State Assessment And Fault Diagnosis Method For Electric Power Transformers

With the development of society and the improvement of automation,the requirement for the reliability of power grid is becoming higher and higher.As an important power hub equipment,the reliability of power transformer directly affects the stable operation of power system.Therefore,it can effectively grasp the status of power transformer,predict the potential faults of power transformer,and provide theoretical guidance and auxiliary decision-making for the maintenance and operation of power transformer,so as to reduce the faults of transformer and reduce power system accidents.In this paper,a variety of data processing methods,based on transformer operation data and experimental index data,according to different evaluation needs to select effective evaluation indicators,establish the corresponding power transformer condition evaluation system,through intelligent data processing method to build a transformer condition evaluation model.When a power transformer fails,this paper takes dissolved gas in oil as the fault index and establishes a transformer fault state assessment model based on Intelligent algorithm.The main work of this paper is as follows:When the power transformer is in relatively good condition,in order to have relatively simple practicability in the process of overhaul,a power transformer condition assessment method based on Fuzzy C-means clustering algorithm(FCM)is proposed to use as few experimental indicators as possible.Relatively implementable and important experimental indicators are selected as evaluation indicators in transformers.Data and transformer condition evaluation system are constructed.Relative aging degree of each index is quantified.Weights of each index are set by the method of optimal weights.Based on this,a power transformer condition evaluation model based on fuzzy C-means clustering algorithm is built.Taking transformer oil as an example,the validity of transformer evaluation method based on fuzzy C-means clustering is proved.When the transformer has a certain operating life,the purpose is to have a comprehensive and meticulous assessment of the transformer state when the transformer is under maintenance.A transformer state assessment method based on Set Pair Analysis Theory and Evidence Fusion is proposed.Through various transformer experiments,the indexes of transformer state are selected,and the evaluation system of transformer state is established.The state index of transformer is quantified by relative deterioration degree,and the quantified index is fused by set pair analysis method.The correspondingrelationship between the characteristic quantity and the state grade is obtained.The relationship between the whole transformer and the state grade is obtained by fusing each characteristic quantity with evidence theory.The state of transformer is evaluated by this method and verified by an example.The state of the device is evaluated more accurately.It provides theoretical guidance for actual transformer condition assessment.When the transformer has been faulted,it will become particularly important to identify the type of transformer faults effectively and quickly.The operating environment and external environment of the transformer have a certain impact on the experimental data of the transformer.Based on the local density clustering algorithm,a method for transformer fault state assessment is proposed.With the data of Dissolved Gases Analysis(DGA)in oil as the fault index,the fault index is preprocessed and normalized.A model for transformer fault state assessment based on local density clustering is established.The actual transformer fault data are used to verify the method.It is proved that this method has high fault recognition rate of transformer and can modify the fault evaluation model with input fault data.It provides a new idea for transformer fault state assessment.