Prediction of Uncertain Factors on Both Supply and Demand Sides of Microgrid Based on CNN-BiGRU Model

Wind power, photovoltaic output and power load in microgrids are highly random and volatile, and the accuracy of their prediction is crucial to providing microgrid power supply reliability and operational economy. Based on this, this paper proposes a microgrid supply and demand uncertainty prediction model based on the CNN-BiGRU model. Firstly, the key factors affecting the wind power, photovoltaic output and power load of the microgrid are determined based on Pearson correlation. Furthermore, a CNN-BiGRU-based probability density prediction model for uncertainty factors in microgrids is constructed by comprehensively adopting deep learning, multi-task learning, quantile regression and other methods. Finally, a case study is carried out to verify the proposed model in a typical microgrid.