Presentations

Abstract

Accurate Annual Energy Production (AEP) prediction and robust uncertainty quantification are critical for reducing financial risk as the wind industry scales toward larger, complex farm clusters. Current validation efforts are typically limited to single sites or public datasets, lacking the statistical weight required for general conclusions. This work presents an ongoing massive validation initiative within the Horizon Europe FLOW project, systematically benchmarking a comprehensive matrix of wind farm models against real-world operational data. To standardize assessment across varying fidelities - ranging from fast engineering codes (e.g., PyWake, FOXES), atmospheric perturbation models (e.g., WAYVE), to high-fidelity CFD solvers (e.g., code_saturne) - we developed the integrated Wind Farm modeling framework (WIFA). Built upon the WindIO ontology, WIFA unifies input/output structures, enabling the consistent execution and benchmarking of diverse solvers within a single environment. The study uses an extensive portfolio of more than 100 commercial wind farms, representing a wide range of topographies, climates, and layout densities. To overcome the proprietary nature of high-fidelity SCADA data, we implement a distributed validation protocol. The WIFA framework is deployed directly within the secure internal environments of industry partners, i.e. Vestas and Siemens Gamesa. This "code-to-data" approach allows models to be validated against sensitive assets while ensuring only aggregated, anonymized error metrics are extracted for analysis. The validation covers two distinct phases: (1) pre-construction workflows, driven by standard wind resource assessment methodologies to evaluate total system performance; and (2) post-construction workflows, driven by upstream turbine signals to specifically isolate and quantify wake and blockage model accuracy. The resulting analysis provides a rigorous quantification of model-chain uncertainty, correlating prediction errors with environmental variables to inform industry best practices.