Research On Fault Early Warning For Transmission Lines Based On Historical Information

China's geographical environment is complicated and changeable.Usually,high voltage transmission line often steps over multiple different natural environment,therefore it is easier to be influenced by gale,heavy rain,lightning stroke,ice coating and other severe meteorological condition.Based on the long-term operating experience of power grid,the transmission line malfunction is associated with severe meteorological conditions to a large extent.With the increasingly huge and complex size of the power grid in China,the damage of the transmission line malfunction caused by severe meteorological condition becomes more serious every day.Its degree of damage is ranging from flashover trip to broken line,tower collapse and other accident.Based on the lots of historical data,the writer studies the occurrence risk of the transmission line malfunction under different meteorological environment,so as to make early warning for the high-risk line and to prevent the occurrence of accidents.Therefore,it has very important theoretical and practical significances in the aspects of strengthening the power grid resistance ability to severe weather,preventing the occurrence of power grid disaster and improving the safe and stable operation of power grid.In this paper,the study is conducted focusing on the ice-coating fault and lightning strike fault of transmission line.The key point is to study the occurrence mechanism of these two types of faults,the influences of meteorological characteristics on these two types of failure risk.And the failure risk is evaluated and predicted via using meteorological characteristics and monitoring data,so as to achieve risk early warning of transmission line malfunction.The data mining method is applied in the early warning of transmission line malfunction.Different from the traditional forecasting and early warning,data mining depends on data driving to study the relationship among the all data.It is mainly based on the grid equipment information,operation information,environmental information,equipment failure and other data to make forecasting and early warning for the operating condition of the power grid equipment.For the risk early warning algorithm of the transmission line malfunction based on historical information proposed in this paper,the writer carries out mining and correlation analysis to the historical data via collecting a lot of historical information instead of establishing specific analytical model,so as to achieve the risk early warning of transmission line malfunction.The main research contents are as follows(1)Study on the risk early warning of the ice-coating fault of transmission line.After carrying out the extraction,cleaning and fuzzy classification processing to the historical icing data,principal component analysis is used to extract icing-related characteristic quantity.The classical algorithm in DM,namely,the FP-growth algorithm is applied to dig out the association rules among the icing growth rate,current icing quality,temperature,humidity and wind speed of transmission line,so as to forecast the growth curve of the icing quality.Through the collection of icing quality values during ice-coating fault,parameters are applied to estimate the icing alarming quality.Combined with the forecast curve of icing quality and icing alarming quality,fault early warning is conducted to the transmission line.And the early warning area is analyzed and determined via the statistics of the regional information of the ice-coating faults in history.(2)Study on the risk early warning of the lightning strike fault of transmission line.In consideration of the instantaneity feature of the lightning strike fault,the writer proposes two kinds of warning algorithms,namely,the first one is the risk early warning for the lightning strike fault of transmission line based on the historical information.And another algorithm is the risk early warning for the lightning strike fault of power transmission line based on the real-time information.For the risk early warning for the lightning strike fault of transmission line based on the historical information,it is mainly based on the Naive Bayesian algorithm to set up Naive Bayesian forecast model via the fuzzy processing to historical characteristics data,so as to carry out the lightning strike fault early warning of transmission line.For the risk early warning for the lightning strike fault of transmission line based on the real-time information,KD tree search algorithm is used to inquire K neighbors away from predicted points via monitoring the real-time characteristic parameter information recorded by the device.Based on the K-neighbor thought,the risk early warning for the lightning strike fault of transmission line is conducted.