A Data-Driven Method for Sub/Super-Synchronous Oscillation Source Location

Sub/super-synchronous oscillation (SSO) induced by power electronic equipment has become a serious challenge to the security and stability of power system. Accurate and timely location of SSO sources is the basis of emergency control. This paper proposes a data-driven SSO source location method. Referring to the transient energy flow method and the sub/super-synchronous power method, the proposed method contains two parts: sub-stations and the master station. The ANN classification model deployed in sub-station judges the relative position of SSO source based on local measurement, and the choice of features is verified by a feature ablation study. After aggregating judgment of all sub-stations, the main station locates the oscillation source. Finally, scalability study is constructed. The results show that the proposed method does not rely on the gird topology, and achieves high accuracy when the grid topology changes, but the generalization ability of the oscillation induced by different types of generating units needs to be further improved.