Transformer winding loosening state identification method based on vibration signal fractal box dimension

Abstract

The invention relates to a transformer winding loosening state identification method based on the vibration signal fractal box dimension. The fractal box dimension of the vibration signal time domain waveform acts as the characteristic quantity based on the phenomenon that the nonlinear characteristic of the vibration is enhanced when the transformer winding is in the loosening state to quantitatively describe the nonlinear characteristic of the winding vibration waveform, and the loosening state of the transformer winding is identified according to the change law of the fractal box dimension. According to the method, the loosening state of the transformer winding can be effectively identified, and state characteristic quantity extraction is performed by directly using the vibration signal time domain waveform so that calculation is simple and the integrity of the vibration signal can be guaranteed.

Description

Technical field: [0001] The invention relates to a transformer winding loose state identification method based on vibration signal fractal box dimension. Background technique: [0002] Long-term operation of transformers may cause loosening of transformer windings due to aging of insulation pads, loose fasteners, short-circuit impact, etc., and loose windings will further reduce the short-circuit resistance of the transformer, resulting in transformer failure. is of great significance. The transformer winding state diagnosis technology based on the vibration signal has received wide attention due to its advantages of no electrical connection with the system and simple operation. The change law of the transformer winding determines the state of the transformer. Existing studies have proposed main feature quantities including the fundamental frequency energy ratio of the vibration signal, the energy spectrum distribution law, and the characteristic frequency response. These ...

AI Technical Summary

Problems solved by technology

Existing studies have proposed the main feature quantities including the fundamental frequency energy ratio of the vibration signal, the energy spectrum distribution law, and the characteristic frequency response. These feature quantities are the frequency domain representation of the vibration signal. All the information of the time-domain signal cannot be retained, and only part of the frequency is selected when extracting the feature quantity, which leads to the fact that the feature quantity extracted based on the frequency domain does not contain the complete information of the time-domain signal, which may be due to the omission of important information. Accuracy and versatility have a greater impact

Images