Microgrid Distributed Power Forecast Taking Multivariable Correlation and Weather Conditions into Account

Optimizing scheduling in accordance with each microgrid's source and load matching is essential to properly reaping the benefits of the microgrid group. The microgrid has a small area, and the network's distributed power sources and sensitive loads are subject to the same meteorological conditions. As a result, it is necessary to simultaneously estimate the microgrid's supply and load power in the same environment. In order to achieve this, the combined effects of weather factors on the microgrid's source and load power are taken into account. The input data is then extracted in order to carry out additional short-term output projections for the microgrid's distributed power sources.Principal component analysis (Principal Component Analysis,PCA) is used to reduce the dimensionality of the extracted factors after the multidimensional maximum correlation coefficient method (multi-variable maximum information coefficient, MvMIC) has been used to screen the features and eliminate factors with less influence. Lastly, a variety of neural network combination prediction models are built. Forecast power. Simulated results indicate that the forecast accuracy is higher when microgrid power and load power are predicted for different seasons as opposed to just taking the weather's effects on the microgrid's distributed power supply into account.