Based on the WSP-Optuna-LightGBM model for wind power prediction

In order to optimize energy dispatch and enhance the predictive performance of wind power forecast, this study proposes a WSP-Optuna-LightGBM mixed regression prediction model based on wind speed-power curve (WSP), Optuna parameter optimization, and Light Gradient Boosting Machine (LightGBM). The first step of this study involves utilizing historical wind speeds and actual power generated by wind turbines to develop a wind speed-power curve. The curve is used to estimate the output power of the turbines at any given moment, serving as inputs for the model. Next, this study processes historical data from the wind farm and continuously optimizes the model using cross-validation and the Optuna algorithm. Finally, the data prediction results under the WSP-Optuna-LightGBM regression prediction model are obtained. Experiments based on two assessment indicators are conducted to analyze the predictive performance of the wind power prediction model. The results demonstrate that the proposed method exhibits greater accuracy in its predicted values when compared with other wind power prediction models.