A method of transformer fault identification based on hybrid intelligent algorithm

Abstract

The invention discloses a transformer fault identification method based on a hybrid intelligent algorithm, comprising the following steps: establishing a sample set, determining the structural scale of the identification model, and presetting parameters; generating a particle swarm according to preset parameters of the recognition model, wherein each particle represents a set of recognition modelparameters to be optimized; the fitness of each particle in the particle swarm is calculated, the evolution mode of the particle is selected, the position and velocity of the particle are updated, theoptimization result is output when the set iteration times are reached, the optimization result is taken as the parameter of the identification model, the identification model is established, and theidentification model is used for the identification of transformer faults. This method uses algorithm to mine and analyze a large amount of data, and then make judgments and predictions. It is not limited by expert experience and subjective cognition, and has strong scalability. It realizes intelligence through self-learning and reasoning ability based on data.

Description

Technical field [0001] The invention relates to a fault identification method, in particular to a transformer fault identification method based on a hybrid intelligent algorithm. Background technique [0002] As an important and expensive equipment in the transmission and distribution link, the current operation and maintenance of transformers mainly relies on regular batch maintenance combined with condition assessment. This time-consuming and labor-intensive method has become a bottleneck in improving the safety, reliability and economy of power grid operation. Improving the accuracy and effectiveness of equipment condition assessment can greatly improve the efficiency of operation and maintenance, and transform the original passive way into active inspection and repair work in response to changes in transformer status, and reduce the cost of equipment operation and maintenance. [0003] Traditional state maintenance is mainly based on the "Guidelines for the Evaluation of the Co...

AI Technical Summary

Problems solved by technology

[0005] As the scale of data becomes larger and larger, the data dimension continues to expand with the investment in inspection instruments and equipment, and the information continues to increase and generate redundancy. The transformer data targeted by intelligent algorithms is becoming more and more complex, and a single algorithm can no longer meet the analysis requirements.

For example, when the support vector machine is dealing with high-dimensional complex nonlinear data, the calculation is slow, time-consuming, low-precision, and weak in generalization; fuzzy C-means clustering is sensitive to samples, and the classification accuracy is low; Mistakes are prone to occur in the determination of , the division of status levels, and the combined classification or stratification of data indicators with complex correlations.

The state association rule analysis method based on mathematical statistics has high requirements for previous statistical data and poor fault tolerance

The genetic algorithm has a large amount of calculation, is complex, and the late convergence speed is slow, and it is easy to fall into premature

The artificial neural network has a complex structure, is sensitive to parameters, has slow convergence speed, and is prone to overfitting

Images