| abstract |
The invention relates to a method and system for distribution network load forecasting. The method includes: obtaining an unsupervised training sample set, a supervised training sample set and a test sample set according to time information according to the distribution network historical load influence factor and the historical load value of the area to be predicted; according to the unsupervised training sample set, the The DBN model layer in the pre-established load forecasting model performs layer-by-layer unsupervised training; the load forecasting model includes: a DBN model layer and a linear neural network layer; the network parameters obtained by the unsupervised training are used as the load forecasting model The initial value of the network parameter; According to the supervised training sample set, the load forecasting model is supervised and trained to obtain the optimal load forecasting model; the test sample set is tested by using the optimal load forecasting model, and the load forecasting model is obtained. The load forecast value for the forecast area. The invention can realize high-precision load forecasting influenced by various factors in the smart grid environment. |