Short Term Load Forecasting Based on Data Driven

It is difficult to process large data volume by using traditional short term power load forecasting model which is provided with strong randomness and the forecasting accuracy is hard to satisfy the requirements of the smart grid. Therefore, this paper proposes a kind of data-driven short term load forecasting method based on the gated recurrent unit (GRU) neural network to predict power load in some area and make comparative analysis on the forecasting results with the results of the traditional method. The comparison results indicate the proposed method takes into full consideration of the time sequence and nonlinear of power load and is feasible in deeply excavating data characteristics of power consumers. Meanwhile, the short term forecasting property of this method is superior to the traditional method and is able to meet the demand of the smart grid for high precision.