7. Carta A. Velázquez S. Cabrera P.: ‘A review of measure-correlate-predict (MCP) methods used to estimate long-term wind characteristics at a target site’, Renew. Sustain. Energy Rev., 2013, 27, pp. 362–400 (doi: 10.1016/j.rser.2013.07.004).
  8. DNV KEMA Energy & Sustainability: ‘Framework for the categorisation of losses and uncertainty for wind energy assessments’ (DNV KEMA, 2013), pp. 1–8.
  9. Risø National Laboratory: ‘WAsP Manual: Wind Analysis and Application Program (WAsP). Vol. 2: Users Guide’, 1993.
  10. Bowen A. Mortensen G.: ‘Exploring the limits of WAsP the wind atlas analysis and application program’. Proc. EWEC-1996, Gothenburg, Sweden, 1996.
  11. Bowen A. Mortensen G.: ‘WAsP prediction errors due to site orography’ (Risø National Laboratory, Risø-R-995(EN)), p. 65.
  12. Pinto P. Guedes R. Ferreira M. et al.: ‘Wind prediction deviations in complex terrain’. Proc. European Wind Energy Conf., Dublin, Ireland, October 2004, pp. 317–320.
  13. Pedersen T. F. Dahlberg J. A. Busche P.: ‘ACCUWIND – classification of five cup anemometers according to IEC 61400-12-1’ (Risø National Laboratory).

Case study

To demonstrate the benefit of investing in additional measurements, thereby reducing uncertainty, a real-world case study of a wind farm project is now presented.

Baseline scenario

When considering the impact of any improvements to a wind resource assessment campaign, it is necessary to have a baseline in order to allow comparisons to be made. In this case study, the project characteristics are

  • number of wind turbines: 12
  • wind turbine model: Enercon E70
  • wind turbine rating: 2.3 MW
  • wind farm rated capacity: 27.6 MW
  • wind turbine hub height: 64 m

Within the development of this project the following measurements were undertaken in order to assess the site wind resource:

  • 1 × 40 m meteorological mast installed
  • 2 × NRG Max#40 anemometers at heights of 40 and 10 m
  • 1 × NRG #200P wind vane at a height of 40 m

The setup of the measurement campaign is subject to several limitations, including non-compliance of the mounting setup with industry best practice approaches [3–5], low measurement equipment quality and low measurement height relative to the proposed turbine hub height.

Furthermore, the site is in complex terrain with some slopes greater than 17°. The original analysis was undertaken based on the WAsP wind flow model [9], which has limitations in complex terrain which are well documented [10–12]. These limitations within the measurement campaign and modelling approach resulted in elevated levels of uncertainty within the wind resource assessment.

Baseline scenario uncertainty

The baseline scenario measurements and analysis approach are sufficient to produce a wind resource assessment by following the processes as detailed in the section Context: wind resource assessment. The resulting pie chart of contributing uncertainties is shown in Fig 4 this is a typical breakdown for wind resource assessments with some variation depending on site specifics following the process briefly described in the section Uncertainty assessment.

Fig 4: Breakdown of contributors to project uncertainty

Based on this chart, it is possible to identify possible areas for reductions in uncertainty in order to improve confidence in the analysis, which are discussed in the following sections.

Site measurements – equipment quality

The quality of the anemometers used to record wind speed on the mast could be increased. The NRG Max#40 anemometers used on the mast are classified as IEC Class 2.4A [13], if more accurate anemometers were deployed, for example class 1.8 anemometers (common Thies and Vector models), then the uncertainty associated with site measurements would have been reduced (from 2.6 to 2.0% wind speed uncertainty).

Vertical extrapolation – mast height

The uncertainty associated with the vertical extrapolation of wind speeds can be decreased with increasing height of measurements relative to the proposed wind turbine hub height. To minimise the required vertical extrapolation, anemometers should be installed on a mast at heights representative of the wind turbine rotor – increasing the mast height reduces this uncertainty contributor.

Flow modelling – detailed flow physics

The site is characterised by some areas of steep slopes. The standard WAsP wind flow model has limitations in complex terrain, but alternative computational fluid dynamics (CFD) models address many of these limitations. CFD models use a method based on Reynolds-averaged Navier-Stokes equations to simulate fluid flows. These allow for an assessment of the flow physics and include the ability to model flow separation and recirculation over hills. This approach is more computationally expensive, requires detailed scientific knowledge and training to use and also requires more time to achieve a converged set of results when compared to the standard WAsP wind flow model. However, it can be used to reduce the uncertainty in wind flow modelling of a site.

Flow modelling – multiple measurement locations

The uncertainty associated with spatial variation and flow modelling can be reduced by taking multiple measurements on-site. Installing several additional measurement locations across the site increases the number of inputs and reduces the required extrapolation distances, whilst also allowing for flow model validation. This can be achieved in a cost effective manner by deployment of a fixed location meteorological mast along with a moveable remote sensing device. LiDAR units are ground-based remote sensing devices that can be moved and provide wind speed readings of comparable accuracy to those from anemometry installed on masts.

Cost–benefit analysis

The above options are the four primary methods for reducing the uncertainty for this site. Within Table 1, the approximate cost and benefit of each of these methods are presented. Cumulative costs and benefits are presented in order to truly represent the benefits of each improvement to the measurement approach. From an investment perspective the additional initial expenditure must be weighed against the potential returns. The simplest measure of this value is the return on investment with values >100 for the first three suggested improvements and a value of 35 if all suggestions are taken into account. The return on investment has been calculated using a typical financial model for a project-financed wind farm and comparing the predicted returns after 10 years of operation. The improved project P90 leads to financing based on a higher yield prediction, therefore the value of the asset is increased, leading to a greater attraction for equity investors. This in turn yields an improved equity internal rate of return.

Wind resource assessment approach

Total cost

Additional cost

P90 (10-year) improvement, >#/b###

Return on investment (10-year)

baseline setup


improved equipment quality





increased mast height





CFD modelling of the site





multiple measurement locations





Table 1: Cost–benefit analysis of improvements to the baseline measurement campaign

The figures presented are valid for the specific site and measurement campaign presented, and will vary significantly depending on site characteristics and the baseline approach considered. For the site under consideration the return on investment figures may be considered to be high due to the poor quality of the baseline scenario – it is anticipated that additional investments would typically lead to diminishing returns – this should be assessed on a case by case basis.


Go to the profile of David Pullinger

David Pullinger

Technical lead , Energy Resource Services at Lloyd's Register

No comments yet.