Research On Transformer Fault Diagnosis Based On Immune Particle Swarm Optimization Learning Vector Quantization

In the power supply system,transformer is one of the main electrical equipment.Once it breaks down,it will bring serious consequences to enterprises and users.When the transformer collapses,it is very important for the technicians to make a reasonable response to the transformer fault in the quickest time.Compared with the causes of other electrical equipment failures,the causes of transformer failures are more complex and changeable,and the single fault testing method has been difficult to be applied in the transformer fault testing.Therefore,this paper will put forward a method combining a variety of intelligent.algorithms,that is,the method combining learning vector quantization neural network algorithm and immune particle swarm optimization algorithm for transformer fault diagnosis.In this paper,the author firstly introduces the electrical equipment of transformer,and analyzes the necessity and importance of fault diagnosis.At the same time,the characteristic gas and transformer fault types are described in detail when the transformer fails.Then it uses the chemical method of transformer detection-dissolved gas analysis in oil,that is,the gas generated from the transformer to realize the detection of transformer faults.The common treatment method of generated gas-three ratio method is introduced.Then from the principle,coding and other aspects,the three ratio method is elaborated in detail,but also explains the shortcomings of this method.In this paper,a new processing method,namely the organic combination of learning vector quantization neural network and immune particle swarm,is presented.Finally,learning vector quantization neural network algorithm and immune particle swarm optimization algorithm are described in detail.The selection of parameters and eigenvectors of the learning vector quantization neural network algorithm improves the accuracy of fault diagnosis.The optimization of learning vector quantization neural network algorithm by immune particle swarm optimization algorithm increases the stability of learning vector quantization neural network algorithm.The combination of learning vector quantization neural network algorithm and immune particle swarm optimization algorithm is used to diagnose the gas generated by transformer,and the accuracy of diagnosis results is as high as 93%.In order to prove the superiority of the algorithm and verify the correctness and reliability of the experiment,under the same conditions,the diagnosis model of immune particle swarm optimization algorithm combined with learning vector quantization neural network algorithm is established in the software.The learning vector quantization neural network classification model and the immune particle swarm optimization neural network classification model are established,and they are compared and analyzed.By comparing and analyzing the correct rate of fault identification,the running time of the model and the stability of fault diagnosis,the generated gas is processed by the combination of learning vector quantization neural network algorithm and immune particle swarm optimization algorithm,and the results show that this method has a good effect on transformer fault classification.Figure[32]table[13]reference[80]...