Back to the programme printer.gif Print




Delegates are invited to meet and discuss with the poster presenters in this topic directly after the session 'Wind speed predictions: Are we at the limit of our knowledge or can we improve?' taking place on Wednesday, 12 March 2014 at 11:15-12:45. The meet-the-authors will take place in the poster area.

Fernando Mourão INEGI, Portugal
Co-authors:
Fernando Mourão (1) F P José Matos (1)
(1) INEGI, Porto, Portugal

Printer friendly version: printer.gif Print

Abstract

Achieving long term reference data by statistical downscaling

Introduction

Information from the several available reanalysis datasets are commonly used in energy yield and site assessment studies to account the long-term significance of local measurements. However, this data is seldom available for specific sites being that in such cases, it is accepted that the geographical point of this data is located at a site with the same wind regime as the location of the meteorological station. This is not always true, especially at complex locations. Efforts can, nonetheless, be done to bypass this handicap and create location dependent data series with long term representativeness providing some local data is available.

Approach

An approach based on Empirical Orthogonal Functions is here employed to create weather patterns from long-term meteorological data considered as the reference one. A training period for the model is defined based on the observations performed at a specific location. This training period allows to extend the significance of the observed data when compared to the long term expected, which is then used in resource assessment and energy yield calculations.

Main body of abstract

A statistical downscaling method allowing, with a minimum of a year worth of meteorological observations at a specific site, to downscale information from reanalysis datasets or from mesoscale simulation models to any site under appreciation for wind energy project development purposes is here presented. Ultimately, this exercise results in the expansion of the time significance of the observed data, both in terms of wind speed and direction, up to a level closer to the long-term expected, depending on the used reference data.
The output time series as attained from the downscaling model can then be used to, using conventional tools, generate wind resource maps and estimate the energy yield of a wind farm. The paper will be focused on presenting the methodology and assessing how both wind resource assessment and energy yield calculations respond to the usage of this data and, as importantly, how is uncertainty affected by analyzing several test cases.


Conclusion

It is hoped that with the here proposed methodology, the long term significance of data commonly used in resource assessment can be improved with positive impact on the accuracy of energy yield estimates and site assessment studies enabling, simultaneously, a reduction on the associated uncertainty figures. It is also believed that it is, based on the presented methodology, to establish an interesting benchmark that can be widely accepted by the wind energy industry.


Learning objectives
The audience will learn of a novel approach on statistical downscaling used for wind resource assessment purposes and potentially use this technique in their own studies or even developing further approaches based taking the here presented information as a starting point.