Ann Mary Toms

Ann Mary Toms, Xingpeng Li

The global transition towards renewable energy has accelerated the deployment of utility-scale wind farms, increasing the need for accurate performance and economic assessments. Although wind energy offers substantial potential for carbon emission reduction, investment decisions are highly sensitive to predicted annual energy production and economic profitability. Conventionally wind farm analyses often estimate turbine power output based solely on incoming wind conditions, neglecting wake interactions between turbines. These wake effects can significantly reduce downstream turbine performance, leading to overestimation of energy yield and financial returns. This study proposes WAKE-NET a wake-aware optimization framework that incorporates both turbine layout optimization and hub height diversification across turbines of varying capacities. Unlike traditional approaches that assume a uniform hub height or ignore wake dynamics, the proposed methodology accounts for wake-induced power losses in its framework. Results indicate that the benchmark model that neglects wake effects can overestimate annual profits, while the use of multiple hub heights reduces wake overlap and associated power losses. Overall, the findings demonstrate that wake-aware design and hub height diversity improve energy yield accuracy and economic viability, offering a valuable guidance for wind farm developers and investors seeking to invest in renewable energy systems.

Linhan Fang, Fan Jiang, Ann Mary Toms et al.

Accurate wind speed prediction is crucial for designing and selecting sites for offshore wind farms. This paper investigates the effectiveness of various machine learning models in predicting offshore wind power for a site near the Gulf of Mexico by analyzing meteorological data. After collecting and preprocessing meteorological data, nine different input feature combinations were designed to assess their impact on wind power predictions at multiple heights. The results show that using wind speed as the output feature improves prediction accuracy by approximately 10% compared to using wind power as the output. In addition, the improvement of multi-feature input compared with single-feature input is not obvious mainly due to the poor correlation among key features and limited generalization ability of models. These findings underscore the importance of selecting appropriate output features and highlight considerations for using machine learning in wind power forecasting, offering insights that could guide future wind power prediction models and conversion techniques.