Open AccessThe Application of Improved Neural Network Algorithm Based on Particle Group in Short-term Load Prediction
Author(s) -
Naichang Yuan,
Tianyin Zhang,
Teng Zhang
Publication year - 2021
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/632/4/042045
Subject(s) - artificial neural network , computer science , scheduling (production processes) , electric power system , term (time) , approximation error , algorithm , power (physics) , simulation , real time computing , artificial intelligence , mathematical optimization , mathematics , physics , quantum mechanics
Short-term load forecasting is the basis of power system operation and analysis, and is of great significance to unit composition, economic scheduling, safety verification and so on. In modern power system, the influence of meteorological factors on power system load is becoming more and more prominent. However, the traditional model fails to take into account the weather factors that affect the load change, and when the weather changes greatly, the model prediction error is large. Therefore, based on the factor analysis of five meteorological factors, a neural network prediction model combining particle group algorithm is established, and the working day load of a region is predicted, compared with the actual load, the average relative error is 1.24% and the maximum relative error is 6.50%. It shows that the model has high precision. The prediction results of the model also provide the basis for the load adjustment of the power system.