Power prediction algorithm for wind solar integration into DC power grid under trend analysis

Abstract After wind and solar power are connected to the DC grid, there are complex interactions between them and the grid. Fluctuations in voltage and frequency of the power grid, as well as changes in other power sources and loads, will all have an impact on the power of wind and solar power generation. Meanwhile, the control strategies and operational characteristics of the DC power grid will also alter the output of wind and solar power generation. This mutual influence increases the uncertainty of power prediction, and it is necessary to accurately analyze the variation patterns of wind and solar power to achieve relatively accurate prediction. Therefore, this paper explores the power prediction algorithm for the integration of wind and solar power into the DC grid under power flow analysis. The Savitzky–Golay filtering and linear interpolation method were used to complete the preprocessing of wind and solar power grid access data. The Newton-Laffson method is adopted for power flow analysis to construct the power flow model of the DC power grid, and the node voltage and power distribution are solved through iterative algorithms, providing real-time data support for power prediction. Based on this, the variational mode decomposition (VMD) method is utilized to extract data features, and the original data is decomposed into modal components of different frequencies to reflect the variation law of wind and solar power. A power prediction model for wind and solar power integrated into the DC grid based on BP neural network is established. The extracted features are taken as input, and the power of wind and solar power integrated into the DC grid is taken as output. Through training and adjusting the network weights and thresholds, accurate power prediction is achieved. The experimental results show that the proposed method has high accuracy in DC grid power prediction, with an average absolute error only between 0.11 and 0.25. The operation time of DC grid power prediction is within 5 s, which has practical application value.