Research On Lightning Location Algorithm Based On Clustering Analysis

Lightning location system can monitor lightning parameters such as the time,position and number of lightning return strokes in real time through detection stations,which is the scientific basis for lightning warning.Due to the lightning data received by the detection stations are affected by various complex factors,it is difficult for the current lightning location systems to accurately correct the errors caused by these factors.Therefore,studying lightning location algorithm which is against error interference is meaningful for improving the detection efficiency and disaster reduction by meteorological departments.On the basis of summarizing the principle and algorithms of lightning location,this paper deeply studies the application of clustering analysis in lightning location,the theory and method of error analysis,and evaluates the performance of national business network.The main contents are as follows:Firstly,the principle of lightning location and basic location algorithms are summarized,and the main research object is Cloud-to-Ground lightning location.The commonly used lightning location methods are introduced,including the direction location method,the time difference location method and the combined location method.The differences and applicable scenes between the planar location model and ellipsoid location model are discussed.The basic equations of two-station combined location method,multi-stations time difference location method and Taylor series expansion method are listed.Secondly,the lightning location algorithm based on clustering analysis is studied.Different clustering algorithms are introduced into the location calculation according to the characteristics of lightning data,mainly include K-Means,FCM and DBSCAN.In view of the applicable scene of lightning location,improved algorithms are proposed to determine the optimal parameters for the algorithm.The effectiveness of the algorithm is verified by examples and regional simulation experiments.Thirdly,the lightning location error is quantitatively evaluated by Monte Carlo analysis method,mainly discuss the influence of the number and layout of detection stations in the Cloud-to-Ground lightning location.The grid search optimization which improves the location accuracy of Cloud-to-Ground lightning location are studied to satisfied the demand of practical applications.Finally,the performance of national business network is evaluated.The lightning location algorithm based on clustering analysis is used to relocate the historical data.The comparative analysis is conducted from three aspects:lightning frequency and time distribution,lightning spatial distribution and regional spatial distribution.The results validate the effectiveness of the proposed algorithm and the feasibility of its application to practical application.