Wind resource variation over time is the main cause of the variation of the production output of a wind farm.

Wind energy indexes allow to quantify the variation of the wind resource on a given period and thus constitute an efficient tool to get an overview of the behavior of a wind farm portfolio for a competitive cost. Using wind energy indexes as a macro scale tool can be considered as the first step before deciding or not to carry out a deep analysis of SCADA data per turbine.

irec indexes are mainly based on data issued from ERA5 reanalysis.
ERA5 is a climate reanalysis dataset, covering the period 1950 to today, which is developed through the Copernicus Climate Change Service (C3S). ERA5 is the fifth major global reanalysis produced by ECMWF. Data from ERA5 cover the Earth on a 0.25° grid and on an hourly time step.
These data have been converted into production data to ensure that irec indexes are directly comparable to your wind farm production output.

Cumul irec indexes provide the ratio between the wind energy available over a given time period and over the long-term average resource in that same time period.
As an example, a Cumul irec of 95% for the period January-March of the current year means that, in the region considered, the production expected should be 5% below what can be expected for the period January-March on a long-term average.

Dealing with monthly databases, a direct correlation between these data and your production data (corrected from availability issues) can be carried out, in order to assess the actual production capacity of a wind farm and ensure the stability of its performance over time as a macro scale tool. To understand all the possibilities offered by such data and make the most of wind energy indexes, we propose dedicated training sessions (see Windex sessions).

Note that pre-calibrated index for your analysis can be calculated on a specific demand (customized index).

A feedback experience on more than 200 operational wind farms around the world (especially across Europe) allowed to validate the reliability of irec indexes. The average correlation coefficient between irec indexes and actual production data is about 97.5% prior to any filtering, and it exceeds 94% for 9 out of 10 wind farms tested..
The lowest correlation levels encountered are often in quite complex areas subject to quick variation in space of the wind regime (small predefined regions).

Let’s keep in mind that the level of uncertainty month by month remains quite high. This uncertainty is inherent in wind energy indexes themselves (not made from on-site wind measurements) but could also be linked to each wind farm specificities (unusual wake effects due to a particular monthly wind rose…) . Thus, a conclusion cannot be drawn from a single month. However, a trend analysis or a look at cumulated periods can lead to a quite finer analysis.

Globally, the use of wind energy indexes can allow to ensure the stability of the production capacity of a farm within a ± 3 % to 5 % range while considering cumulative production periods. A drift in performance can thus be suspected outside this range (more details on the poster presented in the WindEurope Technology Workshop).

Learn more about this topic attending Windex training sessions.

First of all, did you make sure of the following points?
1- Is your wind farm inside the predefined region?
Use the tool on the left-hand menu to make sure, ask for a customized index if not.
2- Are your production data adjusted to 100% availability (Ideal production)?
Wind energy indexes reflect the wind resource that can be harnessed by a wind farm with no availability issues. That is why the production output should be corrected from production losses encountered by the wind farm. All causes of downtime, except for lack of wind should be taken into account.
3- Is your wind farm submitted to a significant curtailment?
If yes, corresponding production losses should be accounted for.

Once ruled out these causes, note that all modelled data as ERA5 are subject to uncertainties, especially in complex areas and/or in areas with a very local wind regime. Such uncertainty on modelled data can lead to lower correlation levels in a few cases.

Need further clarifications? This topic will be discussed in Windex training sessions.

Predefined regions proposed on irec index website correspond to areas within which the wind regime is homogeneous (i.e. similar wind speed variation over time). These regions correspond also to areas with a high density of operating wind farms. The regions contours were determined based on an algorithm allowing to aggregate all areas offering similar wind trends over time.

	Energy index < 97% (wind resource on the period lower than on average for the same period)
	Energy index 97%-103% (wind resource on the period close to the average for the same period)
	Energy index > 103% (wind resource on the period higher than on average for the same period)

See the list of predefined regions

Regions presented as dots on the map (“+” sign) correspond to areas with very local wind regimes, with a high probability of variation over short distances.
On the opposite, big predefined regions reflect the fact that the wind regime is similar on large areas.

Predefined regions correspond to circles as far as possible in the interest of simplification. However, in several areas the wind regime did not permit to define circles, and ellipses were favored to match the specificity of the wind resource.

Use the dedicated tool of the desktop version of the website.

Learn more about our indexes on customized locations.
Be careful if you are outside predefined regions, especially small ones. The wind regimes can vary on very short distances in some places.

All irec indexes on predefined area take into account the period January 2011-December 2020 (10 years) as the long-term reference period. In the wind industry, a decade is a standard reference period, and not going far back in time allow to limit the risk of inconsistency of the data over time. More than the duration of the reference period, one of the key issues when comparing indexes year after year is the stability of this reference period. Thus, in order to maintain values comparable between each other from one year to another, this reference period will remain constant for upcoming years.
Each new year a conversion factor will be provided to allow a translation of the energy indexes from the fixed long term reference period to the most recent decade.
Note that indexes on customized location can allow you to get indexes on the long term reference period of your choice

It is commonly accepted that the longer the reference period, the better the long-term prediction of wind resource should be. However, statistically speaking if a trend is noticed over time (decrease or increase of the wind resource), then extending the reference period in the past could generate a bias for future prediction.
Thus, according to us the best compromise to limit the uncertainty on the long term prediction of the wind resource, regardless a possible trend over time, is to consider a significant period but relatively recent such as 10 years, or possibly 15 years.

If re-analysis data (such as ERA5) allow to get satisfactory information regarding the evolution of the wind resource over periods of about 10 years, some deviations with ground measurements can be observed over longer periods on several locations (see Eoltech publications on this topic). Thus, the relevance of these modelled data can be subject of debate when a trend (upwards or downwards) is observed on the wind resource on periods widely exceeding 10 years. For this reason, in order to limit a risk of bias we have limited to 15 years the historical data available, and we recommend to consider 10 years.

Within a region considered homogeneous in terms of wind regime, the amplitude of variation of production can differ from one farm to another due to their specific characteristics (turbine type, exposure level…).
Providing a range allows to cover the major part of indexes that can be associated to different types of wind farms within each region. The range can be exceeded in some specific cases (unusual wind farm exposure and/or turbine type in the region).

Unless specified otherwise before December 1st of the current year via an email to irecindex@eoltech.fr, your subscription to monthly or Cumul irec will automatically be renewed for the coming year. The prices invoiced in this case will be the prices mentioned on this website. If these prices are different from the ones on the previous year, you will be informed by email before reception of the invoice for the following year.

In just one day, this training session will allow you a better understanding of the key aspects linked to the use of wind energy indexes, in order to make the most of this tool based on practical cases.

The session can allow you to answer to the following questions:
- How to explain the production variations of my wind farm?
- What are the main applications for wind energy indexes?
- How to use an annual / a monthly index?
- How to check the stability of my production capacity?

Get pratical information about the sessions here: Windex, a training session proposed by Eoltech

Any other question? Contact us