Research on Precise Control of Distribution Network Based on Multivariate Load Forecasting

This paper proposes a mathematical model based on variational mode decomposition to achieve precise control of distribution network power flow, in response to the non-stationary, temporal correlation, and nonlinear characteristics of time series of multivariate power load data. The model takes the active power of multiple loads as the original data and acquires the analysis results by improving the variational mode decomposition of the stopping criterion. Extract the energy values of each component as features, and classify the signal through a deep confidence network. The simulation results prove that the model can effectively predict multiple load data.