Research On Fault Diagnosis Of Dry-type Transformer Based On Finite Element Analysis Of Temperature Field

Dry-type transformer is used in power systems due to their small size,low noise,and environmental friendliness,and their usage has increased year by year.Partial or even large-scale power outages in power systems are caused by transformer failures,which have a serious impact on production and life.Therefore,research on fault diagnosis of dry-type transformer has important theoretical value and application prospects for power systems.The change of temperature field is caused by the fault types,so the temperature is selected as the main feature of dry-type transformer fault diagnosis in this dissertation.The three-dimensional solid model of dry-type transformer is established by finite element analysis method.The fluid-solid coupling numerical calculation of the electromagnetic field,temperature field and fluid field of the model is carried out to obtain the temperature field distribution and hot spot position of the dry-type transformer under stable load operation.The accuracy of the model is verified by comparing with the measured values.On this basis,various faults of dry-type transformer are simulated.After acquiring the transformer temperature information sample set under fault condition,a dry-type transformer fault diagnosis model based on improved particle swarm optimization BP neural network is established.On the basis of previous studies,the clips are added to the dry-type transformer simulation model in this dissertation.At the same time,the key parameters affected by temperature change such as air density,specific heat capacity and viscous coefficient are added to the dry-type transformer temperature field modeling process,and the accuracy of the simulation model is improved.Aiming at the shortcomings of BP neural network which is easy to fall into local optimum and slow convergence speed,the traditional particle swarm optimization algorithm is improved in this dissertation,which is combined with BP neural network.The nonlinear approximation ability of BP neural network and the improved global search ability of particle swarm optimization algorithm are both considered by this method.Therefore,the method is applied to the dry-type transformer fault diagnosis model.In this dissertation,after establishing the three-dimensional temperature field simulation model of dry-type transformer,the calculated temperature data is compared with the actual operating temperature,and the error is less than 1.5%.The BP neural network with improved particle swarm optimization algorithm is compared with that before optimization.It is found that the minimum mean square error is 8.2%before optimization under the same iteration number.When the same error is reached,the number of iterations after algorithm optimization is less than 31.6%before optimization.The dry-type transformer fault diagnosis model based on particle swarm optimization BP neural network is verified,and the accuracy is 95.83%.