Regional power distribution network short-term load prediction method

Abstract

The invention belongs to the technical field of power load prediction, and particularly discloses a regional power distribution network short-term load prediction method based on a twin network classification method, a grey wolf algorithm and a long-short-term memory network. The load prediction method comprises the following steps: firstly, extracting and classifying historical load data by applying the twin network classification method, determining the category of input characteristics according to the load condition of a prediction day, and extracting a load with strong correlation to theload of the prediction day as input; the problems that the current effective input variable is difficult to select and the calculation steps are tedious are solved; and then parameter optimization selection is performed on the long-short-term memory network prediction model by applying a grey wolf algorithm, inputting an optimal parameter, and training and testing the long-short-term memory network prediction model to obtain a load prediction value under the prediction day. The method effectively solves the problems that an existing prediction model is low in efficiency, high in operation costand lack of self-adaptive capacity.

Description

technical field [0001] The invention belongs to the technical field of electric load forecasting, in particular to a short-term load forecasting method for a regional distribution network. Background technique [0002] Accurate short-term load forecasting of distribution network is an important guarantee for its safe and stable operation, dispatch optimization and reduced operating costs. In the process of short-term load forecasting, it will be affected by nonlinear characteristic factors such as severe weather, emergencies, grid failures, etc. If all these factors are used as input features, the complexity of load forecasting will increase, and the basis of these factors The data scenarios are diverse and the regularity is weak, which leads to a great reduction in the prediction accuracy based on data-driven methods. [0003] The historical data of distribution network load is the basic data for the operation of power grid enterprises. It has the characteristics of large ...

AI Technical Summary

Problems solved by technology

However, the above prediction methods have disadvantages such as slow prediction speed, lack of adaptability, high cost and long time required for model integration when processing time series data.

Images