Novel interpretable wake model using Machine Learning based symbolic regression for wind turbines

The escalating demand for energy, dwindling fossil fuel reserves, shifting fuel prices, and worries about climate change and pollution have prompted the investigation and assessment of renewable energy sources. Among these, wind power has emerged as a particularly promising candidate. However, due to limited suitable locations, wind turbines are often clustered together, necessitating the study of wake aerodynamics to optimize wind farm performance. Understanding and predicting wake characteristics that includes decreased velocity deficit and increased turbulence intensity are crucial for improving the efficiency of downstream turbines. Traditionally, low-fidelity models were used to predict wake characteristics, but recent advancements in computational fluid dynamics (CFD) have enabled more accurate predictions, albeit at the cost of increased computational time and resources. Machine learning (ML) has proven to be a valuable tool in addressing complex problems in various engineering domains, leading to the development of ML-based wake models that leverage CFD data for rapid predictions. However, these models have been limited by their reliance on a single wind turbine’s CFD results and their lack of generalizability to other turbines with different blade geometries and operating conditions. In response to these limitations, this study presents a novel wake prediction model using the symbolic regression technique. This innovative approach aims to offer precise aerodynamic performance predictions while maintaining simplicity, interpretability, and generalizability across various wind turbine designs. The proposed model holds significant potential for enhancing wind farm performance by accurately predicting wake characteristics and facilitating better decision-making in turbine placement and operation.