Voltage sag classification method based on SSA-RF algorithm

The invention discloses a voltage sag classification method based on an SSA-RF algorithm, and the method comprises the following steps: S1, carrying out the processing of three-phase voltage data containing a voltage sag occurrence time period, obtaining a root mean square value, carrying out the Clark transformation, and converting the three-phase voltage into a group of orthogonal quantities; s2, processing the orthogonal quantity obtained in the S1 to obtain parameters of a voltage sag vector ellipse, including a long axis value dm, a short axis value dn and an inclination angle theta, and taking the parameters as characteristic quantities of voltage sag; and S3, by taking the characteristic quantity of the voltage sag as an input value, establishing a fuzzy relationship between the characteristic quantity and the voltage sag type through a random forest RF algorithm optimized by a sparrow SSA, so as to classify the voltage sag. The voltage sag ellipse parameters can effectively reflect the characteristics