ABBA Optimization Framework

Introduction

This simulation and optimization framework was developed at the University of Stuttgart as part of the research project ABBA, which was financed by the German Federal Ministry for Economic Affairs and Energy. It contains tools to simulate and optimize the operation of wind turbines, i.e. to improve the ratio of revenue generation to fatigue damage accumulation.

Content

Simulation Management Framework

Located in the folder simdriver.

Provides pre- and postprocessing as well as multi-threaded orchestration of OpenFAST simulations. Damage equivalent loads are calculated automatically using fatpack. The routines are implemented in Python and integrated into a Python module. Further instructions can be found in simdriver/README.md.

Data Processing

Located in the folder data_processing.

Various data processing routines to support the optimization tools.

Shutdown-based Optimizer

Located in the folder shutdown_opt.

This optimizer is based on the idea of shutting down the turbine at certain times to reduce fatigue damage. Analysis of the response to errors in the wind speed forecast or large price changes is supported. Additionally, some machine learning approaches to the optimization have been implemented.

MMKP Optimization

Located in the folder MMKP_opt.

This optimizer is based on the idea of adjusting the turbine's operating point to reduce fatigue damage. The optimization problem is formulated as a multidimensional multiple-choice knapsack problem (MMKP) and solved using a compositional pareto heuristic. Please refer to the publications below for more details.

The optimizer is written in Rust and integrated into a Python module. The Rust code is located in the folder MMKP_opt/mmkp_heuristic. Various scripts for data preparation and evaluation of the results are provided.

Low-frequency Fatigue Analysis

Located in the folder low_frequency_fatigue.

Extension of the usual fatigue analysis methodology to account for low-frequency fatigue. The method is based on the work of Schmelter and Cheng (2026) and implemented in Python. Example scripts for the application of the methods are provided.

Publications

Flexible Operation of Wind Turbines as a Multidimensional Multiple-Choice Knapsack Problem

Julius Schmelter, Nico Ruck and Po Wen Cheng, Wind Energy Science Conference 2025, Nates, France

https://doi.org/10.5281/zenodo.15798195

Comparison of optimization approaches for price-dependent wind turbine operational strategies

Niklas Requate, Julius Schmelter and Nico Ruck, DeepWind 2026, Trondheim, Norway

Impact of flexible wind turbine operation on low-frequency fatigue

Julius Schmelter and Po Wen Cheng, Torque 2026, Brussels, Belgium

Accepted for publication, in press.

Acknowledgements

This work is part of the research project ABBA, funded by the German Federal Ministry for Economic Affairs and Energy under grant number 03EE3068B.