Litcius/Paper detail

Tuning Parameters of Genetic Algorithms for Wind Farm Optimization Using the Design of Experiments Method

Wahiba El Mestari, Nawal Cheggaga, Feriel Adli, Abdellah Benallal, Adrian Ilinca

2025Sustainability12 citationsDOIOpen Access PDF

Abstract

Wind energy is a vital renewable resource with substantial economic and environmental benefits, yet its spatial variability poses significant optimization challenges. This study advances wind farm layout optimization by employing a systematic genetic algorithm (GA) tuning approach using the design of experiments (DOE) method. Specifically, a full factorial 22 DOE was utilized to optimize crossover and mutation coefficients, enhancing convergence speed and overall algorithm performance. The methodology was applied to a hypothetical wind farm with unidirectional wind flow and spatial constraints, using a fitness function that incorporates wake effects and maximizes energy production. The results demonstrated a 4.50% increase in power generation and a 4.87% improvement in fitness value compared to prior studies. Additionally, the optimized GA parameters enabled the placement of additional turbines, enhancing site utilization while maintaining cost-effectiveness. ANOVA and response surface analysis confirmed the significant interaction effects between GA parameters, highlighting the importance of systematic tuning over conventional trial-and-error approaches. This study establishes a foundation for real-world applications, including smart grid integration and adaptive renewable energy systems, by providing a robust, data-driven framework for wind farm optimization. The findings reinforce the crucial role of systematic parameter tuning in improving wind farm efficiency, energy output, and economic feasibility.

Topics & Concepts

Genetic algorithmAlgorithmMathematical optimizationOptimization algorithmComputer scienceEngineeringMathematicsWind Energy Research and DevelopmentIcing and De-icing TechnologiesSolar Radiation and Photovoltaics