Power system load prediction method and system based on wavelet noise reduction and EMD-ARIMA

Abstract

The invention discloses a power system load prediction method and system based on wavelet noise reduction and EMD-ARIMA, and belongs to the field of power system load prediction. The method comprisesthe following steps: firstly, acquiring original load data of a power system, and then performing noise reduction processing on the load data by utilizing wavelet analysis; further processing the denoised data by using an empirical mode decomposition (EMD) method to obtain different load components; and finally, constructing a corresponding differential autoregressive moving average ARIMA model for different load components. And meanwhile, the ARIMA model is optimized by utilizing a red pool information criterion and a Bayesian information criterion, and finally, load components predicted by different ARIMA models are reconstructed to obtain a final prediction result, so that the accuracy of load prediction is effectively improved.

Description

technical field [0001] The invention belongs to the field of power system load forecasting, and more specifically relates to a power system load forecasting method and system based on wavelet noise reduction and EMD-ARIMA. Background technique [0002] Power system short-term load forecasting is one of the core parts of smart grid integrated intelligent energy management system. An accurate short-term load forecasting model facilitates rational planning of ongoing grid operations with efficient resource management. The randomness, non-stationarity and nonlinearity of the load curve make the accurate modeling of short-term load forecasting a rather challenging task. [0003] At present, there are many related studies on power system load forecasting, including differential autoregressive moving average (Autoregressive Integrated Moving Average Model, ARIMA) model, Kalman filter forecasting, Markov forecasting, support vector machine (Support Vector Machine, SVM), Artificial...

AI Technical Summary

Problems solved by technology

First, use two time series models to match the original data, detect the types of outliers, and then repair the noise points and missing values ​​in it, and then construct a prediction model based on SVM, but the prediction method used lacks further optimization.

Images