A Short-Term Load Forecasting Model with a Modified Particle Swarm Optimization Algorithm and Least Squares Support Vector Machine Based on the Denoising Method of Empirical Mode Decomposition and Grey Relational Analysis

Abstract

Abstract:As an important part of power system planning and the basis of economic operation of power systems, the main work of power load forecasting is to predict the time distribution and spatial distribution of future power loads. The accuracy of load forecasting will directly influence the reliability of the power system. In this paper, a novel short-term Empirical Mode Decomposition-Grey Relational Analysis-Modified Particle Swarm Optimization-Least Squares Support Vector Machine (EMD-GRA-MPSO-LSSVM) load forecasting model is proposed. The model uses the de-noising method combining empirical mode decomposition and grey relational analysis to process the original load series and forecasts the processed subsequences by the algorithm of modified particle swarm optimization and least square support vector machine. Then, the final forecasting results can be obtained after reconstructing the forecasting series. This paper takes the Jibei area as an example to produce an empirical analysis for load forecasting. The model input includes the hourly load one week before the forecasting day and the daily maximum temperature, daily minimum temperature, daily average temperature, relative humidity, wind force, date type of the forecasting day. The model output is the hourly load of the forecasting day. The models of BP neural network, SVM (Support vector machine), LSSVM (Least squares support vector machine), PSO-LSSVM (Particle swarm optimization-Least squares support vector machine), MPSO-LSSVM (Modified particle swarm optimization-Least squares support vector machine), EMD-MPSO-LSSVM are selected to compare with the model of EMD-GRA-MPSO-LSSVM using the same sample. The comparison results verify that the short-term load forecasting model of EMD-GRA-MPSO-LSSVM proposed in this paper is superior to other models and has strong generalization ability and robustness. It can achieve good forecasting effect with high forecasting accuracy, providing a new idea and reference for accurate short-term load forecasting.

Case Library Construction Technology of Energy Loss in Distribution Networks Considering Regional Differentiation Theory

“…From a data-processing perspective, to assess the influence degree between influence factors and the major variable, some data analysis methods could be used. The method of GRA is applied to case analysis in this paper, which is a basic method of the grey system theory for systems analysis [14].…”

Probabilistic Forecasting of Short-Term Electric Load Demand: An Integration Scheme Based on Correlation Analysis and Improved Weighted Extreme Learning Machine

“…Consequently, AI approaches have become the mainstream of LF methods. Traditional LF studies apply AI-based methods of neural network [7][8][9], support vector machine [10][11][12] and other methods [13,14] to forecast the load demand with exactly values. Although the performance of traditional LF models can be improved by optimizing the AI model, the forecasting uncertainty at each load point is unknown [15].…”

A Weekend Load Forecasting Model Based on Semi-Parametric Regression Analysis Considering Weather and Load Interaction

“…For a long time, many experts have been focusing on research on short-term load forecasting. The forecasting methods such as the time series method [3,4], support vector machine method [5][6][7][8][9][10], random forest models [11][12][13][14], artificial neural network method [5,[15][16][17][18][19][20] and grey theory [21][22][23] could be applied to general weekday scenes and obtain good results. However, the difference between characteristics on weekend load and working days, as well as the interactive coupling relationship with external weather information, have become the shortcoming factors that restrict the accuracy of weekend load forecasting.…”