Research On Fault Diagnosis Of Power Transformer Based On Grey Neural Network Model

The transformer plays the role of boost and depressurization in the power system.It is the key node of the power transmission.If the transformer fails,it will affect the transmission of electric energy and cause a large area blackout.So if we want to make the power system safe to run,it is necessary to diagnose and predict the transformer faults in advance.If an effective method of transformer fault diagnosis and predic tion is found,the frequency of transformer fault can be reduced,and the potential faults and hidden dangers existing in the transformer can be found in advance.There are several methods of transformer fault diagnosis and prediction,but these methods ha ve not been able to meet the requirements of the accuracy of the current prediction,so this paper puts forward the method of transformer fault prediction based on grey neural network model to improve the accuracy of prediction and carry out deep research.Through the study of the standard grey prediction model and the BP neural network model,it is found that if only one of these models is used to predict the model,it can still improve the model itself,but there are still limitations.So the grey neural network prediction model is established in this paper.In order to avoid the limitation of only one model,this paper uses two kinds of models to combine effectively.Due to the complex characteristics of the transformer fault type and many undetermined fa ctors in it,this paper adopts the principle of dissolved gas analysis(DGA)in transformer oil,and takes the dissolved gas dissolved in transformer oil as the research object,and combines the common faults of the voltage transformer to find out the transformer.The internal ratio between the fault gas dissolved in the oil and the common faults is coded by the three ratio method.First,the grey GM(1,1)model is set up.Based on the GM(1,1)model,the grey GM(1,n)model can be built more accurately to analyze a variety of transformer fault input and to optimize the grey GM(1,n)model.By optimizing the background value,the stability of the body is reduced and the error is reduced.Poor.In order to make up the limitations of the grey GM(1,n)model and further improve the accuracy of transformer prediction,the BP neural network model is introduced in this paper.The prediction model based on the grey neural network is established through the effective combination of two models,and the output dat a of the grey GM(1,n)model preprocessing are used as the BP neural network.The input of the collaterals is trained by the L-M adaptive learning rate method,and the MATLAB software is used to simulate the transformer fault diagnosis model,and the simulation results are analyzed.Using the grey neural network model,the four kinds of common faults of the transformer are simulated and analyzed,which improves the stability and accuracy of the transformer fault prediction.