Research on traction load forecasting of electrified railway

Research on traction load forecasting of electrified railway

Blog Article

In order to ensure the power quality of traction power supply system and reduce the costs for purchasing electricity by traction substation, a hus726040als214 probabilistic forecasting model of traction load based on sparse Gaussian process (SGP) is proposed.Firstly, the model constructed the input feature vector based on the historical characteristics of traction load and time information.Then, the mapping relationship between input characteristics and traction load was established, and SGP was used to fit the mapping relationship.Finally, a rolling forecasting methodology was used to predict the traction load.

The comparative experiments carried out on the real operation data of a traction substation on Shuohuang railway verify the advantage of the proposed method, where the point rivolis las vegas forecasting can obtain the prediction value with an error of about 7%, and the probabilistic forecasting can obtain the reliable prediction interval under different confidence coefficients.