Transformer winding fault diagnosis method based on morphological image processing

Abstract

The transformer winding fault diagnosis method based on morphological image processing comprises the steps: establishing a lumped parameter model for a transformer winding, and acquiring amplitude-frequency curve data of the transformer winding under normal conditions and various faults of the transformer winding; for the obtained amplitude-frequency curve data, selecting two frequency bands withobvious changes, and storing the two frequency bands in an image form; establishing an image library for the obtained image, and establishing a label; importing the images of the to-be-tested transformer winding fault under the two frequency bands, carrying out the preprocessing and morphological method with the images in the built image library, and finally obtaining the area value; and sorting the area values, and diagnosing the transformer winding fault according to the sorting result. The method provided by the invention has very high accuracy, not only can identify the fault type, but also can present the fault degree of the fault type, and is beneficial for maintainers to comprehensively evaluate the operation state of the to-be-tested transformer.

Description

technical field [0001] The invention relates to the technical field of transformer fault detection, in particular to a transformer winding fault diagnosis method based on morphological image processing, which is used for offline detection of transformer windings. Background technique [0002] As the hub of the power grid, the transformer undertakes the function of voltage level conversion and plays the role of energy transmission. The safe and stable operation of the power network is inseparable from its normal working state. Therefore, the fault detection of the transformer is particularly important. At present, the detection methods for transformer faults can be roughly divided into non-electric quantity detection and electric quantity detection. Non-electrical quantity detection includes DGA analysis, ultrasonic detection method, infrared temperature measurement, etc. Electrical quantity detection includes frequency response method, partial discharge detection, etc. T...

AI Technical Summary

Problems solved by technology

However, the lack of utilization of the frequency response method in the prior art is mainly manifested in: ①. Partial differences in the frequency response curve cannot be used to effectively diagnose the fault of the transformer

② Existing methods cannot be used to mine frequency response data, and too much reliance on perceived judgments

It is impossible to process the data obtained by existing methods to obtain more accurate diagnostic results and higher recognition accuracy

Images