When the low-frequency oscillation occurs

the accurate oscillation type discrimination is crucial to determine the oscillation causes and take the reasonable suppression measures. This paper subdivides the low-frequency oscillation into local oscillation

interval oscillation and local-interval coupled oscillation. Based on the basic principle of entropy and the oscillation energy temporal and spatial distribution characteristics of different types of low-frequency oscillation

the discrimination criteria based on the energy distribution entropy is proposed. The electrical quantity data of dominant oscillation mode is extracted by Prony decomposition firstly

and the criteria of spatial distribution entropy and normalized standard deviation can be obtained by calculating the unit oscillation energy. Then

the trend of oscillation energy is reflected by adding the sliding windows and the temporal and spatial distribution characteristics are manifested by curve fitting. Finally

different types of low-frequency oscillations are distinguished by the entropy-deviation classification criterion. Simulation results of different type low-frequency oscillations in North China power grid demonstrate the validation of this method. This work is supported by State Grid Corporation of China(No. 524606160015).