Comparative evaluation of weibull parameter estimation methods for wind energy forecasting: A case study of the tetouan wind farm with SCADA-based availability integration

Abstract

• A comparative study of 11 Weibull parameter estimation methods was conducted. • Real SCADA data from the Tétouan wind farm (2019–2020) were used for validation. • The Empirical Method of Lysen (EML) provided the most accurate energy forecasts. • Integration of technical availability significantly improved prediction reliability. • The proposed methodology enhances wind farm management under real conditions. The reliability of wind power forecasts relies on accurate estimation of the Weibull distribution parameters, which is widely used to model wind variability. This study presents a comprehensive comparison of eleven Weibull parameter estimation methods (AMLM, EEM, EMJ, EML, EPFM, GM, MLM, MM, MMLM, PDM, STDM) applied to real wind data from the Tétouan wind farm in northern Morocco. Wind speed measurements, collected at an 80-meter height between 2019 and 2020, were analyzed to assess the forecasting performance of each method. Results demonstrate that the Empirical Method of Lysen (EML)provides the most accurate predictions, with corrected deviations between theoretical and actual energy production reduced to +2.30s% in 2019 and -0.24 % in 2020. Incorporating technical availability data from the SCADA system significantly improved forecast reliability. This dual approach, combining statistical modeling and operational constraints, proves to be an effective tool for optimizing the management and performance of wind farms under real-world conditions.