Research On Data-Driven Temperature Rise Early Warning And Economic Life Assessment Of Distribution Transformers

Focusing on the background of the times when the country promotes the development of the distribution network towards the direction of digitalization and intelligence,this paper comprehensively considers how to improve the distribution transformers from three aspects:the aging of the joints of oil-immersed distribution transformers,the aging of insulating oil and the economic life evaluation of the transformers.Long service life,improve the safety and economic benefits of distribution network operation.The equivalent circuit model of the three-phase four-wire distribution transformer and the equivalent equation system for calculating the contact impedance are proposed to solve the hidden problem of local temperature rise in the distribution transformer in the platform area.Starting from the contact impedance of the transformer joint,based on the electrical parameters of the TTU upstream,the equation solving problem is equivalent to an optimization problem of finding the minimum value;a short-term oil temperature prediction model of the transformer is proposed using the electrical data collected by the TTU without input parameters,using the KS test to compare and analyze the multi-parameter sensitivity of the factors affecting the oil temperature rise of the oil-immersed distribution transformer,and use the LSTM algorithm to overcome the problem of the timeliness of the missing information in the oil temperature prediction of the GRNN algorithm.Solve the defect that the top oil temperature prediction of oil-immersed distribution transformers does not contain time-sensitive information data processing,analyze the internal thermal effect of oil-immersed transformers,expound the heating and heat dissipation mechanism of transformers,and use the top oil temperature of transformers to replace transformer hot spots The temperature reflects the overall insulation condition inside the transformer.Combined with the research on contact impedance and oil temperature,we continue to study the cost of oil-immersed distribution transformers over time after they are put into operation,and give a transformer investment strategy model based on LCC theory.Compare the cost of its investment strategy with the cost of using an investment model that uses traditional service time to treat transformers.The optimal solution after the iteration of the BFGS algorithm is used to replace the solution of the original equation system,and the long-term heating model under the maximum load current is used to establish the equivalent relationship between the contact impedance and the temperature,and form a graded early warning system for the contact impedance.The comparison of multi-parameter sensitivity using the KS test shows that the ambient temperature has the highest influence on the prediction of the oil temperature of the top layer.The LSTM algorithm overcomes the defect of the timeliness of the predicted oil temperature information,and the prediction results are more accurate.Conducting a comprehensive cost analysis of distribution transformers and optimizing the investment decisions of transformers can not only reduce the hidden danger of transformer failures,but also minimize the operation and maintenance costs of power grid investment in transformers.Rapid and stable sustainable development of the industry.