Study on Negative Sequence Source Model of Electrified Railway Traction Load Based on Clustering Tendency Analysis and Stepwise Regression

SUN Qian,YAO Jiangang,LI Xinran,KONG Qiang,HU Jinhong,KANG Tong,XU Zhenchao

2012-01-01

Abstract:To establish the electrical railway traction load negative sequence source model suitable for negative sequence characteristics analysis and power flow calculation is extremely important. Negative sequence source modeling could be regarded as a load modeling problem, and a modeling method based on the classification and synthesis of negative sequence characteristics was proposed. The clustering tendency analysis was conducted on the sample set, with the measured response space of fundamental negative sequence current and the actual running state of locomotive as eigenvectors. On the premise of clustering possibility, the clustering validity function was used, aiming at obtaining the optimal classification result of the sample set. By means of mechanism analysis for the negative sequence characteristics of traction load, the model structure was determined, while the optimal model expression was achieved applying the method of stepwise multiple regression. Besides, the newly added sample was classified into the class within the minimum Euclidean distance from the clustering center, and the related model was tested. Case study shows that the negative sequence source model built is reasonable and effective. © 2012 Chin. Soc. for Elec. Eng.