A Transformer Condition Assessment Method Based On Projected Outlyingness Function

In recent years,my country’s power grid has developed rapidly,the scale of the power grid has continued to expand,and the difficulty of power grid operation and maintenance has increased significantly.Transformers play the role of electric energy transformation and distribution in the power system,and are key equipment in the power system.Once a failure occurs,it will not only affect the normal use of electricity by residents,but may also endanger public safety.Ensuring the safe and reliable operation of transformers is an important task for the power operation and maintenance department.Therefore,an accurate evaluation of the state of the transformer is the first prerequisite for condition-based maintenance.Therefore,in recent years,how to monitor the operating status of transformers and identify possible faults early has become a research topic of interest to scholars.In practice,the abnormal state of the transformer is reflected by monitoring the oil chromatographic data,that is,the content of each gas in the transformer.However,the existing methods do not fully consider the correlation and distribution of the indicators of the transformer oil chromatographic data and the trend items caused by the aging of the transformer.These problems are very likely to affect the evaluation results of the final transformer operation status.In view of this,this thesis combines the projection outlier function,a non-parametric tool,to create different recognition windows,and calculate the value of the transformer state projection outlier function at different times in each window.According to the properties of the projected outlier function,four types of transformer states can be divided based on the projected outlier function value score and "Test Regulations for Condition Maintenance of Power Transmission and Transformation Equipment"(Q/GDW1168—2013): Absolutely safe state;General safe state;Safe state But pay attention to status and abnormal status.Finally,the reliable evaluation of transformer status and grading early warning are realized.The outlier detection method of the existing transformer oil chromatographic data is expanded and innovated;then,this thesis will write the relevant R program for the proposed transformer state evaluation method based on the projection outlier function,and aim at abnormal observations with trend items The evaluation method is introduced to carry out numerical simulation,and the results show that the rating method of transformer status based on projection outlier function has good robustness;finally,this thesis applies the outlier detection method based on projection outlier function to transformer status evaluation to investigate The application of transformer state evaluation method based on projection outlier function in actual data shows that the transformer state evaluation method based on projection outlier function proposed in this thesis is effective and has certain practicability.This thesis introduces the concept of statistical projection outlier function into the field of electric power for the first time,which can not only solve the high-dimensional problems,trend problems and unknown data distribution problems in actual power data.Specifically,the data distribution does not satisfy the normal distribution assumption.It can also realize better functions in the transformer state evaluation method.(1)Hierarchical early warning function:through the four types of areas described in the above method,the transformer status can be graded and early warning.(2)Robustness: The analysis of the monitoring data of the transformer’s normal operating state and fault state shows that this method is not easily affected by sample outliers.(3)It is easy to visualize,has higher recognition sensitivity,and can provide evidence for finding the cause of transformer faults according to the abnormal state of the transformer caused by the abnormal content of certain types of gases.It has good generalization in practical applications.