Condition Assessment And Fault Diagnosis Approaches For Power Transformers Based On Data Mining Technology

The power transformer is the core equipment for transmitting and converting energy in the power system,and plays a vital role in the safe and stable operation of the power grid.With the expansion of the power grid,many transformers have been running at full load all the year round.The operating status of these devices has gradually deteriorated and the failure rate has been continuously improved.Due to the special nature of power transmission,once a fault occurs,it will affect the safe power supply of the entire network,and cause equipment damage and economic loss,and even endanger life.Therefore,it is of great significance to carry out reasonable state assessment and accurate fault diagnosis of the transformer to effectively improve the reliability of power supply.At this stage,the operation and maintenance personnel mainly rely on online monitoring data and preventive tests to judge the operating state of the transformer.For a specific indicator,the state is divided according to the threshold specified by the relevant standards.Although the development of measurement technology improves the accuracy of the test,most of them can only evaluate a certain state,and due to limited indicators,it has a large one-sidedness and does not comprehensively use various types of information to make an overall state assessment.In view of the above problems,based on the relevant standards and procedures,this paper establishes a transformer state assessment and fault diagnosis model based on data mining,and provides guidance for the operation and maintenance personnel to accurately grasp the running state of the transformer and perform state maintenance.The paper refers to a large number of technical regulations and related standards,and considers the experience of various experts to establish a multi-dimensional information evaluation model for transformers.The index system is optimized in terms of practicability,effectiveness and economy,and the evaluation criteria and operation and maintenance strategies are studied.Based on this,the state evaluation and fault diagnosis model of the transformer are established.A method of transformer condition evaluation based on fuzzy analytic hierarchy process is proposed.According to the online monitoring data and historical data,the multi-dimensional information and various factors are combined to evaluate the health of the transformer.The weight relationship between the indicators is studied and the calculation process is optimized.Aiming at the subjective defects in this method,the improved entropy weight method is introduced to calculate its objective weight.The larger the objective entropy of the indicator,the smaller the weight obtained by information entropy.Considering the weight of subjectiveand objective weights to obtain the combined weights,it not only highlights the differences between the data,but also pays enough attention to the expert experience.The BP neural network is used to learn the nonlinear mapping relationship of state evaluation,which simplifies the process of index weighting and calculation,and improves the practicability and convenience of the evaluation model.The fuzzy clustering theory is used to diagnose the fault of the transformer,and the uncertainty and ambiguity of the index are considered.The fuzzy membership degree is used to indicate the fault type to which the transformer belongs,and detailed fault diagnosis steps are given.Fuzzy C-means clustering is a clustering method for unsupervised learning,which can effectively utilize the unlabeled information in a large amount of monitoring data.Aiming at the defect that the clustering result is unstable due to the random setting of the initial clustering center in the clustering method,the clustering center is optimized by the seeker optimization algorithm,which improves the efficiency and accuracy of clustering.Through the engineering example verification,the transformer state evaluation and fault diagnosis model designed in this paper can accurately judge the health status and fault type of the transformer,and provide a reliable decision basis for the differential operation and maintenance of the transformer.