Wind and PV Power Prediction Based on Disclosure Forecast Bias and GRU Model

With the continuous integration of wind power, photovoltaic, and other renewable energy stations, the power market is encountering both increased opportunities and challenges. Accurate prediction of wind and photovoltaic power generation plays a crucial role in guiding the economic dispatch of the grid and informing the strategic decisions of market participants. To enhance the accuracy of power predictions, this paper proposes a bias-correction-based prediction method utilizing disclosed data. Specifically, this method employs a Gated Recurrent Unit (GRU) deep learning model to forecast the prediction bias in disclosed data and subsequently corrects the disclosed data based on these bias predictions, thereby improving the overall accuracy of power predictions. The effectiveness of this method is validated using disclosed and actual data from Zhejiang Province spanning 2022–2024. Experimental results demonstrate that, compared to the prediction accuracy of raw disclosed data, the accuracy of wind and photovoltaic power predictions after bias correction has been significantly improved.