WindKI is a German research project led by LATODA in cooperation with Fraunhofer IWES (Fraunhofer Institute for Wind Energy Systems), focused on applying artificial intelligence to improve performance diagnostics for wind turbines.

Objectives

The project aims to develop an AI-supported diagnostic system that can

• detect performance losses (underperformance) in wind turbines early,

• evaluate effective wind speed vs. electrical output (power curve validation), and

provide insights into likely causes (e.g., blade angle settings, operating conditions).

Approach

• We combine high-resolution operational data (SCADA), simulations, and machine-learning models to identify anomalies and root causes in turbine performance data.

• The first models are built using data from the 8 MW Adwen AD8 research turbine provided by Fraunhofer IWES.
   

Benefits for Industry

• Wind turbine operators can monitor and optimize their assets faster, plan maintenance better, reduce downtime, and optimize energy yield.

• The project also aims to set a framework for future AI-driven wind energy research by bridging domain expertise with modern machine learning.
   

Partners’ Roles

LATODA: Develops the AI-based analysis system, combining sensor data and machine learning to flag and explain anomalies; Fraunhofer IWES: Provides data sets, simulation expertise, validation methods, and deep domain knowledge in wind turbine systems.

This project is funded by BMBF