| Transformer is the key equipment for voltage converting and power transmission in power system,and the insulation performance is directly influence its operation reliability.It is very important for the safe operation of equipment to make an accurate assessment of transformer insulation status,so as to make effective maintenance or replacement strategy.The dielectric response(DR)technique which can diagnose the condition of oil-paper insulation and make up the drawbacks of traditional diagnosis methods attract attention of domestic and foreign scholar.It has made great progress in the testing technique and evaluation method of the moisture condition of the transformer main insulation,and some commercial insulation diagnostic instruments has been developed.However,due to the moisture and polar products of aging can both affect the DR characteristic of oil-paper,the accuracy of the assessment results is influenced by two factors.At present,DR method can not realize the aging status evaluation of transformer main insulation and it usually neglect the influence of the aging status when evaluate the moisture condition.Therefore,put forward a comprehensive method for accurate assessment of the aging condition of transformer oil-paper insulation has important theoretical research and practical application value.Firstly,in the laboratory,a large number of oil-impregnated pressboard samples with different aging and moisture condition were prepared by accelerated thermal aging and moisture absorption experiment,and their frequency domain DR were measured at constant temperature,then investigate their DR properties under different aging and moisture condition.Then,a neural network data fitting method used for the assessment of aging and moisture condition of oil-paper insulation was proposed,the assessment of moisture condition is accurate but it is unable to evaluate aging status.Therefore considering use the oil-paper moisture equilibrium curves under different aging status to assess the aging status of oil-paper insulation.Then complete oil-paper moisture equilibrium experiment,analysis the effect of aging on oil-paper moisture equilibrium curves.A method based on the assessment result of moisture condition,the oil-paper moisture equilibrium curves under different aging status and the moisture content(m.c)of insulation oil was proposed to assess the aging status of oil-paper insulation,and the validity of this method was verified by laboratory samples.The main conclusions of this paper are as follows:(1)This paper established DR spectroscopy database of oil-paper insulation samples under different moisture condition and aging status The effect of moisture and aging status on the frequency domain DR properties of oil-paper insulation is similar,but the influence of moisture is greater than that of aging.The relationship between the characteristic parameters and moisture and aging status was established by neural network fitting method.The results show that this method can be used to quantitatively evaluate the moisture condition of oil-paper insulation,but can not evaluate the aging status.(2)The oil-paper moisture equilibrium curves under different aging status were obtained,as the aging degree increases,the moisture in the paper move gradually to the oil.While the m.c of oil is same,the m.c of aging paper is lower than that of unaged paper.As the equilibrium temperature increase,the m.c of paper would reduce and the m.c of oil would increase.When the temperature is low,the moisture distribution ratio difference caused by aging status is less than that of high temperature.(3)A method by combine the oil-paper moisture equilibrium curves under different aging status and DR spectroscopy database to assess the moisture condition and aging status of oil-paper insulation.Firstly,using the neural network data fitting method to evaluate the m.c of oil-paper samples,then according to the assessment and measured m.c of insulation oil to assess the aging status of oil-paper samples based on the oil-paper moisture equilibrium curves under different aging status.The assessment results of 15 oil-paper samples are in good agreement with the actual values,the feasibility and validity of the method are preliminarily verified. |
PDF Full Text Request