Data-driven wind turbine simulation model and interpretability analysis

The wind turbine simulation model is the core component of the optimal control of wind farms and farm groups. However, due to the complex and changeable operating environment and working conditions, the current simulation model cannot accurately simulate the operating characteristics of the unit. In this regard, this paper proposes a data-driven wind turbine simulation model modeling method, and uses the SHAP method to perform model interpretive analysis. First, based on the measured data, a data-driven method was applied to establish a wind turbine simulation model that takes environmental factors into account. The accuracy of the model reached 96.4%. Compared with the measured data regression power curve and factory power curve, NRMSE, NMAE and MAE were all reduced by more than 60%. The simulation accuracy of the wind turbine power output is significantly improved. In addition, the interpretability analysis of the model is performed through the SHAP method, which enhances the interpretability of the model. It not only proves the effectiveness of the model, but also provides important theoretical support for accurate simulation and further optimization of wind turbines.