Long Term Wind Turbine Performance Analysis Through SCADA Data: A Case Study

Abstract

Performance monitoring of horizontal-axis wind turbines is a complex task because they operate under nonstationary conditions. Furthermore, in real-world applications, there can be data quality issues because the free stream wind speed is reconstructed through a nacelle transfer function from cup anemometers measurements collected behind the rotor span. Given these matters of fact, one of the objectives of the present work is applying an innovative method for correcting the nacelle wind speed measurements, which is based on the manufacturer power curve and statistical considerations. Three operating wind turbines, having 2 MW of rated power and owned by the ENGIE Italia company, are contemplated as test cases. Operation data spanning ten years (2011–2020) are studied: actually, this work aims as well at contributing to the methods for estimating the performance decline with age of wind turbines, basing on long term SCADA data analysis. The raw and corrected wind speed measurements are fed as input to a Support Vector Regression for the power curve: by selecting appropriately the training and validation data sets, it is possible to estimate the average yearly rate of performance decline. Using the corrected wind speed, the estimate obtained in this study is compatible with the most recent findings in the literature, which indicate a -0.17% decline per year.