Posters

Abstract

In the accelerated development of wind energy projects, there are often situations in greenfield projects where local wind measurements are incomplete or entirely absent, or in the case of wind farm reconstructions, limited to SCADA data from existing or decommissioned turbines. These scenarios pose significant challenges for accurate wind resource assessment. To address these challenges, this paper introduces a statistical conflation method applied to enhance wind estimation accuracy by integrating and weighting multiple wind measurements through computational fluid dynamics (CFD) models and statistical compatibility tests. The method evaluates the compatibility of diverse data sources, including distant anemometric stations, virtual-mast satellite data, and SCADA datasets from wind farms, whether operational or decommissioned. Each measurement is extrapolated to the target location or height using a CFD model, incorporating uncertainties from measurement equipment, wind shear, and horizontal extrapolation distance. The uncertainty is further classified into dependent and independent components to account for shared CFD model errors, with measurements of higher uncertainty assigned less weight in the final conflation process. The agreement of all measurements is verified through statistical compatibility tests. Incompatible measurements have their uncertainty adjusted to achieve compatibility, a feature especially valuable for SCADA data, which often includes extensive datasets prone to anomalies such as wake effects and turbine-specific operating issues. A novel method is introduced to assess the increase in uncertainty with increasing distance and variations in orographic complexity. This method, based on the T-RIX methodology, uses a preliminary cross-check among measurements, ensuring the uncertainty assessment is site-specific. Validation across multiple sites with reliable long-term anemometric data demonstrated significant uncertainty reductions. For instance, combining two wind measurements from 8 km distances with virtual-mast satellite data at the site reduced uncertainty to a more acceptable level, enabling improved decision-making in project feasibility studies. This method is particularly beneficial for greenfield projects, where it provides critical preliminary estimates pending the installation of local measurement equipment. Additionally, it optimizes the utilization of SCADA data in wind farm total reconstruction projects, and mitigates the impact of wake effects and turbine-specific operating anomalies. Statistical compatibility tests ensure robust integration by identifying and adjusting incompatible data to achieve agreement among datasets.