Lightning Warning Model Research And System Design Of Transmission Line Towers Based On Panoramic Data Driven

The generation of lightning is the result of various factors such as thundercloud thickness,air temperature and humidity,wind speed,wind direction,topography,surface vegetation,and transmission line running status.Use these panoramic data to analyze the relationship between data and lightning strikes,and to provide early warning of lightning strikes,can effectively reduce the losses caused by lightning strikes.At present,lightning monitoring,forecasting and research are still relatively weak in China.Especially in transmission line warning.The existing lightning warning methods for overhead transmission lines are mostly based on single variables.It ignores the relationship between Panoramic data,such as micro-meteorology,micro-topography and transmission line ontology that leads to lightning strikes.On the other hand,the existing lightning warning method has big granularity,it has a little of significance to the decision-making of power system lightning protection.For the above two problems,this thesis establishes lightning strike warning model for transmission towers based on the method of panoramic data driving that covers micro-meteorology,micro-terrain,and the properties of towers.The specific work is as follows.1)First,the k-Nearest Neighbor Support Vector Machine is used to deeply learn the historical data of the lightning activity accumulated by the power system.Training a supervised lightning warning model for transmission lines.2)Secondly,put forward transmission line towers lightning warning model based on Neighborhood Preserving Embedding(NPE)algorithm.Using the historical data under normal condition of transmission line to train the model and establish the SPE statistical limit.The new data collected in the future is put into the model space to calculate statistical of The SPE.If the SPE statistic of the new data exceeds the statistical limit so the new data is abnormal and the corresponding tower take the risk of being struck by lightning.3)Finally,based on the statistical data of Meng Xi T line provided by the Pu’er Power Supply Bureau in 2016,the experiments was conducted with MATLAB to verify the effectiveness of panoramic data-driven transmission tower lightning strike warning model.according to the algorithm model of this thesis,realize the interface design of JAVA and MATLAB based on transmission tower lightning strike warning model.