Electricity sales prediction method based on combination of grey correlation analysis and SA-PSO-Elman algorithm

Abstract

The invention discloses an electricity sales prediction method based on combination of grey correlation analysis and an SA-PSO-Elman algorithm. Because the power selling quantity of the power grid enterprise is influenced by many factors, the power selling quantity data sequence of the power grid enterprise shows complex nonlinear change and is difficult to describe by an accurate mathematical model. According to the technical scheme adopted by the invention, the method comprises the following steps: 1) objectively selecting main influence factors for selling the electric quantity by adoptinga grey correlation analysis method; 2) taking the influence factor as the input of an Elman neural network, and determining the training parameters of electricity sales prediction through employing anSA-PSO-Elman algorithm; and 3) substituting the training parameters into the electricity selling quantity prediction model, establishing the electricity selling quantity prediction model, and predicting the electricity selling quantity, so that the electricity selling enterprise can accurately and reasonably estimate the electricity selling quantity, and the electricity selling enterprise can conveniently and effectively manage and control the investment and income of the electricity selling enterprise.

Description

technical field [0001] The invention belongs to the field of electricity sales forecasting of grid enterprises, and in particular relates to a method for forecasting electricity sales of grid enterprises. Background technique [0002] Grid electricity sales are an important part of grid enterprise investment and cost control. In order to formulate the best investment strategy, it is necessary to accurately and reasonably estimate grid electricity sales. However, due to the influence of many factors on the power sales of power grid enterprises, the data series of power sales of power grid enterprises show complex nonlinear changes, which are difficult to describe with accurate mathematical models. [0003] Currently, Elman neural network is used to establish electricity sales forecasting model, see Wang Shiting, Zhou Chunlai, Feng Jing, et al. Research on Power Load Forecasting Based on Elman Neural Network [C] / / The Ninth National Signal and Intelligent Information Processing...

AI Technical Summary

Problems solved by technology

However, due to the influence of many factors on the electricity sales of grid enterprises, the data series of electricity sales of grid enterprises show complex nonlinear changes, which are difficult to describe with accurate mathematical models.

[0003] Currently, Elman neural network is used to establish electricity sales forecasting model, see Wang Shiting, Zhou Chunlai, Feng Jing, et al. Research on Power Load Forecasting Based on Elman Neural Network [C] / / The Ninth National Signal and Intelligent Information Processing and Application Academic Conference.0. However, the Elman neural network uses the error backpropagation algorithm to implement weight correction, which has disadvantages such as long training time, slow convergence speed, and easy to fall into local minima.

Images