Operational diagnostics of high voltage instrument current transformers

Abstract

The research contains the results the analysis of oil-filled high-voltage current transformers and selected diagnostic parameters that enable to detect main types of damage. It is shown that in case of the most common faults numerous diagnostic parameters are to be controlled and it is not always possible in conditions of electric power systems modes control by operative dispatching staff of power systems. To reduce the information load on the personal, responsible for decision-making regarding modes control it is proposed to use the parameter of total residual resource, the method of this parameter determination is given, based on the usage of remaining resource ratios, defined by individual diagnostic parameters, taking into account statistical data, concerning taking the current transformers out of service. It is proposed to determine these factors in relative units, that will simplify their usage while determining the ratio of general residual resource that varies from 1 during putting into operation to 0, that corresponds to taking the transformer out of service. Application of mathematical tools of neural- fuzzy modeling, namely Sugeno algorithm, allows to develop mathematical and computer models of total residual resource that enables to determine the impact of diagnostic parameters on the ratio of total residual resource. In training sample of neural -fuzzy model it is suggested to use data, verified by independent experts and taken from the test reports. It is shown that the current prediction error residual life does not exceed 14%. It is expected that the error can be reduced by increasing the training set, and the number of failures in the current transformer will reduce the defects in the early stages of development