Establish Model Of Dielectric Relaxation Response Of Transformer's Oil-paper Insulation And Study On Diagnosis Of Aging

Aging of oil-paper insulation has been the main reason resulting in the malfunction of the oil-paper insulation transformers,hence take the condition of transformers' oil-paper insulation in control timely and accurately will enhance the operational reliability of the power grid in a deep degree.Nowadays,the diagnosis technology of insulation still exist some shortages.There are many interference factors in field measurements,the model of the medium response is difficult to determine and an effective multi-parameter quantitative evaluation method for aging state is in shortage.This thesis aims at researching on analyzing interference factors in tests of transformers' insulation on the spot and put forward improving measurements for increasing testing accuracy of time domain dielectric spectroscopy.Meanwhile,study on polarization theory of oil-paper insulation,and set up a new mathematical model of dielectric relaxation response.Moreover,based on the testing data from the time domain dielectric spectroscopy,the research makes effort to identify characteristic parameters of medium response which can characterize the aging status of oil-paper insulation.Finally,achieve synthetic diagnosis of multi-parameter for aging status of oil-paper insulation.The main work of the thesis is as follows:Firstly,analyze the impact of surroundings when transformers are in the tests,and put forward measurements to response different influencing factors.In the case of abnormal polarization spectrum receiving from tests,the thesis raises an idea of optimizing the testing parameters,and obtains the applicable conditions in tests which are combined with case analysis,in order to achieve the purpose to avoid and reduce the impact of external factors and ensure the accuracy and effectiveness of tests.Next,the thesis investigates equivalent circuit model of dielectric relaxation response of oil-paper insulation and reminds of the shortage of the type of polarization mechanism in extended Debye model.The model introduces interface polarization branches into consideration to show the process of the interaction between the media.Furthermore,form the relationship between the model and the dielectric parameters of insulation aging by means of RC elements.Afterwards,make improvements on traditional extended Debye model and it can raise the accuracy of the equivalence of the model to the insulation condition effectively.Besides,the thesis raises the method of parameters' identification for the equivalence circuit model.First of all,analyze and identify the circuit model,then set up the objective function which is by means of nonlinear optimization to solve out each circuit parameters.Secondly,finish the analysis for dielectric spectrum and obtain the number of polarization branches,hence we can determine the number of circuit parameters which need identification.Lead the same amount of data received from field measurements into the objective function and the optimal value of each circuit parameter is identified and acquired when combined with Adaptive Particle Swarm Optimization.Ultimately,make comparison of accuracy in return voltage polarization spectrum identified by two models with the testing polarization spectrum.It confirms the veracity of method and the superiority of improved model.Last,the thesis is based on the establishment of the model for improved circuit and afterwards proposes characteristics which can express the medium's relaxation institutions and process of response and assess the oil-paper insulation aging status.In addition,with the help of testing information and data,the thesis raises an approach to diagnose insulation aging,which is based on the theory of fuzzy rough set.Furthermore,the characteristics of diagnosis are divided fuzzily through FCM clustering algorithm and the membership function for the properties of insulation evaluation system takes shape.According to discern matrix set up by fuzzy attributes,the thesis achieves the reduction of the fuzzy attribute set and extract the criterion of evaluation of oil-paper insulation states with the interval of membership.At length,according to the criterion of evaluation,the thesis receives the insulation status and confirms the evaluation's feasibility by living examples.In conclusion,the thesis puts forward steps to response influencing factors in field measurements.Then lead in interface polarization branches and increase the circuit model's accuracy of the equivalence.Simultaneously,combine with the characteristics of the model and the information and data of transformer insulation,and then introduce the assessment based on multiple parameters for the diagnosis of transformers'insulation aging status.The research has important practical significance and value of application on grasping the status of transformer's oil-paper insulation timely and accurately,heightening the operational reliability of equipment,and confirming the operational stability and the safety of the power grid.