Power short-term load prediction method and device

Abstract

The invention discloses a power short-term load prediction method and device. The method comprises the steps of dividing day types of historical data based on a daily load curve; Respectively establishing a plurality of multivariate linear regression prediction models aiming at different day types by taking the obtained historical daily load information contained in each day type and the selectedcharacteristic data as input and taking the predicted daily load value as output; And based on the TensorFlow deep learning model, carrying out training, parameter tuning and verification on the plurality of established multivariate linear regression prediction models to obtain short-term load prediction models for different day types. According to the invention, day types can be automatically divided according to information such as holidays and daily load curves; A multivariate linear regression model is adopted, and influences of holidays and weather changes on loads are comprehensively considered; And automatically training and optimizing according to the daily type under a deep learning framework to obtain three short-term load prediction models, and calculating to obtain a relativelyaccurate load prediction value.

Description

Technical field [0001] The invention relates to the technical field of power system load forecasting, and in particular to a short-term power load forecasting method and system. Background technique [0002] Power short-term load forecasting is one of the important components of power system demand-side management, and it is closely related to power system production planning and power dispatch operation. Electric short-term load forecasting requires comprehensive consideration of power load-related factors, and accurate short-term load forecasting is of great significance for optimizing resource allocation and ensuring the safety, reliability and economy of power supply. [0003] However, most of the existing short-term load forecasting methods, such as those based on BP neural network, face the problem of unreasonable feature selection or too many redundant features, resulting in low prediction accuracy. Summary of the invention [0004] The technical problem to be solved by the ...

AI Technical Summary

Problems solved by technology

[0003] However, most of the existing short-term load forecasting methods, such as the forecasting method based on BP neural network, face the problem of low forecasting accuracy caused by unreasonable feature selection or excessive redundant features.

Images