Impact of electric charge compensation on wind energy production: a statistical quantile regression analysis

Abstract

The energy production behavior of wind turbines is generally controlled and regulated by the analysis of the generated energy versus wind speed. The wind speed is referred to as the measurements taken by an anemometer placed in the gondola, after the main plane of each wind turbine. Wind turbines are large in size, with mechanical, electrical, and electronic components, and their cost, installation, and maintenance are high. For these reasons, and due to the environment in which they operate, they require electrical protections, similar to other important installations such as telecommunications towers, airport radars, transmission lines, and substations, among others.The Electronic Charge Compensation Devices (ECCDs) are passive devices that carry electrical currents to the electrical Earth, preventing lightning impacts and derivative electrical current pulses, reducing radiofrequency disturbances in the protected area, and attenuating the corona effect as well.In this work, data collected over more than a year of operation of several wind turbines installed at different locations are analyzed. We compare the energy produced versus wind speed in turbines with and without ECCDs installed, in order to assess the changes in their behavior. Our preliminary work shows a relevant increase in energy production (about 9 %2B- 5 in the Annual Energy Production – AEP – for the different analogous wind turbines) due to the use of ECCDs. Additionally, better maintenance and longer operational life can be expected for turbines with ECCDs installed.Our experience, accumulated over four years of continuous work with various statistical analyses and mathematical fitting programs, shows that statistical analyses are typically based on histograms of data where dynamic ranges or sampling intervals are assumed to fit a Gaussian distribution. However, the data derived from the wind turbines we considered do not follow a normal Gaussian distribution. The introduction of a new parameter or variation due to the use of ECCDs in the wind turbines results in significant changes in the histogram distributions. In particular, there are notable changes in the asymmetry index of the generated energy and wind speed statistical distributions.In this communication, we present a statistical analysis based on a double medians methodology, which we have used in other works, as well as statistical quantile regression and deconvolution histogram tools, to evaluate the inter-correlation mechanisms associated with the presence or absence of ECCDs in the wind turbines.This work was partially developed during the 1ENG08 WindEFCY Project of the EMPIREURAMET Program and has continued with the support of the Dinnteco group of enterprises.