| Electricity metering equipment as an important bridge between users and power measurement department,its accuracy directly affects the accuracy and fairness of trade settlement,which is related to the interests of the majority of electricity consumers and the power department.With the establishment of the electricity collection system,the power department has accumulated a tremendous amount of operational data about metering devices.The development of data mining technology makes it possible to do deep analysis and obtain valuable information by using the above massive operational data.Among them,the decision tree induction has been widely used because of its relatively sma ll computation,high classification accuracy,important decision attributes can be displayed,easy to extract explicit rules and other advantages.Based on the operation of measuring equipment in State Grid JiBei Electric Power Company,three cases of higher false rate of measurement abnormalities are analyzed,includ ing the abnormality of power flow,abnormality of power meter breakage and reverse power meter.The reason and the concrete manifestation of the anomaly are analyzed.And then summarized the feedback information of the wrong order.An improved method is proposed for the judgment rules of each anomaly based on site investigation results and expert discussions.In this paper,the specific process of data mining is researched.Data mining technology is applied to deeply analyze the operating data of measuring equipment in the information collection system of electricity consumption.Combined with the above improved rules,the fault diagnosis of metering is realized by the decision tree classification algorithm.Finally,a large number of field examples are used to verify the accuracy and efficiency of the proposed method. |
