Transformer winding fault diagnosis method based on atomic sparse decomposition

Abstract

The invention provides a transformer winding fault diagnosis method based on atomic sparse decomposition. The method comprises the following steps of using a three-phase transformer test to simulate transformer winding faults under different working conditions; collecting vibration signals, and carrying out atomic sparse decomposition on the vibration signals to obtain an attenuation modal parameter; carrying out data preprocessing on the attenuation modal parameter to obtain characteristic vectors of a transformer winding under different fault types; dividing the characteristic vectors into atraining sample and a test sample; taking the training sample as input, and the fault types of the transformer winding as output, establishing a transformer winding GSO-SOM network fault diagnosis model; training the fault diagnosis model to acquire the trained transformer winding GSO-SOM network fault diagnosis model based on the atomic sparse decomposition; and inputting the test sample into the trained fault diagnosis model to judge a transformer winding fault, and outputting a diagnosis result. The method has characteristics of high reliability, good accuracy and the like and can be widely applied to the fault diagnosis field.

Description

Technical field [0001] The present invention relates to the technical field of fault diagnosis, in particular to a transformer winding fault diagnosis method based on atomic sparse decomposition. Background technique [0002] Transformer winding failure is a major hidden danger to the safe operation of power systems. According to statistics, serious failures caused by mechanical deformation of windings under the action of electrodynamic force account for 70% of total winding failures. The accident rate has risen to the first place. Therefore, it is of great significance to discover hidden troubles of transformer winding faults in time, avoid sudden accidents, and carry out research on transformer winding fault diagnosis. For the fault diagnosis of transformer windings, scholars at home and abroad have proposed a variety of fault diagnosis methods. The main fault diagnosis methods include short-circuit impedance method, frequency response method, rough set theory, support vector ...

AI Technical Summary

Problems solved by technology

Both the short-circuit impedance method and the frequency response method require the transformer to be shut down, and the detection methods are complex and have low precision.

Rough set theory has great advantages in dealing with fuzzy and uncertain information, but its decision rules are very unstable and less accurate, and it is based on a complete information system. When processing data, it often encounters data loss.

Support vector machines have advantages in solving small sample, nonlinear and high-dimensional pattern recognition problems, but the recognition ability is easily affected by its own parameters

The neural network has a simple structure and strong problem-solving ability, and can handle noisy data well, but the algorithm has local optimal problems, poor convergence, and limited reliability

[0003] It can be seen that in the prior art, the transformer winding fault diagnosis method has problems such as low precision, poor reliability, and large deviations in diagnosis results.

Images