Method for identifying and diagnosing inverter overcurrent in real time based on time sequence working condition event set

The invention discloses an inversion overcurrent real-time identification and diagnosis method based on a time sequence working condition event set, and the method comprises the steps: S1, collecting current data, obtaining an inversion current frequency through fast Fourier transform, and obtaining the size and step length of a sliding window through the calculation of the current frequency; s2, offline fault analysis: collecting historical data of a traction transmission system signal to perform inversion overcurrent fault mechanism analysis, then constructing a feature identifier, and performing segmentation processing on the constructed feature identifier under the obtained sliding window to obtain a feature index in a period; s3, generating corresponding time sequence events from the obtained characteristic indexes through a hysteresis module, constructing a fault identification template through the obtained time sequence time, and finally forming an inversion output overcurrent fault positioning and ide