Industrial medium and long term load prediction method based on long and short term memory neural network and support vector regression combination model

Abstract

The invention discloses an industry medium and long term load prediction method based on a long and short term memory neural network and a support vector regression combination model, and the method comprises the steps: firstly, analyzing the correlation degree of an industry medium and long term load influence factor and an industry load based on a Pearson's correlation coefficient; and key influence factors of long-term load prediction in the industry are extracted. Secondly, respectively constructing a long-short-term memory neural network prediction model considering load time sequence change characteristics and a support vector regression prediction model considering load nonlinear characteristics; and then, on the basis of an optimal combination prediction algorithm, constructing a combination prediction model which is based on a long-short-term memory neural network and support vector regression and considers load comprehensive characteristics, and predicting medium-and-long-term loads in the industry. Based on the optimal combination prediction algorithm, the characteristics and advantages of the long and short term memory neural network and the support vector regression prediction model are comprehensively considered, and compared with a single prediction method, the precision of medium and long term load prediction in the industry is effectively improved.

Description

technical field [0001] The present invention relates to the technical field of power systems, and more particularly, to a medium and long-term load forecasting method in the industry based on a long-short-term memory neural network and a support vector regression combined model. Background technique [0002] Power system load forecasting refers to the pre-estimation and prediction of future load development and changes. It is the basic work of power system planning, dispatching and other departments, and plays a vital role in power system planning and operation. According to the different forecasting time scales, load forecasting can be divided into ultra-short-term forecasting, short-term forecasting, medium- and long-term forecasting, etc. Among them, medium- and long-term load forecasting is mainly aimed at forecasting the load on monthly and above time scales. Accurate medium- and long-term load forecasting is beneficial to Power supply companies master the electricity c...

AI Technical Summary

Problems solved by technology

Therefore, load forecasting models based on deep learning neural networks and ensemble trees, which are currently studied more often, often have poor forecasting performance due to overfitting and inability to converge.

[0004] In this context, the integrated combination forecasting model for medium and long-term load forecasting in the industry still needs further research

Images